Japanese Loanwords in Hiragana

From loaning to owning: Japanese loanwords in hiragana

Hannah Kunert

ORCID 0000-0003-2250-6697

Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy

September 2017

School of Languages and Linguistics, and the Asia Institute

The University of Melbourne

Abstract Loanwords1 (外来語・カタカナ語), which are words ‘borrowed’ from other languages, are an integral part of the Japanese language, and are estimated to account for around 10% of the modern Japanese lexicon. While loanwords are conventionally written with the katakana script, recently some examples have been appearing in the hiragana script, which is usually reserved for words of Japanese origin.

This research investigates what kinds of loanwords appear in hiragana, in which genres of text they are typically found, and why hiragana is being used in these cases. A mixed methods research design provided a broad base from which to approach this phenomenon, and consequently four different data sets were utilized: a corpus of hiragana loanwords, a survey, a series of interviews with native Japanese speakers, and four case studies of individual texts. The case studies, in particular, drew on the multimodal nature of these texts, and utilized the ‘visual grammar’ of Kress and van Leeuwen (2006), and the semiotics of typography described by Stöckl (2005) in order to understand how hiragana was being used within a text.

The findings from this research illustrate the wide range of semiotic functions this marked use of script can perform, for example connoting traditional Japanese culture or cuteness; being ‘easier to read’ for perceived audiences; providing a sense of balance with the other scripts used in the text; or highlighting an instance of wordplay. While loanwords in hiragana can be described as a ‘marked’ use of language, another important finding was native Japanese speakers' general level of acceptance of these words within authentic texts, with script having the effect of blurring the line between loanword and ‘Japanese word’. The results of this study therefore extend prior research on Japanese loanwords (Loveday, 1996; Rebuck, 2002; Stanlaw, 2014); typography and graphic design (Kress & van Leeuwen, 2006; Spitzmüller, 2012, 2015); and language play (Gottlieb, 2010; Knospe, 2015). These findings reinforce the often-cited flexibility and adaptability of the Japanese writing system, as well as providing new perspectives on script as a semiotic resource within the Japanese language.

1 It is noted that ‘漢語/kango’ (words borrowed from Chinese), were originally also ‘loanwords’, however they are not part of the current research project. Declaration

This serves to declare that:

(i) the thesis comprises only my original work;

(ii) due acknowledgement has been made in the text to all other material used; and

(iii) the thesis is fewer than the maximum word limit in length, exclusive of tables, maps, bibliographies and appendices.

Signed:

Hannah Kunert

Acknowledgements

I would firstly like to express my gratitude to my supervisors, Dr. Celia Thompson, and Dr. Ikuko Nakane; and the chair of my committee, Associate Professor Dr. Paul Gruba, for their advice and support during the course of my candidature. I am also extremely grateful for all those who gave their time to participate in my research and share their opinions with me. I also acknowledge the support of the Faye Marles Scholarship which has assisted me financially over the last three years.

I would also like to thank my husband, friends, and family for all their extra support during my candidature. A big thank-you is also due to my friends and colleagues in the School of Languages and Linguistics and the Asia Institute for sharing the highs and lows of my research journey with me; especially my fellow travellers, the PhD candidates in room 308.

Table of contents

Chapter 1 Introduction ...... 1 1.1 Background ...... 2 1.1.1 The Japanese writing system ...... 2 1.1.2 Loanwords in Japanese ...... 8 1.2 Rationale for the current study ...... 9 1.3 Aim and research questions ...... 11 1.4 Definition of ‘loanword’ ...... 12 1.5 The use of Japanese in the thesis ...... 14 1.6 Overview ...... 15

Chapter 2 Literature Review ...... 16 2.1 Social semiotics ...... 16 2.2 Multimodality ...... 20 2.3 Systemic functional linguistics ...... 22 2.4 Socio-historical connotations of the Japanese scripts ...... 25 2.5 Loanwords in Japanese ...... 29 2.5.1 A brief history of foreign language contact in Japan ...... 29 2.5.2 Adapting foreign words into Japanese ...... 34 2.5.3 Functions of loanwords ...... 38 2.5.4 Perceptions of loanwords ...... 40 2.6 ‘Non-standard’ orthography...... 44 2.7 Language play ...... 51 2.8 Typography ...... 55 2.9 Translingualism...... 60 2.10 Conclusions from the literature review ...... 63

Chapter 3: Research design and methodology ...... 66 3.1 Terminology ...... 66 3.2 Epistemological background ...... 67 3.3 Design of the current research project ...... 69 3.4 Data collection and analysis...... 73 3.4.1 Layer 1: Descriptive statistical analysis of loanwords in hiragana ...... 73 Layer 1: Data Collection ...... 74 Layer 1: Data Analysis...... 80 Layer 1: Summary of research design ...... 87 3.4.2 Layer 2: Case studies of selected texts ...... 87 Layer 2: Data Collection ...... 89 Layer 2: Data Analysis...... 91 Layer 2: Summary of the research design...... 95 3.4.3 Layer 3: Survey investigating opinions of loanwords in hiragana ...... 96 Layer 3: Data Collection ...... 97 Layer 3: Data Analysis...... 102 Layer 3: Summary of the research design...... 107 3.4.4 Layer 4: Interviews and focus groups ...... 108 Layer 4: Data Collection ...... 109 Layer 4: Data Analysis...... 118 Layer 4: Summary of the research design...... 119

Chapter 4 Findings from the corpus analysis and case studies ...... 120 4.1 Layer 1: Characterising loanwords in hiragana ...... 120 4.1.1 Word frequency ...... 120 4.1.2 Word class ...... 123 4.1.3 Vowel lengthening ...... 124 4.1.4 Referential meaning of loanwords in hiragana ...... 126 4.1.5 Source language ...... 130 4.1.6 Language play ...... 132 4.1.7 Consecutive loanwords in hiragana ...... 135 4.1.8 Instances of hiragana loanword(s) per text ...... 137 4.1.9 Genres containting loanwords in hiragana ...... 139 4.1.10 Inconsistencies in transcription ...... 141 4.1.11 Summary of layer 1: the characteristics of loanwords in hiragana ...... 144 4.2 Layer 2: Loanwords in hiragana in multimodal texts ...... 145 4.2.1 Case 1: ‘Happiness’ Clinic...... 145 4.2.2 Case 2: ‘Honey Lemon’ Drink ...... 150 4.2.3 Case 3: ‘I can fly!!’ sticker ...... 155 4.2.4 Case 4: ‘G’mornin’’ tweet ...... 159 4.2.5 Summary of layer 2: loanwords in hiragana in multimodal texts ...... 163

Chapter 5: Findings from the survey ...... 164 5.1 Respondent background ...... 164 5.2 Script associations ...... 165 5.2.1 Hiragana ...... 165 5.2.2 Katakana ...... 168 5.2.3 Kanji ...... 171 5.2.4 Romaji ...... 173 5.3 Responses to sample texts ...... 175 5.3.1 Appropriateness by example ...... 175 5.3.2 ‘Happiness’ clinic ...... 177 5.3.3 ‘Honey Lemon’ drink ...... 183 5.3.4 ‘I can fly’ sticker ...... 190 5.3.5 ‘Dairy Coffee’ instant coffee ...... 197 5.3.6 ‘Cute Hankachi’ handkerchief ...... 203 5.4 Summary of layer 3: survey responses to scripts and sample texts ...... 209

Chapter 6: Findings from the Interviews and Focus Groups ...... 213 6.1 Participants’ general impressions ...... 214 6.2 Motivations for the use of hiragana for loanwords ...... 216 6.2.1 Unfamiliarity with English spelling ...... 217 6.2.2 Accidental use of hiragana for loanwords ...... 218 6.2.3 Balance and layout of text...... 219 6.2.4 Obscuring word boundaries ...... 221 6.2.5 Expressing ‘un-realness’ ...... 223 6.2.6 Expressing Japanese-ness ...... 225 6.2.7 Expressing familiarity ...... 226 6.2.8 Representing voice quality and pronunciation ...... 229 6.2.9 Highlighting word play ...... 231 6.3 Perceptions of age and gender ...... 232 6.4 Influence of typography ...... 237 6.5 Reactions to inconsistencies ...... 239 6.6 Alternative transcriptions of ‘I can fly’ ...... 244 6.7 Suitability of particular loanwords in hiragana ...... 247 6.8 Fluidity of script use ...... 250 6.9 Summary of layer 4: interpretations of loanwords in hiragana ...... 255

Chapter 7 Discussion ...... 256 7.1 ‘Not wrong’: the acceptability of loanwords in hiragana ...... 257 7.2 Trans-orthography...... 260 7.3 The functions of loanwords in hiragana...... 264 7.3.1 Socio-historical associations of hiragana ...... 265 Japanese-ness ...... 265 Femininity ...... 267 Familiarity ...... 268 7.3.2 Interpreting loanwords in hiragana in context ...... 269 Emphasis ...... 269 Unseriousness and Un-realness...... 271 Creating or obscuring word boundaries ...... 274 Marking language play ...... 275 Pronunciation and script choice ...... 277 7.4 Script as a mode ...... 280 7.5 The sedimentation of script choice ...... 281 7.6 Summary of the discussion chapter ...... 285

Chapter 8 Conclusion ...... 287 8.1 Theoretical implications...... 289 8.2 Critical reflections and suggestions for further research ...... 291 8.3 Epilogue: A return to the Japanese house ...... 293

References ...... 295 Appendices………………………………………………………………………………….306

Index of tables

Table 2.1: Associations of the Japanese scripts ...... 25 Table 2.2: History of language contact in Japan, based on Loveday (1996) ...... 30 Table 3.1: Summary of research design ...... 72 Table 3.2: Corpus analysis in the context of research design ...... 73 Table 3.3: Summary of corpus parameters ...... 79 Table 3.4: Coding categories for the corpus ...... 81 Table 3.5: Case studies in the context of the research design ...... 87 Table 3.6: Texts used as case studies ...... 90 Table 3.7: Survey in the context of the research design...... 96 Table 3.8: Sample responses of associations of hiragana script ...... 103 Table 3.9: Sample coding for associations of hiragana ...... 104 Table 3.10: Sample responses of reasons for appropriacy ...... 105 Table 3.11: Sample coding for reasons of appropriacy ...... 106 Table 3.12: Interviews and focus groups in the context of the research design ...... 108 Table 3.13: Summary of interview participants ...... 110 Table 3.14: Summary of focus group participants ...... 111 Table 3.15: Structure of the interviews and focus groups ...... 113 Table 4.1: Word frequencies ...... 121 Table 4.2: Grammatical class of loanwords in hiragana ...... 123 Table 4.3: Vowel lengthening by method and vowel ...... 124 Table 4.4: Referential meanings of loanwords in hiragana ...... 126 Table 4.5: Product names by type ...... 127 Table 4.6: Business names by type ...... 129 Table 4.7: Source langauge of loanwords in hiragana ...... 131 Table 4.8: Types of language play containing loanwords in hiragana ...... 133 Table 4.9: Consecutive loanwords in hiragana ...... 136 Table 4.10: Units of hiragana loanword(s) per text ...... 138 Table 4.11: Genres of text containing loanwords in hiragana ...... 139 Table 4.12: Types of inconsistency in the corpus…...... …...…...... 141 Table 5.1: Response rates for script associations sections ...... 165 Table 5.2: Common associations of hiragana ...... 166 Table 5.3: Common associations of katakana...... 169 Table 5.4: Common associations of kanji ...... 172 Table 5.5: Common associations of rōmaji ...... 174 Table 5.6: Appropriacy by example...... 176 Table 5.7: Response rates for ‘Happiness’ clinic (Q13-15) ...... 177 Table 5.8: Reasons for appropriacy of 'happiness' in hiragana (Q14) ...... 178 Table 5.9: Reasons for inappropriacy of 'happiness' in hiragana (Q14) ...... 180 Table 5.10: Respondents’ perceived reasons for the use of hiragana for 'happiness' (Q15) .. 182 Table 5.11: Response rates for 'Honey Lemon' drink (Q16-18) ...... 183 Table 5.12: Reasons for appropriacy of 'lemon' in hiragana ...... 184 Table 5.13: Reasons for inappropriacy of 'lemon' in hiragana ...... 186 Table 5.14: Respondents’ perceived reasons for the use of hiragana for 'lemon' ...... 187 Table 5.15: Response rates for 'I can fly!!' (Q19-21) ...... 190 Table 5.16: Reasons for appropriacy of 'fly' in hiragana ...... 191 Table 5.17: Reasons for inappropriacy of 'fly' in hiragana ...... 193 Table 5.18: Respondents’ perceived reasons for the use of hiragana for 'fly' ...... 195 Table 5.19: Response rates for 'Dairy Coffee' instant coffee (Q22-24) ...... 197 Table 5.20: Reasons for appropriacy of 'coffee' in kanji ...... 198 Table 5.21: Reasons for inappropriacy of 'coffee' in kanji ...... 199 Table 5.22: Respondents’ perceived reasons for the use of kanji for 'coffee' ...... 201 Table 5.23: Response rates for 'Cute Hankachi' handkerchief (Q25-27) ...... 203 Table 5.24: Reasons for appropriacy of 'hankachi' in rōmaji ...... 204 Table 5.25: Reasons for inappropriacy of 'hankachi' in rōmaji ...... 206 Table 5.26: Respondents’ perceived reasons for the use of rōmaji for 'hankachi' ...... 208 Table 5.27: Summary of the perceived reasons for the use of hiragana for loanwords …....210

Index of figures

Figure 1.1: Examples of hiragana and katakana ...... 4 Figure 1.2: Sample Japanese sentence ...... 5 Figure 1.3: Script options for ‘kumo’...... 7 Figure 2.1: Language realising context, adapted from Eggins (2004, p.112) ...... 23 Figure 3.1: Language realising context, adapted from Eggins (2004, p.112) [duplicated] ..... 69 Figure 3.2: ‘Hello Kitty’ brand products ...... 77 Figure 3.3: POS text created by same agent ...... 78 Figure 3.4: POS text created by different agent ...... 78 Figure 3.5: ‘Sweet Dog’ website ...... 82 Figure 3.6: ‘Dairy Coffee’ instant coffee mix ...... 100 Figure 3.7: ‘Cute hankachi’ handkerchief ...... 100 Figure 3.8: Screenshot of online survey ...... 101 Figure 3.9: Picture cards used in interviews and focus groups ...... 112 Figure 4.1: Processed food/drink, ‘Red Wine’ ...... 128 Figure 4.2: Omiyage food, ‘Kyoto Green-tea Milk Caramels’ ...... 128 Figure 4.3: Fresh produce, ‘cabbages’ ...... 129 Figure 4.4: Business name, ‘Lion Dentistry’ ...... 130 Figure 4.5: Restaurant name, ‘Steak House Juju’ ...... 130 Figure 4.6: Hostess bar name, ‘Snack Endless’ ...... 130 Figure 4.7: Portuguese loanword, ‘pan’ ...... 131 Figure 4.8: Portuguese loanword, ‘castella’ ...... 131 Figure 4.9: French loanword, ‘Espoir’ ...... 132 Figure 4.10: Italian loanword, ‘Bravo’ ...... 132 Figure 4.11: Example of homonym, ‘Cider’ ...... 134 Figure 4.12: Example of lexical blend, ‘Koedarize’ ...... 134 Figure 4.13: Example of concealed word, ‘Happy Terrace’ ...... 134 Figure 4.14: Single loanword in hiragana, ‘Cheese-in patty’ ...... 136 Figure 4.15: Two consecutive words, ‘Potato House’ ...... 137 Figure 4.16: Three consecutive words, ‘Power of Smile’ ...... 137 Figure 4.17: Text containing two units, ‘Lemon Koritto’ ...... 138 Figure 4.18: Product packaging, ‘Alpha’ ...... 140 Figure 4.19: Business sign, ‘Chicken-Pork Kitchen’...... 140 Figure 4.20: Product sign, ‘Towelling hats’ ...... 140 Figure 4.21: Inconsistent label, ‘Strawberry milk mochi’ ...... 142 Figure 4.22: Inconsistent label, ‘Mushroom cabbage mix’ ...... 142 Figure 4.23: Inconsistent related texts, 'Milk' ...... 143 Figure 4.24: Inconsistent vowel elongation, 'date' ...... 143 Figure 4.25: Inconsistent vowel elongation, ‘Hop choux crème’ ...... 143 Figure 4.26: ‘Happiness’ clinic sign with English translation ...... 145 Figure 4.27: ‘Honey Lemon’ drink label with English translation ...... 151 Figure 4.28: ‘I can fly!!’ sticker with English translation ...... 156 Figure 4.29: 'G'mornin' ' tweet with English translation ...... 160 Figure 6.1: Social networking 'sticker', ‘Fight!!’ ...... 216 Figure 6.2: Manga cover, 'K-on!!' ...... 232 Figure 6.3: Typeface used for 'Ice cocoa' ...... 237 Figure 6.4: Typeface used for 'Tomatos' ...... 237 Figure 6.5: Lettering used for 'Happy Valentine' ...... 238 Figure 6.6: Typeface used for 'Floria' ...... 238 Figure 6.7: Typeface used for 'Lounge Chiaki' ...... 238 Figure 6.8:Typeface used for 'Unicorn' ...... 238 Figure 6.9: Typeface used for 'Spring' ...... 238 Figure 6.10: Inconsistent use of script in same text, ‘Spice’ ...... 239 Figure 6.11: Inconsistent use of script in related texts, tomato display ...... 240 Figure 6.12: Inconsistent signage, ‘Cabbage for salads’ ...... 241 Figure 6.13: Inconsistent signage, ‘Uji green-tea castella’ ...... 241 Figure 6.14: Inconsistent signage, cabbage display ...... 243 Figure 7.1: Script options for 'lemon'...... 282 Figure 7.2: Adjusted script options for 'lemon' ...... 283 Figure 7.3: Print club images containing the 'chīpopose'...... 284 Figure 8.1: Japanese style room with bed …………………………………………………..294

Chapter 1: Introduction

The Japanese language is written with a combination of three distinct scripts, the ideographic Chinese characters called kanji, and two phonetic scripts, hiragana and katakana. A typical sentence will usually include a combination of these scripts, with hiragana and kanji typically used to write words of Japanese origin, while words from foreign languages, called ‘外来語 /gairaigo’ or ‘loanwords’, are conventionally written in katakana. These do not include the many ‘loanwords’ of Chinese origin, or ‘漢語/kango’ which have been integrated into the Japanese language, a distinction further clarified in section 1.4, ‘Definition of loanword’ below.

While katakana serves as a straightforward way of importing foreign terms into the written Japanese language without the need for translation (Kay, 1995; Rebuck, 2002; Stanlaw, 2004), at the same time, it brands them orthographically as non-native words. Loveday (1996) notes the significance of this practice, arguing that it serves to encode Japanese perceptions of these introduced words, the katakana script ‘symbolising and emphasising the foreignness of the transferred item’ (1996, p.48). Kay (1995) also notes the role of script in segregating loanwords from the rest of the Japanese lexicon, describing the use of scripts as serving to compartmentalize the Japanese lexicon as neatly as the rooms of a Japanese house are divided into ‘Japanese style’ and ‘western style’. The relationship between loanwords and the katakana script conventionally used to write them is cemented in the term ‘カタカナ語 /katakana-go’, (katakana-words) which is used synonymously with ‘外来語/gairaigo’ or ‘loanwords’, as noted by Hosokawa (2015). These foreign terms are thus orthographically distinguished from other lexical items in the Japanese language, and as Irwin (2011, p 179) notes, ‘That loanwords are thus rendered so conspicuous on a page of printed matter, and in consequence their loanword status perpetually reinforced in the mind of the reader, is unique among the writing systems of the world.’

The topic of the current study, however, is loanwords that are written in hiragana, not the conventional katakana script. It therefore problematises the relationship between the katakana script and these foreign terms, including the idea that ‘loanwords’ and ‘ カタカナ語

/katakana-go’ are synonymous, by discussing examples of loanwords that are not visually marked and segregated through the katakana script as described above. The loss of such a significant orthographical marker of foreign heritage, it is argued, may be evidence of a sense of ‘owning’ rather than ‘loaning’ these once foreign words. If, as Kay suggests, katakana represents the ‘incomplete linguistic assimilation’ (1995, p. 73) of loanwords within the Japanese language, the shift to hiragana may represent a move toward acceptance of these foreign terms. The current research explores how the conventional associations of particular lexical items with particular scripts may be giving way to a more flexible use of the scripts as semiotic resources within the Japanese writing system.

1.1 Background

This section will provide a brief introduction to the Japanese writing system, and loanwords in Japanese, the two areas which form the foundation of the current study.

1.1.1 The Japanese writing system

This section will give an explanation of the historical development of the Japanese writing system, and the factors which have influenced the form it takes today. It will also give a brief description of Japanese literacy education in Japan, and of the effect of technology on written Japanese. This section will end with reference to the importance of the writing system within the national consciousness.

Japan had no native writing system before contact with China, through Korea, from the end of the 6th century (Loveday, 1996). Many other aspects of Chinese culture, technology and religion (Buddhism) were also introduced at this time. In adopting the writing system, the Chinese language was also imported, leading to a diglossic bilingual setting, where Chinese (漢語/kango) represented the High variety and was used for written texts such as official documents at court, and sutras and prayers in the monasteries, while native Japanese (和語 wago) was spoken by the general population. Loveday (1996) notes that it is difficult to estimate the extent of oral communication in Chinese, but believes it was likely to have occurred in formal, ceremonial speech events.

Gradually a system of ‘instantaneous translation’ of Chinese texts was developed, allowing written Chinese to be ‘read’ in Japanese and thus the characters to become associated with native Japanese words. The legacy of this practice can be seen in the two sets of readings that most kanji have today; an ‘音読み/on-yomi’ (on reading) from Chinese and a ‘訓読み /kun- yomi’ (kun reading) from Japanese.

As each kanji carried a semantic meaning, these characters could be combined to represent native Japanese words, for example the characters for ‘earth’ and ‘produce’ were combined to form the word ‘土産/miyage’ (souvenir), and local food specialities remain a typical souvenir today. These types of calques where characters were selected based on their meanings are called ‘熟字訓 /jukujikun’. At other times, characters were chosen not for their semantic content but their phonetic properties: the Chinese also used a set of phonetic kanji to spell out the names of foreign people and places, and this practice was adopted by the Japanese in the 8th century in order to write their own native words. For example the phonetic kanji representing the sounds ‘fu’ and ‘ro’ were combined to write the word ‘風呂/furo’ (bath); these types of kanji calques are called ‘当て字/ateji’ (Kess & Miyamoto, 1999).

The phonetic kanji were often the basis of the syllabic scripts, hiragana and katakana. Although many are simplified versions of the same kanji, they were simplified in different ways and for different purposes. Hiragana is sometimes referred to as the ‘cursive’ Japanese syllabary, as the characters represent their phonetic kanji counterparts in a connected or simplified form, as they might appear when being written quickly with a brush and ink. The phonetic kanji for the sound ‘ka’ (加) became ‘か’ in hiragana, the square on the right hand side of the original kanji being reduced to a dash in the hiragana character. The ‘hira’ in ‘hiragana’ means ‘ordinary’, or ‘common’, as this script was considered a writing system for general use or non-government documents (Shibatani, 1990; Akizuki 2005). Due to its non- official status, hiragana was enthusiastically taken up by literary-minded women, who were discouraged from learning the more complicated kanji characters which were associated with official documents and other sources of male power (Shibatani, 1990). Hiragana was therefore used for private letters, poetry and other secular texts (Kess & Miyamoto, 1999; Akizuki, 2005), and the script itself is sometimes referred to as ‘女手/onna de’ (‘women’s hand’) (Ishikawa, 2007).

Around the same time, katakana was developed in monasteries where it was used for annotating Buddhist scriptures in order to aid the monks in pronouncing the Chinese characters (Akizuki, 2005). The ‘kata’ in katakana means ‘part’ or ‘fragment’, and these characters represent part of the original kanji, for example katakana ‘ka’ (カ) represents the left half of the original phonetic kanji for ‘ka’ (加). Both hiragana and katakana were widely used by the upper classes from the 10th century, and were important in allowing the Japanese language to be represented without resorting to Chinese characters (Loveday, 1996).

Today, written Japanese consists of a combination of all three scripts: kanji, the ideographic characters adopted from Chinese1; and the two syllabic scripts, hiragana and katakana (figure 1.1 below). The Roman alphabet is often considered a fourth script used in Japan, and is referred to as ‘ローマ字/rōmaji’ (roman letters), and Arabic numerals are also used.

Roman alphabet A I U E O Hiragana あいうえお

Katakana アイウエオ

Figure 1.1: Examples of hiragana and katakana

All three scripts can be found in a single sentence, as each script is associated with different words or functions. Kanji are usually used for the lexical morphemes or ‘content words’ of a sentence, such as nouns and verb stems, and are associated with ‘漢語/kango’.2 Hiragana is usually used for the inflexive endings of words which have a stem written in kanji, such as the tenses of verbs, as well as connective particles, and some words of Japanese origin that are not assigned a kanji, which are termed ‘和語/wago’. Katakana is usually used for foreign words as noted above, and is therefore associated with ‘外来語/gairaigo’ (loanwords) or ‘外国語 gaikokugo’ (foreign words). Katakana is also used as a form of emphasis for Japanese words, sometimes compared to how italics are used in English, and is also used for onomatopoeic expressions, and plant and animal names. These conventions are reflected in Japan’s current language policies, developed by the governmental National Language

1 There are a small number of such characters developed in Japan, and referred to as ‘国字/kokuji’ (national- letters) or ‘和製漢字/wasei kanji’ (Made-in-Japan kanji). See Buck (1969) for more details of such characters. 2 It is noted, however, that not all words which contain a kanji are 漢語/kango. Kango do not include the stems of verbs and adjectives which inflect, only those verbs created by the addition of ‘する/suru’, and ‘な/na’ adjectives. 4

Council, and any changes to this involve a complex process of research and debate, as well as community discussion.

The sentence in figure 1.2 below shows how hiragana, katakana, and kanji fulfil their functional roles in the following simple sentence, ‘My name is Hannah’. Kanji is used for the words ‘ 私 /watashi’ (I, myself), and ‘ 名前/namae’ (name); katakana for the phoetic approximation of the foreign name ‘ハナ/hana’ (Hannah); and hiragana for the possessive particle ‘の/no’, subject marker ‘は/wa’, and the final verb ‘です/desu’ (to be). 3

私の名前はハナです。

watashi no namae wa hana desu. My name is Hannah.

Figure 1.2: Sample Japanese sentence (Kanji hiragana katakana) While languages such as English use spaces to separate words, the allocation of particular lexical items to particular scripts serves a similar function in Japanese. For example, many kanji-based or katakana lexical items are separated by hiragana particles, as in the above example.

In the Japanese education system, hiragana is the first script learned by children, and therefore books aimed at very young children are an exception to these script use conventions, being written completely in hiragana, with spaces to indicate word boundaries. Loanwords in such books are also written in hiragana, or at least glossed with these

ばなな characters, for example ‘バナナ/banana’ (banana). Katakana is also typically learned in the first year of school, and from the time this script is introduced, children are expected to differentiate loanwords from other Japanese words by writing them in katakana, reflecting the government policy titled ‘外来語の表記/gairaigo no hyōki’ (the transcription of loanwords) established by the Ministry for Education, Culture, Sports, Science and Technology (‘文部科

3 The use of multiple scripts within the one sentence or even phrase therefore separates the Japanese writing situation from other cases of digraphia, which are distinguished along national or cultural lines. These include the use of Cyrillic script for Serbian, and Latin script for Croatian; and Devanagari script for Hindi, and Perso- Arabic script for Urdu. 5

学省/mombukagakushō’, or MEXT), (Ministry of Education, Culture, Sports, Science and Technology, 1991).

The more complex kanji characters are introduced gradually over the years of compulsory schooling. The Japanese Ministry of Education specifies a list of 1,006 kanji to be learned at primary school, with an additional 1,130 characters for secondary schooling, which together form the ‘常用漢字/jyōyō kanji’ (regular-use kanji). Official government documents are limited to these kanji to ensure accessibility to all those who complete compulsory schooling,4 and when specialized terms that fall outside this set occur they are required to be

きりん glossed with hiragana, for example ‘麒麟/kirin’ (giraffe).

While MEXT’s policies and the associations of hiragana, katakana, and kanji with differnent lexical items suggests little deviation in the way Japanese words can be written, the presence of multiple scripts does provide fertile ground for variety. For example, Joyce, Hodošček, & Nishina (2012) list five orthographic variants for the word ‘tamanegi’ (onion) in the ‘Balanced Corpus of Contemporary Written Japanese’ (BCCWJ), which utilize kanji, katakana, and hiragana, as well as combinations of the three Japanese scripts5 (2012; p. 271). Even within the kanji script, variation can occur with homonymous near-synonyms, for example while the verb ‘to meet’ is generally written as ‘会う/au’; Joyce et al. note orthographic variations including ‘ 遇う/au’ (to meet, with a nuance of an unexpected meeting); and ‘逢う/au’ (to meet, the a nuance of a date or tryst). 6 The morphological analyser dictionary developed for the BCCWJ, ‘UniDic’, described by Joyce et al. (2012) lists 198,937 word forms, for which there are 318,690 orthographic types, suggesting that script variation is an integral part of written Japanese.

The resurgence of unusual or more complex kanji, such as as the variants for the verb ‘to meet’ described above, is often believed to be related to the advent of Japanese word processing software (Gottlieb, 1995; Sasahara, 2002). Japanese text is entered via a ‘qwerty’

4 Students are taught to read these characters at ‘中学校/chūgakkō’ (junior high school), and are expected to be able to write them by the time they graduate from ‘高等学校/kōtōgakkō’ (senior high school). Compulsory schooling only includes junior high school, so even the few who finish their formal education at that stage are theoretically able to read these characters. 5 In descending order of frequency, these were 玉ねぎ, タマネギ, たまねぎ, 玉葱, and 玉ネギ. 6 Homonymous near-synonyms are further discussed in chapter 3, methodology and research design, section 3.4.3., as they appeared in some responses to the survey. 6

keyboard by typing the words in rōmaji, for example in order to enter the word ‘cloud’ (kumo), one enters the letters K-U-M-O. These initially appear on the screen in hiragana, and by pressing the space bar, different orthographic options appear, including hiragana (くも /kumo) and katakana (クモ/kumo), the kanji for cloud (雲/kumo), and the kanji for the homonymic word ‘spider’ (蜘蛛/kumo),7 from which the user can select the characters they require (figure 1.3 below).

Figure 1.3: Script options for 'kumo' Word-processing software on computers and mobile phones has therefore given users easy access to kanji they might not be able to write by hand, including obsolete or archaic forms, and those outside the prescribed kanji learned in school. In this way, technology has therefore served to increase the range of characters in popular use, and represents an important development in the Japanese writing system.

Despite, or perhaps because of, the complexity of written Japanese, it is often closely tied to notions of national identity, and the Japanese language is often cited in the pseudo-scientific discipline of ‘日本人論/nihonjinron’ (theories of Japanese-ness) as being the vehicle for, and ultimate embodiment of, the Japanese spirit (Dale, 1986; R. A. Miller, 1982). However, critics of ‘日本人論/nihonjinron’ are quick to point to the irony of the fact that the Japanese writing system was originally adopted from China, as explained earlier in this section, and therefore perhaps not the best representation of Japan’s uniqueness (Dale, 1986; Gottlieb, 2000). Nevertheless, the writing system, and kanji in particular, is felt to be a cultural icon in Japan (Gottlieb, 2010b), and holds a particularly important place in the national

7Homonyms are called ‘同音異義語 /dōonigigo’ in Japanese, translating as ‘same sound, different meaning’ (Gottlieb, 1995) 7

consciousness. For this reason, orthographic variation, especially that surrounding foreign lexical items such as loanwords, is an important topic of investigation.

1.1.2 Loanwords in Japanese

Loanwords play a significant role in the Japanese language, and are often estimated to account for around 10% of the Japanese lexicon (Honna, 1995; Loveday, 1996). Irwin (2011) reviews a number of studies which estimate the proportion of loanwords in a diverse range of spoken and written genres from various time periods, with figures closer to 5%. However, even this proportion of the language as a whole is significant, and the numbers of loanwords in Japanese are generally felt to be increasing (Hosokawa, 2015; Igarashi, 2007). In the early years of this millennium, Rebuck (2002) noted that the number of entries in loanword dictionaries continued to rise, citing figures from 1967, up to the 2000 edition of Sanseido’s ‘Concise dictionary of foreign words’ which contained 45,000 loanwords; the 2010 edition contains 56,300. Loanwords cover a wide range of domains, including fashion (for example, スカート/sukāto [skirt]); sport (テニス/tenisu [tennis]); theoretical concepts (インターナショナライゼーション/intanashonaraizēshon [internationalization]); and even kinship terms (ママ and パパ/ mama and papa). While most loanwords used in Japan have English as their source language, other languages with historical contact with Japan are also represented in the loanword lexicon, including ‘ パン/pan’, the word for bread originating from the Portuguese ‘pão’ and ‘ゴム/gomu’ (rubber, elastic) which comes from the Dutch ‘gom’.

Japan has a long history of importing and adapting foreign terms into the language, starting with words from Chinese which entered the language along with the adoption of the Chinese writing system in the 6th century (see section 1.1.1). From the late 15th century, missionaries from Portugal brought their language along with their teachings to Japan, and during the period of national seclusion (鎖国/sakoku) from 1639, the limited contact with Dutch traders allowed a small number of Dutch words to enter the language. When Japan was finally reopened to trade relations in 1853, a period of drastic modernization occurred and words were needed to describe the new technologies and theoretical concepts introduced from the west. Following Japan’s defeat in World War II, another influx of English words entered the Japanese vocabulary, and the trend of adopting loanwords, primarily from English, continues to this day. The history of language contact in Japan is described in more detail in section 2.4.1. 8

1.2 Rationale for the current study

Because of the close association between loanwords and the katakana script, embodied by the term ‘カタカナ語/katakana go’ (katakana words) being used synonymously with the term ‘ 外来語/gairaigo’ (loanwords), the presence of loanwords in hiragana, a marked orthographical choice, are a significant linguistic and cultural phenomenon worthy of exploration. To date, no research has investigated contemporary use of hiragana for loanwords, and no mentions of this practice are found in the literature on the Japanese language.8

The result of this is that research on the Japanese writing system, such as Smith and Schmidt’s (1996) promisingly titled ‘Variability in written Japanese: Towards a sociolinguistics of script choice’, inevitably overlooks this phenomenon. In their investigation on the relationship between script and lexicon, the authors simply state that loanwords are ‘always encoded in katakana, hence, not subject to writer choice’ (Smith & Schmidt, 1996, p. 64). The current research suggests that the script used for loanwords, like other words in the Japanese language, is subject to writer choice, and therefore an important semiotic resource within the Japanese writing system.

The use of hiragana for loanwords therefore represents an example of how the Japanese are ‘experimenting with the orthographic and visual aspects of their language and writing systems’ (Stanlaw, 2004, p. 33). Stanlaw is speaking of the presence of English in Japan in katakana and rōmaji, and specifically draws parallels between this experimentation with English loanwords and the importation of the Chinese language in the 5th century. Words of Chinese origin (漢語/kango) and the kanji characters themselves are now considered integral parts of the Japanese language, but are also, of course, not native to Japan. It is argued that writing loanwords in hiragana, the script conventionally associated with Japanese words, may represent a similar move towards ownership of these once ‘foreign’ words.

Variation in script use, as in other aspects of language, carries the potential for meaning making, as the following quotation from Sebba (2009) illustrates:

8 Examples of the use of hiragana for selected early loanwords, for example, ‘ たばこ/tabako’ (tobacco, cigarettes), do occur (Irwin, 2011; Loveday, 1996). 9

From this potential for variation, (licensed or unlicensed), comes the notion of writing systems as a social practice: a recurrent activity involving meaningful choices (Sebba 2007). At first this notion may seem paradoxical, when so much to do with writing is standardized and constrained; but…there are in fact many points within writing systems where variation can occur, and where there is variation, there is in practice always social meaning. (Sebba, 2009, p. 39)

Investigating the social meanings attached to the marked use of hiragana script for loanwords is one of the central aims of this study. As Spitzmüller (2015) echoing Blommaert (2013) warns, linguistic landscape studies and others that take photographs of signs as their primary objects of analysis fail to demonstrate what social meanings are attached to those signs by those who encounter them. In a similar vein, corpus studies of variation in written Japanese such as that conducted by Joyce et al. (2012) are valuable in describing the breadth of script variation across large numbers of texts, but as the authors note, they ‘have not touched on the complex sociolinguistic factors that might influence the selection of one orthographic representation over another when writing in Japanese…’ (2012, p. 274). It is these motivating factors and nuances of script choice that this study aims to understand and describe.

The meaning-making potential of a loanword in hiragana is understood as social; mediated by the readers of these texts, and it follows that Saussurean terms used to describe signs such as ‘signifier’ and ‘signified’ are problematic, as they suggest that meaning is simply encoded in the sign and is therefore the same in all contexts and for all individuals. Discussing blackletter typeface as a semiotic resource, Spitzmüller (2012) describes graphic ideologies as ‘both discursively rooted and floating,…both part of a collective knowledge and permanently subject to negotiation’ (2012, p.256), and it is argued that this is also an apt description of the role of script in the Japanese writing system. Hiragana has conventional associations of gentleness and femininity, for example, which forms part of the collective Japanese knowledge of script use, however these associations are subject to negotiation in different contexts and by different readers of those signs. The interaction between the established connotations of the scripts and the specific multimodal contexts in which they are deployed forms a central theme of the current research. 10

The conventional association of loanwords with the katakana script makes the case of loanwords in hiragana a subject worthy of investigation. Loanwords are particularly rich translingual resources, being non-native in their origins and yet firmy integrated into the Japanese language. Loanwords in hiragana therefore provide a useful lens through which to understand issues of cultural ownership of these terms, suggesting the relative place of loanwords on the continuum between ‘native’ and ‘foreign’, as well as variation between particular loanwords. In addition, script as semiotic resource is seen to be used flexibly and purposefully by native speakers to perform a variety of functions, which will be explored in the following chapters.

1.3 Aim and Research questions

The aim of the current research is to understand the role played by loanwords in hiragana as translingual semiotic resources. Social semiotics underpins the theoretical and analytical frameworks developed for this study, as such signs do not have designated ‘meanings’, they are dynamic and context-dependent, as well as being negotiated by the readers themselves. Related to this is the concept of multimodality, which acknowledges the role of other semiotic resources which co-occur within these texts, for example colour, typeface, and image, in the meanings suggested by loanwords in hiragana. Finally, systemic functional linguistics (SFL) is also drawn upon to illustrate the kinds of meanings that can be ascribed to these resources; the ideational, interpersonal, and textual metafunctions. These approaches are described in more detail in the literature review chapter, sections 2.1 (Social semiotics); 2.2 (Multimodality), and 2.3 (Systemic functional linguistics).

In order to address these aims the following research questions are proposed:

1. How can loanwords in hiragana be characterised?

2. In what contexts are loanwords in hiragana usually found?

3. Why do loanwords appear in hiragana?

A four-layer research design has been utilized in order to approach the topic of loanwords in hiragana from various angles, as well as allowing for the triangulation of results between these layers. The first layer involves descriptive statistics of a corpus of texts containing hiragana loanwords collected by the researcher, in order to illustrate the general characteristics of these words and the contexts in which they are found. The second layer of analysis takes selected examples from this corpus as case studies, analysing them multimodally in order to highlight how the meanings ascribed to the use of hiragana for loanwords are supported and sometimes challenged by the other semiotic resources in the text. The third layer seeks the opinions of native Japanese speakers on loanwords in hiragana through an online survey. The survey collects data on the associations or connotations of each of the Japanese scripts, as well as asking respondents to engage with the texts used as case studies in layer two. The final part of the research design, layer four, consists of interviews and focus groups, which expand on the survey by allowing participants to discuss examples of loanwords in hiragana in more depth.

The current research therefore aims to contribute to the field of social semiotics by describing how loanwords in hiragana function as semiotic resources. It also adds to our understanding of how texts function multimodally, by investigating how Japanese script interacts with the other modes used in a text. Finally, the current project highlights a novel aspect of how loanwords operate as translingual resources, being re-purposed in different cultural contexts in the orthographic repertoires of their users.

1.4 Definition of ‘loanword’

The word ‘外来語/gairaigo’ (loanword) itself is telling in its etymology, translating literally as ‘outside-coming-word’. 9 Loveday (1996) notes the existence of alternative academic terminology, such as ‘移入語/inyūgo’ (imported words), and ‘借用語/shakuyōgo’ (loaned words), but admits these are not in common use. The term ‘外来語/gairaigo’ is well established in the Japanese language, and is usually translated as ‘loanword’ in English.

9 Hosokawa (2015) notes the parallel between this term and ‘外来種/gairaishu’, (outside-coming-species), used to describe introduced invasive plants in the ecosystem. 12

However, the term ‘loanword’ is problematic in many ways. Firstly, the metaphor that these words are ‘loaned’ implies some sense that they are freely given for a finite period of time and will someday be returned to their original owner, which is clearly not the case. For this and other reasons, Stanlaw objects to this term, calling the metaphor of loaning or borrowing ‘misleading’ and ‘problematic’ (Stanlaw, 2004, p. 33). He cites the many cases of semantic shift, narrowing and expansion these words undergo in the Japanese context, as well as the issue of ‘和製英語/wasei-eigo’ (‘made-in-Japan English’) to support his objection. ‘Made-in- Japan English’ terms are those constructed from English elements, but which are not usually understood outside Japan, such as ‘O.L’, which refers to an ‘office lady’, or female office clerk.10 As an alternative to ‘loanword’, Stanlaw suggests that ‘English-inspired vocabulary item’ more accurately describes these words. While the sentiment which it is expressed is acknowledged, its unwieldy nature, as well as its failing to acknowledge loanwords from non-English languages, makes it unsuitable for use in this study.

Therefore, despite the issues surrounding the Japanese term, ‘外来語/gairaigo’, and its translation as ‘loanword’ in English, both of which are acknowledged as significant, these terms will be used throughout the thesis. They are widely used and understood, and therefore provide transparency of meaning, as well as connecting this research to the greater body of academic research on ‘外来語/gairaigo’ and loanwords.

The scope of the term ‘loanword,’ as it is used in this thesis, will be understood as describing words contained in a Japanese loanword or ‘katakana-go’ dictionary, in this case, ‘Sanseidō's concise dictionary of katakana words’ (Sanseidō Henshūjo, 2010). This includes place and country names, as well as morphological elements such as ‘-ize’. The definition will further be expanded to include phrases such as ‘yes I am’, and words containing recognizable foreign elements, such as the name of the florist ‘ふろーりあ/furōria’ (Floria), from the Latin-based ‘flora’ and nominal suffix ‘-ia’.

The scope of this definition will further be narrowed in two ways. Firstly, loanwords of Chinese origin will be excluded, as the boundaries of what can be considered a ‘loanword’, as distinct from a ‘Sino-Japanese word’, is particularly fluid in the case of loanwords in hiragana. For example, food terms such as ‘ramen’ are so well established in Japanese

10 These cases are described in detail in section 2.5.4, perceptions of loanwords. 13

language and culture, they are frequently seen in both katakana (ラーメン) and hiragana (ら

ーめん).11 While this is, it is acknowledged, an important element of the phenomenon being described, it is felt that including such words would be outside the scope of this study, which looks at words more recently borrowed into the Japanese language. The transition of Chinese loanwords into Sino-Japanese words is noted as a topic worthy of future study (see section 8.2).

A further type of text is excluded from the scope of this study; texts aimed at young children unfamiliar with the katakana script. As noted above, loanwords are conventionally written in hiragana, or glossed with this script, in books and signage (for example, at a kindergarten) associated with young children. While these examples do, in a sense, contribute to understanding why loanwords are written in hiragana, this is a well-established function, and does therefore not offer any new contributions to the current investigation of loanwords in hiragana.

1.5 The use of Japanese in the thesis

Japanese words are represented in Japanese characters, accompanied by a transliteration (修正ヘボン式 /shūsei hebon-shiki’ [modified Hepburn romanization]), and English translation, for example ‘日本/nihon’ (Japan). However, longer phrases and sentences in Japanese are simply be presented as an English translation, with the original Japanese appended as a footnote. The captions of tables and figures are presented in English only for the sake of brevity.

Loanwords in hiragana, the central topic of this thesis, are further identified by being underlined, allowing readers unfamiliar with the Japanese scripts to recognize them, especially in phrases and sentences. For example, the product name ‘あいすココア/aisu kokoa’ (Ice Cocoa) contains two loanwords, where only the first, ‘あいす/aisu’ (ice), has been written in hiragana. Underlining is not used when hiragana loanwords occur as captions to figures and tables, and within tables themselves, to ensure clarity for the reader.

11 Cases such as these are also difficult for Japanese people to define as ‘loanwords’ or ‘Sino-Japanese words’. A column in the Yomiuri Newspaper was dedicated to the explanation of the appropriate script for the word ‘gyōza’ for this reason (Sekine, 2014). 14

1.6 Overview

The thesis contains eight chapters. The first has provided the relevant background to the topic and stated the rationale and aim of the study, as well as defining the scope of the term ‘loanword’. Chapter two contains a review of the relevant literature, beginning with an explanation of social semiotics, which forms the basis of the theoretical background of the study, and the related fields of multimodality and systemic functional linguistics. The chapter will then review the literature on loanwords in Japanese, as well as other related topics such as non-standard orthography, the role of typography, and finally the translingual flows of language through which loanwords in hiragana can be understood. The third chapter describes the methodological approach and research design, explaining the data collection and analysis techniques for each of the four layers of the research project; namely, the statistical corpus analysis, case studies, survey, and interviews/focus groups. Following this, the findings are presented in three separate chapters; chapter four presents the findings from the statistical corpus analysis and case studies, chapter five discusses the findings from the survey, and finally chapter six contains those from the interviews and focus groups. The findings from all four layers are brought together in chapter seven, the discussion. Finally, chapter eight will present conclusions drawn from the findings in relation to the research questions, and describe the academic contributions made by this project. The thesis concludes with critical reflections on the current study and suggestions for further research.

Chapter 2: Literature Review

This chapter will critically review the literature that forms the basis for the current study. Since this project is the first to investigate the marked use of hiragana for loanwords, it is necessary to engage with a range of diverse fields in order to interpret this hitherto undescribed phenomenon.

First, social semiotics, the theoretical framework which forms the basis of the current research design, will be reported on, along with the related fields of multimodality and systemic functional linguistics (SFL). The chapter moves on to investigate the socio- historical associations of each of the Japanese scripts, in order to provide context for the shift from katakana to hiragana. Next, the literature on loanwords in Japanese will be analysed, noting the history of language contact in Japan, as well as the functions and perceptions of loanwords today. A review of other non-standard orthographical practices and their functions will also be presented, followed by the specific example of non-standard orthography for the purposes of language play. Relevant literature on typography will also be introduced in this chapter, as the choice of script in Japanese has many parallels to the choice of font or typeface. The chapter will end with a brief discussion of translingualism, as loanwords are prime examples of the flows of language across the globe.

2.1 Social semiotics

Social semiotics is based on Halliday’s ‘Language as a social semiotic’ (1978), but scholars in this field look beyond language to investigate which other resources can be described as ‘social semiotics’, for example styles of dress, or the floorplans of offices. As a derivation of semiotics, this discipline also has as its heritage the work of Ferdinand Saussure.

Saussure established the study of what he called ‘semiology’, describing the ‘sign’ as a relationship between a signifier (form) and a meaning (signified), representing a kind of code that users could draw on to express meaning (Saussure, 2015). Saussure’s contribution to establishing semiotics as a field of academic inquiry, while significant, is limited in that it does not acknowledge how language is shaped by its users, nor how readers or receivers of signs can be ‘agentive’ and ‘transformative’ in their readings of 16

signs (Kress, 2010). Saussure (2015) also gave little weight to how context could affect both the choice of particular signifiers, and how these might be interpreted, arguing that signs could be understood without their surrounding context.

Hodge and Kress (1988) were two of the early proponents of social semiotics, and argued for an alternative model of semiotics to that proposed by Saussure. They aimed to acknowledge the social processes and effect of context involved in meaning-making, and extended Halliday’s metafunctions to areas other than language as mentioned above. Social semiotics is therefore a post-structuralist approach, and considers language and other sign systems to be constantly evolving, and influenced by individual users. This is a key difference to Saussure, who, as a structuralist, saw semiosis as a stable system kept in place by the conventional use of these signs by society. In social semiotics, forms and meanings are not organized in an abstract series of relationships waiting to be drawn upon by individuals; rather, signs are created by individuals in and for particular social contexts, and therefore situated in time and place:

Traditional semiotics likes to assume that the relevant meanings are frozen and fixed in the text itself, to be extracted and decoded by the analyst by reference to a coding system that is impersonal and neutral, and universal for users of the code. Social semiotics cannot assume that texts produce exactly the meanings and effects that their authors hope for: it is precisely the struggles and their uncertain outcomes that must be studied at the level of social action, and their effects in the production of meaning. (Hodge and Kress, 1988, p. 12)

The above quotation clearly distances social semiotics from its structuralist beginnings, and this distancing is also visible in the terminology chosen by some social semioticians. For example, van Leeuwen (2005) expresses a preference for the term ‘semiotic resource’ over Saussure’s ‘signifier’; and ‘semiotic potential’ over ‘signified’, arguing that the use of Saussurean terms implies that ‘what a sign “stands for” is somehow pre-given, and not affected by its use’ (2005, p. 3). Kress (2010), on the other hand, does not seem to think that a break with Saussurean terminology is necessary to establish social semiotics as a separate field, as he uses the term ‘sign’ extensively, describing it as ‘the core unit of semiotics’, and

also the terms ‘signifier’ and ‘signified’. Kress is also working under the same post- structuralist, social semiotic framework as van Leeuwen, and his use of these terms simply encompasses this approach, for example, when Kress uses the word ‘sign’, he is not referring to pre-determined structuralist sign, but one that is created by social actors for a particular context. In this thesis, where the terms ‘sign’, ‘signifier’ and ‘signified’ are used, they are to be understood as Kress (2010) has used them, as part of a post-structuralist social semiotics framework. The terms ‘form’ and ‘meaning’ are also used within this social semiotics approach, following Kress and van Leeuwen (2006, p. 6).

When signs are understood as being created by individuals, they can be seen as ‘motivated’, marking another break with Saussure who stressed the arbitrariness of combinations of signifiers and signifieds (Kress, 2010). In social semiotics, the combination of form and meaning is not random, but represents a process where a sign maker ‘brings meaning into an apt conjunction with a form, a selection/choice shaped by the sign-maker’s interest’ (Kress, 2010). In other words, forms and meanings are relatively independent until they are brought together by a sign-maker (Kress & Leeuwen, 2006, p. 8). Since signs are motivated, it is possible to ask what influences a sign-maker to combine particular forms with particular meanings. The key influences are the ‘interest’ of the sign maker, and the social context. The ‘interest’ of the sign maker (Kress, 2011, p. 62) refers to what a sign maker considers criterial at the moment of bringing together a signifier and signified.

While the interest of the sign-maker contributes to the choice of which form is selected for combination with which particular meaning, in a social semiotic framework the context in which the sign making occurs is also a major influence. As signs are created by individual at the moment of sign-making, they are intrinsically linked to a particular time, place and communicative event. As Jewitt (2009) explains, it is the situated use of semiotic resources which allows us to investigate…

…the non-arbitrary and motivated character of the relationship between language and social context... The context shapes the resources available for meaning-making and how these are selected and designed. (Jewitt 2009, p.30)

A socially-situated view of meaning-making allows us to consider which semiotic resources are available to a sign-maker in a given context, as the use of semiotic resources by an individual may be restrained by rules or best practices concerning how they are used (van Leeuwen, 2005). These ‘rules’ are the product of convention; while we have at our disposal the culturally produced semiotic resources of the societies we live in, these come with the dialogic histories of their use which may limit or constrain our use of them. Kress and van Leeuwen argue that while convention does not negate new meaning making, it does ‘attempt to limit and constrain the semiotic scope of the combinations’ (Kress & Leeuwen, 2006, p. 12). The conventional use of semiotic resources leads to what Pennycook (2010, p. 47) calls the ‘sedimentation’ of particular language forms, where repeated use over time becomes habitual and resists further change. The extent to which certain loanwords in hiragana are showing signs of sedimentation is discussed in section 7.5.

As mentioned above, social semiotics takes Halliday’s ‘Language as a social semiotic’ as a starting point and asks what other resources can be considered ‘social semiotics’. Kress and van Leeuwen’s (2006) best known contribution in this area is their analysis of multimodal texts, often containing both images and written words. In their seminal work, ‘Reading images: The grammar of visual design’ 12 (2006), Kress and van Leeuwen argue that individuals ‘read’ multimodal texts using a visual ‘grammar’ to interpret the meanings of and connections between the various elements represented. 13 Semiotic resources included in Kress and van Leeuwen’s ‘visual grammar’ include the vectors in images and the gaze of the represented participants; the salience of the elements in the text realized through colour, proximity, or focus; framing devices used to separate or associate elements within the text; and the placement within ‘information zones’ (top/bottom, left/right, centre/periphery).

While Spitzmüller (2012, p. 258) is skeptical of attempts to describe a ‘context-abstract “grammar” of visual elements’, arguing that the dynamic nature of these signs prevents such attempts succeeding, it appears that Kress and van Leeuwen are well aware of the role context plays in the messages suggested by various elements in the text. The importance of context in interpreting meaning is central to social semiotics, and also plays a role in

12 The authors admitted that they hesitated over the choice of the word ‘grammar’ in the title as it suggests a structure or set of rules, but decided that it best expressed how elements combine into meaningful wholes. (2006, p.1). However, writers such as Machin (2011) still question the usefulness of applying this metaphor to images. 13 Walker (2000) notes the work of Kress and van Leeuwen as being among the few whose research integrates the fields of linguistics and typography. 19

multimodality, described in the following section (2.2). The visual grammar proposed by Kress and van Leeuwen will therefore be drawn upon in interpreting the case studies (layer 2) to investigate how loanwords in hiragana are integrated into their multimodal texts. A detailed explanation of Kress and van Leeuwen’s framework, along with an explanation of how this will be applied to Japanese language texts, can be found in the research design and methodology chapter, section 3.4.2.

In summary, a social semiotics approach to meaning making involves three main tenets. Firstly, signs are motivated, not arbitrary combinations of form and meaning. Secondly, the relation between forms and meanings arises from the interest of the sign maker and is based on what they consider to be ‘criterial’ at the point of sign-making. Finally, the repeated use of particular form-meaning combinations in certain social contexts can result in the ‘sedimentation’ of such forms, which may limit or constrain the potential use of alternative semiotic resources.

2.2 Multimodality

A multimodal approach to text analysis requires understanding meaning-making as a process that occurs through the interaction of multiple channels of communication (Jewitt, 2011). For example, in oral communication, a listener derives a speaker’s meaning not only from their words but also through other modes such as their tone of voice, facial expression, and use of gesture. In written language, ‘meanings’ are realized though modes such as size, font, placement, and relation to other features such as images. Multimodality therefore acknowledges that meaning is made through the interaction of semiotic resources, each of which has its own affordances and limitations.

Like social semiotics, multimodality is also influenced by Halliday’s (1978) ‘Language as a social semiotic’, in which he argues that the grammar of a language is a ‘semiotic resource’, rather than a code or set of rules for producing accurate sentences. Halliday describes the communicative functions of language in terms of three metafunctions; the ideational metafunction is used to describe the world around us. The interpersonal metafunction creates relationships between the sign maker and receiver, as well as the object being represented.

Finally, the textual metafunction of language provides cohesion and structure to the text14 (Halliday 1978; p. 40-41). Multimodality extends these metafunctions to resources other than language, for example gesture or font. Furthermore, Halliday’s emphasis on the social nature of communication takes into account the context in which meaning is made, which is also an integral aspect of studies on multimodality.

Within a multimodal framework, scholars differ in their use and definition of the word ‘mode’; some, such as O’Halloran and O’Toole, prefer the term ‘semiotic resource’ (Jewitt, 2011), and others avoid these terms altogether, in order to abstain from ‘drawing strong boundaries between different resources, highlighting instead the significance of the multimodal whole’ (Jewitt, Bezemer, & O’Halloran, 2016, p. 6). Kress (2011, p. 59) proposes a formal test of what can be considered a ‘mode’, arguing that if a particular sign- system can be shown to demonstrate each of Halliday’s three metafunctions (ideational, interpersonal, and textual); it is operating as a mode.

Kress (2011) gives a detailed demonstration of this ‘test’ by arguing that ‘layout’ should be considered a mode, using the example of a page from a high school science textbook, and showing how changes in the layout of the text can vary the ideational, interpersonal and textual functions. Other modes noted by Kress include speech, writing, music, and both still and moving images. It should be noted that ‘language’ is broken down into ‘speech’ and ‘writing’ by Kress, as he argues that the significant differences between verbal and written language mean it cannot be considered a single mode, but should be represented individually (2011). This distinction is useful for the current study, as the orthographic shift from katakana to hiragana is only present in the written medium, but, as will be seen, can have an impact on the sound or pronunciation of written words (see section 7.3.2).

Modes are a product of their culture, and this affects what they are commonly used to express; this is called the ‘affordance’ of a mode, and combines what is possible to represent and express easily with a mode, as well as how it is used by a particular culture (Jewitt, 2011, p. 293). For example, the Japanese writing system, with its different script options (hiragana, katakana and kanji) for writing has an affordance that the English writing system does not have. ‘Culture’ is, of course, a product of the sedimentation of particular practices over time,

14 These metafunctions are revisited in the following section, ‘2.3 Systemic functional linguistics’. 21

and therefore affordance 15 is ‘shaped by how a mode has been used, what it has been repeatedly used to do, and the social conventions that inform its use in context’ (Jewitt, 2011, p. 24). This reinforces the idea of a social semiotic approach to meaning making, acknowledging that ‘culture’ is created by individuals and their use of language over time.

As discussed above, multimodality focuses on what meanings are being made from the interaction of the various modes in a multimodal text, and how these meanings are created. Its key premises are that meaning is made through different modes or semiotic resources which combine to form multimodal wholes; and further that meaning cannot be fully understood without attending to all the semiotic resources involved (Jewitt et al., 2016). Multimodality is therefore useful in understanding loanwords in hiragana, as the functions of script as a mode or semiotic resource must be taken into account when understanding these orthographic choices within the Japanese writing context. Within a multimodal ensemble, particular modes can be ‘foregrounded’, which means they carry major ‘informational weight’ or ‘functional load’ (Kress, 2011, p. 60), and it is argued that hiragana, as a marked script for loanwords, is being foregrounded in this way.

2.3 Systemic functional linguistics

Like multimodality, Systemic Functional Linguistics (SFL) has also grown out of the work of Halliday (1978). Halliday described language as ‘systemic’, because users make choices from sets of options, although these choices may be constrained by other factors, such as the context in which these choices are made (Eggins, 2004). As mentioned in the previous sections, Halliday argued that language makes three kinds of meaning simultaneously; ideational, interpersonal and textual.

The ‘systemic’ nature of language is illustrated by Eggins (2004) through the example of a traffic light. One of the defining traits of a ‘system’ is having a finite set of choices of oppositions, for example, the three colours of a traffic light: red, amber and green. Furthermore, the choices in the system must be discrete, in other words, a traffic light must display only one colour at a time. Additionally it is the oppositions (not the substances) in the

15 Jewitt (2011, p. 24) notes that while ‘affordance’ is used by Kress, van Leeuwen uses the term ‘meaning potential’ for a similar concept. 22

system that are meaningful, for example, it is not the shade of green that is important, but simply that it is not red or amber (Eggins, 2004). These systems contain ‘forms’ that also have ‘meanings’, a red light means ‘stop’ for example. Script in the Japanese language can be seen as such a system; there are a finite set of options (hiragana, katakana, kanji, and rōmaji), which are associated with different meanings and/or lexical items, discussed in section 2.4.2, the socio-historical connotations of the Japanese scripts.

The connection between form and meaning within these systems is best understood by considering the language realising context shown in figure 2.1 below, based on Eggins (2004).

Figure 2.1: Language realising context, adapted from Eggins (2004, p.112)

In SFL, ‘forms’ refer to the graphemes of written language, which then realize words and sentences at the lexico-grammar level. The meanings of these elements of language are brought out at the discourse semantics level, and the three types of meaning, (ideational, interpersonal and textual) are represented as the three points of the white triangle in figure 2.1 above. Ideational meaning relates to information or ideas about the world; interpersonal meaning represents the relationship between those in the act of communication, including degrees of power or familiarity as well as modality (truth value); while textual meaning refers to the makeup of the text, including ordering and placement.

However, the meanings created at the discourse semantics level are also shaped by the context of situation they occur in, the ‘register’. In SFL, three variables are identified that have ‘significant and predictable impacts on language use’ (Eggins, 2004). These are the register variables of ‘field’, the topic or focus of the activity; ‘tenor’, the role relations of power or solidarity; and ‘mode’, the amount of feedback and role of language.

The ‘field’ or topic of the text aligns with the ideational metafunction, the kinds of ‘literal’ information being communicated. Fields can include planning a social activity, or selling a product, for example. The ‘tenor’ is related to the interpersonal metafunction, and describes the relative power differences between people, for example an employer and a job applicant; and also the extent to which they share similar views and opinions, for example, their political persuasion. The ‘mode’ represents the textual metafunction, of which the potential for feedback is element, and might involve a high amount of feedback, for example in a conversation, or a low amount, for example to a published book, which could occur via a book review.

A genre comes about when particular values for register variables regularly co-occur and become ‘typical’ in the culture. Genres can therefore be described as sedimented language practices, and texts are usually recognizable as belonging to a particular genre by members of that culture (for example, a romance novel, or a horoscope) (Eggins, 2004).

Beyond this, the greater context of culture also plays a role in the meanings associated with different forms. Context is vital in interpreting meanings in SFL, as it is in other fields connected with Halliday’s work, including social semiotics, and multimodality. Without context, meanings may not be recoverable from the language given.

The understanding of how language makes meaning provided by SFL demonstrates how texts can be internally ‘cohesive’, when the elements at the levels of graphology, lexico-grammar, and discourse semantics bind it together as a unified whole; as well as externally ‘coherent’ through their relationship with the social/cultural context. SFL provides a useful backdrop through which to describe the meanings created by the use of hiragana script for loanwords in Japanese, as it links a graphemic choice, namely ‘hiragana’, with a way of describing ‘meaning’ through the metafunctions, when understood in their individual contexts.

2.4 Socio-historical connotations of the Japanese scripts

As explained in section 1.1.1, contemporary Japanese is written with a mix of hiragana, katakana, and kanji characters, and rōmaji (the roman alphabet) is increasingly considered a fourth Japanese script, (e.g. Inoue, 2005; Tranter, 2008). Each of the scripts has particular connotations or associations for users of the language, which are products of the scripts’ individual histories and uses. In the summary of connotations presented in table 2.1 below, the consistency with which certain traits are associated with certain scripts can be seen.

Hiragana is often described as ‘soft’ and ‘round’, having more curved shapes than the angular katakana or complex kanji, and many sources also refer to the historical association with women’s writing as giving this script a ‘feminine’ nuance. Hiragana’s historical use for personal communication also relates to the connotations of ‘intimacy’ for this script. Hiragana is also associated with terms such as ‘easy’ or ‘simple’, being the first script mastered by learners of the language, which also relates to the connotation of being ‘childish.’

The other phonetic script, katakana, is described as ‘hard’ or ‘harsh’, often in relation to its ‘angular’ shape, being primarily made up of straight lines. It is frequently associated with the related topics of ‘modernity’ and ‘foreign words/loanwords’, as the west was the source of innovation and technology at key points in Japan’s history such as the Meiji restoration and the period following World War II. The relationship between katakana and foreign (western) terms is reinforced by it being the typical script for writing loanwords today.

The kanji script, due to its complexity, is described as ‘elite’ or ‘intellectual’, and is associated with ‘formal’ writing. Historically, texts in the male-dominated domains of politics and government were written exclusively in kanji, which also relates to the script’s connotation of being ‘masculine’, relative to the hiragana used in the private domain of women’s letter writing and poetry. While kanji has a long history as a ‘Japanese’ script, its origins in China are also represented as connotations in table 2.1.

The collected associations of rōmaji in table 2.1 below are fewer and thus harder to generalize; ‘cool’ and ‘international’ featured more than once, for example. As the most recent addition to the Japanese script repertoire, words in this script may seem ‘cool’ where katakana would feel hackneyed and overused. Rōmaji also allows a more authentic rendition of many foreign terms, such as those from English.

Table 2.1: Associations of the Japanese scripts

Source Hiragana Katakana Kanji Rōmaji

Soft Foreign Sherry & Camargo, Formal Intimate Crisp - 1987 Stiff Feminine Direct

Softness Modernity Smith & Schmidt, Erudition Commerciality Femininity Pop culture 1996

Loanwords Easy Slang Childlike Kataoka, 1997 Elite Female Onomatopoeia Non-authentic Emphasis

Soft Hard Round Hard Difficult Iwahara, Hatta, & Tender Angular Intellectual - Maehara, 2003 Simple Foreign China Feminine Calligraphy

Inoue, 2005 ひろしまヒロシマ広島 HIROSHIMA Soft Hard Neutral Modern Examples of Entertaining Harsh Formal Cool ‘Hiroshima’ Tourism Atomic bomb Administrative International

Simplified Smoothed Angular Softness Miyake, 2007 - - Used by women Simplest of Japanese writers in Heian scripts period (8-12th C.)

男らしい女らしい(feminine) (masculine) 新しい(new) やさしい(gentle) むずかしい外国（欧米）風やわらかい(soft) (difficult) Akizuki, 2005 (foreign [western] - 簡単複雑 (complex) (simple) style) 子供っぽい大人っぽい (adult- 鋭い(sharp) (childish) like) 古い (old)

scientific, rigid, novel, foreign, elite, masculine, feminine, soft, imitative, prestigious, formal, hard, smooth, round, emphasising, hard, global, difficult, tender, simple, simple, inorganic, decorative, intellectual, Robertson, 2015 childish, lovely, fake, marked, young, international, conspicuous, unmarked, intimate, male, futuristic, eye-catching, learned, visual, private, nice, elegant, neutral, sharp, fresh, symbolic, cool, adult, Chinese, poetic, Japanese jarring, precise, sophisticated substantial, angular Japanese

Despite the ‘easy’ and ‘simple’ associations of hiragana shown in table 2.1 above, a study by Tamaoka, Hatsuzuka, Kess & Bogdan (1998) suggests that the brain does not process words in hiragana any faster than in katakana. Reaction times in a lexical judgement test, where participants were asked if the presented stimuli were real words or not, resulted in no significant differences for script. However, a ‘script familiarity’ effect was found; real Japanese words, such as ‘きょういく/kyōiku’ (education’) were processed faster in hiragana; and real loanwords, for example ‘エクスプレス/ekusupuresu’ (express) were processed faster in katakana. This is of relevance to loanwords in hiragana, which could be said to represent cases of ‘script unfamiliarity’.

As for the relationship between loanwords and katakana, Rebuck (2002) echoes Fukumitsu’s (2001) thoughts on the relative familiarity of hiragana as making it less suitable for loanwords, as it is learned before the katakana script, explaining ‘it is katakana’s relative unfamiliarity that enables it to deliver a loanwords with greater impact than could the more familiar and gently flowing hiragana’ (2002, p. 48). The notion of ‘impact’ is also alluded to by Fukuda (2006), who notes that katakana serves to make loanwords stand out (目立つ /medatsu) from words in either hiragana or kanji. He also notes that katakana brings attention to a different or particular kind of sound, which is why this script is used for both foreign loanwords and onomatopoeia. These themes of familiarity, impact, and sound, reappear at various points in the data for the current study, and are brought together in the discussion chapter, section 7.3.

Akizuki (2005) describes the use of katakana for words other than loanwords, ones which would normally be written in hiragana, such as ‘スミマセン/sumimasen’ (sorry); or kanji, such as ‘セートカイ/seitokai’ (student council). He terms these words ‘特殊カタカナ語 /tokushu katakanago’ (special katakana words) (2005, p. 164), noting their popularity in girls comics and blogs. He argues that as the script changes the image of the word itself, writers may be trying to avoid the associations of hiragana or kanji respectively, as described in table 2.1. Akizuki argues that hiragana and kanji, and the associations of these scripts, may no longer suit the heroines of today’s girls comics, and the ‘エッジ感/ejjisa’ (edginess) or ‘自分らしさ/jibunrashisa’ (individuality) (2005, p.174) afforded by these special katakana words may help them break away from traditional gendered roles.

Akizuki also draws attention to how the associations of a particular script change over time, using the example of the names of singers, music groups, comedians, and authors to illustrate his point. He notes that katakana was used to give an ‘international’ air to rock groups in the 1960s, who emulated the sounds of western artists as well as their names, such as ‘ザ・テンプターズ/za temputāzu’ (The Tempters). Over the following decades, Japanese versions of ‘western’ music styles such as rock and pop established themselves as genres, and by the 1990s, katakana had lost its catchiness, and rōmaji had become the script choice. Rōmaji accounted for 45% of the names singers or groups in the 2004 sample compiled by Akizuki (2005, p.179), which included names made up of English words, such as ‘Mr. Children’, as well as Japanese names, such as ‘aiko’. Akizuki also sees the precedent for the next shift as the impact of rōmaji names wears out, citing the mixing of upper and lower case letters, such as ‘YeLLOW Generation’, and other non-standard typographical devices as being typical of the new generation of musicians. Although singer and group names are a specific case, Akizuki’s (2005) findings are a reminder of the shifting associations and connotations of the Japanese scripts.

The associations of each Japanese script may also translate to their relative use in authentic texts, according the findings of a script proportion analysis conducted by Smith and Schmidt (1996) on a variety of popular ‘light reading’ texts. They found that texts aimed at women such as women’s comics and romance novels contained more hiragana than other genres, while genres such as ‘business novels’ aimed at men contained more kanji. A follow up analysis demonstrated that these associations could not be explained by lexis or grammar constraints of the word choices.

This section has described how each of the Japanese scripts holds distinct nuances and associations for users of the language, which are a product of socio-historical factors relating to their development and use. Because of the different and often opposing connotations of hiragana and katakana in particular, the examples of loanwords in hiragana investigated in this study are significant examples of a marked orthographic choice, and a worthy topic of linguistic research.

2.5 Loanwords in Japanese

The influx of foreign words (overwhelmingly from English) is a cause of both interest and concern for both the Japanese public and scholars of the Japanese language, especially given the the country’s strong belief in the link between language and (national) identity (Dale, 1986; Gottlieb, 1995; R. A. Miller, 1982). Loanwords pervade almost all aspects of life; food, medicine, sport, technology, culture, even loaned kinship terms such as ‘mama’ and ‘papa’ as described in section 1.1. It is often said it is difficult to have a conversation in Japanese without loanwords, and they feature in tests of Japanese language for non-native speakers, such as the ‘日本語能力試験/nihongo nōryoku shiken’ (Japanese Language Proficiency Test [JLPT]).

This section will be divided into four parts. The first section reviews the history of foreign language contact in Japan, in order to place the phenomenon of loanwords in hiragana within its historical context. The second describes the changes necessary in adapting foreign words into loanwords in the Japanese language. The third section reviews some commonly cited functions of loanwords within Japanese, which provides an important background from which to investigate the functions of loanwords in hiragana, the topic of the third research question. Finally, the fourth section discusses some commonly held perceptions of these foreign words, which will also be taken into account in interpreting the reactions of the native speakers of Japanese on the topic of loanwords in hiragana.

2.5.1 A brief history of foreign language contact in Japan

Today, katakana is the standard script for transcribing loanwords in Japanese, but this has not always been the case. Different systems were favoured at different times and for different sociolinguistic reasons, although many of these changes were gradual, and alternative ways of writing were likely to coexist at any one time. The different periods of language contact are summarized in table 2.2 below, which is based on descriptions by Loveday (1996).

Japan had a trade relationship with Portugal from 1542 to 1639, and words associated with European food and clothing, as well as the religious vocabulary of the missionaries who accompanied the traders, entered the Japanese language. These loanwords were written in

Table 2.2: History of language contact in Japan, based on Loveday (1996)

Major Donor Time Background Script Example Language

European traders and たばこ/tabako 1542–1639 Portuguese Hiragana missionaries (tobacco)

Dutch traders on Dutch Hiragana 硝子/garasu 1609–1853 Dejima Kanji calques (glass)

Meiji Restoration: Hiragana プリン/purin 1853–1920s British and American English Katakana traders Kanji calques (pudding)

ジャズ/jazu 1945–present Post-World War II English Katakana (jazz)

Globalization, high Katakana インターネット/ Present status of English in English Rōmaji/English intānetto education (internet)

hiragana, and the ones which date from this period are still often written in hiragana today. These include ‘たばこ/tabako’(tobacco, cigarettes); ‘ぱん/pan’ (bread); ‘かすてら/kasutera’

(castella);16 and even the dish ‘てんぷら/tempura’ (battered fried food),17 which is often thought to be a ‘native’ Japanese dish. The Tokugawa government, feeling threatened by the work of the missionaries in converting the Japanese population and fearing a political takeover by foreign powers, closed the country to almost all foreign contact in 1639. They also banned Catholicism and destroyed bibles and other texts, and for this reason very few loanwords associated with religion survive today.

During the period of ‘鎖国/sakoku’, or ‘national isolation’ from 1639 to 1853, contact with the world outside Japan was limited to strictly monitored transactions with Dutch traders, who were not allowed to live in Japan but were based on the artificial island Dejima in the bay of Nagasaki. Since the Dutch were the sole transmitters of knowledge of the outside

16 ‘Castella’ is a type of sponge cake originating from Portugal, and is quite common in Japan. 17 This is believed to derive from the Portuguese ‘tempero’ (to season). 30

world, it is difficult to distinguish words of Dutch origin from words which entered the language through Dutch. For example the word for cup or beaker, ‘コップ/koppu’, is reminiscent of both Dutch ‘kop’ and Portuguese ‘copo’, as well as the English word ‘cup’ to today’s users. It is possibly for this third reason that it usually appears in katakana today. Other Dutch words were written with phonetic kanji calques (当て字/ateji). For example the word ‘glass’ (glas) was written with characters pronounced ‘gara’ and ‘su’ (硝子/ garasu), and this transcription is still in use today, although it is also seen in katakana and occasionally hiragana.

In 1853, Japan ended its 200 years of self-imposed seclusion, and capitulated to Commodore Perry’s demands for trade relations. Upon re-opening its doors to the world, a period of mass modernization and an enthusiasm for all that the west could offer followed. Phonetic kanji calques were still used; at other times kanji calques were created which represented the meaning of the term (熟字訓/jukujikun), rather than the sound, such as the Meiji period word for camera, ‘写真機/ shashinki’, which consists of the kanji characters for ‘copy-reality- machine’. Although this word has since been replaced by the English loan ‘カメラ/kamera’ (camera), the word for photo, ‘写真/shashin’ is still used. At this time, the Japanese kanji lexicon was so extensive that new coinages could be easily made, and while the influx of new concepts from the west was great, it was still gradual enough for these new kanji calques to be created, disseminated, and understood. Loans entering the language in the Meiji period could also take other forms; hiragana, as the ‘common hand’ was also used; katakana use for loanwords also started at this time, (Fukuda, 2006); and English was also sometimes used in labels for western-style products made in Japan, for example ‘beer’ and ‘tooth powder’ (Miyoshi, 2013). As mentioned above, the changes in the preferred script were gradual, and many orthographic variants were likely to have been present at any one time.

Hida and Sato (2002), for example, draw attention to the changing orthographic variants for the word ‘coffee’ in various English-Japanese dictionaries, from its first appearance as ‘コツフヒー/koffuhī’ in 1887, through examples of ‘コーヒー/kōhī’ familiar today (1892, 1911), to the kanji ‘珈琲[こーひー]/kōhī’ in 1913, and the hiragana variants ‘こーふいー/kōfuī’ and ‘こーひー/kōhī’ in 1928 and 1928 respectively. From 1931, the word returns to the familiar katakana form, ‘コーヒー/kōhī’. This varying history of transcription is perhaps one of the reasons this word was a frequent topic of discussion in the interviews (see section 6.7).

During other historical periods, the origins of the loanwords rather than the scripts used to write them were the main issue. The positive associations of English words being fashionable or modern were considered a problem by the Japanese government when war was declared with America. A process of ‘linguistic cleansing’ was enacted to remove enemy words from the language, which involved creating meaningful kanji calques (熟字訓/jukujikun) for words that were usually written in kana. These new and sometimes clumsy kanji calques were unpopular and not many survived the war period, however one that did was the word coined for baseball, ‘野球/yakyū’, made from the characters for ‘field’ and ‘ball’. At this time newspapers also reverted to the phonetic kanji readings for the names of countries (Loveday, 1996). This habit continues in the press today, but perhaps as a space-saving device rather than from patriotic motivation, as a single kanji character such as ‘米/bei’ can act as shorthand for ‘ 米国/beikoku’ (America), taking up less space than the four-character katakana ‘アメリカ/amerika’ (America).

During the occupation and post-war period, another major period of loanword uptake began. However, at this time loans were almost exclusively written in katakana, rather than with kanji calques. Loveday (1996) suggests that the sheer volume of loans introduced at this time meant calquing new compounds for each word was impractical, while Honna (1995) reasons that following the war, kanji compounds were unpopular as they had associations of military propaganda and wartime slogans. Both Honna (1995) and Loveday (1996) also cite the effects of the introduction of English as a compulsory subject at schools; this meant people were more familiar and comfortable with English words, whereas in the past, a kanji calque that represented the meaning of the new word was more transparent (Inoue, 2005).

In addition, during the post-war period, kanji use in public domains became limited to a prescribed 1,850 characters, which were to be taught in compulsory schooling (Honna, 1995; Loveday, 1996), this was the predecessor of the list of ‘常用漢字/jōyō kanji’ used by schools today. Advocates of script reform and the limitation of the numbers of characters in use had been active in Japan since the Meiji restoration (Akizuki, 2005), however the ultra-nationalist spirit of the war years had made it difficult and dangerous for these sentiments to be expressed too openly (Gottlieb, 1995). However, with the support of the occupying forces who also perceived difficulties with the Japanese writing system, language reformers argued that the previously infinite number of kanji enforced semi-literacy on the lower classes, and

saw the restriction of the number of kanji as a move to ensure access to information for all people.

Honna (1995) and Loveday (1996) suggest that with this decrease in the number of characters available for use in official texts, it was therefore also more difficult to create kanji calques for new western concepts. At the same time, the growing familiarity with English meant that katakana loanwords were in an ideal position to take over as the preferred method of rendering these new terms. While these factors were no doubt influential, other authors note that the use of kanji for loanwords was already in decline; Ishī’s (2013) study of loanword orthography in the magazine ‘婦人公論/fujin kōron’ published in the years 1916 to 1926 showed a gradual shift from a preference for kanji loanwords with katakana ruby, such as

インド ‘印度 /indo’ (India) to katakana loanwords, for example ‘インド/indo’ (India). Fukuda

(2006) also notes that new kanji were rarely made in the Taisho (1912-1926) and Showa 1926–1989 periods. It is quite likely that both sides are correct, and that the gradual disfavour for rendering loanwords in kanji was cemented by the language changes brought about at the end of the Second World War. The habit of transcribing loanwords in katakana has continued until today, where it is considered the standard script for these foreign terms, cemented by MEXT’s language policy on writing loanwords, ‘外来語の表記/gairaigo no hyōki’ (the transcription of loanwords) (1991).

While this policy covers written materials produced by government departments and the education system, style guides for publishing houses and large organizations also consistently state that foreign words/loanwords should be written in katakana, including the Japanese Psychological Association (日本心理学会/nihon shinrigaku kai), and the ‘Japan Technical Communicators Association [JTCA] (一般財団法人テクニカルコミュニケーター協会 /ippan zaidan hōjin tekunikaru komyunikētā kyōkai’).

However, despite or perhaps because of the widespread consensus of the use of katakana for loanwords, alternative orthographic choices are often the topic of general interest articles in newspapers and language-related blogs. For example, a post on the blog ‘ことば会議室 /kotoba kaigishitsu’ (Language meeting room) lists examples of loanwords in hiragana from novels of the 1950s and 60s, and also notes the presences of signs for ‘きゃべつ/kyabetsu’ (cabbage) and ‘とまと/tomato’ (tomato) in hiragana in contemporary supermarkets (Yeemar, 2005). A question addressed to the website ‘よくある「ことば」の質問/yoku aru ‘kotoba’

no shitsumon’ (Common questions about language) administered by 国立国語研究所 /kokuritsu kokugo kenkyūjo’ (National Institute for Japanese Language and Linguistics [NINJAL]) asks if it is permissible for their children, who have not learned katakana yet, to write loanwords such as ‘ぼおる/bōru’ (ball) and ‘そおす/sōsu’ (sauce) in hiragana (Yamada, 2013).18 An article from the column ‘なぜなに日本語/naze nani nihongo’ (The ‘what’s and ‘why’s of Japanese) titled ‘ギョーザは外来語/gyōza wa gairaigo’ (‘Gyōza’ is a loanword) explains that despite its Chinese origins, this term is technically a ‘loanword’ rather than a Sino-Japanese word (漢語/kango), as it entered the Japanese language after the Muromachi period (1336–1573) (Sekine, 2014).

These kinds of general-interest texts highlight two important and seemingly contradictory points. Firstly, variation does occur in the way loanwords are transcribed; and secondly, katakana is the conventional script for loanwords, as it is only against this backdrop that the variation described above is noteworthy and topical. In other words, it is only because katakana is so widely accepted as the appropriate medium for loanwords that parents are concerned when their children cannot write them in this way, or that an entire newspaper column is dedicated to untangling the etymology and pronunciation of the word ‘gyōza’ in order to establish the appropriate script in which to write it. The departure from convention in writing loanwords in hiragana is another such variation.

2.5.2 Adapting foreign words into Japanese

When foreign words are adapted phonologically to the Japanese kana system, they usually undergo a number of changes to account for differences in the sounds of the source language and Japanese (Akizuki, 2005; Hida & Sato, 2002). Phonetically, extra syllables are often added to loanwords in order to account for the Japanese mora system, meaning an English word with one syllable, such as ‘text’, becomes the four-mora loanword ‘テキスト/tekisuto’. Other adaptions are made for sounds whichare not found in the Japanese kana system, for example both ‘la’ and ‘ra’ sounds are approximated by the Japanese sound ‘ラ’, anglicized as ‘ra’ but actually somewhere in between the two English phonemes. The policies of MEXT,

18 The answer provided explains that it is acceptable for this stage in their linguistic development, and they will gradually master the rules of script use as they learn katakana, and later, kanji. 34

the JPA, and the JTCA referred to in section 2.5.1 above further define permissible sound sequences of kana, which are not conventionally used for words of Japanese or Chinese origin (和語/wago or 漢語/kango), but which may be necessary for transliterating foreign terms, such as the combination of a voiced ‘ウ/u’ with a half size ‘イ/i’ to represent the /v/ sound ‘ヴィ/vi’. These permissible sound sequences and guidelines for writing loanwords have undergone many changes as described by Gottlieb (1995); and MEXT often lists multiple permitted forms, for example ‘ビーナス/bīnasu’ and ‘ヴィーナス/vīnasu’ for ‘Venus’, with the recommendation that whichever is chosen should be used consistently throughout the document (Ministry of Education, Culture, Sports, Science, and Technology Japan, 1991).

Kay (1995, p. 69), also writing in the mid-90s, notes a growing acceptance of these ‘special’ combinations of kana that allow a more accurate representation of foreign sounds, such as the ‘ファ/fa’ in ‘ファックス/fakkusu’ (fax). The combinations she mentions are now common in transcribing loanwords, although Akizuki (2005) notes a revival of pre-war forms such as ‘プリチー/purichī’ (pretty) over the more native-like ‘ プリティー/puritī’. Akizuki also describes a number of examples of loanwords featured in manga which he argues represent a closer approximation to the English pronunciation, despite moving away from the standard transliteration rules, for example ‘ポインツ/pointsu’ over the more common ‘ポイント /pointo’ for ‘point’, and ‘チェケ/chekke’ for ‘チェック[イット] /chekku [itto]’ meaning to ‘check [it out]’. Ironically, many of his examples, which he terms ‘ニュータイプ外来語 /nyū taipu gairaigo’ (new type loanwords), have fallen out of contemporary use, however, similar examples appeared in the data for the current study. The use of hiragana to highlight a non-standard transcription of a loanword which more closely reflects a native-English norm is discussed in sections 6.2.8 and 7.3.2.

While katakana and hiragana both represent the same set of sounds, a difference in the way vowels are lengthened in each script occurs. In Japanese words in hiragana, an extra vowel character is added, for example ‘大阪/おおさか/ōsaka’ for the place name ‘Osaka’. 19 However, for loanwords in katakana, a bar is used to lengthen a vowel, for example ‘チーズ /chīzu (cheese)’. Because this represents a significant difference between the transliteration of

19 It is noted that the extra vowel character is often ‘う/u’ when an ‘お/o’ sound is lengthened. For a detailed explanation see Irwin (2011). 35

words in hiragana and katakana, the method of lengthening vowels in the case of loanwords in hiragana is investigated as part of the present research (see section 3.4.1).

A number of ‘types’ or traits of loanwords, representing both linguistic as well as semantic features, have been described by a number of scholars (Hoffer, 1980; Honna, 1995; Ishikawa & Rubrecht, 2008; Kay, 1995; Loveday, 1996). There is much overlap between their individual taxonomies, and therefore only those which are considered most commonly occurring and/or relevant to the current study will be mentioned here.

Loanwords, especially long ones, sometimes undergo ‘back-clipping’, for example the word for ‘department store’ is shortened to ‘デパート/depāto (depart[ment store])’. This can also happen in two-word compounds, where both words are back-clipped, for example in the word for computer, ‘パソコン/pasokon’ (perso[nal] com[puter]). Front-clipping is also possible but much less common; one example is the word for train platform, ‘ホーム/hōmu’ from ‘プラットホーム/purattohōmu’, often incorrectly believed to be derived from the English word ‘home’.

Loanwords can be combined with other Japanese morphemes to form new words, for example ‘歯ブラシ/ha-burashi’ (tooth-brush), and can take Japanese grammatical suffixes appropriate to their word class, for example ‘キュートな/kyūto na’ (cute + na-adjective stem),20 or ‘ショッピングする/shoppingu suru’ (shopping + ‘to do’). It is also possible to change the word class of a loanword by adding the appropriate Japanese stem, for example ‘ナウい/nau-i’ (now + i-adjective stem = current, trendy) (Hoffer, 1990). More recent innovations include ‘ググる/guguru’ (goog[le] + verb stem = to google)21 and ‘ディスる /disuru (to ‘dis’, to disrespect), a borrowing from hip hop culture. Backhaus (2012) notes further examples such as ‘メモる/memoru’ (to take a memo), and ‘スタンバる/sutanbaru’ (to stand by).

Other features of loanwords in Japan, also common in other language contact situations, are semantic narrowing, semantic widening, and semantic shift (Honna, 1995; Kay, 1995). For

20 It is noted, however, that when English adjectives are borrowed into Japanese they usually become ‘な/na’ (noun) adjectives rather than ‘い/i’ adjectives, for example ‘クールな/kūru na’ (cool), and ‘シャイな/ shai na’ (shy), though counterexamples such as ‘グロい/guroi’ (grotesque) and ‘エロい/eroi’ (erotic) also exist. 21 This example mirrors English use where the name of the search engine ‘Google’ is now used as a transitive verb, for example ‘to google something’. However, the Japanese examples are interesting in using hiragana to mark the grammatical stem, while the rest of the word remains in katakana. This is further discussed in section 2.8, ‘Typography’ below. 36

example, semantic narrowing can be seen in the loanword ‘ツナ/tsuna’ (tuna), which refers to canned, but not fresh tuna (Kay, 1995). Semantic widening occurs in the use of the word ‘ジュース/jyūsu’ (juice) to refer not only to fruit juices but to soft drinks such as lemonade and cola (Olah, 2007). A further example is the word ‘rare’ (uncooked, lightly cooked) which can be applied to foods other than meats, for example a ‘rare cheesecake’ (レアチーズケーキ/rea chīzu kēki) is a cold-set cheesecake, rather than a baked one (Ishikawa & Rubrecht, 2008). Semantic shift, where the meaning in the source language differs from its use in the loaned context, occurs in terms such as ‘ハンドル/handoru’ (handle) for the steering wheel of a car, and ‘マンション/manshon’ (mansion) to refer not to a large, luxurious house as it does in English, but to an apartment.

A variation on semantic shift is the phenomenon of ‘made-in-Japan English’ (和製英語 /wasei-eigo), which are Japan-made terms constructed from English words, such as ‘ペーパードライバー/pēpā doraibā’ (paper driver), used to describe someone who has a car licence but rarely drives, or ‘モーニングセット/mōningu setto’ (morning set), which refers to a set ‘western-style’ breakfast at a restaurant, typically consisting of bread, salad, and coffee. A further subset of this category are Japan-made acronyms based on English words, with the pronunciation of each letter adapted to the Japanese sound system as described above. These include the term ‘O.L./ō eru’ described previously, which comes from the English words ‘office lady’, used to refer to female office clerks, and ‘L.D.K/eru dī kē’, representing a combined lounge-dining-kitchen in a small apartment. The extent to which examples of ‘made-in-Japan English’ can be considered loanwords is discussed in section 2.5.4 below.

There has been much discussion on the role of katakana script in the adaption of loanwords into the Japanese language. Many scholars acknowledge the ease with which katakana allows the Japanese to introduce foreign words into their native language (Kay, 1995; Rebuck, 2002; Stanlaw, 2004). There is no need for consensus or official approval of a kanji calque before a word becomes widely used, as there was in the past (see section 2.4.1), and guidelines such as those described above provide suggestions for transcribing sounds not found in Japanese. Honna (1995, p.52) also suggests that the fact that loanwords can be combined with Japanese morphemes is evidence that ‘they are structurally and semantically treated as Japanese words… They have become part of the Japanese lexicon and grammar’.

However, despite the efficiency with which these foreign terms can be presented in written Japanese, the use of the katakana script has also been argued to prevent the complete integration of these same terms (Hosokawa, 2015; Kay, 1995; Loveday, 1996; Rebuck, 2002). It is argued that the use of katakana brands loanwords as ‘foreign terms’, and therefore differentiates them from the two other main categories of the Japanese lexicon, native Japanese words (和語/wago) and Sino-Japanese words (漢語/kango). Heitani (1993, p.43) cited in Rebuck (2002, p. 54), notes the anomaly of a language that ‘allows everything in, but in fact gets by without anything entering’. Because of the perceived role of katakana in preventing loanwords from fully assimilating into Japanese, the presence of loanwords in hiragana represent an important area of study. Like katakana, hiragana offers a straightforward method of representing foreign words in Japanese, but these words no longer retain this orthographic marker of foreignness, being written in the script associated with Japanese words.

2.5.3 Functions of loanwords

Loanwords perform a number of functions in Japanese language; the most commonly cited ones include filling a lexical gap, acting as a euphemism, projecting an image of modern/western lifestyle, and creating word play.

Perhaps one of the earliest motivations to incorporate English words into Japanese was to fill a lexical gap (Hoffer, 1980; Loveday, 1996; Rebuck, 2002). This accounts for the large number of loanwords that describe food and clothing that were not typical in Japan before contact with the west during the Meiji period, for example ‘バター /batā’ (butter) and ‘ネックタイ/nekkutai’ (necktie). Many technology-based words also fill lexical gaps, such as ‘キーボード/kībōdo’ (keyboard), and ‘インターネット/intānetto’ (internet) (Honna, 1995).The function of filling a lexical gap also refers to concepts that are not necessarily new, but are newly arrived in public discourse, such as ‘ホームレス/hōmuresu’ (homeless) and the back- clipped compound ‘セクハラ/sekuhara’ (sex[ual]) hara[ssment]) (Rebuck, 2002). It can be argued that these words, representing social issues, may also have a euphemistic function over a more literal or direct kanji equivalent.

The most commonly cited euphemistic loanword is the word for bank loan (ローン/rōn) (Hoffer, 1980; Honna, 1995; Rebuck, 2002). It is often claimed to be preferable to the Japanese equivalent ‘ 借金/shakkin’, which literally means ‘borrow money’. While the Japanese word has negative connotations of spending more than one can afford to, it is argued that the loanword is more opaque and carries positive, modern connotations, suggesting a planned financial strategy. Another interesting example of the euphemistic function of loanwords is suggested by Stanlaw (2004), who argues that the use of the loanword ‘my’ in compounds such as ‘ マイホーム /mai hōmu’ (my home) serves a euphemistic function. As the Japanese translations of ‘my’ (‘私の/watashi no’ or ‘自分の /jibun no’) would be more direct and thus at odds with society which prioritizes the needs of the group over those of the individual, the use of the loanword softens what could be seen as a selfish term.

Loanwords can also be used to give a modern or international image to an item, even when there is a Japanese equivalent. For example, both Hoffer (1990) and Rebuck (2002) notes the preference for the loanword ‘キッチン/kitchin’ (kitchen) over the Japanese ‘台所/daidokoro’ (kitchen), as the loanword connotes an air of modernity and luxury. Loanwords are also used to distinguish between western and Japanese versions of the same referent; Loveday (1996) notes that when rice is served in a rice-bowl as part of a Japanese meal, it is referred to as ‘ご飯/gohan’, but served on a plate as part of a non-Japanese meal such as curry or steak, it becomes ‘ライス/raisu’ (rice). He also suggest that the image created by the word ‘ウエディング/uedingu’ (wedding) is likely to involve a church with the bride wearing a white dress, but a ‘結婚式/kekkon shiki’ (wedding) would suggest a Shinto shrine with the bridal party wearing kimono. For this reason, loanwords in hiragana are a worthy topic of investigation, as they problematise the loanword/katakana and native word/hiragana distinction.

A final function of loanwords relevant to the current study is that of word or language play, which is also often listed in the functional taxonomies of loanwords along with the categories described above by researchers such as Hoffer (1980) and Honna (1995). This topic is of direct relevance to one of the functions of loanwords in hiragana, and is therefore explored in a separate section, 2.7 below.

2.5.4 Perceptions of loanwords

After this illustration of the changes words from English (and other languages) undergo phonetically and semantically when becoming ‘loanwords’, it may be asked whether these words can really be considered ‘English’ at all. It may be hard for a native speaker of English unfamiliar with Japanese phonology to recover the word ‘wedding’ from ‘uedingu’, and the semantic changes illustrated above show how these words may not even have the same meaning as their source terms. Examples of ‘made-in-Japan English’ (和製英語/wasei eigo), including acronyms such as ‘O.L’ (office lady) described above, further complicate matters, and the use of English words in rōmaji (such as ‘club’ rather than the loanword ‘クラブ/kurabu’) is becoming increasingly used as a further orthographic alternative. While these words in rōmaji script would not usually be considered ‘loanwords’, Backhaus (2007) shows how many of these terms can only be understood in their Japanese contexts, such as the words ‘Beauty Menu’ heading the list of services available at a salon. These rōmaji terms, the meanings of which may not be immediately understood by an English speaker unfamiliar with Japanese, therefore function more like loanwords, even if their orthographic form suggests a direct English borrowing.

Stanlaw (2004) has therefore argued that the term ‘loanword’ is misleading because most of the ‘English’ used in Japan is in fact ‘made-in-Japan English (wasei-eigo)’ and incomprehensible to an English speaker with no knowledge of Japanese. While it is true that there are a large number of these colourful made-in-Japan English words in use in Japan today, it would be difficult to show that these represent the majority of cases as Stanlaw claims, and he also acknowledges the difficulty in compiling accurate figures representing the proportions of genuine ‘loanwords’ to ‘made-in Japan English’. This was attempted, however, by Ishikawa and Rubrecht (2008) in their analysis of loanwords used in television programs. Of the 18 codes used in their analysis, the most common was ‘direct borrowings’ comprising 63% of their examples, while those representing ‘made-in-Japan English’ represented much smaller proportions, such as ‘Japanese phrasings in English’ (0.9%), and ‘semantic narrowing or shift’ (3.7%). For the purposes of the present study, it is sufficient to note that both ‘made- in-Japan English’ as well as more direct borrowings with equivalent meanings are aspects of the lexicon referred to as ‘loanwords’ in Japanese.

Due to the vast number of loanwords in use in everyday Japanese, negative attitudes to these words are not uncommon; Otake (2007) describes the ‘torrential influx of foreign words’ into the Japanese language, for example. These negative attitudes are based on a number of reasons including their unintelligibility; the danger they pose to ‘pure’ Japanese language; and even the difficulties they create for both native Japanese speakers learning English (Olah, 2007), and native English speakers learning Japanese (Igarashi, 2007) . The debate is also a long running one, and the attitudes present in studies conducted in the 1990s such as those by Loveday (1996) and Tomoda (1999) are still present today. In 2013, Takahashi Hōji attempted to sue Japan’s national broadcaster, NHK, for its overuse of loanwords, arguing that the ‘deluge’ of loanwords was causing him ‘great emotional stress’ (Osaki, 2013). However, Irwin (2011) questions the notion of a ‘deluge’, pointing out that while the number of types of loanwords may be relatively high and continues to rise, the low number of tokens highlights the fact that many loanwords are short lived, being replaced by others; or are otherwise technical terms limited to specialist usage.

Hosokawa (2015) investigated attitudes to loanwords by analysing over 200 letters to the editors on this topic, which were published in two major newspapers. She noted that both proponents and opponents of loanwords based their arguments on the history and tradition of the Japanese language, but interpreted in very different ways. Those against the use of loanwords described the beauty and tradition of the Japanese language, and how it needed to be protected from an ‘inundation’ of foreignisms. Those for the use of loanwords argued that part of the beauty of the Japanese language was its flexibility and long tradition of adapting foreignisms, and used the parallel water metaphor of ‘absorption’ to describe how the language integrates these new words.

In Hosokawa’s study, the reasons given by those opposing the use of loanwords included opaqueness of meaning, the word ‘コンセンサス/konsensasu’ (consensus) was cited as a representative example (Hosokawa, 2015). While older generations are often said to be those who have the greatest difficulty in understanding loanwords, a study by Daulton (2004) included only young Japanese adults, and still found that that a quarter of the loanwords used in the Mainichi newspaper were not understood by this research sample. Of the university students in Olah’s (2007) study, 43% reported they did not understand more than half the loanwords used in the media, and 60% agreed with the statement ‘There are too many loanwords in Japanese.’ The difficulty in understanding loanwords was the study of a report

produced by the National Institute for Japanese Language and Linguistics (NINJAL), which investigated the comprehensibility of 450 loanwords. Many, such as ‘パブリックインボルブメント/paburikku imborubumento’ (public involvement) were found to be understood by only a small minority of those surveyed, leading Otake (2007) to wonder how many citizens would make use of opportunities to be ‘involved’ if the term was unclear to them. NINJAL further produced a report which made suggestions for existing Japanese words to replace commonly misunderstood loanwords, for example, in the case of ‘コンセンサス/consensasu’ as raised by Hosokawa’s respondent, ‘合意/gōi’ (consensus) was proposed as a replacement (National Institute for Japanese Language and Linguistics, 2006, p. 63).

Comments that loanwords with Japanese lexical equivalents are unnecessary are represented in data such as Loveday’s (1996) as well as Hosokawa’s (2015). Loveday (1996, p.216) for example cites the jargonistic ‘コーティングスタッフ/kōtingu sutaffu’ (coating staff) to refer to a simple ‘wall painter’ (ペンキ屋/penkiya). However, Hosokawa’s data is valuable as it contains attitudes from proponents of loanwords as well as opponents, and one showed an understanding of the subtle differences between seemingly ‘equivalent’ terms, citing the different connotations of the words ‘ショッピング/shoppingu’ (shopping) and ‘買い物 /kaimono’ (shopping) (Hosokawa, 2015, p. 61).

These subtle differences were the focus of an interesting study by Kikuchi (1994) which investigated the connotations of loanwords as well as semantically similar native Japanese (和語/wago) and Sino-Japanese (漢語/kango) terms, for example the words, ‘ホテル/hoteru’, ‘やどや/yadoya’, and ‘旅館/ryokan’, which all refer to a hotel or inn. He found that both wago and kango have similar associations of being nostalgic, familiar, and old-fashioned; while the loanwords were frequently associated with modernity. However, he also noted that these associations varied with the particular word, as well as the individual respondent. As noted previously in the discussion on loaned and native ‘equivalent’ terms by Hoffer (1990) and Rebuck (2002), loanwords in hiragana problematise this dichotomy between script association and word type.

Aside from the usefulness of these introduced terms, another frequently given reason by those in favour of loanwords in Hosokawa’s (2015) study was how the absorption of loanwords today was not unlike the absorption of Sino-Japanese words (漢語/kango) in earlier centuries, an idea that was also mentioned previously by Hoffer (1980). These words, it is argued, were

also originally ‘foreign’ and are now considered an integral part of the Japanese language. The use of hiragana for loanwords investigated in this study, it is argued, can be seen as a further nativising device.

The perceptions of loanword by both language experts and the general public described above highlight the complexities surrounding the place of these words within the Japanese language. The term ‘loanword’ can include terms closely related in form and meaning to their source language counterparts, as well as ‘made-in-Japan English’, which represents the other end of the spectrum. Loanwords in Japanese perform a number of important functions, and yet the reliance on these terms is seen by some as a threat to the integrity of the Japanese language. As the current research investigates a particular class of these loanwords, the aforementioned attitudes and perceptions need to be taken into account when understanding this topic.

To conclude this section, loanwords occupy an important and well-established place within the Japanese language. Looking to their origins, they can be seen as ‘foreign words’, but they are also very much ‘Japanese words’ in their functional use, pronunciation, and orthography. While some object to the jargon-like nature of these words and express concern for the numbers of these foreign terms entering the language, others draw parallels between the adoption of Sino-Japanese words (漢語/kango) and the kanji characters to argue that this flexibility represents a core strength of the Japanese language.

Katakana is often cited as being an effective method for adapting foreign words into Japanese (Kay, 1995; Rebuck, 2002; Stanlaw, 2004), although it is also noted as preventing the complete integration of these terms into the Japanese language (Hosokawa, 2015; Kay, 1995; Loveday, 1996; Rebuck, 2002). These authors do not, however, consider hiragana as an option for the transcription of loanwords, which is equally effective, and does not carry the connotations of ‘foreignness’ associated with katakana. This thesis aims to show that writing loanwords in hiragana is not uncommon in particular genres, and therefore a topic worthy of investigation. Furthermore, the dichotomy commonly drawn between ‘native’ words in hiragana or kanji, and ‘foreign’ words in katakana, as described by Hoffer (1990), Kikuchi (1994), and Rebuck (2002), is also problematised by the current study, which focuses on examples of foreign words represented in a script associated with native Japanese words.

2.6 ‘Non-standard’ orthography

Atypical or ‘non-standard’ forms of written language can be used for a number of functions, including to represent a particular identity within a sub-culture; as an aid to representing linguistic cues from speech in written language; or simply to perform a decorative or eye- catching function. Examples of non-standard orthography are often found in informal texts outside the control of schools and other domains of ‘standard’ language use, including personal letters, and computer mediated communication (CMC).22 This section will discuss examples of non-standard orthography in product and business names; informal handwritten texts including letters and print-club images; as well as examples from manga and CMC.

Unusual or marked orthography can often be found in the names of products or businesses; this ‘expected’ transgressing of norms is noted to be particularly typical in relation to items aimed at young people (Blommaert, 2013). The mixing of Latin and Cyrillic scripts in Ukrainian business names was also noted to be associated with ‘youth-oriented’ business (Dickinson, 2015, p. 517). Jaworski (2015, p.217) describes how unusual letterforms, punctuation, and diacritics create ‘novel, foreignized, visual-linguistic forms’ which he dubs ‘globalese’, noting its frequent use in business names. One of the most well-know examples of this is the American ice-cream brand, ‘Häagen-Dazs’; its unusual spelling and diacritics (umlaut) connoting an exotic or European air.23

Davies (1987) identifies a number of functions of unconventional spellings commonly used in business names, many of which are also typical of the examples of business names collected in this study. For example, unusual spelling may be used to differentiate the brand name from a simple description of the product, for example ‘Loctite’ (lock-tight) glue. It may also be used to differentiate products with similar names, for example ‘Quick’ washing up brushes, ‘Quik’ sponges, and ‘Kwik’ cuff links. Davies suggests that the popularity of the letter ‘x’ in business names may be due to its rarity in conventional spelling, making it an effective attention-getting device. It is often used to replace the letters ‘cks’, for example in ‘Stix’ glue sticks. A further trait identified by Davies was the use of unconventional spelling to connote another time or place, for example archaic spellings such as ‘The Olde Taverne’

22 This term is used in this paper to also include text messages sent from mobile phones and posts on social media. 23 An umlaut is a set of two dots that can be placed above some vowels, for example ‘ü’, to alter the pronunciation of the letter. The umlaut as a marker of ‘German-ness’ is discussed further in section 2.8, ‘Typography’. 44

are typically used to suggest nostalgia or history, while the use of ‘crème’ for ‘cream’ can give products an air of French-ness, and the associated connotations of fashion and luxury. Davies notes two general trends in the unconventional spellings she discusses; one toward simplification, by removing unnecessary letters, for example; and the other towards elaboration, using unusual letters or combinations of them. In either case, she notes ‘it is essential that the innovation in spelling should not be so extreme as to obscure the origin of the name completely’ (Davies, 1987, p. 49). These traits identified by Davies (1987) will be discussed in relation to the product and business names in the corpus in chapter 4, section 4.1.4.

Non-standard orthography can also be found in personal writing; the characters produced by young Japanese women have taken various forms in recent decades, and have become a topic of linguistic as well as public interest. For example, the craze for the handwriting style called ‘丸文字/maru-moji’ (round letters) characterised by puffy, rounded shapes and oversized diacritic marks started in the 1970s, and became widespread in the 1980s. Maru-moji was even banned in some middle and high schools, the purposeful flouting of conventional writing style being suggestive of ‘delinquency’ according to some teachers (Gottlieb, 2010a, p. 403). Although the boom for maru-moji has passed, it is still in use today, and other orthographic features which co-occurred with the style such as pictorial marks including hearts (♡) and tears ( ), excessive use of punctuation marks such as exclamation points (!!!!) and characters used to represent voice quality, such as bars and tildes for elongated sounds (eg. だーい好き/dāisuki [looooove]), continue in popularity, and were found to be typical of the personal letters exchanged between young women in the 1990s in Kataoka’s studies (1997, 2003). He suggests that young women use these strategies to convey an emotional stance and reciprocity of feeling in their writing. Kataoka’s data also contains one example of a loanword in hiragana, the greeting ‘はろー/harō’ (hello). It is included as an example of ‘anomalous or idiosyncratic script use’ followed by the brief comment ‘Loanwords are supposed to be written in katakana’ (1997, p.115). Maru-moji and the other paralinguistic features described by Kataoka (1997, 2003) were also found in some of the hand-written texts containing hiragana loanwords in the corpus, such as hand-made signs and print-club images.

L. Miller (2003) notes that creative use of colours and script continue to be used in young women’s writing in what she terms ‘graffiti photos’, now commonly called ‘print-club’. Print

club booths are common in game centres in Japan, and allow the user to have their photo taken, and then decorate the image with writing, clip-art like ‘stamps’, and various backgrounds via a stylus connected to the screen. After a couple of minutes, stickers of the decorated images are produced by the booth, and more recently digital copies of the images can also be sent to an email address and/or automatically uploaded to an online gallery. The images discussed by L. Miller include not only humorous and risqué poses, but captions in ‘deviant’ orthography, characterised by uneven, malformed characters, and unconventional choices of script. She cites two examples of a loanwords in hiragana; the phrase ‘SEXY しょっと/SEXY shotto’ (sexy shot)24, and the Hindu greeting ‘なますて/namasute’ (namaste) thus departing from the convention that loanwords be written in katakana, although this is not explored by the author. L. Miller argues that the non-standard use of script which frequently occurs in the print club images she discusses is particularly rebellious because of the association of women with the refined art of calligraphy; ‘when they “degrade” the art of calligraphy, girls are defying a constraining model of gender, and are refusing to be caretakers of a cultural tradition that they associate with adult expectations of gender conformity’ (2003, p 35). Print club are therefore not just a site of girlish playfulness, but a way for young women to critique and rebel against conventional models of femininity. L. Miller’s identification of print club images as a site of unconventional orthography motivated the investigation of online print-club galleries in this study.

Another type of non-standard orthography often discussed in the Japanese context is referred to as ‘ギャル文字/gyaru-moji’ (‘girl-letters’), but in contrast to their maru-moji precedents, are associated with text input via a mobile phone keypad or computer, rather than by hand (L. Miller, 2011; Miyake, 2007; Tranter, 2008). As text-messaging, or emails sent via mobile phone became widespread in Japan during the 1990s, a new way of encoding Japanese characters because common with young, fashionable and tech-savvy ‘gyaru’ (girls/gals). Gyaru-moji took advantage of the full range of characters and symbols available in CMC, including letters of the roman, Greek and Cyrillic alphabets, and mathematical symbols to recreate Japanese hiragana, katakana, and sometime kanji characters. For example, the katakana character for ‘ キ /ki ’ could be represented by the mathematical symbol for unequal ‘≠’, and hiragana ‘し/shi’ with the capital letter ‘U’. Scholars such as L. Miller argue

24 This example further mixes a word in rōmaji/English (SEXY) with a hiragana loanword; a similar example from the corpus is analysed in case study 2, ‘I can ふらい/furai/fly!!’, see section 4.2.3. 46

that gyaru-moji was both a form of group membership and identity for young women, as well as a form of resistance to the mainstream adult/male-dominated culture (2004, 2011).

A particular orthographic aspect of ‘gyaru-moji’ which was also found in the corpus and commented on by focus group participants was the use of the half-size character, for example, the extra character at the end of the phrase ‘終わりましたかぁ/owarimashita kā’ (Is it over?) as cited by Nishimura (2003). These characters act as an emphasising device, making the syllable in which they occur longer and stronger, in the previous example, something like ‘overrrrrr?’ in English. While Nishimura’s data contains additional half-size characters, Miyake (2007) notes cases where those which form part of the word are represented in half- size characters, for example the final character in the word ‘ありがとぅ/arigatō’ (thank you) (2007, p. 60). Miyake notes this habit as being particularly popular amongst female high school students, and describes a number of functions of these half-size letters. These include giving the message a ‘soft, rhythmic and causal feel’ (2007, p. 60); expressing prosodic features when found in final positions; and finally simply making the messages appear more ‘kawaii’ (cute).

‘Gyaru-moji’ shows many similarities to the cipher-like ‘leet/l337’25 language of chatrooms, originally developed by hackers to avoid their messages being intercepted by authorities. Lange (2015) cites the phrase ‘N 3 1 g0+ n 3 /\/4r3z(?)’ (Anyone got any ‘wares’ [pirated software]?’) being used in a chatroom, for example. However, the enthusiastic uptake of ‘leet’ by novices posing as hackers has resulted in a fall in the perceived ‘coolness’ of this orthographic style.

Other forms of non-standard orthography used in chat rooms, text messages, and more recently instant-messaging services, are now well-established, such as phonemic substitutions of letters for words (such as ‘B’ for ‘be’, and ‘4’ for ‘for’), and acronyms such as ‘LOL’ (laughing out loud), and ‘BRB’ (be right back) (Jones & Schieffelin, 2009; Ross, 2006). This ‘text speak’ not only saves time allowing for faster responses, but creates a community of practice for users who may be separated by vast distances and have never met face to face. English text speak is also a common feature of CMC between speakers of languages other than English, for example Thai and Serbian, where it has additional connotations of worldliness and modernity (James, 2014; Seargeant & Tagg, 2011).

25 The etymology of the term ‘leet’ is derived from ‘elite’. See http://www.urbandictionary.com/define.php?term=leet 47

‘Emoji’26 or ‘emoticons’ are another aspect of CMC which are widely used today. Many represent faces expressing different emotions, but also include an expansive range of icons including those representing sports, modes of transport, and food. The website ‘Emojipedia’ lists 2,666 emoji in the Unicode Standard as of May 2017, 27 and their establishment in contemporary culture even lead to one particular emoji, representing a face crying with laughter, being named the ‘Word of the year’ by Oxford Dictionaries in 2015.28 Danesi (2016) notes a wide range of functions of emoji, and argues that they are useful in filling the gap left by the absence of visual and audio cues in CMC.

Features of spoken language, such as a particular pronunciation, or accent, can also be encoded with non-standard spelling. Olivo (2001) noted how the non-standard spellings used in the lyrics to rap songs were used to graphically represent features of African American Vernacular English (AAVE), for example. In the Japanese corpus investigated by Joyce et al. (2012, p. 270), the authors also noted that some unusual orthographic forms seemed to be attempts to mimic the pronunciation of a certain situation, citing the variants of ‘ぜんぜん /zen zen’ (totally), including those with an elongated initial syllable ‘ぜーんぜん/zēnzen’ (‘tooootally’), and the childlike ‘じぇんじぇん/jen jen’ (‘totawy’)29 (Joyce et al., 2012, p. 270).

One aspect of spoken language that orthography can be used to represent is the presence of a non-native accent. In Japan, the use of katakana for words usually transcribed in hiragana or kanji has become a device for marking the speech of a non-native speaker of Japanese, both in the telop30 on Japanese television, as well as in manga. In an analysis of manga featuring non-native speakers, Roberston (2015) notes the use of katakana to index less fluent speakers of the language. For example, in one scene where two characters who are speakers of Japanese as a second language greet each other with the phrase ‘konnichi wa’, the higher level student’s utterance in transcribed in standard hiragana (こんにちは), while less fluent

26 Emoji (絵文字) is a Japanese coinage made from the characters for ‘picture’ (絵/e) and ‘letter’ (文字/moji). It is sometimes incorrectly thought to derive from the word ‘emotion’, which is uses in the English lexical blend, ‘emoticon’ (emotion + icon). 27 https://emojipedia.org/stats/ 28 http://blog.oxforddictionaries.com/2015/11/word-of-the-year-2015-emoji/ 29 These approximated translations are those of the current author for the benefit of readers unfamiliar with Japanese. 30 Telop (テロップ/teroppu [from Television Opaque Projector]) are the transcribed utterances of those who appear on Japanese light entertainment and variety shows, and sometimes also in news reportage, which appear at the bottom of the screen. They differ from subtitles aimed at the hard of hearing in being only a selection of the spoken dialogue, and the variety of colours, fonts and sizes utilized for these transcriptions. 48

speaker is shown saying the same utterance in katakana (コンニチハ). Robertson (2015) argues that the use of katakana is only one of various ways script can be used to index non- nativeness, and indeed, the data in the current study highlights the varying associations of katakana in relation to ideas of ‘foreign-ness’ and ‘Japanese-ness’ (see section 7.3.2, pronunciation and script choice).

The orthographic representations of foreign-ness and nativeness were also explored in a study of German punk fanzines 31 conducted by Androutsopoulos (2000, p. 521). He noted the appearance of loanwords in indigenized spelling variants which he terms ‘interlingual spellings’, for example the Germanized ‘Sitt Vissches’ for the punk musician ‘Sid Vicious’. He describes how these indigenized spellings are used for two quite different purposes; in the case of ‘Sitt Vissches’, these spellings can be interpreted as a kind of cultural appropriation of culturally-relevant loanwords, establishing a sense of ownership of referents which have their origins in a foreign culture. On the other hand, the same strategy of indigenized orthography can be used to mark the speech of the uninitiated of the subculture, representing a cultural gap between the in-group and the out-group. Androutsopoulos notes a fanzine article in which an author complains of ‘...having to explain to two natives what a ‘Fähnziehn’ is…’ using the Germanized spelling of ‘fanzine’ to signify the unfamiliarity of this term for the two ‘out-group’ interlocutors. These two distinct and opposing uses of ‘interlingual spellings’, it is argued, have parallels in the use of hiragana for loanwords in Japanese, in being used to signify either cultural ownership of a foreign item, or a Japanized pronunciation of these foreign words. This is discussed in section 6.2.6 (Expressing Japanese-ness), and 6.2.8 (Representing voice quality and pronunciation).

The examples of non-standard orthography described in this section cover a range of different aspects of writing systems and appear in different genres. With those associated with CMC in particular, there are often fears that these unconventional written forms pose a danger to ‘proper’ forms of the language, a debate that is used to humorous effect in the commercials laden with ‘text speak’ described by Jones and Schieffelin (2009). However, the sophisticated linguistic commentary on these commercials by young viewers suggested that this demographic held a nuanced understanding of the stylistic affordances of text-speak, rather than being evidence of language decline.

31 A blend of ‘fan’ and ‘(maga)zine’, a type of alternative print media produced by individuals or small groups, characterised by non-professional, low-tech, low budget production methods, and underground distribution. 49

A report on the use of CMC in the Canadian context also suggested that users of non-standard orthographic forms were aware of when they would be considered inappropriate, with those surveyed reporting being less likely to use emoticons and abbreviations with business associates or teachers (Randall, 2002). Gottlieb (2010b), in writing about the Japanese context of CMC in particular, also suggests that individuals are able to compartmentalize their language practices, and know what kinds of language use are appropriate in different circumstances (Gottlieb, 2010b). This is supported by a letter to the editor of the Asahi newspaper from a 17 year old high school student, who explains that to be ‘for’ or ‘against’ gyaru-moji and other non-standard orthographic devices is beside the point; while she likes to use these creative orthographic representations with her friends, she understands it would not be appropriate for a sign in a public place (Nakamura, 2012).

However, there are cases where the line between appropriate and inappropriate contexts for the use of non-standard orthographical devices is not so clearly defined. The police in the northern Japanese prefecture of Aomori, for example, use twitter to spread notifications of bear-sightings, and have taken to including ‘kao-moji’ (face-letters, or ‘smileys’) in the shape of bears in their tweets. For example, the following string of characters,

‘≡(ö·(ェ)·)ö ~(/;>_<)/’ represents a bear with paws raised chasing a person with their hands in the air, and is accompanied by a written warning not to approach a bear with cubs. While these important safety announcements could be argued to be a genre of text where the use of playful kao-moji should be avoided, the police report that this strategy has been successful in increasing the number of re-tweets, and therefore the spread of these important posts (‘Tsubuyaku’ kuma shustubō jyōhō, 2014).

To conclude this section, non-standard orthographic choices are used in a number of different genres of text, to perform a variety of different functions. In commercial product and business names, they may be used to invoke particular connotations or to give their referents an eye- catching quality. In texts produced by individuals, they may align the writer with a particular social group, as in the case of the maru-moji or gyaru-moji used by young Japnaese women. Non-standard orthography can also be used to approximate verbal and facial cues missing from CMC, or represent an accent or particular voice quality. Although prescriptivists express concern that these non-standard orthographic practices may be evidence of linguistic

decline, many scholars believe that the use of these practices in particular contexts and not others conversely shows a strong understanding of the conventions of writing systems, as well as when these rules can be transgressed (Gottlieb, 2010b; Jones & Schieffelin, 2009; Randall, 2002). In fact, it is because individuals understand the standard writing practices that these examples are clever, interesting, or funny (Crystal, 1998a; Gottlieb, 2010a). The use of hiragana for loanwords can also be thought of as a choice of ‘non-standard orthography’, and was found to be used in a number of genres of text discussed in this section, including product and business names, social media posts, and print-club images. The use of hiragana for loanwords specifically was also identified as a ‘non-standard’ orthographic practice in the young women’s letters analysed by Kataoka (1997), and the print club images investigated by L.Miller (2003), although neither of these authors explored the meanings attached to these forms.

2.7 Language play

Written representations of language that are considered ‘non-standard’ or ‘atypical’ can be found in many cultures, and for a variety of reasons. Sometimes these deviations can be a form of language play used for ludic or humorous effect. However, the ‘deviant form’ cannot stray too far from its original form before it becomes incomprehensible; it must remain within the ‘zone of social meaning’ to be a form of meaningful communication (Sebba, 2012, p. 6) This section will describe examples of language play from Japan and around the world.

Gottlieb (2010a) cites the long history of word play in Japan, including the puns in the 8th century collection of Japanese verse called the ‘万葉集/man’ yōshu’, as well as in more recent literature. She argues that the large number of homophones in the Japanese language provide a rich basis for puns and word play, often utilising substitutions of kanji to add extra layers of meaning. She cites the well-known example of the word for ‘rubbish bin’, conventionally written as ‘ゴミ箱/gomi-bako’ (rubbish-box) being transcribed as ‘御美箱 /go-mi-bako’, or ‘honourable-beauty-box’. This homonymic alternative transliteration is therefore a reminder that placing rubbish in the bin is necessary to keep the area ‘beautiful’.

Nagashima (2006) also comments on the large number of homonyms in Japanese, and provides some additional historical and cultural background to the wide use of ‘sha-re’ 51

(puns) in Japanese. He cites the traditional poetic form ‘短歌/tanka’, consisting of just 31 syllables in five lines, and argues that the use of homonyms, or ‘pivot words’ ( 掛詞 /kakekotoba), in these short poems allows the writer to incorporate two separate meanings in one word or expression. Nagashima also explains the cultural preference for puns over the kind of ‘jokes’ which are heard and retold in the English-speaking world. He argues that telling an obvious ‘joke’ to a fellow Japanese might embarrass the listener if they were not to grasp the meaning, but puns, with both a literal and a humorous interpretation, do not run such a risk; they can be laughed at or safely ignored. Since Nagashima’s main research focus is puns in television commercials, which tend to be verbal, he cites Nomura Masa’aki on how a pun is marked in spoken language: ‘…some part at least of the spoken sentence will draw attention to deliberate artifice by a kind of uneasiness, either in the pronunciation, the intonation, the grammar or the syntax’ (Nomura, cited in Nagashima, 2006, p.77). These verbal strategies described by Nomura can be compared to the use of marked hiragana script for loanwords to highlight the presence of a pun, operating as a similar attention-drawing device, and are discussed in section 7.3.2.

At other times humorous puns are created accidentally; Sunakawa (2008) reports on a text message conversation where the loanword ‘サンキュー/sankyū’ (thank you) is mistyped32 resulting in ‘産油/sanyu’ (oil-producing); the interlocutor humorously provides three more homonyms for the correct input ‘sankyū’; ‘産休’ (maternal leave), ‘三級’ (level three), and ‘三球’ (three balls). The word ‘thank you’ seems to be particularly fertile ground for language play in Japanese; Kallen and Ni Dhonnacha (2014) describe the name of the shop ‘Thank you Mart’ as being a pun on the numbers ‘三/さん/san’ (three) and ‘九/きゅう/kyū’ (nine), as each product sold by this shop costs exactly 390 yen. The data for the current study also contains an izakaya called ‘三九/さんきゅう/sankyū’ (three-nine), which also appears to play on the loanword ‘thank you’.

Another example of an accidental pun relates to the Nestle’s chocolate bar ‘Kit-Kat’, which has become a lucky charm for students in Japan attempting the highly-competitive school and university entrance exams (Okada, 2012). When pronounced with a Japanese (Kyūshū) accent, Kit-Kat sounds much like the phrase ‘きっと勝っと/kitto katto’, meaning ‘you will

32 See section 1.1.1 for an explanation of how Japanese script is input via a keyboard. 52

certainly pass/win’. Nestle Japan’s executives, becoming aware of this practice, started producing special editions of the chocolate bar during the exam-preparation season, and have also collaborated with stationery companies and the national postal service to produce lucky exam merchandise. Okada notes that this example is unusual in that it started ‘naturally among the people’ (Okada, 2012, p. 175) before being taken up by the company, instead of being an intentional marketing strategy.

Other instances of word play appear in marketing slogans or business names; Honna (1995, p. 51) described the Japan Rail (JR) campaign to the snowfields, ‘JR Ski’,33 as representing such an example, as the Japanese word for ‘like’ (好き/suki) sounds similar to the loanword ‘ski’ (スキー/sukī ). Hoffer (1980, p. 10) cites the example of a chain of clubs called ‘友＆愛/yū & ai’; while the meanings of the kanji characters are ‘friends’ and ‘love’, they also mimic the sound of the English phrase ‘you & I’.

Puns can also be found representing English sentences; Hoffer (1990) notes two examples, which he sees as significant as the reading follows the English grammatical order. The first, a sign designed to reduce noise on the train, reads ‘美サイレント/bi sairento’, and contains the kanji for ‘beautiful’ (美/bi) followed by the loanword ‘silent’ (サイレント/sairento). One approximate reading would therefore be ‘silence is beautiful’, but since the reading for ‘美/bi’ is homophonous with the English ‘be’, an alternative reading is the imperative command ‘Be silent’. Hoffer’s second example uses Japanese characters within an English sentence in an advertisement for a tourism campaign to the Izu (伊豆/izu) district. The text reads ‘This 伊豆 Map’, therefore playing on the similarity between the place name ‘伊豆/izu’ and a katakana- ized pronunciation of the English ‘is/izu’. It can therefore be understood as the English sentence ‘This is (a) map’, as well as ‘This (is a) map of Izu’.

Similar bilingual puns have been found in German newspapers. Knospe (2015) investigates bilingual German/English puns, arguing that the ‘gradual bilingualism’ of the German public makes this a frequent source of language play. His examples include a tourism campaign which utilized the pseudo-word ‘Cool-tur’, combining the English word ‘cool’ with the German culture ‘kultur’; which attempted to give a ‘cultural’ destination a ‘cool’ image. He

33 Honna does not describe in which script(s) this phrase was written. 53

also cites a newspaper article titled ‘MISS-Stimmung’, highlighting the English word ‘Miss’ within the German word ‘Missstimmung’ (discord), which describes conflicts between contestants in the ‘Miss World’ pageant. Knospe (2015, p.177) notes that ‘additional typographical means’ may be used to signal the presence of a pun; the two examples cited above both use a hyphen where one would not normally occur, and the latter example additionally employs capital letters for the English component ‘MISS’. It is argued that the use of hiragana for loanwords sometimes functions as a similar ‘additional typographical means’ to signal a pun, to draw the readers’ attention to the marked form and encourage them to decipher it. In this way, it echoes the ‘kind of uneasiness’ in intonation used to mark the presence of a pun in spoken language cited in Nagashima (2006) above.34

Renner’s (2015) study of language play focuses on lexical blends from many languages, not including Japanese, although many such examples were found in the current study (see section 4.1.6). Renner argues that all instances of lexical blending, such as the terms ‘netiquette’ (‘net’ + ‘etiquette’), and ‘chillax’ (chill + relax) are examples of wordplay due to their very formation process. Renner introduces five features which he argues increases the ‘wordplayfulness’ (2015, p. 119) of an item, of which the two most relevant to the examples of lexical blends in the corpus are described here. Firstly, ‘structural transgressions’, which relate to ‘the violation of structural well-formedness rules’ (Renner, 2015, p. 126) concern illicit phonotactic combinations and grammatical compounding. Loanwords in hiragana can be said to be a kind of ‘structural transgression’, as katakana is the conventional script for writing loanwords, and therefore perhaps better described as a ‘graphological transgression’. This also relates to the second aspect of blending most relevant to the currentstudy, described as a ‘graphic play on words’ (Renner, 2015, p. 127). Graphic word play involves blends which are only identifiable in their written forms, being homophonous with other existing words, such as ‘pharming’, defined as the ‘production of pharmaceuticals from genetically altered plants or animals’ (Renner, 2015, p. 128). Some of the loanwords in hiragana collected for this study also contain this type of word play, which is only perceptible in the written form, and are described in section 4.1.6. The role of hiragana loanwords as markers of language play is discussed across all layers of the findings in section 7.3.2.

34 Nagashima’s (2006, p. 81) data also includes a ‘bilingual pun’ in the same fashion as the ones cited by Knospe, in a jingle containing the line ‘We 好き whiskey’ (Pronounced wī suki wuisukī /We like whiskey). This example also has parallels to the alliterative ‘JR Ski’ campaign described by Honna (1995) above. 54

2.8 Typography

While the Japanese language is unique in having a variety of scripts, each with established but flexible domains of use and a variety of connotations, it is helpful to look at how typefaces (families of fonts) are used in other languages as a parallel example. Like the choice of script in Japanese, the choice of typeface can add an extra layer of meaning to the linguistic information encoded in the word. Serif fonts, for example ‘Times New Roman’, are felt to have a more formal or traditional feeling, while sans serif fonts such as ‘Arial’ can give a text a more modern or casual look. Of course, creators of Japanese texts can also choose from a variety of typefaces besides their choice of script; these choices may support or re- contextualize the connotations associated with the script used as highlighted by the case studies described in section 4.2.

While typography is not often considered relevant to the field of linguistics, some scholars have argued that these disciplines have much to offer each other. Crystal (1998b), for example, suggests the possibility of a ‘typographical linguistics’, and Walker (2000) has argued that both fields could benefit from exchanging methods and approaches. The use of font as a carrier of meaning has been noted by van Leeuwen (2005, 2006) who also suggests the possibility of a semiotics of typography, noting that much of the cohesive ‘work’ previously done by language is now being realized through non-linguistic resources such as font (2006, p.139). This is consistent with the understanding of texts as ‘multimodal’; as information is conveyed not only through the words chosen but through the use of other devices such as colour, layout, image and framing (see section 2.2, multimodality). By considering a range of typefaces, van Leeuwen (2006) constructs a SFL-style system network diagram for the distinctive features of fonts, for example weight, slope, and curvature, opening up the potential for a ‘grammar’ of typography.

Stöckl (2005) also addresses the question of a ‘grammar of typography’, suggesting that compared to language, typography has a ‘weak grammar’ (p.209), but this would also seem to be true of other non-language semiotic systems such as layout and colour, and does not reduce the importance of the information such a system encodes. Stöckl takes up Kress’ (2010, 2011) notion that for a semiotic resource to be considered a ‘mode’35 in a multimodal text, it must be able to demonstrate the three metafunctions outlined by Halliday, the

35 Stöckl uses the word ‘code’, sometimes in combination with ‘mode’ (‘mode/code’, p. 205) but the term ‘mode’ is more consistent with modes as described in section 2.2. 55

referential, the interpersonal and the textual. Stöckl argues that typography works on the ideational level by commenting on or reinforcing the message of the text, as shown by the connotations of serif and sans-serif fonts described above. Typography also operates on the interpersonal level; it can indicate the nature of the communicative contact between writer and recipient, or refer to the emotional state of the writer. For example, some typefaces suggest handwriting and therefore may connote a personal letter, while words or phrases in capital letters in the body of a text are generally felt to be ‘shouting’. Finally, typography has a textual function as it serves to visually structure a message and its constituent parts, what Walker (2000) refers to as the ‘articulation’ of a text. Headings in a document are often marked by larger and more creative font, for example.

In some situations, the connotative power of a typeface may even carry more meaning than the words it depicts. This is suggested by Huang and Archer’s (2014) research on English translations of Japanese comic books. They noted that many onomatopoeic expressions formed part of the image rather than being contained in speech bubbles, and were therefore left untranslated in the English versions. Huang and Archer argue that this practice is unlikely to pose a problem for readers unfamiliar with Japanese, as the meanings of these expressions are expressed via the choice of lettering, for example the jagged or puffy shapes, rather than the actual words.

The connotative power of particular typefaces has not been lost on those in the marketing industry, with many companies taking the choice of a font to represent their business or product very carefully. This care is well justified, as Doyle and Bottomley (2006) found that not only do people perceive appropriate and inappropriate matches between hypothetical combinations of product names and fonts, but the connotations of a typeface (such as positive or powerful) are transferred to the actual product (Doyle & Bottomley, 2009). Schorn, Brunner-Sperdin and Ploner’s (2014) research on the effect of font formatting on consumers’ perceptions of logos also supports this, with those in bold fonts making companies appear more competent and secure.

Typography has also been used in branding tourist districts, representing them as authentic sites for cultural experiences. Coupland and Garret (2010) report on the use of ‘Celtic font’ as one of many visual markers of ‘Welshness’ in the Welsh colonial town of Patagonia in Argentina. Moriarty (2013) notes a similar trend in the use of ‘Gaelic’ or ‘Celtic’ font, as well as the Irish language, being used to connote ‘tradition’ in the linguistic landscape of the Irish

tourist town of Dingle. In Japan, the use of a bold calligraphic font is often used in advertisements promoting tourism in areas associated with Japanese history and tradition. Bartal (2013) notes its use in a campaign for Mount Kōya, for example, which is well-known in Japan as the seat of Shingon Buddhism.

Besides the typeface itself, superfluous diacritics have often been used to create a foreign or exotic brand image, described by Jaworski (2015, p. 217) as ‘globalese’, as mentioned in section 2.6, in relation to the much discussed American ice-cream brand, ‘Häagen-Dazs’. The motivations ascribed to this particular umlaut include producing an air of foreignness or ‘German-ness’ (Spitzmüller, 2012), connotations of the ‘Nordic-esque’ (Jaworski, 2015), or a simple attention-getting strategy (James, 2014). These are not far from the intentions of the Jewish founder of the brand, Reuben Mattus, who created the fictional name hoping it would sound Danish, as both a marketing strategy and a nod to Denmark and its sympathetic treatment of Jews during World War II (Nathan, 2012).

Context is vital in understanding the connotations of a typeface or typographical accent such as a diacritic mark, leading Spitzmüller (2012, p. 256) to argue that ‘graphic ideologies are both discursively rooted and floating’; in other words, these meanings form part of the collective knowledge of a culture, but are also subject to negotiation in different times and places. This reflects the social semiotic understanding of meaning making described in section 2.1, which informs the basis of this study, and also the importance which context plays in this framework.

Spitzmüller (2012, 2015) reports on the use of ‘blackletter’ or ‘Gothic’ style fonts in particular, and diacritics such as the umlaut. He points out that these graphic variants do not ‘have’ a meaning, but a social meaning is ascribed by the reader, although some meanings are more widely ascribed than others. What is particularly interesting about blackletter typefaces is the variety of associations ascribed to them. Within Germany, they can represent tradition; nationalism or even Nazism;36 while globally they are further used to index heavy metal music culture, and more recently the hip hop music scene (Spitzmüller, 2012). The names of heavy metal bands in particular are often transcribed with what has become known as the ‘heavy metal umlaut’ or ‘röck döts’, for example ‘Motörhead’ and ‘Mötley Crüe’. These

36 Hitler was actually persuaded that gothic (Fraktur) types were of Jewish origin, and ordered all documents to be printed in roman typefaces (Flood, 1996). However, the prevalence of gothic types in Hitler’s Germany overall has led to the association of this typeface with Nazism. 57

diacritics do not affect the pronunciation of the bands names, but are simply stylistic additions representing their alignment with the genre of heavy metal music.

Spitzmüller reminds us that terms such as ‘text’ and ‘context’ are only meaningful when there is a ‘reader’ involved, as an audience is essential in making the connections between form and context:

Graphic forms and place contextualize each other…they only do so, however, because there is a reader with a specific graphic, local and historiographical ‘knowledge’, a reader who assigns social value, personae and practice both to forms and to locations and thereby aligns the forms with the location.

(Spitzmüller, 2015, p. 137)

Like the range of connotations associated with blackletter typefaces, the range of meanings ascribed to the hiragana script, such as being ‘childlike’, ‘traditional’ or ‘feminine’, are quite diverse (see section 2.4.2). However, as Spitzmüller reminds us, the orthographic forms themselves do not have instrinsic meanings, and thus the role of both the reader and context in which these forms appear must be taken into account in interpreting the meaning of a graphic sign. An investigation of these factors is therefore an integral part of the current research design, described in section 3.3.

Besides established typefaces, ‘pictorial typography’, defined as cases where ‘the medium and its techniques/instruments become obvious to the reader’ (Stöckl, 2005, p. 208) are a common feature of advertising. Stöckl cites an example of the acronym ‘D.I.Y’ (do it yourself), in which the letterforms appear to be pieces of wood nailed together, with screw heads for full stops. The creative ‘tittles’ (the dots on letters such as ‘i’ and ‘j’) can also be seen as a variation of pictorial typography, when replaced with icons such as hearts or stars (Jaworski, 2015).

Bringing attention to the medium of the typeface, Stöckl (2005) argues, can give the text authenticity, allowing the viewer to ‘encounter texts are real objects associated with their original contexts’ (p.208). Calligraphy, having a long tradition in Japan, is a kind of pictorial font often used in marketing texts for items or places associated with Japanese history and tradition, such as Mount Kōya as described above. This reinforces Stöckl’s suggestion that pictorial fonts, such as calligraphy-style ones, can lead us to associate the texts with their

original contexts, for example the calligraphic tourist poster with the art as practiced by the monks on Mount Kōya. The use of calligraphic typefaces to reinforce traditional connotations for hiragana loanwords is discussed in section 4.2.2 (Case study 2: ‘Honey Lemon’ Drink).

Japanese letterforms, of course, can also be pictorial, and Shelton and Okayama (2006) argue that the ideographic kanji characters in particular blur the boundary between pictures and writing. The pictorial nature of the kanji used in Japanese writing also provides rich ground for visual word play; Shelton and Okayama (2006) note many examples, including an advertisement with the character ‘to run’ (走) being made to look like a person running; and a poster featuring the character ‘賞’ (prize), where the element of the character meaning ‘seashell’ (貝) in the lower half of the character had been replaced with a realistic illustration of a nautilus shell. These examples, the authors argue, are more than simple rebuses, they highlight the ambivalent status of Japanese kanji characters between pictures and language. Bartal (2013) provides more examples which play with the pictorial nature of kanji, such as the poster produced by Yamashiro Ryuichi, which makes use of the iconic forms of the kanji for ‘grove’ (林) and ‘forest’ (森); these tree-like characters fill up the space giving an impression of something more like a picture than a text. Examples such as these, which invite ‘viewing’ rather than ‘reading’, reflect Shelton and Okayama’s (2006) suggestion that Japanese characters occupy a position on the spectrum between images and letters.

As stated at the beginning of this section, an investigation of the connotative uses of typography is at best a parallel case to the associations produced by the Japanese scripts. However, there is at least one example of a particular typeface being used to mark foreign words, in the same way katakana is used for loanwords in Japan. Flood (1996, p. 189) notes that in 17th century Germany, nationalistic sentiments in typography lead to the ‘ugly habit’ of printing not only foreign words, but foreign elements within German words, in roman script while the rest of the text remained in gothic. This is, in fact, very similar to what happens when foreign loanwords are combined with Japanese morphological endings, such as the terms ‘ググる/guguru’ (to google), ‘ディスる/disuru’ (to ‘dis’, to disrespect), and ‘ナウい/naui’ (‘now’, trendy) described in section 2.5.2 above. In these cases, the foreign elements are marked by the katakana script, while the native Japanese elements appear in hiragana. Despite combining different scripts, these words do not have the clumsiness of Flood’s (1996) German examples, perhaps because it is quite common for a Japanese word to contain a kanji stem and a morphological suffix in hiragana.

While it is hard to imagine how a text with foreign morphemes marked by typefaces as different as roman and gothic as described by Flood (1996) would look today, style manuals such as those of the A.P.A and M.L.A often suggest that foreign words be set in an italic typeface to signal to the reader that they are not a misprint. This quite banal practice is possibly the closest modern example of how typography is used to distinguish words of foreign origin, in a similar way to how katakana is used to distinguish foreign words in the Japanese writing system.

2.9 Translingualism

English as it is used and spoken in the world is a topic of interest for those from fields as diverse as linguistics, sociology, literature and business studies. While Kachru’s model of inner, expanding, and outer circles of English use was influential in establishing the study of varieties of English (Kachru, 1986), it is important to recognise that the way English is being used in the world today is increasingly trans-national and continually evolving, and therefore cannot fit neatly into geographical frames of use (Blommaert, 2012; Jenkins, 2003). However, the global spread of English, and its interaction with other languages used in the community, has proved a slippery term to define; Blommaert (2012) points out that alternatives such as ‘multilingualism’ suggest languages co-existing side by side, rather than the mixing and mutual influences that flow between languages. Other terms include ‘metrolingualism’ (Otsuji & Pennycook, 2010); and ‘translingualism’ (Canagarajah, 2013), for example. The latter is the one which is drawn upon in the current research project, as the prefix ‘trans-’, suggesting ‘across’, ‘beyond’ or ‘through’ language, is felt to best capture the sense of movement involved in the way English, and other languages, are being used and adapted in diverse locations.

Two domains of use most relevant to the current research project are computer-mediated communication (CMC) and marketing. The use of English by non-native speakers in social media exchanges has been the focus of investigation in countries such as Thailand (Seargeant & Tagg, 2011); Nepal (Sharma, 2012); Mongolia and Bangladesh (Dovchin, Sultana, & Pennycook, 2016); and Japan (Sunakawa, 2008). In many of these contexts, the use of English is used as a marker of an international outlook, referencing global youth culture and ‘coolness’.

Similar connotations are drawn upon when English is used for marketing purposes in non- English dominant countries, for example in the names of products and business, or in advertising. Piller notes that ‘English has largely become a non-national language and has been appropriated by advertisers in non-English-speaking countries to index a social stereotype’, one associated with ‘modernity, progress, and globalization’ (2003, p. 170). Earlier, Piller (2001) noted that more than 70% of the German advertisements she investigated incorporated a language other than German, usually English. In Japan, English/rōmaji words are frequently found in marketing texts such as advertising slogans and brand names (Honna, 1995; Hyde, 2002; Loveday, 1996; Sase, 2009; Stanlaw, 2004), and are frequently cited to be giving their referents an international, modern, or cool image. James further notes that it is ‘the look of the English, its appearance, its very graphicity as opposed to its actual linguistic content that signals these social meanings’ (2014, p. 20, emphasis in the original). This echoes the register of ‘globalese’ described by Jaworski (2015), where the ‘look’ of diacritics connoted foreign-ness or exoticism as part of business or brand names.

While the visual presence of English/roman letters described by James (2014) is likely to be one factor in their frequent use in such contexts, other factors such as their potential in creating bilingual puns as described by Knospe (2015) goes beyond the ‘look’ of these letters, and many examples of bilingual English/Japanese puns were described in section 2.7. The loanwords in hiragana which are the focus of this study also pose an interesting counterexample to this argument for the ‘look’ of English being paramount, as they are (usually) English words which are stripped of their air of foreign-ness and exoticism by being written in the script associated with Japanese words, rather than in katakana or rōmaji/English, scripts associated with internationalism and modernity.

There are other examples of the translingual use of English use in Japan; as early as 1986, for example, Haarmann (1986) noted that English was used in fashion magazines even in connection with items associated with Japanese culture such as kimonos. More recently, Kallen and Ni Dhonnacha (2014) argue that English is ‘repurposed’ in Japan, citing examples such as the ‘春 mac/haru mac’ (Spring Mac) campaign from McDonalds Japan, which referenced the importance of the seasons in Japanese culture with the ubiquitous ‘mac’ of the western fast-food chain. This ‘cultural hybridity’ (Kallen & Ni Dhonnacha, 2014, p. 31), they suggest, is a typical feature of globalization.

The idea of a repurposed, localized use of English is the focus of Pennycook’s ‘Language as a local practice’ (2010), in which he looks at the ways the English language is being reclaimed and re-localized by those who speak it as a second or foreign language. Pennycook argues that the appropriation of the English language is not imperfect copying from a native- speaker standard, but that through the act of repetition, new local meanings situated in local practices are being created. He further suggests that the global spread of English cannot simply be viewed as a homogenising force, but that this ‘repetition’ of language necessarily involves ‘an act of difference, relocation, and renewal’ (2010, p.36). This finds many parallels in the adoption of foreign words into Japanese, leading scholars such as Stanlaw (2004) to problematise the term ‘loanword’ altogether, as described in section 2.5.4. Pennycook emphasizes the idea of divergence from a common core in the term ‘fertile mimesis’, ‘copying that goes slightly wrong, repetition that is something else, sameness that is difference’ (2010, p.37).

It is clear that viewing the use of English and/or loanwords in Japan as inaccurate attempts to copy ‘norm-providing’ native speaker models is to miss the creative, localized ways in which these words are being used. Language is not simply ‘copied’ or ‘borrowed’, it is re-purposed, reclaimed, and renewed in local, social situations:

…it is not that we use language as a pre-given entity in context, but rather that we produce language in our repeated local activities. Furthermore, these activities are parts of bundled practices, and as such they are always social, always historical, and always local.

(Pennycook, 2010, p.46)

The social, the historical, and the local are key concepts in interpreting loanwords in hiragana in this study; their use as part of their social contexts such as advertising or CMC is investigated against the background of the historical associations of each script as described in section 2.4. Furthermore, the alternatives provided by the Japanese writing system make this necessarily a very local act of language use, as the Japanese situation is unusual in having a variety of ‘native’ script options (katakana, hiragana, and sometimes kanji) for these imported terms, in addition to their native English/rōmaji transcription. Loanwords in hiragana therefore represent a significant aspect of translingual language practice, a theme which will be drawn upon throughout the current research project, and revisited in the discussion chapter, section 7.2.

2.10 Conclusions from the literature review

This chapter has discussed theoretical as well as empirical research from a number of disciplines that are relevant to the current project’s focus on loanwords which are written in hiragana.

Social semiotics, and the related fields of multimodality and SFL, all share an emphasis on the social basis of meaning making, as well as the importance of context in understanding the meanings being made. Semiotic resources such as script do not simply ‘carry’ meanings; these meanings are mediated through human agents who interact with these texts. For this reason, it was important to investigate the opinions of native Japanese language users, as the intended audience of these texts, in understanding the meanings created by loanwords in hiragana.

In addition, an understanding of multimodality acknowledges how texts convey meaning through many ways other than just the choice of words on a page. Modes such as image, layout, and colour have been recognized for the meaning-making work they do, and a multimodal framework allows script choice in Japanese writing to also be considered a ‘mode’. Loanwords in hiragana represent a marked use of this semiotic resource, and therefore the particular connotations and meanings associated with this script are brought into focus in the texts investigated in this study.

Systemic functional linguistics offers a model of language from the smallest units of meaning, graphemes; through lexico-grammar structures, to discourse-semantics, as well as illustrating the relationships to the contexts of situation and culture. While an SFL analysis generally starts with language, and looks ‘out’ to context and ‘in’ to lexico-grammar structures, the current study takes the smallest unit of meaning, a particular graphemic choice (hiragana) as its starting point. Beginning with script choice, the current study looks both ‘out’ to the discourse semantic functions loanwords in hiragana are performing, as well as ‘out’ to the context of situation and culture to describe where this marked language use is occurring.

The literature on the Japanese writing system and on loanwords in particular serves to situate this study in its historic and linguistic context. While script use is relatively prescribed in Japan today, it is important to remember that loanwords have been represented in many different ways throughout the course of Japanese history. In addition, they have been shown

to fulfil a variety of functions within the language, and this research aims to demonstrate the specific functions which loanwords in hiragana can perform.

A review of ‘non-standard’ orthography both in Japan and around the world has shed light on the diverse contexts in which it is found and the motivations behind selected examples. Marked language can show group solidarity in personal communications, or be used to highlight particular pronunciation or voice quality in written texts. It can also be used to draw attention to the presence of language play, as the large number of homonyms in Japanese make the language fertile ground for puns, and are reflected in texts as diverse as classical literature and contemporary marketing slogans. Of particular relevance to the current study are the ‘additional typographical means’ used to highlight an instance of word play (Knospe, 2005, p.177) where typographic devices such as capital letters and hyphens draw attention to the site of language play, as it is argued that the use of hiragana for loanwords may be fulfilling a similar function in some cases. ‘Non-standard’ orthography can also be used to distinguish a particular product or business when used as a branding device, which also a potential function of loanwords in hiragana.

While Japan is unusual in that it has a variety of scripts to choose from, each with different histories and associations, a review of the typography literature was useful in demonstrating a parallel context in which letter forms can be pictorial, and influence the associations of a piece of text. An examination of the functions performed by typography is another way of problematising the distinction between a ‘linguistic/verbal’ and ‘visual’ messages, as represented by the case of loanwords in hiragana.

Finally, the changing status of English in the increasingly globalized world represents the greater context in which loanwords in hiragana are occurring. As English becomes increasingly used around the world, indigenized varieties gain acceptance, and embrace their local traits with more confidence. Hiragana, as a script associated with native Japanese words (和語/wago) is a strong marker of Japanese-ness, and the fact that foreign words sometimes appear in this script can be seen as a significant example of ‘translingualism’; specifically, the localization of these foreign words within the Japanese context.

Given the theoretical concerns outlined above, this study seeks to address the research questions stated previously in section 1.3, namely to describe the typical traits of loanwords in hiragana, to understand the contexts in which they are often found, and to explain why

these words appear in marked hiragana, rather than the conventional katakana script. The research design and methodology that have been developed to address these research questions are outlined in the following chapter.

Chapter 3: Research design and methodology

This chapter will describe the research design and methodology used to address the research questions proposed in the introduction (section 1.3). While Jewitt (2011, p.26) warns that one of the weaknesses of multimodal research is its subjectivity, she suggests strategies to avoid this including using a range of data; utilising high levels of participant involvement; and ensuring that the meanings described are linked back to context and social function. These suggestions have guided the design of the current research project, which will be illustrated in the following sections.

Firstly, a brief explanation of relevant terminology is given. This is followed by the epistemological background of the study, which justifies the choice of a mixed-methods approach, and describes its application in the current research project. An overview of the research design as a whole is then given, before the data collection and analysis procedures for each of the layers are described. These include a corpus analysis of a collection of texts containing loanwords in hiragana (layer 1), a set of case studies of individual texts (layer 2), an online survey (layer 3), and a series of interviews (layer 4).

3.1 Terminology

As mentioned in the introduction, it is acknowledged that the term ‘loanword’ is not unproblematic, as the words are not ‘loaned’ to be given back, and they are not unequivocally ‘foreign’ words as the term suggests. However, ‘loanword’ will be used as it is felt to be the best translation of ‘外来語/gairaigo’, and is used to refer to words that have their origins in other languages, usually English, but also from languages such as Portuguese and French, which are widely used and understood in Japan, and are commonly written in the katakana syllabary.

As this study takes a social semiotic approach to understanding how meanings are made, terms describing signs and their forms should be interpreted within this context. The author follows Kress (2010) in believing even ‘Saussurean’ terms such as ‘signifier’ and ‘signified’

can be used as part of a post-structuralist, social semiotic understanding of the way meanings are made (see section 2.4).

Within a social semiotics framework, it is understood that signs are ‘made’ by individuals, and that the word ‘use’ can suggest that connections between form and meaning are pre- established and structuralist. However, as above, we follow Kress in believing that ‘making’ is part of ‘using’ in social semiotics, and so when the term ‘use’ appears in this thesis, it is to be understood in this manner.

‘Choice’ is a similarly contested item; both social semiotics and SFL consider choice to be constrained by context, other factors such as power may be relevant too. This means the same range of ‘choices’ may not be available to all people in all situations. When ‘choice’ is mentioned in this thesis, it is to be understood as meaning ‘choice within a particular context’.

Finally, as some of the texts discussed in the following sections contained a single loanword in hiragana, while others contained up to four, expressions such as ‘the number of texts containing (a) loanword(s) in hiragana…’ would be the most accurate wording. However, for ease of reading all such examples will be referred to as ‘texts containing a loanword in hiragana’ unless cases of multiple words are being described specifically.

3.2 Epistemological background

One of the first decisions to be made in a research design is whether the research is to be based on an objectivist or constructivist epistemology (Bloomberg & Volpe, 2015). Knowing which of these epistemologies the researcher identifies with will indicate the research paradigms relevant to the study, and suggest methodologies to best achieve the goals of the research project.

An objectivist position holds that reality exists independently of consciousness; research aligned with this epistemology sets out to find an objective truth. A research paradigm related to the objectivist position is ‘positivism’, which holds that knowledge is discovered by scientific observation, and empirical enquiry (Gray, 2009). Results of positivist research are usually quantitative, and presented as facts (Bloomberg & Volpe, 2015); they aim to be generalizable to larger populations and replicable by the same or other researchers.

The idea of a single ‘truth’ is rejected by constructionists, who argue that meanings are created (constructed) by the way individuals interact with the world. There can therefore be multiple interpretations reflecting the same object or phenomenon (Gray 2009). The related research paradigm for constructionists is ‘interpretivism’, whereby reality is conceived as socially constructed, and therefore context-specific. Researchers interested in investigating how different individuals understand the worlds in which they live would be likely to adopt a qualitatively oriented approach to research design, so that the nature of individual differences can be explored in depth. Knowledge from an interpretivist perspective is not perceived in positivist terms as an objective reality, but rather as a subjectively constructed phenomenon (Bloomberg & Volpe, 2015). Social semiotics (section 2.1), with its emphasis on both the context of the signs as well as the backgrounds of the readers, is closely aligned with an interpretivist epistemology.

While the tension between quantitative research approaches from a positivist stance, and qualitative research approaches from an interpretivist stance continues, a third approach called ‘mixed methods’ attempts to combine elements of each. Despite claims of their fundamental irreconcilability (see Gray [2009] for examples), combining qualitative and quantitative methodologies through the use of a mixed-methods approach can provide both breadth and depth of understanding with respect to the phenomena under investigation. In other words, mixed methods research allows ‘between-methods triangulation’, as the biases and shortcomings of one method can be counterbalanced by the inclusion of another, alternative method (Flick, 2006).

A mixed-methods research design which contained elements of both quantitative and qualitative research approaches was therefore chosen for the current study as it would provide the most comprehensive understanding of issues of the representation and interpretation of loanwords in hiragana. With no previous research having been conducted on this phenomenon, it was vital to approach this topic from various angles, to obtain the fullest understanding of this practice. Tashakkori and Creswell (2007) outline a number of ways that a study can be considered as consisting of ‘mixed methods’, for example having two types of research questions (with qualitative and quantitative approaches), two types of data collection procedures (eg. interviews and surveys); or two types of data analysis (statistical and thematic). All of these elements are included in the current study, and will be described in the following section.

3.3 Design of the current research project

As stated in the introduction, the current research project aims to investigate the phenomenon of loanwords which are written in hiragana, rather than katakana, which is the standard script used for these words.

The following research questions are proposed:

1. How can loanwords in hiragana be characterised?

2. In what contexts are loanwords in hiragana usually found?

3. Why do loanwords appear in hiragana?

The model of language proposed by Halliday (described in Eggins, 2004) provides a useful contextualization of the research questions, and is reproduced again below in figure 3.1. The hiragana characters, or graphemes, in combination with particular lexical items, loanwords, form the focus of this study.

Figure 3.1: Language realising context, adapted from Eggins (2004, p.112)

The first research question (‘How can loanwords in hiragana be characterised?’) can be partially addressed by identifying features such as the grammatical classes of these words, the number of words occurring consecutively, their referential meanings, and what source languages they come from. Identifying particular loanwords which frequently occur in

hiragana also contributes to answering this question. These aspects relate to the lexico- grammatical level of the above diagram. At the innermost level, another graphological feature of loanwords in hiragana can be examined, which contributes to describing typical features of these words. This relates to the lengthening of vowel sounds; namely whether they are lengthened with an extra vowel, in the manner of Japanese words in hiragana, or with a bar as loanwords in katakana usually are.

In order to investigate these graphological and lexico-grammatical features of loanwords in hiragana, a corpus of texts containing these words was assembled and examined. Descriptive statistics were the main form of data output from this part of the analysis, making this a quantitative aspect of the overall study. This part of the analysis is referred to as ‘Layer 1’, and is explained in detail in section 3.4.1.

The second research question is designed to ascertain the contexts in which loanwords in hiragana are usually found. This relates to the outermost triangle in diagram 3.1 above, that relating to genre. SFL defines genres as sets of register variables which co-occur regularly in a culture and have become sedimented and are thus easily identifiable by members of that culture (Eggins, 2004). The genres of the texts collected for the corpus described above (layer 1) were coded for their respective genres, such as ‘product packaging’ or ‘business signage’, in order to describe the contexts in which loanwords in hiragana were typically found.

The third research question addresses the reasons loanwords appear in hiragana. This relates to the ‘discourse semantics’ level of the diagram in figure 3.1 above, which refers to the ideational, interpersonal and textual functions language is used to perform. Corpus based studies such as those conducted by Joyce et al. (2012) are useful in describing orthographic variation, but as the authors note, they do not address ‘the complex sociolinguistic factors that might influence the selection of one orthographic representation over another’ (Joyce et al., 2012, p. 274).

One method of investigating the use of hiragana for loanwords was the use of case studies of individual texts from the corpus (layer 1). Case studies allow a deeper understanding of a given process or text than other methods such as surveys can provide (Woodside, 2010). However, the purpose of the analysis is not to generalize to other cases, but rather to allow a rich description of the topic in order to understand its complexity (Bloomberg & Volpe, 2015). Case studies provided an in-depth analysis of the semiotic resources used in a text to

understand how loanwords in hiragana were functioning as one part of a multimodal whole. The case study data therefore complements the quantitative descriptive statistics of the corpus produced in layer 1, which could not take these individual contexts into account. The case studies are referred to as layer 2, and are described in section 3.4.2.

It was also necessary to examine native speakers’ opinions on the perceived reasons for the use of hiragana for loanwords. As Jewitt (2011) suggests, participant involvement increases the validity of studies in multimodality; and Spitzmüller’s (2015) criticism of semiotic landscape studies that only focus on the texts, and not the people who interact with them, is noted. Spitzmüller reminds us that the text and context only make meaning when there is a ‘reader’ involved, as an audience is essential in making the connections between form and context; ‘Graphic forms and place contextualize each other…they only do so, however, because there is a reader with a specific graphic, local and historiographical “knowledge”, a reader who assigns social value, personae and practice both to forms and to locations and thereby aligns the forms with the location’ (Spitzmüller, 2015, p. 137).

The involvement of members of the Japanese public in interpreting the meanings behind loanwords in hiragana also aligns with the social semiotic understanding of meaning making that is central to this research, and for this reason it was important to look beyond the texts themselves to their intended audience. The opinions of native Japanese speakers were gathered via two data sources, an online survey (layer 3) and interviews and focus groups (layer 4).

The main purpose of the online survey (layer 3) was to gather opinions from a large number of respondents on three texts containing loanwords in hiragana. These texts were selected from those used in the case studies (layer 2) in order to triangulate the findings between these two layers. The survey findings represent quantitative data, and illustrate common associations of each of the Japanese scripts, as well as the kinds of meanings suggested by the use of hiragana for loanwords. The survey is explained in detail in section 3.4.3.

To complement the survey data, interviews and focus groups (layer 4) were also conducted, which collected data from a smaller number of people about a larger number of examples. Bestor. T, Steinhoff, & Bestor V.’s ‘Doing fieldwork in Japan’ (2003) was consulted to ensure these sessions were conducted in a manner appropriate to the culture. The design of the questions used in the interviews and focus groups drew on the findings from the previous layers, namely the corpus study, case studies, and survey. This layer of analysis allowed for

more extensive responses than the survey, and, in the case of focus groups, for the perceptions of loanwords in hiragana to be discussed and explored amongst the participants. The interviews therefore generated qualitative data, and were used to complement the quantitative findings of the survey. Details of the interview design and procedure can be found in section 3.4.4.

The final research question, ‘why do loanwords appear in hiragana?’ was therefore approached in two main ways, quantitatively through the online survey (layer 3), and qualitatively through the interview (layer 4). The case studies (layer 2) also contributed to answering this question.

To summarize this section, the current study uses four different approaches to investigate loanwords in hiragana; a corpus analysis, four multimodal case studies, an online survey, and a series of interviews and focus groups. The strengths of a mixed method design include the fact that statistics drawn from large numbers of texts or participants can illustrate general trends, while qualitative analyses allow a deeper understanding of the topic as important details do not go unnoticed. Furthermore, the research design approaches the phenomenon of loanwords in hiragana from two distinct angles; firstly by investigating the texts themselves, and secondly by investigating the opinions of the population who interacts with those texts.

The mixed-methods approach to the multiple data sources can be summarized in table 3.1 below:

Table 3.1: Summary of research design

QUANTITATIVE DATA QUALITATIVE DATA

Layer 1 Layer 2 TEXTS Corpus analysis Case studies

Layer 3 Layer 4 AUDIENCE Survey Interviews and focus groups

While the preceding paragraphs have linked each layer specifically to one or more of the research questions, the ability to triangulate findings between data sources was also an important consideration in the research design. For example, three of the four case study texts

(layer 2) were used in the online survey (layer 3), and the findings on the typical features of loanwords in hiragana gained from the corpus (layer 1) could be compared with participants’ reactions when these examples were produced in the interviews and focus groups (layer 4). For this reason, the term ‘layers’ and not ‘steps’ has been used, because rather than being a linear progression from one data set to the next, the corpus analysis, case studies, survey, and interviews and focus group are linked to each other, allowing each data set to be interpreted in relation to the other layers of the research design.

3.4 Data collection and analysis

The four layers of analysis used in the study will be explained in detail in the following sections. For each layer, a justification for the particular research tool will be given, followed by an explanation of the data gathering procedure, and a description of how the data was analysed.

3.4.1 Layer 1: Descriptive statistical analysis of loanwords in hiragana

The corpus analysis, referred to as layer 1 of the research design, was used to provide quantitative data that was gained from an analysis of a collection of texts containing loanwords in hiragana. The place of layer 1 in the research design as a whole is indicated in table 3.2 below.

Table 3.2: Corpus analysis in the context of research design

QUANTITATIVE DATA QUALITATIVE DATA

Layer 1 Layer 2 TEXTS Corpus analysis Case studies

Layer 3 Layer 4 AUDIENCE Survey Interviews and focus groups

The descriptive statistics gained from the photo corpus were most useful in addressing research question 1, ‘How can loanwords in hiragana be characterised?’ Corpus studies can be used to describe patterns of form, but moreover are capable of highlighting how form and meaning are inseparable (Cheng, 2012). This connection between form and meaning is central to understanding the use of hiragana for loanwords.

Corpus studies can reveal general trends in the ways a particular language feature is being used, and are therefore useful in describing typical features of the object of investigation, in this case, loanwords in hiragana. This quantitative analysis of the corpus can describe features such as the domains loanwords in hiragana typically occur in, how many words usually occur at one time, and what foreign languages are represented in hiragana. This baseline information was used to inform the selection of texts for the following layers of analysis, including the case studies (layer 2), survey (layer 3) and interviews (layer 4).

The corpus collected for analysis as layer 1 of the research design contained 497 texts with a total of 628 loanwords in hiragana, as some texts contained more than one loanword. The data collection procedure and analysis for the descriptive statistical analysis of the corpus is described in the following sections.

Layer 1: Data Collection

To identify the typical features of loanwords in hiragana, it was first necessary to gather a large corpus of texts which contained these words. A pilot study was used to identify potential sites of loanwords in hiragana, such as the labels of products, and the names of businesses.

The texts were collected in one of two ways, either from online sources or through fieldwork in Japan. Online data was collected through 24 one-hour sessions of online browsing that took place during 2015. The online sources included commercial websites and social media sites, primarily facebook and twitter. Online ‘print club galleries’ were also investigated, as examples of hiragana loanwords appearing in these annotated photographs popular with young women had been reported by L. Miller (2003). As explained in chapter 2, many of these sticker booths now allow the users to input their email address to receive digital versions of their images, which can then be uploaded to print club galleries or shared on

social media. As the print club texts for the current study were collected from online sources they will be referred to as ‘print club images’ rather than ‘print club stickers’.

The print club images were also important as they were one of the few sources of handwritten texts in the corpus, as the images are decorated with a stylus on the screen of the booth37. As Walker (2000) notes, handwriting has a very high degree of flexibility, ‘any character and its style, weight, colour and size can be made, dependant only on the dexterity of the writer’, meaning these texts were an important site of marked hiragana orthography.

Social networking sites were another useful resource as the data-sharing culture of social media meant that besides the message posts themselves, which sometimes contained loanwords in hiragana, users included links to other websites such as those related to anime series, restaurants or celebrities, for example, which were also followed up during these browsing sessions. In this way, a relatively broad range of text genres could be accessed remotely from Australia. Data sourced in this way were recorded by taking a screenshot of the text, or downloading an image of the item.

The majority of the data came from the second source, fieldwork expeditions. The first was a ten day pilot study which was conducted in June 2014 and covered Osaka, Wakayama, Hyogo and Tokyo prefectures. This initial broad sweep was vital in refining the fieldwork plan for the main data collection, leading to a more structured list of locations in which to seek loanwords in hiragana.

The main data collection took place in February 2015, which took 18 days and covered Osaka, Hyogo, Kyoto, Chiba and Tokyo prefectures. The fieldwork plan was designed to more fully encompass a range of locations representing different demographics in terms of age, gender, and socio-economic status in order to cover the widest range of possible sources. Data was also collected at other non-focused times, such as on the walk from the station to the researcher’s hotel, or signs viewed from a bus window.

The fieldwork data was collected by taking photographs with a digital camera. Efforts were made to keep disruption to the public to a minimum; for example, when collecting data in supermarkets, the busy afternoon times were avoided, and the researcher ensured that data collection did not interrupt other shoppers or staff. When taking photos of shop signage, the

37It is noted that text can also be added to the photos by selecting characters via the stylus, so not all text that appears on print club images is necessarily ‘hand-written’. 75

camera was deliberately held high above the researcher’s head, to make it clear to bystanders that their photo was not being taken. This was particularly important when collecting examples of bar signage in the red light districts, as customers may not have wished to have their visit on record. These areas were avoided during their busy evening hours for this reason, as well as considerations of the personal safety of the researcher.

‘Data collection’ also includes an element of ‘data selection’, as decisions must be made about what to record for the corpus. The following sections describe what was included in the corpus and how these decisions were justified.

As stated in the introduction (section 1.4), loanwords with Chinese as a source language were excluded from the data set as it was difficult to differentiate them from ‘漢語/kango’ (Sino- Japanese words). Many of these words relate to food, and while more recent borrowings such as ‘わんたん/wantan’ (wonton) might be considered loanwords, other examples such as ‘らーめん/rāmen’ would not be considered an obvious example of a loanword (外来語 /gairaigo) to most Japanese.

A decision was made to focus on types rather than tokens of loanwords in hiragana, in order to more fully account for the breadth of use rather than the popularity of certain texts which used this device. Since the pilot study suggested that the largest categories of referential meaning in the data would be occupied by product names and shop names, these kinds of texts were likely to be met in fieldwork expeditions on a number of occasions, for example the same product being sold in different shops, or a chain restaurant being encountered in more than one location. Conflating types and tokens would therefore have led to some words, through their association with marketing texts, being over represented in the data. While the extent of the popularity of businesses or products with hiragana names is lost by not counting tokens, it was felt to be outside the scope of this study.

A further example concerned cases where the same brand name appeared on different products, for example, the ubiquitous ‘Hello Kitty’ character, whose name appears in hiragana, ‘はろうきてぃ/harō kitei’ (Hello Kitty), on traditional-style or souvenir products, for example the squid crackers, photo album, and green tea shown in Figure 3.2 below.

Figure 3.2: ‘Hello Kitty’ brand products

In this case, the type/token distinction is less clearly defined: if the brand name ‘Hello Kitty’ is considered the type, then these are three tokens of that type, and only one should be included in the data set. However, these are clearly different products, seemingly produced under licence by different companies. On the other hand, if these individual products are considered tokens, then ‘Hello Kitty’ merchandise would be over-represented in the corpus. Although these cases were not common in the data set, they did occur with several food products where variations in the product or flavour were released, such as the ‘さくさくぱんだ/sakusaku panda’ (Crispy panda) range of snacks.

In these cases, a decision was made to include up to 3 variations on the same type. This aimed to capture some of the variation of these product lines without allowing the data to be swamped with examples of any one instance.

A related issue concerned cases where a product with a brand name containing a hiragana loanword was accompanied by POS (point of sale) promotional material which also featured the brand name, such as the display of ‘がーぜたおる/gāze taoru’ (gauze [hand] towels) pictured in figure 3.3 below. A decision was made to exclude POS material produced by the manufacturer of the product (such as in figure 3.3) from the data set, as it had been created by the same agent, and its inclusion was felt to artificially inflate the occurrence of that particular example in the corpus. A related example was where a business’ name containing a hiragana loanword could be found in a number of texts associated with the business including signs, flyers, and menus, for example; these additional texts were also excluded.

Figure 3.3: POS text created by same agent Figure 3.4: POS text created by different agents

However, POS texts created by independent agents, such as the price label for the ‘ひなかすてら/hina kasutera’ (Girls’ festival castella) pictured in figure 3.4 above, were included as a separate instance of a loanword in hiragana in the corpus, as the packaging of the product and the label created by the supermarket in which it was sold were created by different agents, and were found to not always be orthographically consistent (see section 4.1.10, ‘Inconsistencies’).

From the above explanation, it can be seen that data selection was already part of the collecting process. The parameters of what should be included in the corpus of ‘loanwords in hiragana’ are summarized in table 3.3 below.

Table 3.3: Summary of corpus parameters

Example Resolution Justification

The distinction between ‘kango’ and Loanwords of Chinese origin Exclude ‘gairaigo’ is unclear.

These are various tokens of the same Chain restaurant appears in Count once type, and are produced by the same various locations agent.

These are various tokens of the same Product appears in various Count once type and are produced by the same shops agent. Name of shop/restaurant These are produced by the same agent, appears multiple times within and inclusion of all texts would over- Count once the establishment (eg. flyers, represent the appearance of this menus) example.

Variation in products is relevant, but Brand name appears across Count first 3 including all instances will cause some various products instances only brands to be over-represented.

POS material produced by the Including POS material created by the same agent accompanies a Exclude same agent would over-inflate the product appearance of this example.

POS promotional material POS material created by a different created by a different agent Include agent is of relevance. Variation in the accompanies a product use of script occurs.

The online browsing sessions and fieldwork expeditions therefore served to collect examples of loanwords in hiragana from a broad range of sources. However, other potential sites of loanwords in hiragana such as manga, song lyrics, and personal communication were not investigated. These are suggested as potential sites of further research in chapter 8.

Layer 1: Data Analysis

The screenshots from the online sources and the photos taken during fieldwork were uploaded to a software package called NVivo38 for analysis.39 Nvivo was used to both store and code the photographs and screenshots in this layer, allowing the researcher to tag images or parts of images, and display all items coded at a particular category to investigate further connections.

Data collected on the trip to conduct follow-up interviews (layer 4), as well as two private visits to Japan were also included in order to capture data across the greatest range of time, and from different seasons.40

The data was coded to a variety of categories which were designed to capture a number of features of loanwords in hiragana. These arose from breaking down the first research question, ‘How can loanwords in hiragana be characterised?’ into the following sub-questions:

What are loanwords in hiragana used to refer to?

What grammatical class do they take?

What donor language are they from?

Are there any instances of language play?41

How many loanwords in hiragana usually appear in a row?

How many occur in a whole text?

Are vowels lengthened with a bar like katakana words, or with an extra vowel like hiragana words?

Which loanwords most frequently occur in hiragana?

Do the texts display any other inconsistencies in orthography of loanwords?

38 NVivo is produced by QSR international (www.qsrinternational.com). 39 This included the pilot study data, which was felt to be collected under circumstances that were similar enough to the main study to justify combining them. Furthermore, some prefectures such as Wakayama were only represented in the pilot study data. 40 The seasons are very important in Japanese culture, and consequently many ‘季節限定商品/ kisetsu gentei shōhin’ (limited edition seasonal products), are only available at certain times of the year. 41 As mentioned in section 2.7, the large number of homonyms in Japanese provide ample opportunity for puns and other types of word play (Gottlieb, 2010a; Nagashima, 2006; Okada, 2012). Language play has also been specifically cited as a function of loanwords in Japanese (Hoffer, 1980; Honna, 1995). 80

The sub-questions above led to the following coding categories for the corpus (table 3.4).

Table 3.4: Coding categories for the corpus

Language Domain Coding category Example Coding options

 Bar Graphemes Vowel lengthening  Additional vowel

Word frequency (number of occurrences within corpus)

 Noun Words Word class  Verb  Adjective  Product name Referential meaning  Business name  Greeting  English Source language  French  German Word/phrase units  homonym Language play  paronym  lexical blend  Single word Number of consecutive words  2 words in a row  3 words in a row  1 unit per text Number of units per text  2 units per text  3 units per text Texts  Product packaging Genre  Restaurant menu  Social media post  Alternative script for a particular loanword within a text Within texts Inconsistency  Alternative script for a particular Between texts loanword between texts  Vowel inconsistency

Some coding categories apply to individual words within the corpus, while others to word/phrase units, and still others to whole texts, as indicated by the first column. The language domains described are adapted from the SFL approach to language and meaning making described by Eggins (2004), with important differences as described below. The final column in table 3.4 shows representative coding options, while the full list of codes available for each category is presented in appendix 1.

The structure of language domains differs from Eggins’ (2004) representation of language in some ways. The levels of ‘word’ and ‘word/phrase unit’ were distinguished, which would fall under Halliday’s term ‘lexico-grammar’. This was necessary because some coding categories, such as word class, concerned individual words, while others such as ‘referential meaning’ related to a phrase, in cases where two consecutive loanwords in hiragana occurred. Halliday’s ‘discourse semantics’ might fall mainly within the category of word/phrase units, but categories connected with topics other than semantics were also investigated at this level.

However, these categories share with the SFL structure an understanding that meaning is created and interpreted through all levels of language, so while the levels of grapheme, word, word/phrase unit, and text seem self-contained, they are considered parts of a greater whole, which is consistent with Halliday’s model of language, as well as multimodality and social semiotics, which together form the basis of the theoretical framework or this research.

The coding categories will be demonstrated using the text pictured in figure 3.5 below as an example. This figure shows the homepage for a pet-centre called ‘すいーとどっぐ/suīto doggu’ (Sweet Dog) in Yamagata prefecture (http://www.sweetdog.info/).

Figure 3.5: 'Sweet Dog' website

Vowel Lengthening

At the graphological level, the lengthening of vowels was a trait of hiragana loanwords that exhibited variation within the corpus, and therefore loanwords containing long vowels (such as ‘sweet’ or ‘suīto’) were coded according to whether the long vowel was represented by a bar as it is in loanwords (eg. すいーと), or with an extra vowel as it is with Japanese words (すいいと). Tildes (~) were considered a type of bar, so in the example above, ‘すい～と /suīto’ (sweet) was coded as ‘bar lengthening’. Within each category, ‘bar’ or ‘extra vowel’, the number of vowel sounds (eg. a, e, i, o, u) was also calculated, to investigate whether this had any effect on the likelihood of a particular method being used to lengthen a vowel sound.

Word Class

The grammatical class of a word (for example, noun, verb or adjective) was investigated at the word level because the data contained not only single words examples, but sometimes consecutive loanwords in hiragana, which necessitated coding for each word. They were interpreted in the context of their phrase units or texts, for example, ‘sweet’ can be an adjective (lovely or cute) or a noun (a sweet or lolly); but when occurring before another noun (eg. ‘dog’) it can be inferred that ‘sweet’ is functioning as an adjective. The category of ‘adjective’ also included ‘adjectival nouns’, for example, ‘ わいん /wain’ (wine) in the compound ‘わいんばー/wain bā’ (wine bar). Besides adjectives, nouns, and verbs, a further category ‘other’ was included for the small numbers of prepositions and other grammatical units.

Word frequency

Frequency counts were made at the word level to see if certain words appeared in the corpus more regularly than others. The most frequently occurring words were also examined for patterns such as thematic content and period of introduction to the Japanese language. The word ‘sweet’ occurred three times in the corpus, and ‘dog’ occurred twice.

Referential Meaning

Just as grammatical class was understood within its word/phrase unit, the coding for referential meaning also needed to be interpreted within its (con)text. For example, the referential meaning of the phrase ‘すいーとどっぐ/suīto doggu’ (Sweet Dog) is ambiguous without its context, it could be a description of a particular dog, or the name of a hot-dog stall, 83

for example. Taken in the context of the website it appears in, it seems clear that it is the name of a particular pet centre, and was therefore coded as ‘business name’. Other referential meaning categories included ‘menu item’, and ‘publication title’, for example.

Donor Language

Donor language was coded at the word/phrase unit level rather than the word level, firstly because no texts were found that mixed language at the phrase level, and also because examples such as ‘ あらかると/arakaruto’ (a la carte) were felt to be more usefully understood as a word/phrase unit of French, rather than the choice to include three French words. The unit ‘すいーとどっぐ/suīto doggu’ (Sweet Dog) was therefore coded as English. Languages represented by more than one unit in the corpus were coded into the categories English, Dutch, French, German, Italian, Latin, and Portuguese. A further category, ‘other’, contained any language represented only once in the corpus, such as Spanish.

Language play

Word/phrase units were coded for the presence or absence of language play; the business name in this example does not include any puns or hidden meaning. Four types of language play appeared in other examples in the corpus. The first two, homonyms and paronyms (near homonyms), are quite similar; representing cases where two different meanings arise from the same (or similar) sets of phonemes. A common example was the use of the word ‘ほっと /hotto’ as part of a brand name of a food product, such as ‘ほっとコーン/hotto kōn’ (hot corn), a variety of instant soup mix. Besides referencing the loanword ‘hot’, the Japanese word ‘ほっと/hotto’ expresses a feeling of relief or relaxation, both meanings linking well with the idea of a hot, relaxing cup of soup. Another type of language play was a lexical blend, where two words, typically a Japanese words and a loanword, were clipped and combined into a new lexical item.42 The name of the online store for office-supplies, ‘たのめ

ーる/ tanomēru’, was one such example, representing a blend of the Japanese verb ‘たのめる /tanomeru’ (to ask, to order) with the hiragana loanword ‘ めーる/mēru’ ([e]mail), combining to express a sense of ‘being able to order [office supplies] via email/online’. A final type of language play evident in the corpus was the existence of a concealed word, for

42 Renner (2015) argues that all lexical blends are instances of wordplay due to their intrinsic formation process, as described in section 2.7. 84

example in the name of the aged care facility ‘明日見らいふ /asumi raifu’ (Asumi Life), the word for ‘the future’ (みらい/mirai) being created at the boundary of these two words.43

Number of consecutive words

The number of loanwords in hiragana that appeared in a row was also coded. Examples ranged from one to three words in a row, with ‘すいーとどっぐ/suīto doggu’ (Sweet Dog) being coded as two.

A category of ‘partial word’ was created to account for parts of loanwords that were in hiragana, usually the result of a lexical blend or other types of language play. For example, the name of the snack ‘こめっぷす/komeppusu’, made from the Japanese word ‘米/こめ /kome’ (rice) and the loanword ‘ちっぷす/chippusu’ (chips), contains only half of the word ‘chips’ in hiragana.

Truncating of loanwords is common in Japan (see section 2.5.2), and those that appear in the truncated form in a Japanese loanword dictionary44 such as ‘こすめ/kosume’ for ‘cosmetics’, were coded as single words rather than partial words. This ensured an emic view of the data, as the truncated versions are often more familiar in Japan than their complete versions.

Number of word/phrase units in a text

This category was included to ascertain how many word/phrase units appeared in a text. In the ‘Sweet Dog’ homepage, there are no other loanwords(s) in hiragana in the text, so it was coded as ‘1 unit’. A print club image containing the hiragana loanwords ‘さんた/santa’ (Santa) and ‘めりーくりすます/merī kurisumasu’ (Merry Christmas) in different parts of the image was therefore coded at two units.

However, in the ‘Sweet Dog’ homepage, the same unit (‘すいーとどっぐ/suīto doggu) is also contained in the body of the text, and was therefore coded as ‘repetition within text’. Coding repetition separately to units allowed for a more fine-grained understanding the extent of hiragana loanword use within a single text.

43 This example is especially clever, since rather than the standard kanji for ‘future’ (未来/mirai), the word is created with the kanji ‘to see’ (見らい/mirai), therefore also giving the suggestion of ‘looking to the future’. 44 Sanseido’s Concise Katakana-go Dictionary (Sanseidō Henshūjo, 2010) 85

Together with the calculations for consecutive words, the figures for word/phrase units per text provided complimentary measures of the extent to which loanwords in hiragana appear in a text.

Genre

This coding category related to the second research question, which is designed to ascertain in what contexts loanwords in hiragana are usually found. Multimodality, SFL, and social semiotics all stress the importance of context in constructing meaning; therefore it was important to understand the types in genre in which loanwords in hiragana commonly appeared. A genre represents a set of register variables sedimented though use within a particular culture, and a given text can be recognized as being ‘typical’ of that genre by members of the culture in which it occurs (Eggins, 2004). Genres identified in the corpus included product packaging, menus, and business signs, for example. The homepage above was coded as ‘Webpage-business’.

Inconsistency

This category was used to investigate any inconsistencies in how loanwords in hiragana were transcribed, and involved three different types of inconsistency found within the corpus, although none occurred in the web page shown above.

In cases where a loanword in hiragana was represented elsewhere in the text in another script (such as katakana, kanji, or occasionally English/rōmaji), the text was coded as ‘alternative script for a particular loanword within a text’.

The second type of inconsistency related to the same kind of variation, but between related texts, usually product packaging and POS signage. This was coded as ‘alternative script for a particular loanword between texts’.

The final type of inconsistency related to the transliteration of vowels, and included cases where both vowel lengthening techniques, the bar association with katakana, and the extra vowel association with hiragana, were used in the one word/phrase unit; for example ‘でぇーと/deeeto’ (date).

Each of the texts in the corpus was coded for the above categories. The full list of coding categories can be found in appendix 1.

Layer 1: Summary of research design

The findings from this stage of the research were used to address the first research question, which seeks to identify the typical characteristics of loanwords which appear in hiragana. The results on the genres represented in the corpus also relate to the second research question, which aims to describe the contexts in which these words commonly occur. Aspects of a particular function of loanwords in hiragana, language play, can also be found in the corpus analysis, and therefore contribute to addressing the third research question. The results of the descriptive statistical analysis of the corpus are presented in chapter 4.

Finally, this layer of analysis was also used to guide those which followed; as the findings informed the selection of prompts for the case studies (layer 2), survey (layer 3) and interviews (layer 4). A range of texts that were both typical and atypical of the corpus were therefore able to be included at each level, and the reasons for the unusualness or infrequency of particular examples examined in the interviews and focus groups.

3.4.2 Layer 2: Case studies of selected texts

The case studies, referred to as layer 2 of the research design, are comprised of in-depth analyses of four texts containing loanwords in hiragana. The relevance of the case studies to the other layers of analysis is represented in table 3.5 below.

Table 3.5: Case studies in the context of the research design

QUANTITATIVE DATA QUALITATIVE DATA

Layer 1 Layer 2 TEXTS Corpus analysis Case studies

Layer 3 Layer 4 AUDIENCE Survey Interviews and focus groups

Based on the results of the corpus analysis, texts were chosen for the case studies which represented both typical and atypical examples of loanwords in hiragana in different respects.

These cases represent unique ensembles of specific multimodal, orthographic and other meaning-making features. As Bloomberg and Volpe explain, case studies are ‘not for purposes of generalising beyond the case but rather for rich description of the case in order to understand the complexity thereof’ (Bloomberg & Volpe, 2015, p. 31).

The case studies thus complement the corpus study by providing insights that go beyond figures and percentages, and looking more deeply into the process of meaning making within individual texts. The framework used for the case studies (explained in detail in the section ‘data analysis’ below) combine elements of Kress and van Leeuwen’s (2006) ‘visual grammar’, as well as Stöckl’s (2005) semiotics of typography. The case studies were most relevant to addressing the third research question, which seeks to understand why loanwords appear in hiragana.

The particular strength of a multimodal analysis is that it takes into account the interaction between modes of the texts that are overlooked in a traditional corpus study, such as colour, font, and the other scripts used in the text, as well as images and the layout of the text as a whole (see section 2.2). As Kress & van Leeuwen (2006) explain, these non-language elements of a text are also ‘read’ by the viewer, and thus should be taken into account when discussing how meaning is conveyed through a text. Walker (2000) also notes that there are many cases where ‘consideration of graphic features could usefully complement any linguistic description’ (2000, p.8), and, it is argued, a social semiotic understanding of the meanings of loanwords in hiragana is one such case.

The first two layers of the research design therefore include both quantitative data (the descriptive statistical analysis), and qualitative data (the case studies) from the corpus of loanwords in hiragana. The case studies are also used to triangulate the findings from the other layers of the analysis, for example the non-language cues such as font and image identified in the case studies also mentioned by survey respondents (layer 3) and discussed by participants in the interviews and focus groups (layer 4).

The data collection and analysis procedures for the case studies will now be described in the following sections.

Layer 2: Data Collection

Four texts were chosen to form case studies in order to illustrate how loanwords in hiragana work together with other semiotic resources to form a coherent message. Business signs, product packaging and social media posts were common genres within the corpus, leading to the selection of one example of each text for the case studies. Print club images were also common in the corpus, but since these texts are associated with a specific demographic (young women), a more generally accessible social media text was chosen as these same texts would also be used in the survey (layer 3).The selection of individual texts within the corpus was based on their typicality and atypically in terms of the findings of the statistical analysis (layer 1). The texts used in the case studies are shown in table 3.6 below, and are reproduced in a larger format in section 4.2.

Product packaging and business signage were two of the most common genres of text represented in the corpus according to layer 1 of the analysis, and the texts chosen show other typical features such as using the loanword in hiragana as the business name, containing a single loanword in hiragana, and having English as its source language. The social media post was also from English, but featured an atypical transliteration of the well-known greeting. It featured two loanwords in hiragana, but a single word/phrase unit. The sticker was from an unusual genre of text where the words themselves formed part of the product, rather than the packaging, for example. This text was also unusual in featuring a verb in hiragana, as nouns and adjectives were more common overall, and using the hiragana loanword as part of an English sentence.

Table 3.6: Texts used as case studies

Genre Text Hiragana Loanword Content

はぴねす鍼灸整骨院 hapinesu shinkyūseikotsuin Business sign Happiness acupuncture moxibustion45 and osteopathic clinic.46

蜂蜜れもん Product hachimitsu remon packaging Honey Lemon

I can ふらい!! Text on product I can furai!! (Sticker) I can fly!!

ぐっもーにん Social media gummōnin post G’mornin’

45 Moxibustion is a form of eastern medicine where plant materials are burned on or near the skin. 46 Because of the relatively complex name of this business, it will be referred to as ‘Happiness Clinic’ in future. 90

Layer 2: Data Analysis

As previously mentioned, elements of two semiotic frameworks, those of Stöckl (2005) which has typography as its focus; and Kress and van Leeuwen’s (2006) which analyses texts containing visual elements, were applied to the case study texts. The strengths of each approach will be described below.

Stöckl (2005) outlines an ‘analytical toolkit’ (p. 205) to break down the elements of typography into four domains, describing the ‘typographical building blocks’ they contain, and their properties. As explained in section 2.8, the choice of script in Japanese is not identical to the choice of typeface, but it does provide a way of understanding how connotations can be layered on to a word through the way it is written. Stöckl’s framework therefore provides a method for considering the choice of script in Japanese in a similar way to the choice of typeface, as well as looking at the actual typeface used for the words under investigation.

The four levels of typography identified are microtypography, mesotypography, macrotypography, and paratypography. At the ‘microtypography’ level, the design of the font and the individual graphic signs are involved, such as the size, colour, and style. The choice of Japanese script (hiragana, katakana, or kanji) would also be the domain of microtypography. ‘Mesotypography’ relates to how the individual graphemes form lines and blocks of text; most relevant to the texts in this study include the direction of the lines (vertical/horizontal) and the mixing of fonts, and, in the case of Japanese, scripts. ‘Macrotypography’ takes the graphic structure of the whole document into account, and can include the arrangement of text and graphics, ornamentation devices, and typographic emphasis such as underlining. Finally, ‘paratypography’ relates to the materials and instruments used in making the text, such as the kind of paper used, or whether the graphemes were created with a typewriter, or pen and ink, for example. Being able to break down the elements of typography helps to unpack the meaning-making which occurs at this often overlooked or under-analysed level, since every loanword ‘written’ in hiragana involves typographical choices as described above.

However, other than the reference to the assembling of text and graphics in ‘macrotypography’, Stöckl’s framework, with its focus on lettering, does not provide a comprehensive way to understand the other modes at work in a multimodal text, such as

layout and the types of images which can appear. Kress and van Leeuwen (2006) address these issues in their ‘visual grammar’, and show how the types of images used, their placement and framing, all play a part in communicating a message to the reader. The authors (2006) describe how the visual elements in a multimodal text can contribute to the message given by the text as a whole, and link these back to the metafunctions of language described by Halliday (1978).

An example of ideational meaning, giving information about the state of the world, in visual elements are ‘classification processes’, where taxonomies of items are shown either overtly (such as tree diagrams) or covertly (such as advertising images). An image showing an arrangement of all the products in a company’s range has the effect of classifying them as similar or of the same class. Analytical processes show the relationship between elements in terms of a part/whole structure; again these can be overt (such as a pie diagram) or covert, such as an image from a fashion magazine which breaks down an outfit into its constituent items of clothing and accessories. The use of ‘vectors’, lines created through gaze, outstretched arms, or abstract elements, can also play a part in ideational meanings, suggesting the actions of a represented participant towards a goal, for example, a represented participant looking at a product in an advertising image.

The interpersonal metafunction, which represents relationships between entities, can be seen from the way the viewer is positioned, for example whether represented participants are shown looking directly at the viewer (‘demand’), inviting an imaginary social relationship; or whether they are looking away from the viewer (‘offer’), allowing themselves to be objects of contemplation. The size of the frame and the closeness of the shot can also be used to indicate social distance; when the represented participant’s face fills the whole frame, it suggests a relatively intimate social distance such as a romantic partner, while a head-and-shoulders shot could represent a friend, and a full-body shot a stranger. The same use of distance can be seen in the depiction of objects or buildings; the degree of closeness to the viewer can suggests intimacy or distance. Also relevant to objects is the angle from which they are depicted; many images of products in advertising texts are taken from a close high angle, suggesting the item is at arm’s reach on a table in front of the viewer. Angle can also suggest relative power; a high angle puts the viewer of the text in a privileged position, while a low angle suggests the power or status of the item depicted.

Another way the interpersonal function can be communicated in images is through ‘modality’, where various cues can suggest the degree of truth or reality a text is designed to make. Colour can be used in this way, for example colour saturation can be reduced to black and white, or intensified to an ‘unnatural’ degree (both low modality), while a use of colour that reflects the natural environment represents the highest modality. Images can also be ‘contextualized’, shown in a particular surrounding and representing a high degree of modality; or ‘decontextualized’, where, for example in advertising, products are shown against a plain background, which reduces their modality. Another aspect of modality is the degree of abstraction, which refers to how realistically a subject is portrayed. Abstract art, for example, can represent an extreme of this continuum. Images with high modality can be said to suggest they are to be taken more literally, as true representations, that those with lower modality.

Finally, textual function can be expressed through the composition of an image, and applies both to single pictures as well as composite texts made up of smaller visual elements. The first of the three elements of composition described by Kress and van Leeuwen is ‘information value’, and refers to the placement of elements in the various ‘zones’ of the image; top or bottom, left or right, as well as centre or margin. The top of the image suggests the ‘ideal’, the idealized or generalized essence of the message, for example the glamourous lifestyle of those who use a particular product in an advertising image. The bottom represents the ‘real’, and therefore specific details, or practical information, for example. In advertising, this is frequently where a picture of the actual product is shown, as well as information such as the ingredients or where it can be purchased. The left and right zones represent the ‘given’ and ‘new’ information respectively, the ‘given’ is the known, accepted, or established information, while the ‘new’ represents information unfamiliar to the viewer, or problematic in some way.

Kress and van Leeuwen are quick to note the given/new distinction is based on a left-to-right tradition of writing, and may not apply to texts in cultures with other writing systems. They note the webpages of the electronics company Sony in their English and Arabic versions, in which the elements appearing in the ‘given’ and ‘new’ zones are reversed. Japanese can be written both horizontally from left to right; and vertically from top to bottom, starting on the right, and this will be taken into account when applying Kress and van Leeuwen’s visual grammar to the texts in this study. This framework has been utilized with Japanese texts in the past (eg. Oyama, 1998; Maynard, 2008) and the fact that it is based on a western writing 93

tradition is not felt to prohibit its application to texts in Japanese. Kress and van Leeuwen also note a ‘centre-margin’ layout which, they suggest, is not common in modern western societies, but was found to be not infrequently used in Asia. The centre suggests importance or permanence, while the marginal elements suggest ancillary or dependent elements.

Salience is another feature which can suggest the textual metafunction, and refers to the degree to which the various elements attract the viewers’ attention. This can be done in a number of ways, including size, larger elements being more noticeable; placement, as items in the foreground are more salient than those in the background; and contrasts in tone or colour, for example a bright splash of colour in an image of reduced tonal contrast can catch our attention. Combinations of the above features can represent a part of an image or text as more noticeable, and therefore carrying information of more relative importance.

Framing is a further element which contributes to the textual metafunction; frames can be made with abstract lines or with elements of an image, such as a window frame. Framing suggests that elements within it belong together in some way; a dividing frame line, on the other hand, shows elements that are distinct or different from each other. Background colour can also serve a dividing function. Frames can also differ in their degree of salience, they may be clear cut or the parts of the image may blur into each other; these cues also make statements about the relationships between the elements as either permanent or permeable.

Kress and van Leeuwen’s ‘visual grammar’ therefore shows how different parts of a visual text can be used to encode ideational, interpersonal and textual information. The cues described can combine to deliver more complex messages, for example an element of a visual such as a celebrity user of a product in an advertising image which is both large and in the upper portion of the whole suggests through the textual function that this element is important (salient), and represents the ‘ideal’, a lifestyle or image which those who purchase the product can aspire to. In addition, different functions can be achieved by the same element, the celebrity image may be shot at close range, suggesting a close or friendly relationship with the viewer though the interpersonal function. Kress and van Leeuwen’s visual grammar is therefore useful in understanding the meaning-making work that is done by the visual elements of a text, and how this can be achieved through a number of interacting features.

To summarize this section, the analytical approaches of Stöckl (2005), and Kress and van Leeuwen (2006) will be drawn upon to analyse the texts in the case studies. These approaches both refer to the work of Halliday, in particular the metafunctions that language is designed to

make, and expand these functions to other semiotic resources. Stöckl makes the case that typography can be considered a ‘mode’ as it is capable of expressing these same referential, ideational, and interpersonal functions; and Kress and van Leeuwen make the same arguments for the various visual elements of a text that they discuss. These approaches are therefore felt to be complementary in expanding the social semiotics of language described by Halliday to other semiotic resources in a written text, in this case, to Japanese texts containing loanwords in hiragana. As the analytical approaches described above have come from writing traditions outside of Japan, the findings of the case studies were all reviewed by a native speaking Japanese researcher to ensure the applicability of the framework to the current data, and the validity of the interpretations.

Layer 2: Summary of research design

To summarize this section, the case studies allowed a detailed analysis of how loanwords in hiragana function as parts of cohesive texts, which complemented the quantitative findings of the statistical analysis of the corpus (layer 1). This involved recognising loanwords in hiragana as one semiotic resource amongst many used in the text, including typefaces, images and layout, and analysing what meanings each of these items contributed to the multimodal whole. The findings from this layer were most useful in answering the third research question, which investigates why loanwords appear in hiragana. As explained in section 2.2, the messages of multimodal texts are more than the sum of their parts, and so the particular semiotic resource that is the object of this research (loanwords in hiragana), must be interpreted in conjunction with the other resources in the text and the interactions between these resources.

While the case study texts (layer 2) were analysed methodologically through the framework described above, many of the survey respondents (layer 3) and interview participants (layer 4) also drew on the effects of these other resources such as colour and typeface in explaining their views on loanwords in hiragana. This reinforced the multimodal nature of meaning making, and also assisted in triangulating the findings between these layers.

3.4.3 Layer 3: Survey investigating opinions of loanwords in hiragana

The survey, referred to as layer 3 of the research design, comprises of quantitative data gained from the opinions of Japanese people regarding loanwords in hiragana. Its position within the research framework is depicted in table 3.7 below.

Table 3.7: Survey in the context of the research design.

QUANTITATIVE DATA QUALITATIVE DATA

Layer 1 Layer 2 TEXTS Corpus analysis Case studies

Layer 3 Layer 4 AUDIENCE Survey Interviews and focus groups

Like the case studies (layer 2), the survey (layer 3) was also primarily used to address the third research question, ‘why do loanwords appear in hiragana?’ In order to both triangulate and complement the findings of the case studies in relation to this question, the survey sought the opinions and perceptions of native speakers of Japanese on texts which included loanwords in hiragana. Selected findings of the survey were also integrated into the questions used in the interviews and focus groups (layer 4), further allowing the findings relating to this research question to be explored in a variety of ways.

The advantages of using surveys to gather data are that they are efficient in terms of both time and cost in collecting data from a large number of respondents. In addition, there is a lack of interviewer bias, and they can be completed at a time and place that suits the individual (Gray, 2009). This final point was vital to the current study, as it meant that an online survey could be created and administered from the researcher’s location in Australia, and be instantly available to respondents in Japan to complete. However, a critical approach to the perception data from both the survey and interviews (layer 4) needed to be maintained, as respondents may have felt pressured to provide an answer, and invented a plausible reply. Triangulation of the findings from the different layers of the analysis assisted in reducing the effect of such ‘made up’ responses. The data collection procedure and analysis for the survey is described in the following sections. 96

Layer 3: Data Collection

This section describes the data collection procedure for the online survey. It is presented in two sub-sections; survey design and survey administration.

Survey design

The following sections describe the questions and the rationale for their inclusion as part of the survey. Where relevant, changes made after the initial pilot study are also described. The survey was drafted in English, and then translated to Japanese in consultation with two native speaking Japanese researchers to ensure the survey appeared clear and natural to the Japanese research population. A hard copy version of the online survey, in Japanese with an English translation, can be found in appendix 2. For the sake of brevity, only English translations of the questions will be used in this section.

The survey was comprised of three sections, the first collected the respondents’ background information, the second asked what connotations or associations each of the Japanese scripts held, and the third asked specifically about atypical orthographic choices for loanwords, including hiragana. Each of these sections will be described below.

When potential respondents accessed the link to the online survey, the initial screen gave a brief overview of the research goals, where it was stated that the aim of the research was to investigate what people thought of different ways of writing loanwords in Japanese. The estimated time to complete the survey was given as 10 minutes, and participants were assured their answers and personal details would be kept secure and not be passed on to third parties. They were then asked whether they wished to participate in the research (Q1). Participants who selected ‘no’ were taken to a page which thanked them for their interest in the project, while those who selected ‘yes’ continued to the next section. This question played the role of the ‘consent form’ for the online survey.

The first section, ‘questions about you’, required the respondents to supply their background information, including confirming their status as a native speaker (Q2), gender (Q3), age bracket (Q4), and education level (Q5). The questions on age range and native language status were necessary to ensure that respondents were both over 18 years of age and identified as native speakers of the language, which were the target population for the survey. The term ‘identify as a native speaker’ was chosen in order to include people of non-Japanese

heritage, as well as those of mixed ethnic backgrounds. Respondents whose answers indicated that they were less than 18 years of age and/or non-native speakers of Japanese were taken to a ‘disqualification page’ which thanked them for their interested but explained that they fell outside the intended demographic of this study.

Respondents were also asked if they would be interested in participating in an interview on the same topic at a later stage (Q8). Those who selected ‘yes’ were then asked to supply their email address and preferred location, while respondents who selected ‘no’ were taken to the second section of the survey.

The second section of the survey,47 titled ‘impressions of the scripts’, was designed to collect information on respondents’ impressions of the scripts used in Japanese. The word ‘impressions’ (印象/inshō) was chosen over other candidates such as ‘associations’ and ‘connotations’ as being the most natural for the context, after consultation with two native speaking Japanese researchers. While the word ‘impression’ was used in the Japanese version, the word ‘association’ will be used interchangeably hereafter, as respondents’ answers combined a mix of impressions that the various scripts gave them, as well as people and items associated with them.

Participants were asked:

What impression do you have of hiragana? When you see hiragana, what kinds of things do you think of? Write your answer in the free space below. (Q9)

Below this prompt a text box was provided for them to input their answer. This question was repeated with each of the scripts used in Japanese; katakana (Q10), kanji (Q11), and rōmaji (Q12). Although these kinds of script impression studies have been conducted by other researchers in the past (eg. Iwahara, Hatta, & Maehara, 2003), it was necessary to collect this data concurrently with the main questions on the appropriateness of hiragana, and by the same participants.

The third section, titled ‘examples of writing’, presented participants with pictures of texts featuring loanwords in various scripts. Three texts were taken from the case-study examples; the business ‘はぴねす鍼灸整骨院/ hapinesu shinkyūseikotsuin’ (Happiness Clinic); the

47 In the pilot version of the survey, this section came after that on ‘examples of writing’. The order was changed because it appeared that the marked script examples which appeared in this section may have had an effect of the respondents’ answers in the script associations section. 98

drink called ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon); and the sticker containing the phrase ‘I can ふらい/furai (fly)!!’ (all of which are pictured in section 3.4.2, figure 3.6). As described in section 3.4.2, the first two texts were chosen as they were typical of the corpus according to the descriptive statistics of layer 1 of the analysis. They represented a business name on shop signage, and a product name on product packaging, which accounted for the two most frequently occurring types of referent and genre of text. They were also typical in that they were both nouns, from English, appeared singularly (see results section 4.1). The final text containing a loanword in hiragana, the sticker, was chosen conversely because it was unusual in these respects; it appeared on the product itself, it was a verb, and furthermore it was embedded in an English sentence (‘I can…’). As described above, using the case study texts in the survey allowed for triangulation of the findings between layers 2 and 3 of the analysis.

Two further texts were chosen which represented other atypical orthographic choices for loanwords to remove the ostensive focus of the research from ‘loanwords in hiragana’. In addition, the data from these questions provided an insight into the connotations of other scripts used for loanwords. The first was an example of a loanword in kanji ‘珈琲/ kōhī ’ (coffee) which appeared in the name of the instant coffee mix ‘ 牛乳屋さんの珈琲 /gyūnyūyasan no kōhī’ (Dairy Coffee). This word was chosen because, like ‘檸檬/remon’ (lemon), it represented one of the few contemporary loanwords that can be written in kanji.

The second was a loanword in rōmaji, the word ‘hankachi’ (handkerchief), which appeared as part of the product name ‘かわいい hankachi/kawaii hankachi’ (Cute handkerchief). This example was chosen because due to its spelling, it represented a clear case of a loanword, rather than an English word, in rōmaji.48 These two examples not included in the case studies can be seen in figures 3.6 and 3.7 below.49

48 Other loanwords in romaji, such as ‘トマト/tomato’, would be indistinguishable from their English equivalents, and examples of interference between loanword and English spellings are not uncommon, for example those cited by Backhaus (2007). Besides being adapted into the Japanese mora sound system, this particular example seems to have entered the language through aural transmission, as many Meiji period loans were, rather than through the written medium, leading to its unusual transliteration. 49 One further example was included in the pilot study, which contained a loanword in standard katakana. The acceptability rating for this item was 100%, and as it was not felt to generate any useful data, it was removed in the final version. 99

Figure 3.6: ‘Dairy Coffee’ instant coffee mix Figure 3.7: 'Cute hankachi' handkerchief

For each of the five images described above, three questions were asked. Using the clinic called ‘はぴねす/hapinesu’ (Happiness), as an example, the first of these questions asked:

This shop’s name includes the word ‘hapinesu’. Writing ‘hapinesu’ (happiness) in hiragana is… (Q13)

☐Appropriate

☐Inappropriate

This was therefore a closed question, and participants could check the box that indicated their answer. Although judgements of appropriateness were not the main focus of the current research, this question was designed to prompt participants to engage with their reactions to the texts, in the lead up to the follow up question:

Please tell us the reason for your answer. (Q14)

This was therefore an open-ended question, and participants could type their answer into the open-response box below. The final question for each example asked:

Why do you think the shop decided to write ‘happiness’ in hiragana? (Q15)

This was also followed by an open-response box. While some participants who answered ‘appropriate’ may have felt this question repetitive, it was included to allow participants who

100

felt the script use inappropriate to reflect on why the shop decided to use this particular orthographic choice. A screenshot of part of the survey can be seen in figure 3.8 below.

Figure 3.8: Screenshot of online survey

These questions were repeated for the five images described above; namely the drink 蜂蜜れもん/hachimitsu remon (Honey Lemon) (Q16-18); the sticker containing the phrase ‘I can ふらい/furai (fly)!!’ (Q19-21); the coffee mix ‘牛乳さんの珈琲/gyūnyūsan no kōhī’ (Dairy Coffee); and the product ‘かわいい hankachi/kawaii hankachi’ (Cute handkerchief) (Q25- 27). As explained above, the survey in its Japanese version and English translation can be found in appendix 2. The online survey was then formatted through the online survey software platform ‘Survey Monkey’.50

50 More information on Survey Monkey can be found at https://www.surveymonkey.com/ 101

Survey administration procedure

The survey was piloted on a small group of the researcher’s colleagues, and after minor adjustments (see footnotes 48 and 50), the final version was launched on 22nd June 2015, using the ‘snowball method’ which involved sending the link to the survey to the researcher’s contacts in Japan, requesting that they forward it through their networks. Potential participants could then access the survey through the link and complete it on their computers or mobile phones. No time limit was set for answering each question. A progress bar at the top of each page showed how much of the survey had been completed (see figure 3.8 above), in order to encourage participants to continue to the end of the survey. Results were received up until 3rd August, and the survey remained open until 25th October. The data was collected in cloud-based software, and the results downloaded in spreadsheet files. Response rates for each question varied between 112 and 165 for the questions on the associations of the scripts and the sample texts, and therefore the number of responses for each question is reported in the findings in chapter 5.

Layer 3: Data Analysis

This section will describe how the survey data was analysed, and is divided into the three sections of the survey; background information, impressions of the scripts, and examples of writing.

The questions on the participants’ background (Q2-6), were all closed multiple choice questions, meaning the responses were simply tabulated and are presented in the results section 5.2.1. Email addresses for those who expressed an interest in participating in an interview (Q8) were saved securely according to the research ethics guidelines.

The next section of the survey asked for respondents’ impressions of each of the four scripts used in Japan; hiragana (Q9), katakana (Q10), kanji (Q11), and rōmaji (Q12). These were open-ended questions, and therefore the responses needed to be coded in order to generate quantitative data. The data was gathered in the respondents’ native language, Japanese, however English translations are provided in the following explanation in order to describe the coding process. The three responses presented in table 3.8 below show associations for hiragana, and are typical of those received for the questions on each of the scripts (Q9-12) in

102

that while some respondents listed multiple associations or connotations, such as respondents 1 and 3, others only gave a single word, such as respondent 2.

Table 3.8: Sample responses of associations of hiragana script

Respondent Response for ‘hiragana’ (Q9) Translation

柔らかい、丸い、易しい、 1 Soft, round, easy, weak, childlike 弱い、子供っぽい

2 小学校 Primary school

3 やさしい、簡単、わかりやすい ‘Yasashī’, easy, easy to understand

The codes arose partly from the literature and partly from the responses themselves, in a grounded theory approach. As Bloomberg and Volpe (2015, p. 141) suggest, category descriptors can initially be drafted from information gained from the literature review and the researcher’s ‘educated guesses’, and the initial framework should be ‘flexible and emergent’. For example, the previous studies on the associations of the Japanese scripts provided potential codes for this data, such as hiragana being associated with children, (see section 2.5). Respondent 1 in the above table mentions this association in their answer (‘childlike’), as does respondent 2 by mentioning ‘primary school’. These two responses were coded under the category ‘Children’. Other codes arose directly from the data, for example the word ‘weak’, listed by respondent 1.

Some responses contained two tokens for the same code, for example respondent 3 gives ‘easy’ and ‘easy to understand’ as associations, both of which were coded as ‘Easy’. Codes were regularly reviewed, and collapsed or separated where necessary. For example, respondent 1 was the only person who mentioned ‘weakness’ as an association of hiragana, so this response was eventually coded to the category ‘Other’.

Some difficulties arose with coding the word ‘やさしい/yasashī’, which often occurred in the responses for hiragana (Q9), as it does in respondent 3’s answer in table 3.8 above. In Japanese, kanji can be used to express slight differences in meaning between homonyms, for example hiragana was sometimes described as ‘易しい/yasashī’ (easy), or ‘優しい/yasashī’ (gentle). So when it was described as ‘やさしい/yasashī’ in hiragana, it was unclear which of these two meanings were being referred to. After consultation with a native Japanese

103

researcher, it was agreed that the creation of a third code, ‘yasashī (easy or gentle)’, was the best way to account for these responses.

A parallel problem arose with the responses of ‘katai’ which arose in the questions for both the katakana (Q10) and kanji (Q11) scripts. Six different orthographic variants were found for this word; three different kanji ‘堅い’, ‘固い’, and ‘硬い’, the two phonemic scripts, ‘かたい’ (hiragana) and ‘カタイ’ (katakana), as well as a single respondent whose answers were entirely in rōmaji, and included ‘KATAI’. The kanji ‘堅い/katai’ was coded as ‘Formal’, while the two other kanji versions ‘固い’, and ‘硬い’51 were coded as ‘Firm, solid’. Like the case of ‘yasashī’ described above, it was the opinion of the native Japanese researcher that the kana and rōmaji responses of ‘katai’ could should be collected under a third category, ‘Formal or solid’.

The three responses presented in table 3.8 above generated six codes, and a total of nine tokens, as illustrated in table 3.9 below.

Table 3.9: Sample coding for associations of hiragana

Respondent Translated response for hiragana (Q9) Codes No. of tokens

Soft 1 Round 1 1 Soft, round, easy, weak, childlike Easy 1 Other 1 Children 1

2 Primary school Children 1

Yasashī 1 3 ‘Yasashī’, easy, easy to understand Easy 2 TOTAL 6 9

This process of coding was conducted by the researcher and a native-Japanese speaking research assistant for the first 10% of responses to each open-ended survey question. Results were compared, and differences in coding were discussed until a consensus had been reached. The researcher coded the remaining 90% of the data independently, seeking advice from the research assistant when difficulties of translation or coding arose. For example, responses for

51 Although the nuances of these two kanji are noted, their meanings were considered similar enough to be collapsed into a single category for the purposes of this study. 104

hiragana (Q9) included ‘心/kokoro’ (heart, spirit) and ‘感情/kanjō’ (feelings, emotion) both of which later formed the code ‘Emotion’.

Finally, the code names were translated into English with the assistance of a native speaker where necessary, and appear in their English versions in the findings (section 5.2.2).

The third section of the survey, ‘examples of writing’ had three questions for each image prompt, as described in the previous section. For the closed question on the appropriateness of the choice of script, namely for ‘happiness’ (Q13); ‘lemon’ (Q16); ‘fly’ (Q19); ‘coffee’ (Q22); and ‘hankachi’ (Q25); totals for each alternative were calculated. For the open-ended questions on the reason for this choice, (Q14, 17, 20, 23, 26) and the perceived motivation for the use of script (Q15, 18, 21, 24, 27), it was necessary to code the responses. A similar approach used for the script association questions (Q9,10,11,12) was used for these questions and is demonstrated in the following paragraphs.

The first open-ended question asked for the reasons why the respondent thought writing ‘happiness’ in hiragana was (in)appropriate (Q14). The first step was to divide the data into two groups, those who selected ‘appropriate’, and those who selected ‘inappropriate’. Typical responses for those who selected ‘appropriate’ are presented in table 3.10 below. As above, responses were coded in Japanese, however translations are provided here in order to explain the coding process.

Table 3.10: Sample responses of reasons for appropriateness of 'happiness' in hiragana

Respondent Response for ‘happiness’ (Q14) Translation

1 読みやすく、親しみ易いから。 It’s easy to read, and approachable.

なぜかカタカナのハピネスより身近に感 Somehow it feels more familiar than ‘happiness’ 2 じるから written in katakana.

It’s the freedom of the shop owner (to write how 3 店主の自由 they like).

As with the responses to the questions on the associations of each script (Q9,10,11,12), a proportion of the data (10%, half for each of the ‘appropriate’ and ‘inappropriate’ reasons), was also coded by a native speaking research assistant. The results of this process were

105

compared, and issues or differences were discussed until a consensus had been reached. The researcher coded the remaining 90% of the data independently, seeking advice from the research assistant when difficulties of translation or coding arose. The three sample responses in table 3.10 above were coded in the following manner (table 3.11).

Table 3.11: Sample coding for reasons of appropriateness

Respondent Translated response for ‘happiness’(Q14) Codes No. of tokens Readable/Easy to read 1 1 It’s easy to read, and approachable Familiar 1 Somehow it feels more familiar than 2 Familiar 1 ‘happiness’ written in katakana. It’s the freedom of the shop owner (to write 3 Freedom on the individual 1 how they like) TOTAL 3 4

Like the questions on script association (Q9, 10, 11, 12), some respondents gave more than one reason for their answer, as respondent 1 has done in table 3.11 above. For this reason, the number of tokens is necessarily more than the number of responses, and both figures are given in the results of the survey in chapter 5.

A difficulty which arose in analysing the data from question asking why the text was (in)appropriate (Q14) was the presence of contradictory answers. For example one respondent who selected ‘inappropriate’ for Q13 gave the following reason (Q14) ‘their business target is defined, and it has a good image.’52 This seems rather to be a reason for the text being ‘appropriate’, and a native speaking research assistant was consulted on these cases. When both the researcher and research assistant agreed that one of these responses had been made in error, the respondent’s answers for this text (for example, Q13, 14, and 15) were removed from the data set. Their answers relating to the other texts, if consistent, were allowed to remain.

The coding process described above was also conducted on the responses for the question that asked for the perceived reasons for writing happiness in hiragana (Q15). Like the responses for Q14, these were first separated into those who selected ‘appropriate’ and those who selected ‘inappropriate’ for Q13 in order to see whether this affected their results for the

52 ターゲットが明確で、イメージも良い 106

perceived motivations (Q15). The differences were few, and are also discussed in the results of the survey in chapter 5.

This coding process was repeated for the remaining texts in the survey, those containing ‘れもん/remon’ (lemon) [Q17,18], and ‘ふらい/furai’ (fly) [Q20,21] in hiragana; ‘珈琲/kōhī’ (coffee) in kanji (Q23, 24); and finally ‘hankachi’(handkerchief) in rōmaji (Q26,27).

Layer 3: Summary of research design

The online survey was used to obtain data directly related to research question 3, which examines the functions of loanwords in hiragana.

The survey responses also triangulated the findings of the descriptive statistics of the photo corpus (layer 1), for example, many respondents mentioned the genre of the text or the referential meaning of the loanword in hiragana in their answer. A reason for a judgement of appropriateness was sometimes linked to the fact that the loanword featured in the name of a product, for example, which according to the results of the corpus analysis was a common use of these words. As the survey used three of the case study (layer 2) texts, it could also be used to triangulate and expand on the findings of this layer.

Finally, the results of the survey were used to inform the design of the interviews (layer 4), with the texts judged to be most appropriate and most inappropriate being discussed as part of the session as described in the following section.

107

3.4.4 Layer 4: Interviews and focus groups

The interviews, referred to as layer 4 of the research design, comprise of qualitative data gained from the opinions of native speakers of Japanese regarding loanwords in hiragana. Their position as part of the research design as a whole is illustrated in table 3.12 below.

As described in the previous section, in order to investigate the functions of loanwords in hiragana, the opinions of native speakers of Japanese were sought through two methods, an online survey (described above in section 3.4.3), and interviews/focus groups. Along with the survey data, they addressed research question 3, ‘Why do loanwords appear in hiragana?’ as well as triangulating the findings of previous layers of analysis.

Table 3.12: Interviews and focus groups in the context of the research design

QUANTITATIVE DATA QUALITATIVE DATA

Layer 1 Layer 2 TEXTS Corpus analysis Case studies

Layer 3 Layer 4 AUDIENCE Survey Interviews and focus groups

Interviews and focus groups have different advantages as research tools. One-to-one interviews may allow people to speak their opinions more freely, as they do not need to worry about what other members in the group may think. This is particularly relevant in the Japanese context where the harmony of the group is often seen as paramount. Interviews also allow more freedom to pursue topics that are of particular interest to individuals. Finally, this format may be preferred by people with shy dispositions.

Focus groups, on the other hand, have the advantage of allowing the topic of discussion to be explored by the participants themselves. Individuals can describe their impressions, and other members can expand on them or disagree, providing alternative impressions. For this reason, the role of the researcher is minimized in this format, as they are less of a participant in the discussion. In this study, participants were given a choice as to whether they would like to be interviewed individually, or take part in a focus group.

108

Interviews and focus groups provide a useful way of collecting data on participants’ feelings and attitudes. Through well-planned and sensitive questioning, they help people to ‘make explicit things that have hitherto been implicit’ (Arksey & Knight, 1999, p. 32), and have an advantage over surveys in that the researcher/interviewer has the opportunity to ask follow up questions and clarify ambiguous answers (Gray, 2009). This was particularly useful in triangulating the results of the survey, as sometimes the brevity of the responses made them unclear.

Interviews and focus groups also allow participants time to reflect on the questions by talking through their answers, rather than feeling pressured to commit a response in writing. This is especially true of group interviews, where participants can explore their answers together.

Layer 4: Data Collection

The data gathering procedure for the interviews and focus groups is described in the following sections: participant recruitment, question design, and procedure.

Participant recruitment

Respondents who indicated on the survey that they would be interested in participating in an interview/focus group were contacted via email, asking them if this was still the case, and if so to nominate their preferred format (‘individual interview’ or ‘focus group’), location, day of the week, and time of day. They were also asked whether they would like to bring a friend or colleague to the interview, either creating a 2-person focus group or adding to a larger group. This final question was added with the aim of reducing anxiety for the participants as well as increasing numbers in the interviews. Responses were requested after a period of 10 days, and a further 10 days after that a follow up email was sent to those who had not yet replied.

Calls for participants were also sent out through the researcher’s personal networks. These were targeted to counteract the imbalanced demographics of respondents to the survey, as more females than males responded, and the majority were in their 20s and 30s (see section 5.1 for further details). For this reason, some participants in the interviews and focus groups had not taken the online survey. Once responses were received, participants were allocated to an interview/focus group and sent the details of when and where it would occur. They were

109

reassured that their personal information would be kept secure and that they could withdraw at any time.

Details of the interview and focus group participants can be found in the following tables (3.13 and 3.14). The gender (M/F) and age bracket (for example, ‘30s’) is indicated beside the participant’s pseudonym, as well as their home prefecture in cases where the individual was not born in the prefecture where the interview/focus group took place. When participants are referred to in the following sections, their session will also be mentioned for example

‘(G2)’ represents group interview 2, and ‘(S4)’ interview session 4.

Table 3.13: Summary of interview participants

Interview Location Participant

S1 Tokushima Jirō (M 50s)

S2 Osaka Rina (F 20s/Wakayama)

S3 Osaka Atsushi (M 20s)

S4 Kobe Naoya (M 50s)

S5 Tokyo Ryūichi (M 50s)

S6 Tokyo Yūsuke (M 20s)

110

Table 3.14: Summary of focus group participants

Focus Group Location Participants

Yuka (F 20s) Kazu (M 30s) G1 Tokyo Tomi (M 20s) Natsumi (F 20s)

Yui (F 30s) Ryō (M 30s) G2 Tokyo Natsuki (F 30s) Daisuke (M 30s)

Jun (30s) Tokyo G3 Miho (F 20s /Aomori)

Nagisa (F 30s /Saitama)

Yūta (M 50s) Kei (M 30s/Gifu) Nagoya Naoto (M 20s) G4 Gensuke (M 20s/Shizuoka)

Tarō (M 20s) Takuya (M30 s)

Taiyō (M 20s) G5 Kyoto Daichi (M 20s)

Ayano (F 30s/Shiga) Kyoto G6 Satomi (F 30s/Yamaguchi)

Chisato (F 30s/Fukuoka)

Osaka Kanako (F 20s/Wakayama) G7 Tomomi (F 20s/Kyoto)

Tokushima Hiromasa (M 50s) G8 Mio (F 40s/Kōchi)

Tokushima Shinji (M 30s/Nagoya) G9 Shōta (M 50s/Wakayama)

111

Interview Design

A semi-structured format was used for the interviews/focus groups, combining pre-prepared questions and stimuli, with a flexibility that allowed participants to spend more time on the questions that interested them, or to bring up related topics not included in the questions. This format results in qualitative data that which complements the quantitative data obtained from the survey. Each topic had a ‘key question’ which was used consistently for each group, however some additional follow-up questions were prepared for use when the key question did not generate as much discussion as predicted. In the interview/focus group script provided in appendix 4, key questions are marked with a star (★).

Each session contained 7 topics, and each topic was accompanied by colour photographic images or screenshots of examples of loanwords in hiragana related to that topic on laminated A5 cards, which were taken from the corpus. Like those used in the survey, images which made the context clear to the viewer were selected wherever possible, such as a product appearing on a supermarket shelf. Context is vital in understanding how meaning is made, as research in social semiotics, multimodality, and SFL and suggests (see sections 2.1, 2.2, 2.3 of the literature review respectively). A range of these picture prompts can be seen in figure 3.9 below.

Figure 3.9: Picture cards used in interviews and focus groups

The list of interview topics and the relevant text from the picture prompts used is summarized in table 3.15 below. The section ‘concluding questions’, did not utilize any additional picture prompts, but rather contained questions as described in the relevant section.

112

Table 3.15: Structure of the interviews and focus groups

Text contained in picture prompt Text contained in picture Interview topic (Japanese) prompt (English translation)

1. Refine 1. れふぁいん Introductory questions 2. Yuzu Lemon 2. ゆずれもん

3. Various soups 3. いろいろすーぷ Topic 1 4. Towel Muffler 4. たおるマフラー Product names 5. Splash Island 5. スプラッシュあいらんどう

6. らいふアシスト馬車道 6. Life Assist Bashamichi 7. はぐみぃー 7. Hug Me 8.ふろーりあ 8. Floria

9. Snack Shin Topic 2 9.すなっく新 10. Unicorn Business names 10. ゆにこーん 11. Lounge Chiaki 11. らうんじちあき 12. Spring すぷりんぐ 12. 13. Penguin 13. ぺんぎん 14. Idol 14. あいどる

15. Overall Man 15. おーばーおーるまん 16. I love you Topic 3 16. あいらびゅ 17. Volleyball Print club 17. ばれーぼーる 18. Fuck you 18. ふぁっきゅー

19. Family 19. ふぁみりー 20. New Zealand Topic 4 20. にゅーずぃーらんどぅ 21. Yes I am Social media 21. いえすあいあむ 22. Happy birthday 22. はぴばーすでい

113

23. Happy valentine Topic 5 23. はっぴーばれんたいん 24. Mix juice sour Number of consecutive words 24. みっくちゅじゅーちゅさわー

25. すぱいす 25. Spice あいすココア 26. 26. Ice Cocoa

27. とまと 27. Tomato Topic 6 28. りっぷくりーむ 28. Lip cream Inconsistencies 29. サラダようきゃべつ 29. Cabbage for salads 30. 宇治抹茶かすてら 30. Uji Matcha Castella

31. Cabbage display 31. キャベツ台

32. 蜂蜜れもん 32. Honey Lemon Topic 7 33. I can ふらい!! 33. I can fly!! ‘Appropriate’ and 34. [Alternatives transcriptions for ‘I 34. [Alternatives transcriptions for ‘inappropriate’ examples can ふらい!!’] ‘I can ふらい!!’]

Concluding questions - -

The full list of questions used in the interviews and focus groups, along with the pictures used, can be found in appendix 4, however the content of each of the topics is briefly described in the following paragraphs.

The introductory questions explained the topic of the interview, and two examples (pictures 1-2), representing a product and a business, two of the most frequently appearing referential uses of loanwords in hiragana in the corpus, were presented to provide an example of the kinds of texts to be discussed. Participants were asked if they had noticed loanwords being written in hiragana, and if they could remember any particular examples.

114

For topic 1, ‘product names’, three additional examples of products with names containing loanwords in hiragana were produced (pictures 3-5), and participants were asked the key question, which was why they believed the companies decided to use hiragana for these loanwords. For topic 1, the optional follow-up questions included whether the participants thought it created a particular image for the product, and whether the product would seem different if the word was written in katakana. This format of a key question and optional follow-up questions was used for each of the topics, however only the key questions are reported for the remaining sections.53

Topic 2, business names, followed a similar format, with examples of businesses with names including loanwords in hiragana (pictures 6-8) being provided, and participants being asked why they though this particular script choice was used. It was then revealed that the results of the corpus analysis showed that of all the businesses using hiragana loanwords for their names, hostess bars were one of the most frequent (see section 4.1.4), and additional examples of these kinds of businesses were presented (pictures 9-14). Participants were asked why they believed hostess bars were so well-represented in the corpus.

In topic 3, print club, examples of these texts were produced (pictures 15-18), and respondents were again asked why they thought the people wrote these loanwords in hiragana.

For topic 4, social media, participants were asked if they had used hiragana for loanwords in any of their posts, or if they remembered seeing loanwords written in that way. Then, examples of social media posts containing loanwords in hiragana (pictures 19-22) were presented, and participants were asked why they thought the writers chose to use hiragana for these loanwords. The follow-up questions for this topic included whether participants thought they could guess the gender of the writers from their posts, and how.

Topic 5, on the number of consecutive loanwords in hiragana, related to the results of layer 1, in which it was found that single loanwords in hiragana occurred more frequently that consecutive words (see section 4.1.7). Participants were presented with examples of two and three consecutive loanwords in hiragana (pictures 23 -24), asked why they thought these kinds of examples were less frequent.

53 As mentioned above, a full list of the key and follow-up questions can be found in appendix 4. 115

Topic 6 was designed to elicit responses to the different kinds of inconsistencies found in the corpus (see section 4.1.10). The first set of inconsistencies (pictures 25-26) showed examples of products with names including a loanword in hiragana, while the same word was written in katakana elsewhere in the text. Participants were asked why they thought the same script had not been used consistently in each case. The second set of inconsistencies (pictures 27-28) showed products with hiragana loanwords in their names, while the label created by the shop in which they were sold showed the same words in katakana. Participants were asked why they thought the makers of the labels had not used the same script as was used the product names. The third set of inconsistencies (pictures 29 -30) showed examples of the opposite case; products which had names containing loanwords in standard katakana, which were accompanied by labels which showed the same words in hiragana. Participants were asked why they thought the makers of the labels had chosen to use hiragana for these loanwords. For the final example of ‘inconsistency’, (picture 31), participants were asked for their reactions to a display of cabbages in a supermarket in which the quarter cabbages on the upper shelf were labelled in katakana, while the half cabbages on the lower shelf were labelled in hiragana.

Topic 7 related to the findings of the survey in which some of the sample texts containing loanwords in hiragana were judged to be more appropriate than others. The two texts which represented the upper and lower end of the spectrum were presented (pictures 32 -33), and participants were asked why they thought this difference in opinion occurred. In addition, they were asked about alternative transcriptions of the phrase ‘I can ふらい/furai (fly)!!’ which appeared in the text judged to be inappropriate by the greatest number of survey respondents (see section 4.2.3). Participants were presented with seven alternatives, such as having the entire phrase in hiragana ‘あいきゃんふらい/ai kyan furai’, which were designed to investigate why the survey respondents reported that this example in particular was ‘hard to understand’, and which aspect of the ‘mixing of scripts’ was felt to be problematic.

The concluding questions covered two aspects of the transliteration of loanwords; firstly participants were asked whether they considered examples such as the ones they had been shown ‘correct’ Japanese. The follow up questions included whether they would correct a 10- year-old child, who could be expected to know the conventions of script use, who wrote

116

‘lemon’ in hiragana; or a non-Japanese friend who was learning the language. The second aspect investigated the relationship between the script used and the consciousness of a word being ‘Japanese’, and included loanwords commonly written in hiragana from the corpus such as ‘れもん/remon’(lemon), ‘きゃべつ/kyabetsu’ (cabbage) and ‘ぱん/pan’ (bread); as well as ‘たばこ/tabako’ which is also not infrequently written in hiragana. Participants were also asked about loanwords which could be written in kanji such as ‘珈琲/kōhī’ (coffee), and ‘硝子/garasu’ (glass), since the presence of a kanji alternative was frequently given as a reason for appropriacy in the survey in relation to ‘れもん/remon’ (lemon) (see section 5.3.3).

A pilot interview was conducted in Melbourne with two Japanese participants to gauge the efficacy of the interview questions. No major changes to the interview questions or format followed the pilot interview, however it was found that when the official interview questions were exhausted and the interview was brought to a close by the researcher, participants were keen to add additional comments or return to previous examples they found interesting or surprising. For this reason, in future interviews the recording devices were not turned off immediately after the final question. A sample transcript of an entire interview can be found in appendix 5.

Procedure

Interviews and focus groups were conducted in a range of locations that most suited the group and location. These included private meeting rooms rented by the researcher or those in companies and universities, and participant’s homes for small or individual interviews.

The interviews/focus groups were conducted in Japanese by the researcher. At the beginning of each session, the researcher introduced herself and initiated an informal discussion on topics such as her stay in Japan so far, and whether the participants had been to Australia, to break the ice and set the informal tenor of the session. In focus groups, participants were asked to introduce themselves to each other, if they had not done so already and did not know each other.

The researcher then distributed the plain language statement and consent form, and explained that these documents contained details of the research, and that they should read them carefully and ask if they had any questions. These documents can be found in appendix 3. They were asked to sign the consent form and return it to the researcher. The researcher

117

confirmed that the interview/focus group would be recorded, either via a video recorder if the interview was held in a meeting room with more than two participants, or via a small audio recorder for smaller groups or public locations, and the recording was started at this point.

A short background questionnaire was distributed (see also appendix 3), and participants were asked to fill in personal details such as name, sex, and age bracket, as well listing the languages they spoke, and any experiences they had of living overseas. The reason the recording was started prior to this point was twofold; firstly it was hoped that this time would allow participants to become more used to the presence of the recording device, and secondly that any complex answers to the background questionnaire which the participants may have wanted to ask would be captured on the recording and could be used to supplement participants responses to the questionnaire later, if necessary. For example, participants may have wanted to discuss their proficiency in any foreign languages they spoke to determine what level they should select on their questionnaire.

After the questionnaires had been completed and returned to the researcher, the interview/focus group proper was started.

At soon as possible after the interview/focus group, the recorded data was backed up both to the researcher’s laptop computer and external hard drive, which was stored separately from the computer. The participants’ completed questionnaires and consent forms were also stored securely.

Layer 4: Data Analysis

An interview template was created listing the interview questions and sample pictures in the order they occurred in each interview. As each interview was played back, notes were taken in Japanese on points of interest that arose, and significant quotationss were transcribed in full. These significant quotations were also translated into English for future use. Both the notes taken from the recordings as well as the translations of the significant quotations were checked by a native-Japanese speaking researcher to ensure accuracy and consistency.

Once the template for each interview had been completed, general themes emerging from each topic of the interviews were considered across groups, and included both a similarities and differences of opinion as to the meaning or function of a particular example. These

118

themes are described in chapter 6, which contains the findings from the interviews and focus groups.

Layer 4: Summary of research design

The interview and focus group data, along with the survey data (layer 3), was most relevant to answering research question 3, which asks why loanwords appear in hiragana. This involved investigating the opinions of Japanese native speakers, in order to understand what kinds of meanings loanwords in hiragana held for them. The interviews complemented the findings of the survey, as a wider range of texts were used, and in many cases participants had opportunities to discuss their impressions with a group. In addition, as the interviews produced qualitative data, they aided in giving a deeper understanding of the quantitative data provided by the survey.

The interviews were also used to triangulate the findings of the layers of analysis which came directly from the corpus of examples, the statistical analysis (layer 1) and the case studies (layer 2). For example, participants’ reactions to the most frequently occurring genres of loanwords in hiragana reflected that they were commonly encountered in the participants daily lives, and many participants drew upon semiotic resources such as font or colour that were discussed in the case studies.

119

Chapter 4: Findings from the corpus analysis and case studies

This chapter will present the findings from the first two layers of analysis, which examined data gained from the corpus of hiragana loanwords itself. The descriptive statistical analysis (layer 1), takes a quantitative approach, investigating the corpus as a whole in order to illustrate some general characteristics of loanwords in hiragana. This information will be most useful in addressing the first research question, which asks how loanwords in hiragana can be characterised, as well as the second, which is designed to ascertain the contexts in which loanwords in hiragana are usually found.

The case studies (layer 2) were used to analyse selected examples from the corpus qualitatively in order to understand the role loanwords in hiragana play in multimodal texts. This information is most applicable to answering the third research question, which relates to why these loanwords appear in hiragana. As explained in section 3.4.2, the inclusion of case studies in the research design allow for a comprehensive understanding of how the different components of loanwords in hiragana interact with each other to produce meaning.

4.1 Layer 1: Characterising loanwords in hiragana

Between March 2014 and July 2016, 497 texts containing a total of 628 loanwords in hiragana were recorded. As described in section 3.4.1, the corpus was coded for the categories illustrated in table 3.4 in order to understand the general characteristics of loanwords in hiragana and the contexts in which they appear. While vowel lengthening, as a graphemic aspect and therefore the smallest unit of analysis comes first in table 3.4, it is discussed after more global traits such as word frequency and grammatical class in the following sections for ease of understanding.

4.1.1 Word Frequency

The corpus was analysed to investigate which words appeared in hiragana most frequently. The following table (4.1), shows the most commonly appearing words.

120

Table 4.1: Word frequencies

Word Frequency

Pan (bread) 19

Lemon, Happy54 15

Milk 14

Cabbage, Hot, Pudding 13

Castella, Tomato, Towel 11

Cream 10

Panda 9

Choc,55 Love56 8

Handkerchief, You 7

Petit, Pose, Potato, Wine 5

Café, Cake, Date, Festival, Hello, House, Lip, 4 Snack Banana, Baum Kuchen57, Cheese, Coffee, Family, 3 France, Gummy, Kitty, Last, Life, Sweet, One

The Portuguese loanword ‘ぱん/pan’ (bread) [19] appeared more frequently than any other in the corpus, and another Portuguese word, ‘かすてら/kasutera’ (castella)[11], also appeared more than ten times. As these words date from the period of Portuguese contact (1542 – 1639), they not only have a long history in Japan, but also entered the Japanese language during a period when the use of katakana for loanwords was not as prescribed as it is today (see section 2.4). A similar case can be made for the Meiji-era loans which appeared frequently in the above analysis, including ‘きゃべつ/kyabetsu’ (cabbage)[13]; ‘ぷりん /purin’ (pudding)[13]; ‘ たおる /taoru’ (towel)[11]; and ‘ はんかち /hankachi’ (handkerchief)[7]. These words can be placed in this historical period as their transcription in

54 The total of 15 represents 12 tokens for ‘はっぴー/はっぴい/happī’ (happy), and three for はぴねす /hapinesu (happiness). 55 The total of eight represents three tokens for ‘ちょこ/choko’ (choco-), two tokens for ‘ちょこれーと /chokorēto’ (chocolate), two tokens for the French ‘しょこら/shokora’ (chocolat), and one for the hybrid French-English sounding ‘しょこらあと/shokorāto’. 56 The total of eight represents eight tokens for ‘らぶ/rabu’ (love), and one for ‘らぶりー/raburī’ (lovely) 57 Baumkuchen is a type of German cake popular in Japan. 121

Japanese follows the pronunciation of the words rather than their spelling, which suggests they entered the language through an aural rather than written medium.

Another theme arising from the list of frequently occurring loanwords in hiragana (table 4.3) is that many of them refer to food and drink. This is again consistent with the analysis of referential meaning, (to be described in the following section, 4.1.4), which shows that these products were well represented in the corpus. Loanwords referring to food and drink may appear frequently in hiragana because they are well-established words; foodstuffs are often the first terms to be borrowed when different cultures come into contact. This may be due to the difficulty or clumsiness in translating these novel terms where no native equivalents exist.

In some cases, however, Japanese alternatives to loaned food terms appear to exist, and these are interesting in light of those words which appear frequently in the corpus of hiragana loanwords. For example, ‘とまと/tomato (tomato) [11] appears more frequently than ‘ぽてと/poteto’ (potato) [5], possibly because there is an alternative Japanese term for ‘potato’ (じゃがいも/jagaimo), for a writer who wishes to avoid the foreign overtones of a katakana word. However, it is noted that the two terms are not synonymous, as ‘ぽてと/poteto’ (potato) can also refer to ‘french fries’, and is rarely used in relation to the raw vegetable.

The frequent appearance of ‘みるく/miruku’ (milk) [14] in the corpus seems at first glance to be a counter-example, having a Japanese alternative, ‘牛乳/gyūnyū’ (milk). However, these terms are also not completely synonymous, and none of the examples of ‘みるく/miruku’ (milk) in the corpus refer to the generic dairy product; instead it is often used as an adjective meaning ‘milky’ or ‘creamy’ (eg. containing milk), or in reference to sweetened condensed milk.

One frequently occurring word was neither a food term nor a historical loan, ‘ぱんだ/panda’ (panda). This word is unlikely to be confused for a Japanese word, because of its connotations to China and Chinese culture, although it is not, in fact, a Chinese word. Interestingly ‘panda’ in hiragana was never used to refer to the animal itself, only cartoon- like panda characters. The forerunner of these panda-characters may have been ‘たれぱんだ /tarepanda/droopy-panda’, which was produced by Hikaru Suemasa for the San-X company in 1995, and featured on products such as notepads and stickers. It may not be a coincidence

122

that many of the semantic connotations of the animal-character, such as cute, soft, and round, are connotations reported for the hiragana script.

The list of frequently occurring words (table 4.2) contains two adjectives, ‘hot’ and ‘happy’. ‘Hot’ frequently occurred in the names of hot beverages, (eg. ‘ ほっとレモン/hotto remon’[hot lemon]), and as ‘ほっと/hotto’ can mean ‘relieving’ or ‘relaxing’ in Japanese, this punning effect may have been a reason for its popularity (see section 4.1.6 for more on language play). ‘はっぴー/happī’(happy) often occurred in generic phrases such as ‘happy birthday’ and ‘happy valentine’.

4.1.2 Word Class

Each word that appeared in hiragana was coded for its word class; noun, verb, adjective/noun modifier; an additional category, ‘other’ was created for other word types such as prepositions. This was calculated based on tokens rather than types, as many types had different word classes depending on their context. For example, ‘towel’ was used as an adjective/noun modifier in the phrase‘たおるマフラー/taoru mafurā’ (towel muffler), but as a noun in ‘手ぬぐいたおる/tenugui taoru’ (hand-drying towel). Many of the frequently occurring words, such as‘ちょこ/choko’ (chocolate), ‘くりーむ/kurīmu’ (cream) and ‘れもん/remon’ (lemon) were also used in this way, perhaps contributing to their frequency in the corpus.

The following table, 4.2, shows the distribution of grammatical word class of the 628 loanwords in hiragana that were collected. Percentages provided here and in the following sections have been rounded to the nearest whole number.

Table 4.2: Grammatical class of loanwords in hiragana

Grammatical class No. of tokens % of total tokens Noun 438 70 Adjective/noun modifier 137 22 Other 39 6 Verb 14 2

Total 628 100%

123

Nouns represented the highest proportion of the data at 70%, with the next largest category being adjectives/noun modifiers at 22%. These results are consistent with the naming or referential functions of loanwords in hiragana, which are described later in section 4.1.4 referential meaning.58 Verbs were the grammatical class least frequently found in hiragana, but this is likely to be also due to their scarcity in the Japanese loanword lexicon in general, rather than their unlikelihood of appearing in hiragana.

4.1.3 Vowel lengthening

Among the 628 words in the corpus, 169 long vowels occured, which as explained in the methodology section 3.4.1, could be lengthened by adding either a bar (consistent with loanwords in katakana) or an extra vowel character (consistent with Japanese words in hiragana). The following table, (4.3), shows the breakdown of the type of lengthening device used by vowel.

Table 4.3: Vowel lengthening by method and vowel

Vowel No. of bars Bar: % of total No. of extra vowels Extra vowel: % of total A 29 72 11 28

E 14 87 2 13 I 45 76 14 24 O 24 83 5 17 U 17 68 8 32

Total 129 76 40 24

Overall, a bar was the most common method of lenthenging vowels within loanwords in hiragana, being used in 76% of cases. However, the breakdown of method by vowel shows some slight differences within this figure. For lengthening ‘e’ sounds, bars were used 87% of the time, in other words in all but two cases, for example ‘ちょこれーと/chokorēto’ (chocolate). This long vowel was also the least frequently occurring, however, so this finding is based on a relatively small number of examples. The vowel which showed the most variation was ‘u’, with 68% of cases utilising a bar, for example, ‘みゅーじあむ/myūjiamu’

58 This is interesting in the context of Joyce et. al’s (2012) findings on the ‘Balanced Corpus of Contemporary Japanese (BCCJ)’, in which nouns were the lexical category which showed the least orthographic variation. 124

(museum), with the remaining 32% of examples adding an extra vowel, for example ‘さんきゅう/sankyū’ (thank you).

Langauge internal factors are likely to have an effect on these preferences, for example ‘e’ is a sound which is not often elongated in Japanese words, and therefore it is perhaps this association with ‘foreign’ words that results in a preference for the bar, also associated with loanwords.59 On the other hand, ‘u’ can be found elongated in Japanese words, which may increase the chances of the Japanese-style extra vowel technique being considered an option.

The consonant attached to the vowel is also likely to have an effect, for example, by comparing the examples of ‘kyū’ and ‘myū’ above. The sound ‘kyū’ can be found in Japanese words such as ‘きゅうり/kyūri’ (cucumber) and ‘階級/kaikyū’ (rank, grade), which perhaps lead to the choice of a vowel in the example of ‘さんきゅう/sankyū’ (thank you) above. On the other hand, ‘myū’, is a sound that does not occur in Japanese words, which may be the reason for the use of a bar in ‘みゅーじあむ/myūjiamu’ (museum).

The sound ‘o’ can be elongated with two different kana in Japanese, an extra ‘お/o’ as in ‘多い/おおい/ōi’ (numerous), or with an extra ‘う/u’, for example ‘応援/おうえん/ōen’ (support), this latter method being more common. Of the five hiragana loanwords which used an extra vowel to lengthen an ‘o’ sound, four used the ‘u’ sound, for example ‘ぼうろ/bōro’ (bolo, a type of Portuguese biscuit), while only one used an extra ‘o’, ‘こおふぃ/kōfi’ (coffee).60

Within these general tendencies, variation did occur. For example, words that appeared more than once in the corpus were sometimes elongated in different ways, for example ‘はっぴー’ and ‘はっぴい’ for ‘happī’ (happy), and ‘しーる’ and ‘しいる’ for ‘shīru’ (seal or sticker). Other types of variation included the use of different methods for different words in a single phrase, for example ‘ほっぷしゅうくりーむ/hoppu shū kurīmu’ (Hop Choux Crème) and the use of both strategies for a single long vowel, as in ‘でぇーと/deēto’ (date) which are discussed in section 4.1.10, ‘Inconsistencies’, below.

59 It is noted that an elongated ‘e’ sound can be represented with the characters ‘えい/ei’ as in ‘英語/eigo’ (English language), although no examples of this method were found in the corpus. 60 This is not the only departure from the more common transliteration, ‘コーヒー/kōhī’, and may relate to a move towards more native like pronunciation (see section 6.2.8), as well as the trans-orthographic nature of this word in particular (see section 7.2). 125

4.1.4 Referential meanings of loanwords in hiragana

This category considered what the loanword in hiragana referred to, for example, the holiday cruise company ‘さんふらわー/sanfurawā’ (Sunflower) has the referential meaning of the name of a business, not the name of a type of plant. The results of the coding for referential meaning can be found in table 4.4 below.

Table 4.4: Referential meanings of loanwords in hiragana

% of total Referential Meaning No. of word/phase units units Product name 230 45

Business name 110 21

Message, wish or greeting 32 6

Publication title 22 4

Relationship or person description 22 4

Person, group or character name 21 4

Concrete noun 16 3

Menu item 12 2

Event 10 2

Product description 9 2

Place name 7 1

Other 6 1

Blog title or post title 5 1

Business description 4 1

Website title or section 4 1

CD or DVD title 3 1

Service name 3 1

Total 516 100%

The loanwords in the corpus were used to refer to product names in just under half of the texts. A further 21% of the data was accounted for by business names. The remaining third of the data was spread over 17 further categories, each representing 6% or less of the data.

126

The product and business names in the corpus show many of the traits that Davies (1987) identified in her analysis of non-standard spellings in such names. For example, loanwords in hiragana were used to differentiate the name of the product from a simple description of it, or its ingredients, for example the instant curry mix called ‘すぱいす/supaisu’ (Spice). It may also be an attempt to distinguish the product from other similar ones, for example in the large number of drinks using ‘れもん/remon’, in their name, for example ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon); ‘ゆずれもん/yuzu remon’ (Yuzu Lemon); and ‘ほっとれもん /hotto remon’ (Hot Lemon). Hiragana may have been utilized in an attempt to distinguish these products from other ‘レモン/remon’ (lemon) flavoured ones, but the popularity of this ‘unusual’ script choice may itself be becoming cliché. Davies also notes that non-standard spellings in product and business names can represent either simplification or elaboration of the original word;61 and loanwords in hiragana can be seen as simultaneously representing both. Hiragana is said to be the ‘simplest’ of the Japanese scripts, but its marked use for loanwords can also be seen as elaboration of the katakana norm.

As the referential meaning categories of ‘product name’ and ‘business name’ represent significant portions of the corpus, they were further broken down into more specific categories. Table 4.5 below shows the types of products found in the corpus.

Table 4.5: Product names by type

Product type No. of word/phrase units % of total product types

Processed food 112 49

Omiyage (processed) food 33 14

Fresh produce 18 8

Personal accessories 17 7

Other product 15 7

Toys and collectables 15 7

Cosmetics, toiletries 10 4

Stationery 10 4

Total 230 100%

61 For example ‘eeeezy’ is an elaboration of the word ‘easy’, while ‘kwik’ is a simplification of ‘quick’. 127

Within product names, the largest category was processed food at 22% of the whole; and included both food and drink products, such as ‘赤わいん/aka wain’ (Red wine) pictured below in figure 4.1. ‘Omiyage’ or ‘souvenir’ food was counted separately but also represented processed food, and contributed a further 6%,62 for example ‘京都抹茶みるくキャラメル/kyōto matcha miruku kyarameru’ (Kyoto Green-tea Milk Caramels, pictured in figure 4.2 below. The association of hiragana script with traditional Japanese culture as described later in the results of the survey (section 5.2.1) are possible motivations for its use in souvenir products. The use of hiragana as a marker of tradition culture across the four layers of the study is discussed in section 7.3.1.

Fresh produce represented the next largest category having 18 different texts but was limited to only two types, ‘きゃべつ/kyabetsu’ (cabbage) with 11 tokens, as pictured in the price label in figure 4.3 below, and ‘とまと/tomato’ (tomato) with seven tokens. Other product names included personal accessories such as hand-towels and scarves, toys such as collectible figures, and toiletries such as hand creams.

Figure 4.1: Processed food/drink, ‘Red Wine’ Figure 4.2: Omiyage food, ‘Kyoto Green tea Milk Caramels’

62 Purchasing souvenirs for friends, family, and work colleagues when on holiday or business trips is an important part of Japanese culture. A large industry has developed in manufacturing items (usually containing local produce) that make suitable gifts. 128

Figure 4.3: Fresh produce, ‘cabbages’

Since business names also accounted for a large proportion of the data, they were further broken down into more specific categories, illustrated in the following table (4.6).

Table 4.6: Business names by type

Business type No. of word/phrase units % of total business types Services 27 24

Restaurant/café 27 24

Hostess bar 25 23

Goods 21 20

Other/unknown 10 9

Total 113 100%

Business names were quite evenly spread across the categories of businesses providing services (5%), such as ‘らいおん歯科/raion shika’ (Lion Dentistry) in figure 4.4; restaurants (5%), including ‘すてーきはうす壽々/sutēki hausu juju’ (Steak house Juju) in figure 4.5,

129

and hostess bars63 (5%), such as ‘Snack えんどれす/Snack endoresu’ (Snack Endless)64 pictured in figure 4.6.

Figure 4.4: Business name, ‘Lion Dentistry’ Figure 4.5: Restaurant name, 'Steak House Juju'

Figure 4.6: Hostess bar name, 'Snack Endless'

To summarize this section, loanwords in hiragana frequently have the referential meaning of the name of a product (45%) or business (21%). Within products, processed foods accounted for the majority of the examples; while in the business category, hostess bars, restaurants, and shops providing services, were the most frequently occurring categories.

4.1.5 Source language

The source language of the loanwords in the corpus was overwhelmingly English (82%), with smaller numbers represented by European languages such as French and Portuguese (6% each), as shown in table 4.7. The proportions of the various donor languages of the loanwords

63 ‘Hostess bars’ are common in Japan, and their customers, who pay inflated prices for their drinks, receive the attention of a ‘hostess’ who will flirt with them and pour their drinks. ‘Snack bars’ are a slightly more informal and cosy version of this arrangement, but for the purposes of this study are combined under the term ‘hostess bars’. 64 Design considerations including the typeface and layout are visible in this example, in the elongated stroke of the character ‘れ/re’ and its connection to the final character ‘す/su’. This would be more difficult to achieve with the angular katakana ‘レ/re’ and ‘ス/su’. 130

in the data in this study is therefore consistent with the proportions of donor languages for loanwords reported elsewhere (see section 2.5). The reason English was the most frequently occurring language in the corpus of hiragana loanwords may be because it is the most frequent donor of loanwords to Japanese in general, however further theories are explored in the discussion chapter (section 7.2).

Table 4.7: Source language of loanwords in hiragana

Source language No. of types % of total types No. of tokens % of total tokens

English 269 85 425 82

Portuguese 4 1 33 6

French 21 7 31 6

Other 10 3 11 2

Italian 7 2 7 1

German 2 <1 5 1

Dutch 1 <1 2 <1

Latin 2 <1 2 <1

Total 316 100% 516 100%

Of the languages other than English, Portuguese words can be dated to the time of contact in the 16th century when missionaries arrived from that country (see section 2.5.1). There is little variety within the Portuguese words, however, as the 33 tokens contain only 4 types; ‘たばこ /tabako’ (tobacco); ‘ぱん/pan’ (bread); a type of biscuit called ‘ぼうろ/bōro’ (bolo); and the sponge cake ‘かすてら/kasutera’ (castella). Examples of Portuguese loanwords in hiragana can be found in figures 4.7 and 4.865 below.

Figure 4.7: Portuguese loanword, ‘pan’ Figure 4.8: Portuguese loanword, 'castella'

65 Note the use of katakana for another loanword, ‘プリン/purin’ (pudding), in a similar sign on the right. 131

French loanwords date from a more recent period of contact, but were still represented in the data (6% of total tokens). They have found their way into Japanese through their associations with fashion, shopping and food; examples include ‘ぷち/puchi’ (petit/small), ‘まかろん /makaron’ (macaron), and ‘まるしぇ/marushe’ (marche/market). A more unusual example is the word ‘えすぽわーる/esupowāru’ (hope), used as the name of the maid-café and bar shown in figure 4.9 below. They represent a wider variety of words than the Portuguese data; 21 types for 31 tokens. Seven Italian words were also present in the corpus, such as the name of the magazine ‘ぶらあぼ/burābo’ (Bravo) pictured below (figure 4.10).

Figure 4.9: French loanword, 'Espoir' Figure 4.10: Italian loanword, 'Bravo'

4.1.6 Language play

Examples of language play within the corpus were investigated based on the long history of puns and word play in Japanese (see section 2.7). In total, 38 examples of language play (7% of the word/phrase units) were found in the corpus of hiragana loanwords collected in layer 1. A summary of the different types of word play are given in table 4.8 below.

The two most common types of word play were homonyms and lexical blends. Homonyms accounted for 16 of the examples, represented cases where the loanword had a homonymous Japanese meaning, for example in the brand of drink called ‘かぼすサイだぁー/sai dā’ (Kabosu66 Cider), pictured in figure 4.11 below, which was written partially in hiragana to suggest the secondary reading ‘it’s a rhinoceros’, the company’s mascot being a cartoon drawing of this animal. In one of these examples, multiple meanings including an English homonym can be seen. The dog trimming salon called ‘ほっと DOG/hotto DOG’ (Hot Dog)

66 ‘Kabosu’ is a kind of Japanese citrus fruit (Citrus sphaerocarpa), with which this cider is flavoured. 132

plays on the American fast food ‘hot dog’, as well as the slang ‘hot’ meaning fashionable or cool. The use of hiragana for the word ‘hot’, however, brings to mind the Japanese meaning ‘ほっと/hotto’, relieved or refreshed, perhaps how one’s dog might feel after visiting the salon.

Table 4.8: Types of language play containing loanwords in hiragana

Type of No. of Japanese English language play Example Meaning(s) examples element element

かぼすサイだぁーサイだぁーサイダー Homonym 16 (Kabosu) saidā sai dā saidā cider/it’s a rhino (drink) It’s a rhino cider

おわらいおん/ おわらいライオン Lexical blend 16 owaraion owarai raion Comedian-lion (character name) comedy lion

はとハートはあと/hāto Paronym 3 hato hāto heart/dove (Biscuit brand) dove heart

葉っぴいてらす葉 / はハッピーテラス Happy (leafy) happī terasu ha happī terasu Concealed word 3 terrace (train station seating leaf happy terrace

area)

133

Figure 4.11: Example of homonym, 'Cider' Figure 4.12: Example of lexical blend, 'Koedarize'

Figure 4.13: Example of concealed word, 'Happy Terrace'

Lexical blends were the other major category of language play, in each case combining a Japanese element with a loanword element. For example, the cartoon character ‘おわらいおん’ (owaraion) combines the Japanese word for stand-up comedy, ‘おわらい/owarai’ with the loanword in hiragana ‘らいおん/raion’ (lion) to create a single word meaning something like ‘comedian-lion’. In three cases, the loaned element was not a word but a morphological ending. Two of these used ‘-ist’, for example the local government publication titled ‘くらしすと/kurashisuto’, combining the Japanese ‘暮らし/kurashi’ (life, livelihood) with ‘いすと /isuto’ (-ist) to perhaps represent a person who was an expert on ‘living’, as the publication contained advice on taking out personal loans and so forth. The other morphological element used was the English ‘-ize’, for example in the name of the collectable figures called ‘こえだらいず/koedaraizu’ pictured in figure 4.12. These figures represented characters from anime

134

which had been given the same bodily proportions as the children’s doll, ‘こえだちゃん /koeda-chan’ (Little Miss Koeda), namely large heads with small faces and bodies. In other words, these characters had been ‘koeda-rized’, or made to look like Koeda-chan.

The corpus also contained three paronyms, or partial homonyms; these represent cases where the two readings contained are similar rather than identical as they are in the case of homonyms. The biscuits called ‘ はあと/hāto’ are a representative example; the transliteration suggests the loanword ‘heart/hāto’, while the hiragana script suggests the word ‘はと/hato’ (dove). The design of the packet combines both these meanings, depicting a dove within a wreath of hearts, carrying a heart-shaped flower.67

The final category of language play was ‘concealed word’, this was a term coined to account for examples which, through the mora and/or script used, suggested a further word within the word/phrase depicted, and have similarities to the ‘graphic play on words’ (Renner, 2015, p. 127) discussed in section 2.7. These were differentiated from lexical blends by having no clipped or shared mora. The rooftop garden at Kyoto Station called ‘葉っぴいてらす/happī terasu’, pictured in figure 4.13 above, is one such example, the sounds or morae alone suggest the loaned phrase ‘happy terrace’, but the syllable ‘は/ha’ being replaced with the kanji for ‘葉/ha’ (leaf) suggests leafiness as well, and is a type of language play only realized in the written form. The loanword ‘terrace’ may also be playing on the homophonic Japanese word ‘照らす/terasu’ (to shine), bringing an element of sunshine to the outdoor terrace.

4.1.7 Consecutive loanwords in hiragana

The corpus was coded to investigate whether loanwords in hiragana were more likely to occur as single words, or as two or three consecutive word combinations. A category ‘partial word’ was created to account for loanwords which appeared partially in hiragana, often as instances of a lexical blend as described above (section 4.1.6). The following table, 4.9, shows the results of this analysis of each of the 516 units (words or phrases) in the corpus.

67 A paronymic example was included in the interview prompts, the children’s rhythmics classroom called ‘はぐみぃー/hagu mīi’ (Hug me), and is discussed in section 6.2.9. 135

Table 4.9: Consecutive loanwords in hiragana

Code No. of units % of total units

Single word 400 78

2 consecutive words 90 17

Partial word 15 3

3 consecutive words 11 2

Total 516 100%

Of all the units of loanwords in hiragana present in the corpus, 78% were examples of a singular loanword in hiragana, such as ‘はろー/harō’ (hello) which appeared on a print club image; some of these occurred in combination with other non-hiragana loanwords, for example the cheese-stuffed meat patty ‘ちいずインつくね/chīzu in tsukune’ (Cheese-in patty) which was found on a restaurant menu (figure 4.14). The number of consecutive words was inversely proportionate to the number of occurrences in the corpus; in other words two consecutive words such as the salad bar called ‘ぽてとはうす/poteto hausu’ (Potato House) shown in figure 4.15 below were rarer, accounting for 17% of the examples, while three word combinations such as the manga titled ‘ぱわーおぶすまいる/pawā obu sumairu’ (Power of smile) in figure 4.16 were rarer still (2%). In this case, a typeface of a different size and colour has been used to distinguish the middle word, ‘おぶ/obu’ (of), suggesting that an unmarked chain of three hiragana words is even less common. Even partial words, such as those described in the context of lexical blend in section 4.1.6 above, were slightly more common than three word combinations, with 3% of all units.

Figure 4.14: Single loanword in hiragana, ‘Cheese-in patty’

136

Figure 4.15: Two consecutive words, ‘Potato House’ Figure 4.16: Three consecutive words, ‘Power of Smile’

The reason for the large number of single units may be related to intelligibility; Sebba (2012), notes that users of non-conventional orthography must ensure their message is still recoverable by the reader, and indeed some participants in the focus groups and interviews noted that consecutive loanwords in hiragana were harder to read or understand (see section 6.2.4). This balancing of innovation and intelligibility in creative orthography has also been mentioned in relation to brand names (Davies, 1987), which is a well-represented genre in the corpus (see section 4.1.9).

4.1.8 Instances of hiragana loanword(s) per text

While the number of consecutive words was one measure of the use of loanwords in hiragana, it was also useful to analyse the number of units (words or phrases) which occurred in a single text. The following table (4.10) shows how many units of loanwords in hiragana were present in each of the 497 texts.

The vast majority (97%) of the texts contained only one loanword or phrase (unit) in hiragana, for example the sign announcing a restaurant’s ‘りにゅーあるおーぷん/rinyūaru ōpun’ (renewal-open; reopening after renovation), where this was the only instance of loanword(s) in hiragana in the text.

137

Table 4.10: Units of hiragana loanword(s) per text

Code No. of texts % of total texts 1 unit 482 97 2 units 11 2 3 units 4 1 Total 497 100%

Texts with two units represented only 2% of the data, for example the packet of lollies called ‘れもんこりっと/remon koritto’ (Lemon Koritto) 68 pictured in figure 4.17 below. The product name contains ‘れもん/remon’ (lemon) in hiragana which constituted one unit, while in the bottom right corner a diagram shows the lollies contain layers of ‘サクサクみるくミルフィーユ/sakusaku miruku mirufīyu’ (crispy milk[y] mille-feuille), the loanword ‘みるく /miruku’ (milk) forming the second. Only four texts containing three units of loanword(s) in hiragana were found, accounting for approximately 1% of the data.

Combining these results with those in the previous section (4.1.7, ‘Consecutive loanwords in hiragana’) suggests that a typical text in the corpus contains a single loanword in hiragana, possibly two, and that these words will appear consecutively, in other words, this unit is the only place in the text where loanwords in hiragana occur. It should be remembered that some of the texts in the corpus were relatively simple, such as many of the hostess bar signs, in which the only linguistic information is the single-word name of the bar, such as ‘ぷらちな /purachina’ (Platinum), which also reduced the probability of multiple words or units within a text.

Figure 4.17: Text containing two units, ‘Lemon Koritto’

68 ‘Koritto’ seems to be an onomatopoeic expression suggesting ‘snap’ or ‘crunch’. 138

4.1.9 Genres containing loanwords in hiragana

As context plays a central role in meaning making in fields associated with social semiotics (see section 2.1), it was important to determine the genres in which loanwords in hiragana were used. The following table represents the breakdown of genres represented in the corpus (table 4.11).

Table 4.11: Genres of text containing loanwords in hiragana

Genre No. of texts % of total texts

Product packaging 155 31

Business sign 82 17

Product label or sign 58 12

Websites 51 10

Print club images 43 9

Publications 42 8

Advertisement 21 4

Menu 17 3

Product 11 2

Other 6 1

Information sign 5 1

Directional sign 4 <1

CD packaging 2 <1

Total 497 100%

The most common genre in the corpus was product packaging, accounting for 31% of data, for example where the name of the brand for the quilting hoop, ‘あるふぁー/arufā (Alpha) was found, as pictured in figure 4.18 below. A further 18% of the texts were classified as business signs, for example that attached to the restaurant ‘鶏豚きっちん/toriton kitchin’ (Chicken-Pork Kitchen) in figure 4.19. The next largest category, product labels (13%) were sometimes found alongside products with hiragana names, which accounts for this category also being well represented in the data, and sometimes independently as in the display of ‘たおるの帽子/taoru no bōshi’ (Towel[ling] hats) in figure 4.20. Smaller numbers of texts were also found representing websites, which included commercial sites as well as social

139

networking sites such as ‘Facebook’; print club images; and publications which included books, brochures, manga, and magazines.

Figure 4.18: Product packaging, ‘Alpha’ Figure 4.19: Business sign, ‘Chicken-Pork Kitchen’

Figure 4.20: Product sign, ‘Towelling hats’

Links between the genres well-represented in the corpus and the largest categories of referential meaning of the loanwords in hiragana found in the corpus can be seen. For example, for genre ‘product packaging’ accounted for 31% of the data and was the most frequently occurring code, and for referential meaning, ‘product name’ at 45% was also the most frequently occurring code. This suggests that product names were often found on product packaging. A similar link can be seen in the next most frequently occurring codes; the genre of ‘business sign’ (16%), and the referential meaning of ‘business name’ (21%).However, it is noted that the genres investigated in this study were necessarily limited in scope, and that an investigation of additional sites of loanwords in hiragana may be able to expand on these findings.

140

4.1.10 Inconsistencies in transcription

Within the texts containing loanwords in hiragana collected for the corpus, different types of inconsistencies were noted. These were coded and the results appear in the table 4.12 below

Table 4.12: Types of inconsistency in the corpus

No. of % of total Type of inconsistency Example examples inconsistencies

ひなちょこ/hina choko (Girl’s festival Alternative script for a chocolates) particular loanword within 65 70 text チョコでくるんだひなあられ/choko de kurunda hina arare (Girl’s festival snacks rolled in chocolate)

いちごみるく餅/ichigo miruku mochi Alternative script for a (strawberry milk mochi: packaging) particular loanword between 19 21 related texts いちごミルク餅/ ichigo miruku mochi (strawberry milk mochi: price label)

Vowel inconsistency でぇーと/deēto (date) 8 9

Total 92 100%

The most common type of inconsistency was the use of an alternative script for the loanwords which appeared in hiragana in the text (70%). The most common format was in which the hiragana loanword appeared as a product name, for example ひなちょこ/hina choko (Girl’s festival chocolates), while the explanation elsewhere on the packaging used standard katakana for the same loanword, チョコでくるんだひなあられ/choko de kurunda hina arare (Girl’s festival snacks coated in chocolate).69 In some cases, the alternative script was kanji or rōmaji/English.

69 ‘Hina-arare’, a snack made of puffed rice, is traditionally eaten at Hina-matsuri (Girl’s Festival/Doll festival). 141

The second most common inconsistency was the use of an alternative script for the hiragana loanwords between related texts (21%), 13 of which related to the product name and the price tag created by the shop in which it was sold. For example, the product called いちごみるく

餅/ichigo miruku mochi (strawberry milk mochi)70 was found accompanied by the label ‘いちごミルク餅/ ichigo miruku mochi’ (strawberry milk mochi), as pictured in figure 4.21 below. In other words, although the word ‘miruku’ (milk) appeared in hiragana in the product name, the label shows the same word in katakana.

Figure 4.21: Inconsistent label, ‘Strawberry milk mochi’ Figure 4.22: Inconsistent label, ‘Mushroom cabbage mix’

In four of the 13 cases, the inverse situation was found, a product with a conventional katakana loanword in the name was presented with a price label which used hiragana for the same loanword. Figure 4.22 above shows an example; the prepared vegetable pack ‘雪国きのこキャベツ MIX/yukiguni kinoko kyabetsu MIX’ (Snow-country mushroom cabbage mix) has the word ‘cabbage’ in standard katakana (キャベツ/kyabetsu), but this word appears in hiragana (きゃべつ/kyabetsu) in the attached price label. Inconsistencies such as these were discussed in the interviews and focus groups (see section 6.5).

The remaining six cases of alternative script use occurred between related texts; for example, the beauty salon shown in figure 4.23 below has two signs containing the same information in different scripts, ‘美容室 MILK/biyōshitsu MILK’ and ‘美容室みるく/biyōshitsu miruku’ (Beauty Salon Milk).

70 ‘Mochi’ is a traditional Japanese sweet made from pounded rice. 142

Figure 4.23: Inconsistent related texts, 'Milk'

Inconsistencies were also found in the method for elongating vowels in the words present in the corpus (9%). Cases were found where both an extra vowel (in the style of Japanese words in hiragana) and a bar (in the style of loanwords in katakana) were used, for example in the word ‘でぇーと/deēto’ (date) which appears on the print-club image in figure 4.24 below. In all cases, the extra vowel was a half-size character, which is not typically used as a lengthening device, and always occurred before the bar. The combination of extra vowel character and bar could have an emphasising effect, depicting a ‘really long’ vowel, or perhaps is simply indicative to the tension between maintaining consistency with the hiragana script, and with the word’s origins as a ‘loanword’. In other vowel inconsistencies, different methods were used for different words in the unit or text, for example in figure 4.25 below, the pastry shop named ‘ほっぷしゅうくりーむ/hoppu shū kurīmu’ (Hop Choux [a la] crème) used an extra vowel for the word ‘choux’ and a bar for the word ‘crème’, although this was the only example of this mix of vowel lengthening occurring in a phrase.

Figure 4.24: Inconsistent vowel elongation, 'date' Figure 4.25: Inconsistent vowel elongation, ‘Hop choux crème’

143

4.1.11 Summary of layer 1: the characteristics of loanwords in hiragana

The findings from the first layer of analysis address the first research question, ‘How can loanwords in hiragana be characterised?’ At the word level, nouns were the most common word class in the corpus, and an investigation of the most frequently occurring words show older loanwords and/or those relating to foodstuffs to be common, in particular ‘ぱん/pan’ (bread), ‘れもん/remon’ (lemon), and ‘きゃべつ/kyabetsu’ (cabbage). When units of text were considered, most loanwords in hiragana occurred as single words and the source language of these units was usually English, although some words of Portuguese origin (ぱん /pan and かすてら/kasutera) occurred with high frequency. An analysis of the referential meaning of such units showed that loanwords in hiragana were often used for product names and business names. Language play was also evident at the unit level, in 7% of cases, with homonyms and lexical blends being most common. At the grapheme level, vowels were found to be usually elongated with a bar, in the style of katakana loanwords.

At the text level, genres such as product packaging and business signage were well- represented. Other genres such as publication names, social media posts, and print club images also occurred in smaller numbers. Within texts, a single unit containing (a) loanword(s) in hiragana was more common than multiple units. In other words, if two hiragana loanwords appeared in a single text, they were most likely to occur consecutively, rather than in different parts of the text.

Inconsistencies within texts were found in terms of the use of script for a particular loanword, with the same word occurring in two different scripts. The use of different scripts in related texts with the same referent (eg. a product’s packaging and its price label) were also found. A final type of inconsistency sometimes occurred with the method for lengthening vowels, which could differ within or between word units in a single text.

The above analysis of 628 loanwords in hiragana from the accumulated 497 texts was essential for establishing loanwords in hiragana as an object of study, and mapping their fundamental characteristics. These characteristics were used to select examples for the following layers of analysis, the case studies (layer 2), survey (layer 3), and interviews and focus groups (layer 4), to ensure a range of examples, both typical and atypical, were included. The findings of the corpus analysis also guided the design of the interviews and

144

focus groups, for example ensuring the most frequently occurring genres of texts were discussed, and including questions which asked about texts containing inconsistencies, and the relative infrequency of two and three consecutive loanwords in hiragana, for example.

4.2 Layer 2: Loanwords in hiragana in multimodal texts

In the following sections, selected individual examples from the corpus will be subjected to a detailed analysis in order to investigate how hiragana loanwords are being used along with other linguistic and visual elements of a text to create meaning. As described in the research design (section 3.3), a combination of Stöckl’s focus on typography as a semiotic resource (2005), as well as the ‘visual grammar’ approach, as described by Kress and van Leeuwen (2006), will be used to analyse each image.

4.2.1 Case 1: ‘Happiness’ Clinic

In this section, the results of the analysis of the sign for the business ‘はぴねす鍼灸整骨院/ hapinesu shinkyūseikotsuin’ (Happiness Clinic) will be presented in order to demonstrate how hiragana functions as a semiotic resource in a multimodal whole. The insights gained from this analysis will be discussed in relation to the findings of the survey (layer 3) in which it featured. The text is reproduced in figure 4.26 below, along with an English translation of the written elements. The text will firstly be analysed in terms of the ‘visual grammar’ of Kress and van Leeuwen (2006).

Lower back (pain), stiff neck, joint pain, headaches

Please consult us about your body worries!

06-6633-6636

Happiness Moxibustion and Acupuncture Clinic

Figure 4.26: ‘Happiness’ clinic sign with English translation 145

Firstly, the colour palette of the text is generally warm, consisting of colours such as orange, yellow, red, and a yellowish green. The warm palette suggests the ‘warmth’ of the atmosphere of the clinic, while the green may represent ‘nature’ or ‘naturalness’, which is supported by the leaf motif underlining the business name. The company logo in the bottom left corner may also represent a plant or flower. Colour is also used to distinguish the different types of pain in the circles at the top of the text, as well as divide the types of information given in the text, the invitation to consult (green), phone number (red), and business name (orange). These colours are also all present in the business logo on bottom left.

The text contains many round shapes, such as the circles showing the types of ailments treated at the top; an oval shape depicting the interior of the clinic; as well as the ovals and circles which make up the company logo. The use of these round shapes suggests an organic, natural order (Kress & Leeuwen, 2006, p. 55), and this feeling of naturalness is also suggested in leaf-like logo, and leaf underline as mentioned above. This presents the treatment options available at the clinic as ‘natural’ forms of pain relief, rather than a drug treatment, and is evidence of the ideational function. In addition, the round shapes give the text a soft, gentle image, which is also reflected in the choice of hiragana for the company name. For the designers of this text, both the circular motifs and the use of hiragana may have been selected because the corners of square shapes and the angular katakana may have been felt to give a sharp or ‘painful’ image – one which a business offering acupuncture would presumably have been keen to avoid.

The four circles at the top of the text show four people suffering from different types of ailments, back pain or headaches, for example. In each case the painful area is highlighted by a light grey circular shape in the background, and jagged orange lines representing the pain. These four images are presented as a taxonomy, being the same size and using the same colours and format (what van Leeuwen calls a ‘rhyme’[2005, p. 9]), they can be read as a set of images illustrating the kinds of pain that the clinic can treat. This message is supported by the linguistic text in green, ‘Lower back (pain), stiff neck, joint pain, headaches. Please consult us about your body worries!’ The circles and the background colours of red, yellow,

146

green and orange may also appear particularly salient in the image as they bear some resemblance to a traffic light – something which demands attention.71

These pictures have reduced modality in terms of abstraction, in other words, they are cartoon-like pictures,72 rather than photographic images. On the other hand, they are not ‘stick figures’, and nor are they decontextualized images of knees and shoulders, for example. This level of abstraction allows potential patients to identify with these images, they can be seen, for example, as a ‘woman-like-me’, rather than a photo of a ‘woman-who-isn’t-me’. The idea that potential patients can identify with these images is also reflected in the fact that the eye gaze of these represented participants does not seem to be at the viewer, their gazes either focus on the painful joint, away into the distance or down at the ground, or the eyes are closed in pain. This is referred to by Kress and van Leeuwen, following the terms used by Halliday, as an ‘offer’, rather than a ‘demand’, which involves visual contact and establishes an imaginary personal relationship with the represented participants (Kress & van Leeuwen, 2006, p. 119). As a ‘demand’, the viewer may be encouraged to sympathize with these represented participants, for example, but the indirect gaze of the ‘offer’ as presented here allows the viewer to scrutinize them, and thereby identify with them. The size of the frame of these images also suggests objectification: two of these images are presented as ‘long shots’, the persons experiencing knee pain and back pain, meaning their whole body is represented in the image.73 These impersonal distances, along with the indirect gazes of the represented participants also reinforces the role of these images as a type of taxonomy; as objective, dispassionate knowledge about the conditions that the clinic can treat.

The other image in the text is clearly not part of this taxonomy of treatable conditions– the image of the treatment beds is a photo, not a cartoon, and is presented in a different shape and size, and appears in a different part of the text. The photographic image has higher modality or ‘truth value’ than the cartoon images, which suggests to the viewer that it can be taken literally, in other words this is exactly where one would be treated at this clinic. The angle of this shot and the distance from which it is taken also suggests the viewpoint of a customer

71 Most traffic lights in Japan appear horizontally, as the circles are represented here, rather than vertically. However, it is noted that the colour representing ‘go’ in Japan is referred to (and perceived as) blue rather than green, which may weaken the likelihood of this interpretation. 72 Cartoon-like pictures are used more widely in Japan than in other western countries and are not necessarily childish. 73 The remaining two images are ‘medium close shots’, the represented participants being cut off at the waist. This in the middle of the continuum between a long shot (impersonal distance) and close up (intimate distance). 147

who has entered the clinic, it is not, for example, taken from the back of the clinic looking out towards the street, nor is it looking down on the treatment beds as a therapist may do. While this images is different to the ones described above in many ways, the tones used in the photo reflect the overall palette of warm colours – the beds and walls are yellow, the floor is a warm brown, and red items can be seem lined up on the shelves above the beds.

Kress and van Leeuwen also refer to ‘information zones’ when understanding the semiotic messages suggested by a text. In this text, the examples of the different types of ailments appear at the top, this represents the ‘ideal’ in the sense of ‘exactly the types of conditions we can treat’. The name of the clinic appears in the bottom third of the text – this is the ‘real’ place you can go for treatment. The image of the treatment beds is in the middle, and therefore represents a mediator between the types of pain (problem) and clinic (solution). The placing of the ‘given’ information on the left and the ‘new’ information on the right, Kress and van Leeuwen note, is related to the western left-to-right reading tradition, however Japanese can also be written in the way, as it is in this text. Therefore, it is significant that the treatment beds are on the left, and therefore represent the given, or established, as this is what a potential patient would expect to see in such a clinic. The invitation to consult and the phone number appear on the right, in the zone for the ‘new’ information.

Next, the typography of the text will be examined in detail through the framework described by Stöckl (2005). In the microtypography domain, three different colours and two different Japanese typefaces, as well as a set of numeric characters, have been used for the three parts of the written text. Both the green and orange character lines are in ‘san serif’ Japanese styles, broadly described as ‘ゴシック/goshikku’ (gothic) rather than the more traditional ‘明朝 /minchō’ (mincho) style, and, like their san-serif alphabet equivalents, these typefaces have a casual and informal feeling. The use of the exclamation mark after ‘Please consult us about your body problems!’ also gives the text a casual, informal feeling. The font used for the business name reinforces these connotations even further, as if it has been written by hand with a marker; the strokes do not taper off but have round ends, and some characters are less than perfectly formed. For example, in the word ‘はぴねす/hapinesu’ (happiness), the final stroke of ‘は/ha’ does not continue as far as it usually does, and the circle on the ‘ぴ/pi’ is quite large in proportion to the character, while in the kanji characters, some of the final ‘はね/hane’ (flicks) are missing, such as on the ‘月’ within ‘骨/kotsu’; and the final stroke of ‘元’

148

within ‘院/in’ is more round rather than having two distinct corners. These markers of informality and casualness are related to the interpersonal metafunction, commenting on the (real or imagined) relationship between the staff and a potential customer. Besides these markers of intimacy, as noted above, the hiragana script is also noted for its roundness and gentleness, and there are no angular katakana characters in this text. Interestingly, the phone number is also made up of numbers with round shapes: 0, 3 and 6.

At the mesotypography level, the mixing of colours and typefaces can also be seen as a marker of informality; consistency in colour and font would give the text a more formal feel. The lettering appears horizontally and from left to right, which may suggest modernity, and/or westernization compared to vertical lettering. The wording is right-aligned, this places it in the information zone for the ‘new’, so while the image of the treatment beds is the ‘given’, what is already established in the viewer’s mind, the invitation to ask about the treatments for various conditions and the phone number of this particular business is presented as new or noteworthy information as mentioned above.

The hiragana loanword ‘はぴねす/hapinesu’ (happiness)74 appears in the line stating the name of the business, and occurs before five kanji characters ‘鍼灸整骨院/shinkyūseikotsuin’ (acupuncture, moxibustion and osteopathic clinic). The simplicity of the hiragana therefore complements the complexity and relative heaviness of the kanji, and may be an additional reason it was chosen over the unmarked katakana alternative. This relates to the textual metafunction, and how elements of a text form cohesive wholes. The simplicity of the hiragana script used in the business name therefore complements the complex kanji characters which follow. The leaf motif described above underlines the business name, reinforcing the importance of this line.

While the design elements related to macrotypography have been described within the visual grammar analysis above, the paratypography of the text can be considered within its situated context, this reinforces one of the central tenets of social semiotics, which is to consider texts within their broader social contexts. This particular text is printed on a large fabric banner which hangs permanently over the footpath outside the clinic, which is located within a group

74 It is noted that ‘はっぴねす/happinesu’ is an alternative spelling variant, one that is less closer to a native- like pronunciation although more faithful to the English spelling. The pronunciation of ‘はぴ/hapi’ and ‘ハッピー/happī’ are discussed in relation to the phrase ‘happy birthday’ in section 6.2.8. 149

of shops in a busy shopping street near a railway station in Abeno, Osaka. The sign is large and salient for persons walking along this street, and the invitation to consult is therefore relevant both to the phone number appearing on the sign, as well as the potential to enter the clinic and enquire in person.

This case study therefore provides another lens through which to interpret why loanwords appear in hiragana. Hiragana has associations of gentleness and roundness, and this semiotic resource therefore contributes to the soft, relaxed, friendly feeling of the business. These associations are supported by other cues in the above analysis which noted warm colours, soft shapes, and casual lettering styles. The images used in the text invite viewers to identify as potential patients of the clinic, and the warm, familiar atmosphere created by the use of hiragana for the business name, as well as the other semiotic resources in the text, encourages the viewer to take up the invitation to consult with the staff via phone or by visiting the clinic, where this text is displayed. Hiragana therefore plays a role in the message this text suggests as a (internally) coherent text, and reinforces the cues given by the other semiotic resources. The interpersonal metafunction, creating or maintaining a warm and friendly relationship between the business and potential customers, is the metafunction that is foregrounded across many of these examples.

This text was one of those used in the survey (layer 3), and, as will be described in section 5.3.2, 73% of respondents felt it was appropriate to write the word ‘はぴねす/hapinesu’ (happiness) in hiragana. Some of the most common perceived motivations for the use of hiragana that were specific to this text were that it made the business seem friendly or approachable, as well as having a warm or gentle feeling. While these are some of the reported connotations of hiragana script in general, evidence for these interpretations can also be found in the other semiotic resources used in the text as described in the current case study.

4.2.2 Case 2: ‘Honey Lemon’ Drink

The second case study analyses the label of the drink ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon). The findings will be used to triangulate those particularly of the survey (layer 3) which also sought interpretations of this particular text. It also appeared in the

150

interviews (layer 4) when the survey findings were discussed. The text is reproduced with an English translation in figure 4.27 below.

Lemon Honey Yamada Apiary’

Made Made with domestic

lemons

Figure 4.27: ‘Honey Lemon’ drink label with English translation

This text is largely made up of written elements in a variety of typefaces, colours and sizes. The most salient part of the text is the loanword ‘れもん/remon’ (lemon), which is written in the marked hiragana script. The relative salience of this word is due to the large size of the font, being around twice as large as the next largest font (that used for ‘蜂蜜/hachimitsu’ [honey]). Salience within an image represents the ideational function, telling the viewer that this information is important or noteworthy, for example.

The picture of the lemon is also highly salient, being the only picture, and differentiated by being yellow in the overall black/white/red colour scheme. This image has low modality; it is not a photo but a painting of a lemon, and is decontextualized (shown without a background or even a shadow). By reducing the ‘truth-value’ of the lemon, this may represent a reduction of the ‘lemon-ness’ in terms of taste, in other words, suggesting that the product does not taste too sour. This ‘unsour’ impression also helps to mitigate the claim made linguistically

151

that it is ‘made with domestic lemons’ (国産レモン使用/kokusan remon shiyō). The picture of the lemon is lit – unusually – from below; this has the effect of making the product seem ‘high up’ and therefore suggesting good quality.

The word ‘蜂蜜/hachimitsu’ (honey) is framed by vertical lines, marking this section of the text as distinct to what appears around it. This word is also separated from the preceding text ‘山田養蜂場の /yamada yōhōba no’ (Yamada apiary’s…) by the use of a different typeface and colour; this serves to remind the viewer that this phase is not part of the product name. The frame lines around the word ‘honey’ also distinguish it from the next word, ‘lemon’. While the colour and possibly the typeface are the same, these words are different both in terms of the kind of word (Sino-Japanese vs. loanword) as well as script (kanji vs. hiragana).

Another use of framing can be seen around the linguistic element ‘国産レモン使用/ kokusan remon shiyō’ (made with domestic lemons), the red background and white lettering mark this as separate from the rest of the text. This phrase gives a different kind of information to the preceding elements, stating details about the ingredients and their source. This element is also different because it contains the katakana script, visible nowhere else in the text on the front label, and furthermore because the katakana word it contains (レモン/remon) is that which is written in hiragana in the name of the product (れもん/remon).

As mentioned in the description of the case study methodology (section 3.4.2), the information zones described by Kress and van Leeuwen are based in a horizontal, left-to-right writing tradition, and while Japanese can be written in this way, this text uses a vertical right- to-left structure. The interpretation of the information zones should therefore be made with caution. If the right side of this text rather than the left represents the ‘given’, this corresponds with the words ‘山田養蜂場の /yamada youhōba no’ (Yamada apiary’s…), presenting it as an ‘established’ business, one that the reader should be familiar with, whether or not this is the case in reality. The word ‘れもん/remon (lemon)’ is further to the left, and therefore in the zone for the ‘new’, in other words, a lemon-flavoured drink is a new or unexpected product for an apiary. The word ‘蜂蜜/hachimitsu’ (honey) appears between these two elements, providing the link between ‘Yamada Apiary’ and the ‘Honey Lemon (drink)’. The phrase ‘国産レモン使用/ kokusan remon shiyō’ (made with domestic lemons),

152

being in the lower left hand corner of the text, still appears in the bottom information zone when the text is read from right to left. It therefore represents an element of the ‘real’, in other words, what the drink is really made from: Japanese lemons.

In the domain of microtypography, the typeface for the phrase ‘山田養蜂場の/yamada yōhōba no’ (Yamada apiary’s…) is presented in a Mincho-style typeface, which, like its serif alphabet counterpart, connotes tradition and authority. This is reinforced by the vertical lettering, which represents tradition and formality in Japan. This element is presented in one of the smaller fonts in the text, lessening its salience and therefore its importance to the reader.

The lettering for the product name ‘蜂蜜れもん/hachimitsu remon (Honey Lemon), as mentioned above, appears much larger, with the characters for ‘lemon’ taking up around a quarter of the whole text. These characters appear to be written with a calligraphy brush, while the black colour suggest ink, further reinforcing the impression of calligraphy. This image is carried through the other domains of text design, for example at the mesotypography level, the vertical lettering and artistically non-centred placement of the characters for ‘lemon’ also call to mind a piece of traditional Japanese calligraphy. In terms of macrotypography, the image of a creative work is reinforced by the image of the lemon, which appears with rough brush strokes as it if were hand painted. At the paratypography level, which concerns the tools and instruments for writing, besides the ‘brushes’ already mentioned, the flecks in the white background of the text suggest handmade paper. These markers of Japanese-ness are evidence of the ideational function, suggesting to the viewer that this is a domestic, Japanese product.75

The mixing of scripts falls under the mesotypography domain, overall the text has rather a large number of kanji, representing 11 of the 18 characters in the text. Kanji has associations of education, intelligence, and literature, which may serve to give this product an elite image, and is also associated with ‘Japanese-ness’, although less so than hiragana. Hiragana, while representing only four characters in the text,76 is shown in the largest font in the text, giving it

75 It is also argued that the interpersonal function can be seen in these cues, as the intended audience are Japanese themselves. This is discussed in section 7.4. 76 Three for ‘lemon’ (れもん) and one for the possessive particle ‘の/no’ (‘s) which is represented in much smaller font. 153

salience (as explained above) despite the small number of characters. Besides ‘Japanese-ness’, hiragana is also associated with simplicity, and therefore complements the relatively complex characters for ‘蜂蜜/hachimitsu’ (honey)77 which precede them. Since ‘lemon’ also has a kanji, the phrase ‘honey lemon’ has a particularly large number of orthographic options, with three alternatives for each of the two words: ‘はちみつ’, ‘ハチミツ’, or ‘蜂蜜’ for ‘hachimitsu/honey’; and ‘れもん’, ‘レモン’, and’檸檬’ for ‘remon/lemon’. If kanji is used for ‘honey’ as it appears in the text, using kanji for the word lemon would appear quite heavy and complex ( 蜂蜜檸檬), while using katakana for lemon ( 蜂蜜レモン), with its associations with foreign words, may have seemed out of place it a text that expresses ‘Japanese-ness’ through so many other semiotic resources, such as the brush-like font and vertical lettering, and allusions to Japanese calligraphy. For this reason, hiragana may have been chosen as the semiotic option most suitable for expressing the intended message of the text, its simplicity balancing with the complex kanji preceding it, with both kanji and hiragana being associated (though their different histories) with traditional Japanese culture.78

The phrase ‘made with domestic lemons’, also shares the brush-like typeface, however, rather than calligraphy, the white characters on the red oval shaped background suggest a personal seal. Seals are often found on a piece of calligraphy, usually in the bottom left corner, where it appears in this text, but are also in common use in place of a signature in most official documents in Japan today. As explained above, in this phrase the unmarked katakana script has been used for the word ‘レモン/remon’ (lemon), possibly to draw attention to the word’s referring to the fruit from which the drink is made, rather than the product name. This is also evidence of the modality or ‘truth value’ the choice of script can give to a word, an aspect of the interpersonal function. The metaphor of the artist’s personal seal on a piece of calligraphy can therefore perhaps be extended to this textual element; just as the artist is the source of the calligraphy, the ‘domestic lemons’ are the source of this product.

77 For this reason, the word often appears in the more simple hiragana and katakana scripts, for example, the products ‘加藤美蜂園本舗サクラ印ハチミツ’ and ‘トップバリュ純粋はちみつ’. 78 Of course, the use of these scripts in the opposite order, ‘はちみつ檸檬/hachimitsu remon’, also fulfils this criteria, but relative numbers of characters would difficult to fit in the current design of the text: the four character ‘honey’ would have to appear in a space half that of the two-character ‘lemon’. The fact that honey is the main product produced by an apiary may be an additional motivation for wanting this word in the more formal kanji script. 154

To conclude, the use of hiragana for the loanword ‘lemon’ is one of many semiotic resources in this text that suggests a high-class, Japanese product. The choice of hiragana may also be used to give a softer impression to the word, and perhaps minimize the sour impression of lemons. The suggestion of its being a ‘top shelf’ product is suggested by the low-angle lighting of the picture of the lemon, the large number of kanji characters, as well as the many allusions to calligraphy, a traditional Japanese art form symbolic of high culture. These include the brush-like font, and handmade paper background. Other semiotic resources also associated with (traditional) Japanese writing in general include the vertical orientation of the lettering, and text resembling a personal seal. Japanese-ness is also inadvertently suggested by the background the product was photographed on, which shows a traditional motif of snow rabbits. In conclusion, the use of hiragana for the word ‘lemon’ gives the text an internal cohesion, reinforcing the messages produced by the other parts of the text, as well as being externally coherent, with other non-related texts it alludes to such as calligraphy.

This text was also used in the survey (layer 3), the results of which will be described in section 5.3.3. This text met with the highest ratings of appropriateness (90%), some of the reasons given for this decision included that hiragana suggested a mellow or delicious taste, an interpretation justified by the low modality of the image of the lemon, but also related to the idea as hiragana loanwords as a marker of ‘un-realness’ (section 7.3.2). Respondents also felt the use of hiragana emphasised that the product was made from domestic lemons, and that gave a feeling of ‘Japanese-ness’, many other markers of which have been illustrated in the above case study.

4.2.3 Case 3: ‘I can fly!!’ sticker

This text was one of the more unusual in the corpus, the loanword in hiragana being part of the actual product (rather than a product name, for example), as well as being a verb, and embedded in an English sentence. The sticker contains the words ‘I can ふらい/furai (fly)!!’, which are attributable to one of the characters in the image as explained below. This text was also used in the survey (layer 3), and the survey results discussed in the interviews (layer 4), allowing findings to be triangulated across these layers of analysis. The image of the sticker is reproduced here in its packaging (figure 4.28), although the text (sticker) itself forms the main unit of analysis.

155

B-SIDE LABEL

WATER AND UV RAYS

PROTECT STICKERS

This is a waterproof and UV light resistant tough sticker.

I can fly!!

oof… B-Side Label

Figure 4.28: ‘I can fly!!’ sticker with English translation

The text shows two anthropomorphic animals rendered in a cartoon-like fashion, in a reduced palette of only three colours besides black and white. The image has heavy black outlines and highly saturated, unmodulated colours. These kinds of colours are particularly appealing to children (Kress & van Leeuwen, 2006, p. 165), although as it was noted in case study 1, cartoons are not associated exclusively with this demographic in Japan.

While this reduces the represented participants’ modality from a ‘naturalistic’ viewpoint (Bernstein’s (1981) ‘Naturalistic coding orientation’, cited in Kress and van Leeuwen 2006), they are still recognizable as different species of animal when taken as part of the context of their situation. The markings on the face of the lower animal suggest a ‘モモンガ/momonga’ or flying squirrel, while the upper character, may be a hamster due to its round ears, short tail and relative size.79 The hamster appears to be pretending or attempting to fly by jumping on the back of the flying squirrel, exclaiming ‘I can ふらい/furai (fly)!!’, while the squirrel is seen flattened and groaning underneath, muttering the onomatopoeic ‘ ぐふ/gufu…’, approximated as ‘oof…’ in English.

79 On the company, ‘B-Side label’s website, this product is called ‘ミキモモンガ (I can ふらい)’ [mikimomonga (I can furai)]. ‘Miki’ is the first name of the artist, and the flying squirrel (momonga) is named rather than the animal depicted above it. 156

The hiragana loanword in the text, ‘ふらい/furai’ (fly), in the speech of a represented participant, therefore provides the ‘punchline’ for the image, because the hamster is not really flying, it has actually pinned the flying squirrel to the ground. While there is no ‘contextualization’ or setting for the image, such as a line to indicate their position on the ground, or clouds in the case that the characters really are flying, this interpretation is supported by the expression of the flying squirrel with its extended tongue, its compressed shape, and its muttered ‘oof…’ expression. The hiragana loanword therefore serves to give a joking, unserious quality to the claim that the hamster can fly, which is evidence of the interpersonal metafunction. This is further discussed in the context of the survey responses (5.3.4) and in the discussion (section 7.3.2).

Besides the cartoon-like style of the image, the main character has a certain ‘cuteness’, an impression given through the round, chubby shapes of its head and body, its rosy cheeks, compact facial features within the area of the head, and large eyes. In addition, the perspective from which the image is depicted encodes relations between the represented participants in the viewer, as aspect of the ‘interpersonal function’ according to Kress and van Leeuwen (2006). In this case, the characters are depicted from an oblique angle, rather than front on, and the neither character appears to be making eye contact with the viewer. Both these factors suggest a lack of involvement with the situation unfolding in the image: the scene is detached, it is there to be looked at. The slightly high angle of the image suggests that the viewer has ‘power’ over the represented participants; in this case, we are invited to look down on the antics of these two characters and be amused by their situation without the need to be involved or perhaps help them.

In terms of layout, this image utilizes the information zones of the ‘ideal’, and the ‘real’ to good effect. The text ‘I can fly!!’ appears in the upmost portion of the image, representing the dream or hope of the hamster in the zone for the ‘ideal’. The space for the ‘real’, however, shows the squirrel being squashed by the weight of the hamster that has jumped on top, the ‘reality’ resulting from the hamster’s attempt to fly. The hamster itself appears in the middle of the text, mediating the ideal and the real: with its outstretched limbs and upturned face, it embodies its claim in the words above, ‘I can fly!!’, but being in direct contact with the flattened squirrel below, the hamster is very much part of the reality that it cannot, in fact, fly.

157

Of the linguistic elements in the text, the words ‘I can ふらい/furai (fly)!!’ appear in the largest font and are the most salient, and therefore important, to the viewer. The muttered ‘oof…’ of the squirrel appears in much smaller characters, and is only noticeable on close inspection. Less salient still is the name of the company who produces the stickers, ‘B-Side Label’, which appears in light brown characters on the hamster’s white underbelly. The relative size of each of the lines spoken by the characters as well as the punctuation suggests the volume with which they are spoken: ‘I can fly!!’ with its double exclamation mark is ‘cried’ or ‘shouted’, while ‘oof…’ and its ellipsis dots is ‘muttered’ or ‘groaned’. The way in which voice quality can be encoded in this way is also evidence of the interpersonal function of typography.

At the microtypography level, all parts of phrase ‘I can fly!!’, the two English (rōmaji) words, the loanword, and the two exclamation marks appear to be part of the same lettering style, although the characters appear in slightly different sizes. In the domain of paratypography, they appear to be written with the same pen that has been used to outline the cartoon image below, rather than an established typeface, having a handwritten effect, with their different sizes, and the way they curve around the image below. The handwritten feel of the words, as well as the repeated exclamation marks, give these words a casual, excited and perhaps youthful feel.

In terms of mesotypography, the most striking part of this phrase is the insertion of a hiragana loanword within an English sentence. While the mixing of scripts within a Japanese sentence is relatively common, this example shows a mixing of languages: the English ‘I can …’, and the Japanese loanword ‘ふらい/furai’. Furthermore, this particular loanword is represented in marked hiragana script, rather than the katakana conventionally used for loanwords.

In the domain of mesotypography, the directions in which the two utterances are written in the text are worth noting. The English beginning of the hamster’s sentence lends itself to a horizontal layout, and if we consider the exclamation marks to be part of the ‘English’ sentence, we can see the loanword ‘fly’ as a case of code-switching where the matrix language is English. The direction of this sentence also gives a sense of dynamism between the words and the hamster who utters them; the sentence leads the viewer’s gaze from left to right, and the leap of the hamster brings us from the right to the left again. The squirrel’s

158

muttered ‘oof…’ is written vertically, while this can be a marker of tradition (as it is used in the ‘Honey lemon’ example in case study 1), in this case it may reflect the downward motion of the squirrel, from a launching or flying position, to being forced to the ground. The different interpretations of vertical writing, just like the different interpretations of the use of hiragana script, are therefore closely tied to the context, and their meaning only recoverable from other cues within the text.

To conclude this section, hiragana has been used in this text as one of many semiotic resources suggesting the cute or naïve image of the character who utters the phrase, and emphasises the punchline of the text, namely, that the hamster’s plan to fly has failed.80 This text was also used in the survey (layer 3), discussed in section 5.3.4, and the reasons given for the designer’s perceived motivation for using hiragana reinforced these findings, including that it appeared cute, it suited the character, and that the product was made for children.

However, of the three examples of loanwords in hiragana presented in the survey (layer 3), this text was the one felt to be inappropriate by the largest number of people (45%), with the most common reason being that it was hard to read. While hiragana has strong associations of being easy to read, the difficulty, in this case, seems to be its unusual appearance within an English sentence, as well as the confusion with the homophonic ‘フライ/furai’ (fry/fried food), as described later in section 5.3.4. Therefore, despite the meaning of this word being tied to its context in a number of ways as described in the case study analysis above, this image resulted in a breakdown of internal cohesion, in SFL terms, for many readers of this text.

4.2.4 Case 4: ‘G’mornin’ ’ tweet

This example represents a text from social media, and is taken from the twitter feed of the Japanese model Chisato Yoshiki, nicknamed ‘Chīpopo’. The opening line of the tweet contains an English phrase in hiragana, the greeting ‘G’mornin’’ (ぐっもーにん/gummōnin). This example was chosen because besides the marked use of the hiragana script, this phrase is

80 A Japanese research assistant also suggested the comical associations this phrase had in connection to the Japanese actor, Kubozuka Yōsuke. In the 2002 film, ‘ピンポン/pin pon’ (Ping-pong), Kubozuka’s character utters the English phrase ‘I can fly!’ in the Japanese-language film before jumping off a bridge into a river; in 2004 Kubozuka himself survived a fall from his 9th floor apartment building. This event led many of his fans to remember this scene from his film, and joke about the actor’s (in)ability to fly. 159

also unusual because of the particular transliteration of this well-known phrase, a more standard version being ‘グッドモーニング/guddo mōningu’. A screen shot of the tweet and an English translation are provided in figure 4.29 below.

G’mornin’ (sunny face)

Today is the Japan Girls Expo in Osaka (lips)

AbemaFRESH! will be broadcasting from there (confused; thinking)

Don’t miss from 14.15 (peace)

Figure 4.29: 'G'mornin' ' tweet with English translation

At the microtypography level, the long vowel in the phrase ‘ぐっもーにん/gummōnin (G’mornin’) is lengthened with a bar. The findings from the corpus show that this method is more common that including an extra vowel, being used in 75% of cases. The use of this method also has the effect of reminding the reader that this is a loanword, as Japanese words in hiragana do not use this particular character.

Many textual considerations such as the font, colour, and layout are built into the twitter platform; the font is a ‘san-serif’ style gothic font, which appears in black on a white background, and gives a more casual impression than a ‘serif’ Mincho-style font. Any links or pictures posted appear below the linguistic text of the tweet.

It may be because of these design limitations (as well as the limitations on the number of characters), that emoji are often used in tweets and other social media posts to fulfil a range

160

of functions, such as to lend a sense of voice or emotive quality (Danesi, 2016). The other graphological resource Japanese twitter users have is, of course, the choice of script.

The mixing of scripts (the domain of mesotypography) and the ornamentation of her message with emoji (macrotypography) have both been used in Yoshiki’s tweet. The first line contains the loaned greeting, ‘ぐっもーにん/gummōnin’ (G’mornin’) in hiragana, and is followed by a sun emoji giving a warm, sunny voice quality to the message. In the second line, Yoshiki announces that the ‘Japan Girls Expo’ will be held in Osaka that day, ending the line with a pair of red lips, perhaps echoing the theme of ‘girl culture’ of the event. This line is marked with a slightly formal register of speech though the final word ‘ です/desu’ and the preposition ‘ にて/nite’ for ‘in (Osaka)’, giving this statement a nuance of an official announcement. In the third line, Yoshiki explains that an online TV channel (AbemaFRESH!, written in rōmaji/English) will be broadcasting the event, and follows this by two emoji suggesting thoughtfulness: a thinking face, and a thought-bubble. This can perhaps be interpreted as representing the thoughts of the reader: if Yoshiki is attending the event, and the event will be broadcast, it may be possible to see Yoshiki in the broadcast. As if in answer to the readers thoughts, Yoshiki ends the tweet with the cryptic ‘Don’t miss from 14.15’, suggesting that she may appear at an event as part of the expo at 2.15pm. Japan’s ubiquitous ‘peace sign’ gesture appears as an emoji at the end of this line, suggesting both a cheeky, playful tone to the message (interpersonal function) as well as perhaps acting as a final salutation (textual metafunction).

This tweet is typical of Yoshiki’s social media presence, her tweets contain many emoji, and play with script through the inclusion of English words in rōmaji, for example ‘BirthdayParty [sic]’. Her tweets also contain many examples of loanwords in hiragana, such as ‘はっぴーはろういん/happī harōin’ (Happy Halloween); ‘らすとてぃーん/rasuto tīn’ (last teen[age year]); and even ‘ついーと/tuīto’ (tweet). Loanwords in hiragana may therefore be part of Yoshiki’s online identity in the twittersphere, and this marked orthography seems to be reproduced by some of her fans and followers, as explained in the discussion (section 7.5).

This particular English phrase in hiragana, ‘ ぐっもーにん /gummōnin (G’mornin’), is especially interesting because of its non-standard transliteration as described above. In terms of script, the choice of hiragana makes the greeting ‘good morning’ stand out. However, this

161

phrase also stands out because the choice of ‘ぐっもーにん/gummōnin’ over ‘ぐっどもーにんぐ/guddo mōningu’ represents a choice of a more accurate rendition of the English phrase. In Yoshiki’s version, the two words are linked phonetically, and the final syllable ‘gu’ has been dropped from ‘morning’, resulting in a three syllable phrase (like the English ‘G’morning’) rather than the standard five syllables (in the ‘katakana-English’ version). Another deviation from standard transliteration practice is the use of a half-size ‘っ/tsu’ rather than an ‘ん/n’ for the /m/ germinate, perhaps to emphasize this sound. The linking of ‘good’ and ‘morning’ allows Yoshiki to express her knowledge of a more native-like pronunciation in a graphological form. It is therefore interesting that this phrase is marked in two ways: by being written in hiragana rather than katakana; and by being written in a ‘non- standard’ transliteration. Other examples of loanwords in hiragana that represent a similar shift to a more ‘native standard’ pronunciation along with its non-standard script were found in the corpus, these are further discussed in section 7.3.2.

Examples of loanwords in hiragana which also represent a shift to more native-like pronunciation are important to consider in relation to the phrase ‘katakana English’, a denigrating term for the English as it is spoken by many Japanese, perceived to be difficult or impossible for speakers of other varieties of English to understand. Phrases like those described above therefore distance their writers from ‘katakana English’ in two ways: by not being written in katakana, and by being closer to a native-like pronunciation. These kinds of loanwords in hiragana may represent a growing awareness of the lack of intelligibility of standard katakana transliterations and a move toward an alternative. This idea was also mentioned by some participants in the interviews (layer 4).

This shift towards a more accurate pronunciation is further highlighted by the choice of the marked hiragana script, interestingly, a script that has associations of Japanese-ness as reported in the survey findings (section 5.2.1). This suggests that not only is the most ‘Japanese’ script capable of expressing foreign loanwords, but it can do so with more accuracy, representing the possibility of a fluent English-speaking Japanese identity. While no firm conclusions can be drawn without establishing whether similar non-standard transliterations also occur in loanwords in katakana, for example, ‘グッモーニン/gummōnin’ which were outside the scope of this study, the link between hiragana script, its associations with ‘sounds’ as reported in the survey (layer 3), and native like pronunciation was also

162

brought up by interview and focus group participants (see section 6.2.8). These findings from across the different layers of the project are brought together in the discussion section (7.3.2).

4.2.5 Summary of layer 2: loanwords in hiragana in multimodal texts

The case studies have provided an in-depth look at how loanwords in hiragana are working as one semiotic resource among many in any given text, and complement the statistical analysis conducted in layer 1. The case studies have been useful in showing how interpretations which frequently occurred in the survey (layer 3) and interviews (layer 4) are supported by other semiotic resources in the text, for example why hiragana is reported to seem ‘Japanese-like’ in the Honey Lemon drink label, but ‘cute’ when it appears as the speech of a cartoon hamster.

The potential meanings of a multimodal text are produced though a particular combination of semiotic resources which compose it, as well as the context and greater socio-cultural environment in which it appears. Hiragana has a range of established connotations and associations, as described in the literature (section 2.4.2), as well as in the current research project (section 5.2.1). However, which of these connotations are foregrounded or perceived to be the motivation for the use of hiragana for a loanword in a particular text depends on the other semiotic resources deployed, and the perceived message and /or function of the text. In terms of an SFL model of language (see figure 2.1), meaning therefore flows not only ‘outwards’ from script choice (graphemes) to the other features of the [con]text; but also ‘inwards’, as the context informs which particular aspect of the script (‘cute’, ’traditional’, or ‘familiar’ and so on) is being drawn upon.

However, besides hiragana as a semiotic resource, and the context in which it is deployed, the reader or viewer of the text also plays an important role in the process of meaning making. For this reason, the data provided by the statistical analysis (layer 1) and case studies (layer 2) only provide part of the picture of how meanings are created through the use of hiragana for loanwords. The following layers of analysis, the survey (layer 3) and focus groups and interviews (layer 4), sought the opinions of native speakers of Japanese living in Japan, in order to more thoroughly understand how this marked use of script is interpreted.

163

Chapter 5: Findings from the survey

This chapter will describe the findings from the third layer of analysis. While the descriptive statistical analysis (layer 1) and case studies (layer 2) took the corpus of hiragana loanwords as their starting point, the survey (layer 3) described in this chapter investigates how these texts are interpreted by Japanese people.

The survey takes a quantitative approach, gathering opinions from a large number of people to a small number of examples of loanwords in hiragana. These findings were most relevant to answering the third research question, which investigates the functions of loanwords in hiragana.

The survey was divided into three parts, the first collected demographic information, the second focused on the associations of each Japanese scripts, and the final section was designed to elicit participants’ responses to images of texts containing loanwords in a variety of marked scripts. The survey can be found in Japanese with an English translation in appendix 2. The number of people who responded varied from 232 to 112 depending on the question, with an average of 150 people responding to each of the central questions on script association and the sample texts. Because of this variation between different items of the survey, the number of respondents for each question is reported in the following sections, which present the findings of the online survey (layer 3).

5.1 Respondent background

Data was collected on the respondents’ backgrounds through multiple choice questions on gender (Q3), age bracket (Q4), and education level (Q5). Around two-thirds (63%) were female, and a third (38%) males. In terms of age, those aged 21-30 (44%) and 31-40 (43%) made up the majority of respondents, with the next highest being 41-50 at 10%. Participants aged younger than 20 and older than 51 each represented only 2.2% of the total number. Because participants in these age groups, as well as males of all ages, were under-represented in the survey, an effort was made to include these groups in the interviews and focus groups, as described in section 3.4.4. The majority of the respondents (65%) had bachelor degrees as their highest level of education. The next largest categories were postgraduate degrees (22%),

164

and high school diplomas (13%). While it is outside the scope of this study to analyse the survey responses in terms of these demographic factors, they constitute one of the suggestions for further study described in the conclusion, section 8.2.

5.2 Script associations

As described in the research design (section 3.4.3), respondents were asked what impressions they had of each of the four scripts in use in Japan: hiragana (Q9), katakana (Q10), kanji (Q11), and rōmaji (Q12).. Table 5.1 summarizes the total number of associations generated, and the average number of associations per respondent.

Table 5.1: Response rates for script associations sections

Total number of Average number of associations Script No. of respondents associations per respondent

Hiragana 112 362 3.2

Katakana 174 480 2.8

Kanji 165 499 3.0

Rōmaji 156 332 2.1

The findings for each script are described in the following sections, in which the ten most frequently occurring codes will be presented and discussed.

5.2.1 Hiragana

The question on hiragana (Q9) was responded to by 112 people, and generated 362 associations, which were coded under 26 codes. The ten most frequently occurring codes are shown in table 5.2, with a representative example response, and the number of tokens for each code.

165

Table 5.2: Common associations of hiragana

Code Example response No. of tokens

1. Soft やわらかい/yawarakai (soft) 64

2. Easy 易しい/yasashī (easy) 49

3. Children 子供っぽい/kodomoppoi (childlike) 45

4. Gentle 優しい/yasashī (gentle) 30

5. ‘Yasashī’ やさしい/yasashī (gentle or easy) 27

6. Cute かわいい/kawaii (cute) 20

7. Round 丸い/marui (round) 17

8. Japan 日本語/nihongo (Japanese-language) 17

9. Feminine 女性/josei (woman) 12

10. Foundation, normal 基礎/kiso (foundation) 8

The following paragraphs will explain key findings from the responses; numbers in brackets for example ‘(7)’, indicate the number of tokens for that code.

The most frequently occurring code for hiragana was ‘soft’ (64), related codes occurred in smaller numbers such as ‘round’ (17), and ‘airy’ (5). Also ranking highly were the codes ‘easy’ (49), ‘gentle’ (30) and the ambiguous ‘やさしい/yasashī’ (27). The code ‘easy’ included responses such as ‘簡単/kantan’ (simple), as well as ‘易しい/yasashī’, while the code ‘gentle’ was made up almost exclusively with ‘優しい/yasashī’. However, the high number of ambiguous hiragana responses suggests that easy/gentle are best understood as encompassing elements of both ease and gentleness; this is necessarily the case when the word is used in the spoken data.

The third highest ranking code was ‘children’ (45), this included responses such as ‘子供 /kodomo’ (child), as well as related items such as ‘絵本/ehon’ (picture book) and ‘小学校 /shōgakkō’ (primary school). Other codes relating to this idea included ‘foundation/normal’ (8), as hiragana is the first script taught at school, and can be thought of as the ‘foundation’ of

166

literacy; and ‘beginner’ (5). Together these can be understood as emphasising the role of hiragana in the Japanese writing system, as the first writing system learned by children, and other ‘beginners’ of the language.

Another high ranking code was ‘cute’ (20), made up exclusively of the word ‘可愛い/かわいい/kawaii’. While ‘cuteness’ could also be related to the code of ‘child’ above, this term has a wider range of use and higher frequency in Japanese, being used to refer to people and items associated with teenagers and also adults. ‘Cute’ can also be seen to be related to the code ‘feminine’ (12), but as above, the usage of the word is not exclusively limited to females in Japan. As the term ‘かわいい/kawaii’ or ‘cute’ holds such a significant place in Japanese culture81 (Brown, 2011; Borggreen, 2011, Nittono, 2016), ‘cute’ was kept as its own code.

Although the research design did not allow demographic factors to be taken into consideration in interpreting the survey responses, it is worth nothing that the high number of responses for ‘cute’ could also be a reflection of the survey population, being composed mostly of women in their 20s and 30s (see section 5.1). ‘Cute’ is a word is often used by them, about them, and about products and clothing aimed at this particular demographic.

The code ‘Japan’ also ranked highly (17), and was made up of words such as ‘日本/nihon (Japan), ‘日本語/nihongo’ (Japanese language), ‘日本的/nipponteki’ (Japan-like), and ‘日本特有/nippon tokuyū’ (characteristic of Japan). There was also one mention of ‘大和言葉 /yamato kotoba’. 82 This underscores hiragana’s reputation as the ‘most’ Japanese script, because of the other Japanese scripts, kanji is often associated with its Chinese origins, and katakana is strongly associated with foreign words (see the following sections). The code ‘Japan’ can be interpreted in relation with the code ‘classic/old fashioned’ (5), which included terms such as ‘古典/koten’ (classical literature), and ‘昔ながら/mukashinagara’ (as it was in the olden days,) to represent traditional Japanese culture. Hiragana as a marker of Japanese-ness in loanwords in hiragana is discussed in the section 7.3.1.

81 In 2009, three ‘Cute Ambassadors’ were even appointed by the Japanese Ministry of Foreign Affairs. Interestingly, the Japanese term ‘カワイイ大使/kawaii taishi’ uses katakana rather than hiragana for the word ‘kawaii’, perhaps hinting at an edgier, modern ‘cuteness’ as suggested by Akizuki (2005) in relation to ‘特殊カタカナ語/tokushu katakanago’ (special katakana words). 82 Like the term ‘和語/wago’, this refers to the Japanese language as it was used before Chinese language contact. 167

A further code of interest, although small (4), was ‘hard to read’ which some participants noted occurred when a number of words in a row were written in hiragana. As previously mentioned, Japanese writing does not use spaces to indicate word boundaries, but as most sentences use of a mixture of scripts (hiragana, katakana, and kanji), the word boundaries are usually identifiable. Conversely, when a number of hiragana words appear in a row it can be difficult to find those boundaries. The difficulty in reading strings of words in hiragana is relevant to the number of hiragana loanwords occurring in a row in the texts which made up the corpus in layer 1 (section 4.1.7).

The findings of this section of the survey are consistent with associations of hiragana noted in previous studies, summarized in section 2.4.2. These include hiragana being soft, feminine, easy, round, simple and tender/gentle. In addition, further associations arose from this study which were less established, and included ‘children’, ‘cuteness’, and ‘Japan’, as described above.

5.2.2 Katakana

For the question on katakana (Q10), 177 respondents listed words or impressions they associated with the script. A total of 480 associations were given which were classified into 31 different codes. The ten most frequently occurring codes are listed in table 5.3 below.

By far the most common code was ‘foreign’ (128), around 70% of participants listed a word such as ‘外来語/gairaigo’ (foreign words or loanwords) or ‘外国/gaikoko’ (foreign country) among their responses. The second most frequently occurring code was ‘angular/sharp’ (46), like ‘foreign’ it also occurred frequently in the study by Iwahara et. al (2003).

As explained in the research design (section 3.4.3), the responses for katakana also contained a homonym, that for ‘katai’ which can mean solid or firm, as well as formal or stiff. While ‘solid/firm’ seems a more likely interpretation for the word ‘katai’ when referring to katakana, the fact that some respondents did give responses coded as ‘formal/stiff’ means that the intended meaning could not be assumed. Katakana may be considered ‘formal/stiff’ as it was the script in which many government documents appeared during World War II, and it is also used in official printed documents such as bank passbooks, receipts, and the transcription of

168

Japanese names in passports. Printed texts such as passbooks and receipts may also be related to the category ‘cold/mechanical’ (20).

Table 5.3: Common associations of katakana

No. of Code Example tokens

1. Foreign 外来語/gairaigo (loanwords) 128

2. Angular, sharp 角張った/kakubatta (angular) 46

3. Difficult 読みにくい/yominikui (hard to read) 31

4. Firm, hard 硬い/katai (solid) 30

5. ‘Katai’ かたい /katai (solid or formal) 20

6. Cold, mechanical ロボット/robotto (robot) 20

7. Emphasis, unique 強調/kyōchō (emphasis) 18

8. Onomatopoeia 擬音語/giongo (onomatopoeia) 17

9. Cool, youth culture かっこいい/kakkoī (cool) 16

10. Stupid, half-baked バカさ/bakasa (stupidity) 15

Katakana was sometimes associated with difficulty (31), of these, 19 people stated ‘difficult to read’, 10 stated ‘difficult to write’, while the remaining two responses were ‘hard to remember’ and ‘easily mixed up’. These associations were not mentioned in any of the previous studies of script association and therefore contributes a new connotation of the katakana script. This ‘difficulty’ may also relate to the number of business or specialist terms (9), and also to the opacity of some loanwords as cited by Tomoda (1999), as well as the National Institute for Japanese Language and Linguistics (NINJAL) (2006).

169

Besides difficulty in reading/writing, a number of negative reactions to katakana were expressed through codes such as ‘stupid/half-baked’ (15), and ‘vague/shady (10), these may also reflect negative attitudes to loanwords as described by Loveday (1996), Tomoda (1999) and Hosokawa (2015).

The code ‘disfluent/alien’ (12) usually referred to broken Japanese, for example ‘外国人の日本語/gaikokujin no nihongo’ (a foreigner’s Japanese) or ‘片言の日本語/katakoto no nihongo (broken Japanese). In manga and telop (television captions), the speech of even very fluent non-native speakers of Japanese is sometimes rendered in katakana to represent their perceived ‘foreign’ accent (see section 2.6). The word ‘宇宙人/uchūjin’ (alien) was also mentioned three times, and was felt to represent another form of ‘broken’ or disfluent speech. The disfluency associated with the katakana script is also discussed in section 7.3.2.

An interesting feature of the data on katakana was the presence of sets of opposites: easy and difficult; direct and vague; cool and uncool; as well as new and old. Rather than indicating the unreliability of the data, these juxtapositions highlight the many-layered histories, uses and meanings of the script. For example, katakana is both easy to write, but can be associated with technical or jargon-like words. It is associated with ‘new’ terms, but since it was widely used for this purpose during historical periods of loanword uptake such as the Meiji restoration (1853) and the end of World War II (1945) (Loveday, 1996), it may also seem ‘old’. Its association with ‘new’ words may give katakana its ‘cool and fashionable’ image, but perhaps its overuse can make it seem ‘stupid or half-baked’.

These findings are vital in interpreting the significance of loanwords in hiragana, as katakana is the unmarked choice for loanwords. For example, because ‘foreignness’ is so frequently associated with katakana, using hiragana may be a way to make them appear less foreign and more Japanese, for example with items associated with traditional Japanese culture. Katakana’s ‘hardness’, and hiragana’s ‘gentleness’, also suggest a shift from one extreme to another, this was drawn upon in two of the case study texts in particular, はぴねす/hapinesu (Happiness Clinic), in section 4.2.1; and 蜂蜜れもん/hachimitsu remon (Honey Lemon) in section 4.2.2; as well as in the survey responses to those texts (5.3.2 and 5.3.3 respectively).

Finally, the association of katakana with ‘disfluency’ may also be a factor in some of the examples of loanwords in hiragana which exhibit a non-standard transliteration, one that is phonetically closer to the English pronunciation. These include the greeting in case study 4

170

(section 4.2.4), ‘ぐっもーにん/gummōnin’ (G’mornin’). The idea that a shift to from katakana may also represent a shift from a perceived ‘disfluency’ will be discussed further in section 7.3.2.

The associations for katakana therefore represent ‘push factors’ in interpreting loanwords in hiragana, while the associations of hiragana can be thought of as ‘pull factors’. These findings, along with those for the other scripts, provide a relevant background for understanding loanwords in hiragana.

5.2.3 Kanji

For the question on kanji (Q11), 165 respondents listed words they associated with this script, and a total of 499 associations were given, which was higher than for either hiragana or katakana. These were classified into 32 codes, the ten most frequent codes appear in the table below (5.4). As with the preceding script associations, the word ‘katai’ was separated into three codes: formality (堅い), firmness or solidity (硬い、固い) and a third category where the meaning was ambiguous (かたい).83

The results for ‘kanji’ also reflected often-cited connotations of difficulty (181), formality (49), and intelligence or education (32).

Both ‘China’ (48) and ‘Japan’ (10) were mentioned by respondents, highlighting kanji’s Chinese origins as well as its central place in the Japanese writing system. This duality is also reflected in the codes for culture/literature (8), philosophy/religion (5), and art /calligraphy (4). Many of the responses in these codes reflect nativized Japanese forms of Chinese cultural artefacts, such as calligraphy (書道/shodō) and even Bushido, ‘the way of the warrior’ (武士道/bushidō), which was based on Confucian texts.

The results for people’s perceptions of kanji also reflected a range of contrasting attitudes, some reflecting positive associations such as ‘intelligent’, ‘beautiful’, ‘interesting’, and ‘cool’ while others represented more negative terms such as ‘difficult’, ‘oppressive’, and ‘disliked’. 84 While the contrasting views for katakana (such as ‘old’ and ‘new’) in the

83 For this question, ‘katai’ was not found in katakana ‘カタイ’ or romaji ‘katai’. 84 Some of these responses occurred in small numbers and are therefore not included in table 5.4. 171

previous section seemed to reflect different standpoints or contexts, those for kanji suggested either a cause-and-effect relationship (eg. kanji is interesting because it is complex), or a sense of kanji being necessary and useful despite its complexity.

Table 5.4: Common associations of kanji

No. of Code Example tokens

1. Difficult 難しい/muzukashī (difficult) 81

2. Formal 堅苦しい/katakurushī (stiff, formal) 49

3. China 中国/chūgoko (China) 48

4. Intelligent 賢い/kashikoi (intelligent) 33

5. Complex, deep 複雑/fukuzatsu (complicated) 27

6. Firm, solid 固い/katai (firm, solid) 25

7. Easy to understand 分かり易い/warakiyasui (easy to understand) 23

8. Pictorial, graphic 象徴的/shōchōteki (symbolic) 20

9. Old, historical 伝統/dentō (tradition) 17

10. Beautiful 美しさ/utsukushisa (beauty) 15

Other responses given by participants suggest a comparison with the other Japanese scripts such as the phonetic hiragana and katakana, such as kanji being ‘pictorial/graphic’ (20), or ‘concise/precise’ (7). The coding difficulties with the word ‘katai’ are a good illustration of the preciseness of kanji, even the two characters meaning ‘solid, firm’ (固い and 硬い) have their own distinct nuances. The ‘pictorial’ nature of kanji is perhaps what makes them ‘easy to understand’ (23), by identifying the elements or ‘部首/bushu’ (radicals) of an unfamiliar kanji or kanji-compound, a Japanese speaker will usually have some idea of the meaning of the word. A parallel can be seen in the way that a knowledge of Greek or Latin roots can help understand an unfamiliar English word.

172

The associations of kanji therefore provide another backdrop through which to interpret loanwords in hiragana. While not many loanwords have kanji that are widely used or recognized, two exceptions are ‘檸檬/remon’ (lemon) and ‘珈琲/kōhī’ (coffee). These words were both present in the survey85 (layer 2), and formed a key topic of discussion in the interviews (layer 4), ‘lemon’ in particular being one of the most frequently occurring words in the corpus (layer 1). These layers of analysis are brought together in the discussion chapter, section 7.2.

5.2.4 Rōmaji

For the final script investigated, rōmaji (Q12), 156 people responded generating 332 associations, this was the fewest associations per respondent of all the scripts (2.1). The responses were classified into 33 codes, the ten most frequently occurring are shown in the following table (5.5).

Rōmaji is not often included in descriptions of the connotations of ‘Japanese scripts’, and was not investigated in the empirical study conducted by Iwahara et al. (2003). The results of this section can therefore add to the inventories of script association with the following findings.

Two frequently occurring codes relate to seemingly similar concepts: ‘Foreign’, and ‘International’, but the responses differ in their relative positioning of Japan and the world. The first code, ‘foreign’ represents ideas and terms that position rōmaji outside of and external to Japan, for example ‘外来語/gairaigo’86 (foreign words/loanwords), ‘英語/eigo’ (English language), and ‘ 西洋/seiyou’ and ‘ 欧米/ōbei’ (the west). The other code, ‘International’, differed from ‘Foreign’ in that it represents a sense of worldliness inclusive of Japan, for example through terms such as ‘世界共通/kokusaikyōtsū’(worldwide), ‘グローバル化/gurōbaruka’ (globalization) and of course ‘ 国際的 /kokusaiteki’ (international). Comparing these two codes, ‘foreign’ was much more frequent (75), with ‘international’ receiving only 10 responses. A further code, ‘hybridity’, was used for responses that positioned rōmaji as a kind of mediator between the Japanese and English, and included

85 ‘Lemon’ appears in hiragana in the text used in the survey, but many respondents drew on the existence of the kanji in their answers (see section 5.3.3) 86 Although a case is being made that loanwords are part of the Japanese language, it is unusual for loanwords to be written in romaji, so ‘gairaigo’ as it appears here is likely to refer to ‘foreign words (in the roman alphabet)’ 173

responses such as ‘Japanese made to look like English’,87 and ‘half-baked: neither English nor Japanese’.88 A further code, ‘Japan’, was mentioned by only 6 respondents, and included words and phrases such as ‘日本的/nihonteki’ (Japan-like) and ‘日本語/nihongo’ (Japanese language). This suggests that rōmaji most commonly represents something ‘outside’ of Japan, rather than something of Japan, or that connects Japan to the world. It is difficult, however, to determine whether respondents were reflecting on the rōmaji script itself, or on English or foreign words which appear in rōmaji.

Table 5.5: Common associations of rōmaji

Number of Code Example tokens

1. Foreign 外来語/gairaigo (loanwords) 75

2. Cool かっこいい/kakkoī (cool) 38

3. Hard (to read/understand) 読みにくい/yominikui (hard to read) 30

4. Stupid, clumsy ダサい/dasai (uncool) 16

5. For specific situations 暗号/angō (password) 15

6. International 世界共通/sekai kyōtsū (worldwide) 10

7. Emphasis 強調/kyōchō (emphasis) 10

8. Pronunciation 音/oto (sound) 10

9. Easy (to read/understand) 分かりやすい/wakariyasui (easy to read) 10

和製英語/wasei eigo (Made-in-Japan 10. Hybridity 9 English)

87 英語っぽく日本語をみせた 88 中途半端（英語でもなく、日本語でもない） 174

Like katakana and kanji, the responses for rōmaji reflected a range of viewpoints, with the positive term ‘cool’ (‘かっこいい/kakkoī’ [cool], or ‘おしゃれ/oshare’ [fashionable]) representing the second most frequently occurring code, while ‘hard to read/understand’ and ‘stupid/uncool’ ranked third and fifth.

The findings on the associations of ‘rōmaji’ provide another lens through which to interpret loanwords in hiragana. As loanwords are words from languages other than Japanese, writing them in rōmaji is another potential alternative to katakana. However, the distinction between loanword-in-rōmaji and English words can sometimes be unclear as alluded to above. For example, a word such as ‘tomato’ is transcribed the same in English as it is as a loanword, while ‘potato/poteto’ is not. This issue is further explored in the rōmaji rendering of ‘handkerchief’ (hankachi) in the following section of the survey (section 5.3.6).

5.3 Responses to sample texts

In this section, respondents were presented with photos of examples of loanwords in different scripts and asked if they thought the use of the script was appropriate and why. They were also asked why they thought the designer/writer of the text had used that particular script as described in the research design section 3.4.3. The complete survey in Japanese and English can be found in appendix 2, and the question numbers used there are referred to in the following sections.

5.3.1 Appropriateness by example

The survey sought opinions on three different texts containing loanwords in hiragana (‘はぴねす/hapinesu’ (happiness) [Q13], ‘れもん/remon’ (lemon) [Q16], and ‘ふらい/furai’ (fly) [Q19]. Overall, loanwords in hiragana were considered appropriate for the context 73% of the time. However, differences between these examples emerged, for example, ‘lemon’ was considered appropriate by a large majority of respondents (90%), while ‘fly’ received an appropriate rating by just over half of respondents (55%).

175

Two more texts offering further orthographic options, kanji for ‘珈琲/kōhī’ (coffee) [Q22], and rōmaji for ‘hankachi’ (handkerchief) [Q25], were also included in the survey. ‘珈琲/kōhī’ (coffee) received the highest percentage of ‘appropriate’ responses of all the examples in the survey with 98%, while only 63% of respondents rated the rōmaji ‘hankachi’ as ‘appropriate’, making it less than the average for the hiragana loanwords (73%). These results are summarized in table 5.6 below.

Table 5.6: Appropriateness by example

Survey question Appropriate Inappropriate Total respondents

‘Happiness’ clinic [Q13] 116 42 158 (はぴねす) (73%) (37%)

‘Honey lemon’ drink [Q16] 136 14 150 (蜂蜜れもん) (90%) (10%)

‘I can fly’ sticker [Q19] 81 67 148 (I can ふらい!!) (55%) (45%)

‘Dairy coffee’ instant coffee [Q22] 139 3 142 (牛乳屋さんの珈琲) (98%) (2%)

‘Cute handkerchief’ [Q25] 90 52 142 (かわいい Hankachi) (63%) (37%)

Average for all loanwords in survey 76% 26% 148

Average for loanwords in hiragana 73% 27% 152 (Happiness, Lemon, and Fly)

The reasons why respondents labelled these examples as appropriate or inappropriate are explored in the following sections. Findings on the perceived motivation for using a particular script are also presented. The tables included in the following sections show the ten most frequently occurring codes for each question.

176

5.3.2 ‘Happiness’ clinic

The first picture question in the survey asked people to respond to the use of hiragana for the loanword ‘happiness’. A breakdown of the response rate is presented in the following table (table 5.7), similar tables are included for each of the following four sample texts used in the survey.

Table 5.7: Response rates for ‘Happiness’ clinic text (Q13-15)

‘Happiness’ Appropriate Inappropriate Total respondents

Q13. Appropriate/inappropriate? 116 42 158

Q14. Why? 111 40 151

Q15. Why do you think the shop used hiragana? 110 44 154

Most respondents (73%) felt that writing ‘happiness’ in hiragana was appropriate. When asked for the reasons for their decision, the responses of these 116 people generated 165 tokens, or reasons. The ten most frequently occurring reasons are presented in the following table (5.8).

The fact that many respondents noted that this example was acceptable because it was a proper noun (22) is significant in the light of the results of the corpus analysis (layer 1). In the corpus, the most frequently occurring word class was ‘noun’, and one of the largest categories of referential meaning were for business names, (see sections 4.1.2 and 4.1.4 respectively). This suggests that respondents may have been familiar with the use of hiragana loanwords for names or proper nouns, making this the most common reason for appropriacy. Non-standard orthography as a common strategy for business names has also been noted in the literature review, (section 2.6).

The second most frequently occurring code related to ‘the freedom of the individual’ (21) to choose in which script to write a loanword. Despite the associations of particular lexical items with particular scripts in written Japanese (eg. katakana with loanwords, and kanji for Sino-Japanese words), the flexibility of these conventions is one of the main findings of the current study, being mentioned across all the samples in the survey and frequently discussed in the interviews. These findings are brought together in section 7.1.

177

Table 5.8: Reasons for appropriacy of 'happiness' in hiragana (Q14)

Number of Code Example tokens

1. Name, proper noun 店名/tenmei (shop name) 22

2. Freedom of individual 書き手の自由/kakite no jiyū (freedom of the writer) 21

不適切とは思えない/futekisetsu to wa omoenai (do 3. Not wrong 20 not think it is inappropriate)

4. Familiar 親しみ/shitashimi (familiar) 17

5. Soft 柔らかい/yawarakai (soft) 12

6. Readable, easy to read 読める/yomeru (readable) 10

お年寄りが多そう/otoshiyori ga ōsō/ (seems to be 7. Suitable for the elderly 10 many elderly people [patronising this shop])

8. Eye-catching, emphasis 目立つ/medatsu (eye-catching) 8

Image or concepts of the お店のコンセプトを反映/omise no konseputo o 9. 7 business hanei (expresses the shop’s concept) 4 10. Gentle 優しい/yasashī (gentle)

The responses stating that writing loanwords in hiragana was ‘not wrong’ (20) were also of interest, although no-one stated that writing loanwords in hiragana was ‘right’ or ‘correct’ for this example. One participant expressed both an awareness of the conventions of script choice as well as an understanding of the dynamic nature of language: ‘The general rule is that katakana is assigned for the writing of loanwords, but the rules of language change flexibly over time.’89

In addition, indirect support for the code ‘not wrong’ was found in the responses of eight individuals who seemed to be interpreting the question as ‘why is it (hiragana) “はぴねす /hapinesu” and not (English) “Happiness”?’, rather than ‘why is it (hiragana) “ はぴねす”

89外来語にカタカナを充てるのが原則だが、言葉の規則は時代とともに柔軟に変わるものだから 178

and not (katakana) “ハピネス”?’. In other words, respondents understood the question of appropriateness as encompassing the expectation that it should be in English/rōmaji, rather than katakana, evidence of the extent to which rōmaji and/or English terms have become part of the Japanese writing system. This resulted in the code ‘English is hard to read’, and included responses such as ‘(Hiragana is appropriate because) even older people who aren’t familiar with English can read it.’90 This code contained five of these types of responses, while a further three simply prefaced their comment with a remark such as ‘Compared to “happiness” in English…’ 91 The translingual context of Japan and the availability of rōmaji/English as an alternative is further investigated in the discussion chapter, section 7.2.

In addition, many of the associations of hiragana noted by respondents in the first section of the survey (Q9) were given as reasons for the appropriacy of writing ‘happiness’ in hiragana, for example that it made the business seem familiar or approachable (17), it gave the business a soft (12) or gentle (4) image, or that it was easy to read (10).92 This suggests how the associations of a script can be transferred to a referent, adding extra layers of meaning to a word or text.

Some participants explained their choice of ‘appropriate’ by considering the script alternatives; one respondent noted that hiragana is softer than katakana, and compared to rōmaji is easier to read for Japanese people of all ages.93 Another respondent also considered Sino-Japanese translations of the business name ‘happiness acupuncture clinic’ such as ‘幸せ鍼灸院/shiawase shinkyūin’ and ‘幸福鍼灸院/koufuku shinkyūin’, noting that the first seemed a bit unpolished and the second had slightly suspicious religious connotations.94 This example shows how users of Japanese are aware of a range of linguistic options available, not only through the choice of script, but of translated Japanese alternatives to the loanword as well.

The remaining 42 respondents selected ‘inappropriate’ for the text containing the word ‘はぴねす/hapinesu’ (Happiness), and of those, 40 people gave reasons for their answer.

90英語に馴染みのない高齢者でも読める 91 英語としての happiness というよりは… 92 The category ‘Readable/easy to read’ was made up of readable (3), understandable (4), easy to read (3). 93 カタカナよりも柔らかく、ローマ字よりも日本人のどの年齢層にも理解できるから。 94 カタカナや英語表記で書いてあると、対象となる高齢者は近寄りにくくなる。かといって、「幸せ鍼灸院」とすると野暮ったくなる。また、「幸福鍼灸院」では少し(あやしい)宗教的な感じもする。 179

Altogether, their reasons generated 53 tokens; these were categorized into seven codes. The results appear in table 5.9 below.

Table 5.9: Reasons for inappropriacy of 'happiness' in hiragana (Q14)

Number of Code Example tokens

元々日本語ではないから/motomoto nihongo de wa nai 1. English, not Japanese 19 kara (not an original Japanese word)

外国語のひらがな表記は正しくない/gairaigo no 2. Incorrect hiragana hyōki wa tadashikunai (not correct to write 13 loanwords in hiragana)

意味がわかりづらい/imi ga wakarizurai (hard to get the 3. Hard to understand/read 10 meaning)

4. Weird 違和感がある/iwakan ga aru (feels weird) 4

5. Unserious, childish 頭が悪そう/ atama ga warusō (looks stupid) 3

Other あざとい/azatoi (sly, cunning) 4

The two most common reasons given were that ‘happiness’ is an English word, not a Japanese word (19), and that it was incorrect to write such words in hiragana (13). Some responses combined these reasons, such as ‘Happiness is an English word, and shouldn’t be written in hiragana.’95 It is interesting that for the respondents who thought this example was appropriate, hiragana being a legitimate choice ranked highly, while for those who chose ‘inappropriate’, hiragana as an illegitimate choice also ranked highly. In each case, reasons of (il)legitimacy were given more frequently than reasons relating to the associations of the script (eg. either soft and gentle in the case of ‘appropriate’, or childish and weird in the case of ‘inappropriate’. This suggests that, while the connotations and visual impact of the script

95 happiness は英語であり本来ひらがなで表示するべきでないから 180

was important, the conventions of script use were cited more frequently by both groups of respondents.

While the hiragana script was associated with the word ‘easy’ as its second most frequent connotation in the first section of the survey (Q9), respondents who selected ‘inappropriate’ for this example also gave ‘hard to read/understand’ as reasons for their answer. This was explained by one respondent as being because of the word being an unusual case of a foreign word in hiragana, ‘It’s hard to read because it’s English.’96 Another participant noted that the habit of seeing particular words in particular scripts was the key, ‘…Words that are originally foreign words are easier to understand in katakana. This is because you can quickly distinguish Japanese words from foreign words.’97 The responses under the code ‘weird’ (4) may also be related to the fact that this example is a foreign word being presented in a script conventionally associated with Japanese words.

The third question for this text was ‘Why do you think that the shop used hiragana for their name?’ (Q15). The answers from the respondents who chose ‘appropriate’ or ‘inappropriate’ were calculated separately, to see whether this initial response affected their answer to the final question, but the results for both groups were very similar: the seven most frequently occurring codes occurred in both groups, and with minor exceptions, in the same order. For this reason, they were combined and the ten most frequently occurring codes for the 154 respondents appear in the following table (5.10). These respondents generated a total of 257 tokens in answer to the question of the perceived motivation of the designer for using hiragana for the word ‘happiness’.

Many associations of hiragana were the most frequent responses, such as soft (42), gentle (27), and approachable or familiar (62). Some respondents also considered the nature of the business suggesting ‘They wanted to appeal (to the public) as being an osteopathic clinic that isn’t painful or scary, as having a gentle image.’98

96英語なので読みづらい 97外国語をもとにした言葉はカタカナにした方が分かりやすい。日本語か外来語かすぐに判別することができるから。

98 痛くない、怖くない整骨院(優しいイメージ)をアピールしたいように思える 181

Table 5.10: Respondents’ perceived reasons for the use of hiragana for 'happiness' (Q15)

No. of Code Example tokens

1. Familiar 親しみやすい/shitashimiyasui (familiar) 62

2. Soft 柔らかさを出す/yawarakasa o dasu (expresses softness) 42

3. Gentle, kind 優しい/yasashī (gentle) 27

Elderly are target 4. 高齢者むけ/kōreisha muke (for the elderly) 18 market

5. Easy to read/understand 分かりやすい/wakariyasui (easy to understand) 18

6. Cute かわいい/kawaii (cute) 10

7. Eye-catching 目立つ/medatsu (eye-catching) 10

8. ‘Yasashī’ (ambiguous) やさしい/yasashī (easy or gentle) 8

9. Healing 癒しの印象/iyashi no inshō (a healing impression) 7

10. Memorable 覚えやすい/oboeyasui (easy to remember) 7

Many respondents believed that the elderly were the target demographic for this business (18), these responses were found amongst all age brackets up until 41-50, although this was not mentioned by the five respondents over 51 who participated in the survey. This interpretation was often found in combination with suggestions that hiragana would be easier to understand (18), for example the response ‘It’s to make it easier to read for elderly people.’99 Three respondents also commented that English and rōmaji were less likely to be understood by this generation, suggesting that there was also a consideration of English/rōmaji as an alternative orthographic choice as mentioned previously. One respondent in his 30s suggested that rōmaji is ‘distant’ for older Japanese, ‘For elderly people who are at a distance from rōmaji, (hiragana) has a feeling of intimacy or familiarity.’100

99 老人に読み易くするため 100 ローマ字を敬遠するお年寄りにも親近感を抱いてもらうため。 182

Four respondents also cited reasons why katakana would be unsuitable for the business name: two suggested it would feel ‘katai’ (硬い/solid, and カタイ/solid or strict), one that it would be ‘きつい/kitsui’ (harsh), and another said ‘If it was in katakana, it would not have the image of being able to relieve your pain.’101 These examples as well as the consideration of English/rōmaji as alternatives above point to an awareness of users of Japanese as to the orthographic options available to them, and how these kinds of association and connotations can be skilfully drawn upon by the user.

5.3.3 ‘Honey Lemon’ Drink

This item asked respondents to reflect on the use of hiragana for the word ‘lemon’ in the name of the product, ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon). The number of people who responded to each question varied, and can be seen in the following table (5.11).

Table 5.11: Response rates for 'Honey Lemon' drink (Q16-18)

Lemon Appropriate Inappropriate Total

Q16. Appropriate/inappropriate? 136 14 150

Q17. Why? 131 14 145

Q18. Why do you think the company used hiragana? 128 16 144

This example showed a higher degree of appropriateness than the previous example, with 90% of respondents indicating that writing ‘lemon’ in hiragana was appropriate in this text. The 136 respondents generated a total of 193 reasons for their answer. The ten most frequently appearing codes for the appropriacy of writing ‘lemon’ in hiragana are shown in the following table (5.12).

Like the previous example, the idea that it was ‘not wrong’, and that it was acceptable as the word was a noun or name were common responses. Further similarities across examples in the survey will be further discussed in the summary at the end of this chapter (5.4).

101 カタカナだとコリが取れるイメージがない。 183

Table 5.12: Reasons for appropriacy of 'lemon' in hiragana

No. of Code Example tokens

1. Not wrong 間違ってはいない/machigatte wa inai (not wrong) 26

2. Noun/name 商品名/shōhinmei (product name) 22

れもんは漢字もあり/remon wa kanji mo ari (lemon also 3. It has a kanji 15 has a kanji)

4. Freedom of individual 個人の自由/kojin no jiyū (freedom of the individual) 14

ひらがなでれもんと書く人も多い/hiragana de remon to 5. Commonly used 13 kaku hito mo ōi (many people write ‘lemon’ in hiragana)

6. Mellow, soft まろやか/maroyaka (mellow)un 11

7. Japanese-ness 和の印象/wa no inshō (a Japanese-like impression) 11

国産レモンを使用している/kokusan remon o shiyōshiteiru 8. Domestic-grown lemons 10 (made with domestic-grown lemons)

9. Sweet, delicious 甘い印象/amai inshō (sweet impression) 8

れもんとして日本語が確立しており/remon toshite 10a. Becoming/is Japanese nihongo ga kakuritsu shiteori (‘lemon’ is established as part 7 of the Japanese language)

れもんの漢字は難しい/ remon no kanji wa muzukashī (the 10b. Kanji is difficult 7 kanji for ‘lemon’ is difficult)

184

Interestingly, many respondents noted the fact that ‘れもん/remon’ (lemon) can also be written in kanji (15), acknowledging that this particular word had an additional orthographic option compared to most loanwords. Furthermore, seven respondents noted that the kanji for ‘lemon’ (檸檬/remon) was particularly difficult. The existence of a kanji character for ‘lemon’ was therefore given as a reason for the appropriacy of writing the word in hiragana from two different angles: firstly, that the existence of a kanji gave some kind of legitimacy to ‘remon’ as a ‘Japanese word’, and therefore it was appropriate to be written in hiragana; and secondly that because the kanji for this word was particularly difficult, the hiragana version might be preferable to the kanji version. Both of these arguments may be the reason that 13 respondents noted that it was common to see ‘lemon’ written in hiragana, and indeed, it was one of the most frequently occurring words in the corpus occurring 15 times (see section 4.1.3).

Related to these ideas, seven respondents also expressed a feeling that ‘lemon’ was making the transition from ‘loanword’ to ‘Japanese word’, and this meant that it was not inappropriate to write it in hiragana. As one respondent explained, ‘“Lemon” isn’t just written in katakana, it has a kanji too, and so I feel it’s a word that’s transitioning to Japanese. And if it’s a Japanese word, then it doesn’t feel weird to see it in hiragana.’102 One respondent contrasted the degree of familiarity of the word ‘lemon’ with ‘happiness’ in the previous question, explaining ‘In Japan, there’s a recognition that ‘Lemon’ [written in English] means ‘lemon [レモン/remon]’, and that degree of recognition far surpasses that for ‘happiness’. I think of ‘lemon’ as part of the Japanese language, so I don’t have any objections (to this example).’ 103 These results show that ‘lemon’ may be an unusual loanword in both having a recognized kanji, and the degree to which it is thought of as a ‘Japanese word’. This was therefore introduced as a question in the next layer of analysis, the interviews (layer 4). The role of kanji in relation to loanwords in hiragana is explored in the discussion chapter (section 7.1).

Other reasons frequently given for the appropriacy of writing ‘lemon’ in hiragana was that it made the product seem ‘mellow/soft’ (11), and ‘sweet/delicious’ (8). This may be an

102 レモンはカタカナだけではなく、漢字で表記することもあり、外来語から日本語へと移行した言葉だと感じる。日本語であればひらがな表示をしても違和感がない。 103 Lemon はレモンと日本では認識されており、その認知度は happiness を遥かに凌駕している。日本語の一部と考えて差し支えないように思われる。 185

application of the ‘soft’ or ‘gentle’ connotations of the hiragana script (section 5.2.1) in the context of a lemon flavoured drink, in other words a ‘soft’ flavour that was not sour.

Of the respondents who selected ‘inappropriate’ (14), each gave relatively short responses resulting in 14 tokens. This resulted in only three codes, all of which are illustrated in the following table (5.13).

Table 5.13: Reasons for inappropriacy of 'lemon' in hiragana

Code Example No. of tokens

1. Loanword 外来語だから/gairaigo dakara (Because it’s a loanword) 6

普通はカタカナ/futsū wa katakana (It’s normally in 2. Incorrect 3 katakana)

3. Other いまいち/imaichi (It’s not so great) 5

The two most common codes reflected the conventions for writing loanwords, with six people simply stating that ‘lemon’ was a loanword while three reported that it was incorrect, or that loanwords should be written in katakana. Five further responses were coded as ‘other’, and included reasons such as it was inconsistent with the word written in katakana elsewhere on the label,104 or even a preference for the kanji version of lemon, stating ‘kanji would have had more impact.’105

Like the previous example, the results for the final question, asking the designer’s motivation for writing ‘lemon’ in hiragana were initially analysed separately to see if any differences occurred. With such a small proportion (10%) of respondents choosing ‘inappropriate’, it was difficult to draw any firm conclusions, although it was noted that the code ‘Japanese-ness’, which was the second most frequent for those choosing ‘appropriate’, was not given by any of the respondents who selected ‘inappropriate’. The 144 respondents’ answers were coded, and generated 236 tokens in their answers. The results for both groups concerning the designer’s motivation for writing lemon in hiragana are given in the following table (5.14).

104 横に「国産レモン使用」と表記してあるから、統一感がない 105 漢字の方がインパクトがあると思うから 186

Table 5.14: Respondents’ perceived reasons for the use of hiragana for 'lemon'

No. of Code Example tokens

1. Made in Japan 国産製品/kokusanseihin (domestic product) 43

2. Japanese-ness 和風/wafū (Japanese-style) 25

カタカナのレモンは果物のレモンである/katakana no 3. Katakana is unsuitable remon wa kudamono no remon de aru (lemon in katakana 17 would be the fruit)

4. Eye-catching インパクトがある/inpakuto ga aru (has an impact) 18

5. Old-fashioned, traditional 懐かしい/natsukashī ( nostalgic) 11

6. Soft 柔らかい/yawarakai (soft) 10

7. Familiar 親しみやすく/shitashimi yasuku (friendly) 11

蜂蜜が漢字で書かれているから、レモンを平仮名で

書くことでバランスが良く/hachimitsu ga kanji de 8. Balance with other scripts kakareteiru kara, remon o hiragana de kaku koto de 10 baransu ga yoku (As ‘honey’ is in kanji, writing ‘lemon’ in hiragana give a good balance.)

酸っぱさを控えめ/suppasa o hikaeme (moderates the 9. Sweetness, not sour 9 sourness)

10. Easy to read/understand 読みやすい/yomiyasui (easy to read) 9

The two most frequently occurring codes, ‘made in Japan’ and ‘Japanese-ness’, suggest how respondents engaged with the text as a whole and not the word in isolation. The most frequent code related to part of the label that said ‘Made with domestically-grown lemons’,106 despite the small size of this element in relation to the rest of the label. Most respondents

106国産レモン使用 187

used these words in their responses, while some simply stated that it was ‘made in Japan’.107 The second most frequent code also related to the text as a whole, with respondents stating that the design had a ‘Japanese style’.108 Some were more specific, noting the ‘characteristic Japanese font’,109 and three mentioned the calligraphic font, for example, ‘it has a good balance with the paint-brush style font.’110 These kinds of responses reinforce the multimodal nature of meaning making, as the respondents demonstrated how their perception of the motivations of the writer/creator of the text extended beyond the word(s) to the other semiotic resources deployed in the text.

Another example of a multimodal approach to this question can be seen in the code ‘balance of scripts’, in which ten respondents commented on the scripts used in other parts of the text to suggest why hiragana was chosen in this instance. Some respondents suggested that the use of hiragana related to preceding word, which was written in (a relatively difficult) kanji, ‘Kanji has been used for “蜂蜜/hachimitsu” (honey), and they used hiragana to contrast with it.’111 This could be because kanji has connotations of ‘difficulty’ while hiragana is associated with ‘ease’.112 Other respondents suggested that hiragana had been used to differentiate the word ‘lemon’ in the name of the product from the ingredient ‘lemons’ used in the phrase ‘made with domestically-grown lemons’. As one respondent explained, ‘To one side it says “Made with domestically-grown lemons.” So I wonder if they’re making the distinction that ‘lemon’ in katakana is the raw materials, while ‘lemon’ in hiragana is the product name, that’s the impression they’re trying to give.’113 This suggests that scripts can be used to distinguish different referents within a text, and is discussed further in the interviews (layer 4) in relation to other texts that contained the same word in both hiragana and katakana variations (section 6.5).

107日本製である 108 和風 or 日本的 109日本固有の文字 110字体(毛筆)とのバランスもいい 111蜂蜜で漢字を使用し、それと対比してひらがなを使ったと思う 112 This phrase has a rather large number of orthographic options as discussed in section 4.3.2, with combinations of the word ‘honey’ (蜂蜜,はちみつ, ハチミツ) and ‘lemon’ ‘檸檬,れもん,レモン) 113横に「国産レモン使用」とある。レモンは原料で、商品名は「れもん」と使い分けて、印象付けようとしているのではないでしょうか？ 188

In an extended answer, a male in his 30s who stated writing ‘lemon’ in hiragana was inappropriate114 explained what he perceived to be the motivations of the designer as well as how this technique was no longer effective:

Today in Japan, this kind of expression (writing lemon in hiragana rather than the normal katakana) is popular. Originally, this was different from the usual version so it had high visibility, and gave a special value to the product, and had the purpose of differentiating the product from those of other companies. Now, because an exceedingly large number of companies are using this type of expression, the truth is that the effectiveness of this technique is wearing thin. The business (who produced this label) is not aware of this.115

This opinion is interesting as it also reflects the feeling that it is not uncommon to see ‘lemon’ written in this way, as explained in the reasons for appropriateness. However, for this respondent, the technique has become overused and hackneyed.

The third most frequent code for the perceived reasons of the designer related to the unsuitability of katakana for this text. For 17 of the respondents, the choice of the hiragana script for the word ‘lemon’, was preferable because katakana would either seem to represent the fruit (5), appear foreign (4), be too normal (2), seem sour (2), be ‘katai’ (2), be angular (1), or official (1). Others felt that kanji would not be suitable (10), with reasons including that it was difficult (7), oppressive (2), or that it would seem ‘self-important’ for a juice (1), although this respondent noted that if it was a type of alcohol, kanji would be acceptable.116

To conclude this section, the bottle of ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) drink received the highest rating of appropriateness of the three loanwords in hiragana presented in the survey. The reasons given for appropriacy and inappropriacy show similarities with the other texts containing loanwords in hiragana, but specifically for this word/text, many respondents drew on the fact that ‘lemon’ has a kanji as a further alternative to hiragana, katakana, and rōmaji.

114 Despite choosing ‘inappropriate’ in the first question, the writer downplays this decision in the reason ‘I don’t know if it’s appropriate or inappropriate but in terms of the design of the bottle label it’s not great’ 適切か不適切かわからないけど、ボトルのラベルデザインとしてはいまいちかな。 115今日本ではこういうあえて表現（れもん←普通はレモンと表記する）が流行っている。本来この表現は、あえて普段とは違う表記をすることで視認性を高め、同時にその商品に特別な価値が含め、他社との差別化を図る目的があった。今は非常に多くの会社がこの表現を用いるので、そのデザイン効果がうすれていっているのが事実。←このことに気付いていない会社。 116ジュースに漢字だと偉そうな感じがするから、酒なら漢字でよいとおもう。 189

5.3.4 ‘I can fly’ sticker

This text received the most similar proportions of ‘appropriate’ and ‘inappropriate’ responses of all the examples of loanwords in hiragana presented in the survey. The number of respondents for each question is show in the table below (5.15).

Table 5.15: Response rates for 'I can fly!!' sticker (Q19-21)

Fly Appropriate Inappropriate Total

Q19. Appropriate/inappropriate? 81 67 148

Q20. Why? 76 66 142

Q21. Why do you think the designer used hiragana? 78 67 145

The reasons given by the 81 respondents who selected ‘appropriate’ were categorized into 15 codes, generating a total of 109 tokens. The ten most frequently appearing codes appear in the following table (5.16).

Besides the freedom of the designer to write words how they felt best, which was the most frequently occurring reason (15), many respondents simply stated that it was appropriate because it was part of the design (14). This was not mentioned for the ‘はぴねす/hapinesu’ (happiness) text, although five respondents did mention this for the ‘蜂蜜れもん/hachimitsu remon (Honey Lemon) text. While it is unclear from these statements exactly which aspect of the design is referred to, it may be related to the code ‘cute’, which received 12 responses, and also the idea that it suited the character, which was mentioned by six respondents. Because this text differs from the previous two in being a type of creative work (sticker) rather than a text associated with a business or product, it is possible that the overall aesthetics of the design were foregrounded for these respondents, leading to the large number of people giving their reasons in terms of the design. The importance of the genre of text was also noted by the five responses who mediated their choice of ‘appropriate’ by saying that it was acceptable for particular texts but not others, for example ‘It would be inappropriate for official documents like a thesis or a contract, but this is (just) a kind of product…’117

117論文や契約書といった公式な文書では不適切だが、これは商品の一部であり… 190

Table 5.16: Reasons for appropriacy of 'fly' in hiragana

Code Example No. of tokens

1. Freedom of individual 表現の自由/hyōgen no jiyū (freedom of expression) 15

2. Design デザインの意匠/desain no ishō (design concept) 13

3. Cute 可愛い/kawaii (cute) 12

Understandable, 4. 理解できる/rikai dekiru (understandable) 10 readable

5. Not wrong 問題ではない/mondai de wa nai (not a problem) 9

6. Soft やわらかさ/yawarakasa (softness) 7

英語表記だとインパクトに欠ける /eigo hyōki da to 7. Impact, emphasis inpakuto ni kakeru (if it were in English the impact would 6 be lost)

Suits character ゆるキャラだから/yurukyara dakara (because it’s a 8. 6 ‘yuru-kyara’)

マンガやイラストであれば別に問題ない / manga ya 9. Suits genre irasuto de areba betsu ni mondanai (if it’s in comics or 5 illustrations it’s not really a problem.)118

カタカタだと機会的な印象を受ける/katakana da to Katakana is 10. kikaitekina inshō o ukeru (if it were in katakana it would 4 unsuitable119 seem mechanical )

This text was also different from the previous two as the loanword in hiragana appeared as the ‘speech’ of a character, rather than a product or business name like the majority of words in the corpus. This was highlighted by the six responses mentioning that it suited the

118 Also coded at ‘not wrong’. 119 Responses were ‘official’, ‘hard’, ‘angular’, and ‘mechanical’. 191

character. While two of responses simply mentioned suitability, a further two said that it suited the youth or immaturity of the character, and two more said it was suitable as it was (like) a ‘yuru-kyara’. This term, clipped from the words ‘ 緩いキャラクター /yurui kyarakutā’ (loose or soft character) usually refers to a prefectural or local government mascot; a person dressed up in the costume often appears at public events. While the character in the sticker is not one of these official mascots, it shares the same round, cute, slow, and slightly helpless image of these characters, expressed through modes such as round shapes of the creature’s face and body, and the slightly high angle from which it is depicted. The concept of the yuru-kyara was also drawn upon by respondents in answering why they thought the designer had chosen to use hiragana in the following question.

Hiragana is also associated with Japan and Japanese culture, which led other respondents to suggest that using hiragana for the word ‘fly’ represented a pronunciation of the word with a Japanese accent, as ‘furai/hurai’ rather than ‘fly’. This, in turn could suggest the childish or inexpert image of the character as reported by the respondents, and supported through other markers of ‘cuteness’ as described in the case study for this text (section 4.2.3).

As mentioned above, the text containing ‘ ふらい/furai’ (fly) in hiragana generated the highest proportion respondents who selected ‘inappropriate’ for an example of a loanword in hiragana at 45% (67 respondents). The reasons given by respondents were categorized into one or more of 12 codes, and totalled 84 tokens. The ten most commonly occurring codes appear in the following table (5.17).

For those who responded that it was inappropriate to write ‘fly’ in hiragana, the most frequently occurring reason given was that it was hard to read or understand (20). This may relate to the script-familiarity effect (Tamaoka et al., 1998) referred to previously, where words appearing in an unfamiliar script take longer to process. The difficulty in reading or understanding this word in hiragana may also relate to the confusion of this word with other similar words, as reported by seven respondents. These people explained that the word ‘fly’, when transliterated as the loanword ‘furai’, becomes homonymic with the well-established loanword ‘fry’. The word ‘フライ/furai’ (fry) is used in Japan to describe deep-fried foods including prawns (エビフライ/ebi-furai). One respondent also pointed out that the insect, ‘a

192

fly’ is another possible referent: ‘When you write ‘furai’ in hiragana, many meanings such as ‘to fly’, ‘a fly’, and ‘fried food’ can end up being understood.’120

Table 5.17: Reasons for inappropriacy of 'fly' in hiragana

No. of Code Example tokens 一見してわかりにくい/ikken shite wakarinikui (hard to 1. Hard to understand 20 understand at first glance)

英語と日本語が混ざっているのわかりづらい/eigo to 2. Mixing is problematic nihongo ga mazatteiru no wakarizurai (mixed English and 13 Japanese is hard to understand)121

It’s English/a loanword, 3. 日本語ではない/nihongo de wa nai (it’s not Japanese) 10 not Japanese

4. Weird 変/hen (weird) 9

Confused with 揚げ物の意味と間違えそう/agemono no imi to machigaisō 5. 7 homophones (it’s likely to be confused with ‘fry’)

動詞をひらがなで書くのに違和感/dōshi o hiragana de kaku 6. ‘Fly’ is a verb 4 no ni iwakan (writing verbs in hiragana is weird)122

7. Uncool, clumsy ダサい/dasai (uncool) 4

8. Hiragana is unnecessary 必要はない/hitsuyō wa nai (not necessary) 3

見たことがないから/mita koto ga nai kara (never seen 9. Unfamiliar, never seen 3 before)

All English would be 全て英語にした方がいい/ subete eigo ni shita hō ga ī (better 10. 2 better to be completely in English)

The second most frequent category drew attention to the mixing of a hiragana (loan)word in an English sentence that was the problem (13), as one respondent explained: ‘Using rōmaji

120 ふらいとひらがなで書くと飛ぶ、ハエ、あげもののように色々な意味で捕らえられてしまうのでは。 121 This example was also coded at ‘hard to understand’. 122 This example was also coded at ‘weird’. 193

(and hiragana) together feels awkward.’123 The mixing of scripts within a single sentence may also be the cause of the ‘weirdness’ stated by nine respondents.

The mixing of scripts also seemed to prompt commentary of the learning and speaking of English in Japan, which were not mentioned in relation to any of the other texts in the survey. One respondent noted ‘I think this is selling the Japanese’ inability to speak English’,124 which suggests a slight criticism of both the designer of the sticker, and the language abilities of the Japanese in general. This idea of English proficiency was taken up by another respondent who labelled the sticker ‘inappropriate’, explaining ‘If hiragana is used to spell out English words, Japanese people will only ever be able to speak English with a Japanese accent.’125 Until recently, katakana was widely used in English classes in Japan to record the pronunciation of English words, and is still used by some teachers today. Despite being easier to read for students unfamiliar with the Roman alphabet, this practice draws criticism for resulting in speakers of ‘katakana English’, felt to be less intelligible to speakers of other varieties of English (Olah, 2007). It is possible that the respondent is objecting to the use of hiragana for loanwords on similar grounds. This concern on the basis of language education may also be related to the large number of respondents who believed this product was designed ‘for children’, in answer to the question about the designers motivation. The topic of pronunciation and katakana English was also brought up in the interviews (6.2.8) and these findings are brought together in the discussion (section 7.3.2).

Like the previous examples, the respondents’ answers to why they thought the designer wrote this word in hiragana were analysed separately to see if there were any differences. More differences did occur than in other examples, and these are described in the following sections.

The 146 respondents’ answers were coded, and generated a total of 223 tokens in answer to the designer’s perceived motivation in using hiragana for the word ‘fly’. The most frequently occurring codes appear in table 5.18 below.

123 ローマ字と混用されているから、拙く感じるから 124 日本人の英語のできなさを売りにしているように思える。 125英語の発音をひらがなで当て字のように当てはめてしまうと、日本人は日本語訛りの英語しか話せなくなる 194

Table 5.18: Respondents’ perceived reasons for the use of hiragana for 'fly'

No. of Code Example tokens

1. Cute 可愛さ/kawaisa (cuteness) 59

このキャラクターには、ひらがな表記が良い 2. Suits character /kono kyarakutā ni wa, hiragana hyōki ga ī (for this 18 character, hiragana would be better)

子供にもわかりやすい/ kodomo ni mo wakariyasui 3. For children 17 (easy to understand even for children)126

4. ‘Yurui’ ゆるい印象/yurui inshō (loose impression) 14

Easy to 日本人に伝わりやすい/nihonjin ni tsutawariyasui 5. 14 read/understand127 (easily understood by Japanese people)

Impressionistic, 個性的に見せるため/koseiteki ni miseru tame (to 6. 14 individual show individuality)

7. Soft やわらかく/yarawakaku (soft) 13

8. Do not know わからない/wakaranai (don’t know) 11

茶目っ気を出す/chamekke o dasu (exhibit 9. Humour/playful 8 playfulness)

10. Childlike/childish 幼稚さ/yōchisa (childish) 8

Overall, the most frequent occurring perceived motivation for the use of hiragana in this text related to ‘cuteness’, and this was the case for both those who selected ‘appropriate’ and those who selected ‘inappropriate’. While the respondents who thought the use of hiragana appropriate cited design motivations as the next most frequent code (12), for those who thought it inappropriate, responses such as ‘わかりません/wakarimasen’ (I don’t know) was

126 Also coded at ‘easy to understand’. 127 Breakdown for total: Easy to read (2), easy to understand (2), readable (6), understandable (4). 195

the next most common code.128 All 11 of the responses for this code came from respondents who selected ‘inappropriate’. This response was interesting in its relatively large numbers since skipping the question was possible, and indeed many respondents did skip one or more questions in the survey. A hint of exasperation was also perceptible in some of these responses, such as ‘I can’t imagine’129, ‘Absolutely no idea’130, and ‘I feel there’s no points in favour of doing this.’131

Both groups frequently mentioned that this product seemed to be aimed at children, many noting that hiragana was easier to read for them, ‘they wanted to make a product aimed at children, and they couldn’t read the word “fly” (in English).’ 132 Eight respondents also mentioned that the use of hiragana made the character itself seem childish, and therefore cute: ‘Writing hiragana shows the character’s low mental age, and slight silliness, which gives it a cute impression.’133

Related to this idea of immaturity or silliness is the adjective ‘yurui’ (緩い or ゆるい), mentioned by some respondents in their reasons of appropriateness, as the word encompasses a variety of meanings including soft, slow, slack, and helpless, as described previously. This word was mentioned 14 times in responses to the questions for the designer’s reasons, making it the fourth highest code overall.

Another code related to silliness was brought up by seven respondents who noted the fact that the character was not in fact flying, as described in case study 3, section 4.2.3. As one respondent explained ‘Because (the hamster) really can’t fly, the mismatched feeling of Japanese and English is (actually) good.’134 Another respondent explained that the hamster’s inability to fly suited its image as a ‘yuru-kyara’: ‘Because the character is a so-called “yuru- kyara”, if it were to say “I can fly” (in English) it would seem the hamster could actually fly, and would be a more serious character. So I think they used hiragana on purpose to give a

128 This response was also found in smaller numbers for 珈琲/kōhī (coffee) [6] and hankachi (handkerchief) [8], but not for the other loanwords in hiragana. 129 理解できない 130全くわかりません 131 わからない。fly を「ふらい」と書く利点がない気がする。 132 fly がわからない子どもたちに向けた作品にしたかったから 133 ひらがなで書くことで、キャラクターの精神年齢が低いこと、少しバカなところを表 1 現し、“かわいい“という印象を見る側に与えたかったから 134 いかにも飛べなさそうで、英語と日本語のミスマッチ感が良い 196

slightly helpless, slack feeling.’135 This was also mentioned by some interview participants, allowing findings from the case study (layer 2), survey (layer 3) and interviews to be triangulated between layers. The joking or unreal function of loanwords in hiragana is explored in the discussion (section 7.3.2).

5.3.5 ‘Dairy Coffee’ instant coffee

This text featured an example of a loanword in kanji, ‘珈琲 /kōhī’ (coffee) rather than katakana. The overwhelming majority found it acceptable as only 2% labelled it inappropriate. The number of respondents for each question for the section on ‘Dairy Coffee’ can be found in the following table (5.19).

Table 5.19: Response rates for 'Dairy Coffee' (Q22-24)

Coffee Appropriate Inappropriate Total

Q22. Appropriate/inappropriate? 139 3 142

Q23. Why? 128 3 131

Q24. Why do you think the company used kanji? 127 6 133

The reasons that respondents gave for their choice of ‘appropriate’ were categorized into one of 21 codes. As in previous examples, many respondents gave more than one reason in their answer, and the data from the 139 respondents generated 169 tokens. The most frequently given codes for the reasons for labelling the use of kanji for the word ‘coffee’ ‘appropriate’ appear in table 5.20 below.

The most commonly given reasons for the appropriacy of writing ‘coffee’ in kanji were that it was widely used (19), and had been used for a long time (19). As one respondent explained, ‘It’s been used since long ago, and you often see it.’136 Interestingly, these responses were more frequent than more direct claims to correctness, with statements that there is a kanji for ‘coffee’ ranking fifth with 14 responses, and that that writing the word in kanji is ‘correct’

135 キャラクターがいわゆる「ゆるキャラ」なので、I can fly とすると本当に飛ぶ、またはしっかりしたキャラクターになってしまうため、少し頼りないゆるさを出すためにわざとひらがなにしていると思う 136 昔から使われていて、よく目にする。 197

ranking 13th with five responses (and therefore not featured in table 5.20). A further four respondents made references to authority in their reasons by noting that ‘coffee’ in kanji appeared in the dictionary (2), the Kanji Aptitude Test (1), and in some novels (1). The balance of reasons based on language use versus those based on rules or authority reinforces a social semiotic notion of meaning making, a descriptivist rather than prescriptivist stance on script use.

Table 5.20: Reasons for appropriacy of 'coffee' in kanji

No. of Code Example tokens

見慣れている/minareru (used to seeing [it written this 1. Widely used, often seen 19 way])

古くから使われている表現/ mukashi kara 2. Used for long time 19 tsukawareteiru hyōgen (an expression used from long ago)

3. Not wrong 問題は無い/mondai wa nai (it’s not a problem) 16

珈琲という漢字が存在する/kōhī to iu kanji ga sonzai 4. The kanji exists 14 suru (the kanji for ‘coffee’ exists)

味が深そうな印象/aji ga fukasōna inshō (the impression 5. Flavour 11 that the taste is richer)

6. High class 高級感/kōkyūkan (a classy feeling) 10

もはや外来語ではなく日本語として扱われている ‘Coffee’ is (becoming ) /mohaya gairaigo de wa naku nihongo toshite 7. 10 Japanese atsukawareteiru (it’s no longer a loanword but is treated as a Japanese word)

Freedom of the 8. 個人の自由/kojin no jiyū (freedom of the individual) 9 individual

喫茶店をイメージしやすい/kissaten o imēji shiyasui 9. Café image 9 (it’s easy to imagine a coffee shop)

コーヒーと読める /kōhī to yomeru (it can be read as 10. Readable 7 ‘coffee’)

198

While ‘not wrong’ was the code with the third-most responses (16), it is interesting that this answer appeared with similar frequency in the examples of arguably more marked loanwords in hiragana, such as ‘はぴねす/hapinesu’ (happiness), ‘れもん/remon’ (lemon), and ‘ふらい /furai’ (fly).

Other respondents suggested that using kanji gave a particular image to the coffee, such as it being a high class product (10), or that the use of kanji added a certain taste to the coffee (11), such as it being delicious (4), rich (4) and mellow (2).137

Ten respondents also expressed a belief that ‘coffee’ was in the process of becoming a Japanese word, or had already become part of the language. As one respondent explained, ‘the kanji version of “coffee” has already completely permeated the Japanese language’,138 while another used the word ‘Japanized’139 citing the fact it was seen in many places and had been used for a long time. The notion of ‘Japanese-ness’ will be further explored in the findings on the perceived motivations for writing this word in kanji.

As explained above, only three participants labelled the use of kanji for ‘珈琲/kōhī’ (coffee) ‘inappropriate’. Two respondents reported that the kanji version of coffee did not suit the text, while the other said they were not used to seeing the word written this way. Because of the small number, all responses are given in the following table (5.21).

Table 5.21: Reasons for inappropriacy of 'coffee' in kanji

No. of Code Example tokens

パッケージのかわいらしさと、漢字があまりあっていない / pakkēji no kawairashisa to kanji ga amari atteinai (kanji doesn’t suit the cuteness of the pack) Does not suit text 2 商品の雰囲気と漢字の固さに違和感がある/shōhin no funiki to kanji no katasa ni iwakan ga aru (the hardness of kanji feels out of place with the image of the product).

Not often seen 見慣れないから/mirarenai kara (I’m not used to seeing it [this way]). 1

137 美味しい / おいしい (delicious), 深い (rich), まろやか (mellow). 138 コーヒーの漢字表記は日本語の中に十分浸透しているため 139 いろんなところで見かけるし、昔からあって日本語化していると思う。 199

The two responses citing the unsuitability of kanji both do so in reference to the text in which it occurs. This can be seen as evidence that these participants are engaging with the texts as multimodal wholes, in other words, they are taking in information from other semiotic resources such as the font, colour and layout of the text, and using this information as the basis of their response. Their responses also suggest that they are considering other script alternatives and their associations in their answers; the first respondent would seem to prefer a script with a ‘cuter’ image, while the second something ‘less hard’.

While the fact that the kanji version of ‘coffee’ was cited as being widely used and often seen by a large number of respondents who felt this example as ‘appropriate’, the respondent who reported they were not used to seeing it this way reminds us of two important points: firstly, while the use of kanji for coffee is not uncommon, it still does not mean that the word appears this way as frequently as the unmarked katakana version, or rōmaji/English. The characters for ‘coffee’ are not part of the ‘常用漢字/ jōyō kanji’ (prescribed characters learned in school), and therefore this word would appear in katakana, not kanji, in newspapers and texts produced by the government. Secondly, whether a person is ‘used’ to seeing something or not depends on what kinds of texts they are exposed to, which will in turn be affected by factors such as age, locality (urban/rural), and socio-economic status.

As in previous examples, the perceived motivations of the designer in using kanji were also coded. Due to the small numbers of people selecting ‘inappropriate’, it was difficult to comment on any significant differences between the groups. The results of both groups are presented together in the following table (5.22), where the 142 respondents generated 189 tokens.

Respondents noted a feeling of a classic, or old style (31) or a ‘high class’ image (26) being a motivating factor for kanji being chosen for the word ‘coffee’. ‘I felt they wanted to give it a bit of an old-fashioned image, and appeal to a sense of traditional flavour’,140 one respondent explained. This also related to the code ‘authentic’ (8), and ‘café-style coffee’(9). In Japanese there are two commonly used words for ‘café’, the loanword ‘カフェ/kafe’ and the sino- Japanese ‘喫茶店/kissaten’. A ‘kissaten’ has more of an old-world feeling, while a ‘kafe’, like many loanword equivalents, has more modern associations. Of the nine statements coded as ‘café (style coffee)’, four used the word ‘kissaten’ with one using the word ‘real’ in

140 ほんの少しだけ古風な印象を与えて、昔ながらの味わいをアピールしたかったのでは、と感じた。 200

conjunction with it to emphasize the sense of tradition and old-world charm. While three respondents used the loanword ‘kafe’, they were all modified by terms that also suggested this traditional image, such as ‘old-style kafe’(1) and ‘long standing kafe’141(2). Many of the above associations relate to the associations of kanji with ‘tradition’ reported in the first part of the survey.

Table 5.22: Respondents’ perceived reasons for the use of kanji for 'coffee'

Code Example No. of tokens

昔ながらの印象/mukashinagara no inshō (an 1. Classic, old 31 impression of being unchanged from the olden days)

2. High class 高級感を出す/kōkyūkan o dasu (to make it ‘high class’) 26

3. Differentiate from others 差別化のため/sabetsuka no tame (to be different) 20

日本っぽさを出したかった/ nihonpossa o 4. Japan 14 dashitakatta (to bring out the Japanese-ness)

美味しそうに見えるから / oishisō ni mieru kara (to 5. Flavour 13 make it look delicious)

こだわりがある印象/ kodawari ga aru inshō (an 6. Made with care 11 impression that it was carefully made)

純喫茶店のコーヒーみたい/ junkissaten no kōhī mitai 7. Café-(style coffee) 9 (like the coffee from a real café)

8. Authentic 本格的な感じ/honkakuteki na kanji (authentic feeling) 8

縦読みだと、コーヒーの「ー」が冗長に見える/ tateyomi da to, kōhī no ‘ー’ ga jyōchō ni mieru (Being 9. Layout 6 written vertically, the ‘ー’ in ‘coffee’ [in katakana] looks verbose)

10. Adult 大人っぽさ/otonaposa (adult-like) 6

141昔ながらのカフェ and 老舗のカフェ respectively. The two remaining terms were 珈琲店 (coffeehouse) and 個人経営のカフェ(independent café). 201

The code ‘Japan’ (14) was made up of three broad types of statements about Japan. The largest (8) reflected a general feeling of Japanese-ness, characterised by words such as ‘和風 /wafū’ (Japanese-style) and ‘日本的な/nippontekina’ (Japan-like). A further four responded that the use of kanji suggested the product was made in Japan, and two more said the coffee seemed to suit Japanese people.142 The reasons coded as ‘Japan’ echo kanji’s connotations of Japanese-ness reported in small numbers (10) in the first section of the survey. While kanji was more frequently associated with China (48) than Japan, only one respondent mentioned this in their answers to this part of the survey, suggesting that the use of kanji was ‘to appeal to a sense of Chinese-ness’.143

While ‘design’ had been given as a suggested motivation for a particular choice of script in previous questions in the survey, for this item the design considerations were made more explicit, and were grouped under the code ‘layout’ (6). Three respondents noted that writing ‘coffee’ in kanji only used two characters, while using katakana would require four. Looking at the layout of the pack, it is clear that if the word appeared in katakana, the font for this word would need to be much smaller; in kanji, the word ‘coffee’ appears in the largest font on the pack, and is significantly larger than the preceding ‘Dairy...’ (牛乳屋さんの/gyūnyūya san no). Besides the space that the word would take up, two participants noted that the bars used to indicate long vowels in katakana loanwords (of which ‘coffee’ has two) look strange when they appear in vertical writing. As one respondent explained, ‘when there are bars in vertical writing, I feel the balance is bad.’144 These bars are one of the few textual features (along with words in alphabetic characters) which necessitate a change of direction when written vertically, from ‘ー’ to ‘|’, which may account for the sense of imbalance. Another participant simply gave ‘vertical writing’ as the motivation for the use of kanji.145

Interestingly, while vertical writing was not mentioned as an association of kanji by any participant in the survey, horizontal writing was mentioned by four respondents in relation to katakana. If katakana is associated with horizontal writing, it may explain both the reference to vertical writing in this case of a loanword in kanji, as well as the strangeness associated with bars appearing in this style of writing. The above comments on layout, in terms of both

142For example, 国産であるイメージ and 日本人向けのコーヒー. 143中国系をアピールしてるため。 144縦書きで伸ばし棒が入るとバランスが悪いと感じたのでは。 145文字を縦書きにしているから。 202

the number and direction of characters, show an awareness of the orthographic possibilities of the Japanese writing system, and their connection to the multimodal whole.

5.3.6 ‘Cute Hankachi’ Handkerchief

This question contained an example of a loanword in rōmaji, or the Roman alphabet. As explained in the research design (section 3.4.3), some loanwords in rōmaji are rendered the same as their source word (for example, トマト/tomato), however words such as ‘handkerchief’ are spelled quite differently in their rōmaji loanword form, ‘hankachi’. The number of people who responded to this question are displayed in the following table (5.23).

Table 5.23: Response rates for 'Cute Hankachi' handkerchief (Q25-27)

Hankachi Appropriate Inappropriate Total

Q25. Appropriate/inappropriate? 90 52 142

Q26. Why? 82 51 133

Q27. Why do you think the designer used rōmaji? 84 50 134

Writing loanwords in rōmaji is unusual, but this script is often used for foreign words with their original spelling, for example in this case, ‘handkerchief’. However, 63% of respondents (90) selected ‘appropriate’ for the use of rōmaji in this text. The reasons given were coded, and the number of tokens totalled 103. The ten most commonly occurring codes are presented in the following table (5.24).

The most frequently given reason for the appropriateness of this choice of script was the perception that the product was aimed at non-Japanese people such as overseas visitors to Japan (15). ‘I think this product was made to target people from foreign countries. Compared to writing in hiragana or katakana, for foreign people who can’t read Japanese, writing in rōmaji is more effective’146 explained one respondent. The belief that this product was aimed at non-Japanese people was also the most frequently given reason as the motivation of the designer for all respondents (55). However, reflections on the effectiveness of this strategy

146外国人を対象にこの商品は作られてるように思います。ひらがなや、片仮名で表記するよりかは、日本語の読めない外国人がいた場合に、ローマ字の使用が効果的かと思います。 203

were also given by respondents who selected ‘inappropriate’, discussed in the following section.

Table 5.24: Reasons for appropriacy of 'hankachi' in rōmaji

No. of Code Example tokens

外国人向けの商品だから/gaikokujin muke no 1. For non-Japanese people shōhin dakara (because it’s a product aimed at foreign 15 people)

2. Not wrong ダメではない/ dame de wa nai (it’s not wrong) 12

何を使おうが自由/ nani o tsukaō ga jiyū ([people 3. Freedom of the individual 10 are] free to use what they want)

意味は通じるから/imi wa tōjirukara (the meaning is 4. Understandable, sounds same 9 transmitted)

商品を判別出来る/shōhin o hanbetsu dekiru (it can 5. Distinguishes the product 5 distinguish the product)

ちょっとおしゃれにしたかった/ chotto oshare ni 6. Fashionable shitakatta (they wanted to make it a little more 5 fashionable)

日本的要素を伝える/nippontekiyōso o tsutaeru 7. Japan-like feeling 5 (transmit a Japanese factor)

ローマ字をデザインとして使っていて/rōmaji o 8. Design dezain toshite tsukatteite (using rōmaji as part of the 4 design)

誰がみてもハンカチだとわかるので/dare ga mite 9. Type of product is clear mo hankachi da to wakaru node (because anyone who 4 looks at it can see it’s a handkerchief)

10. Product name, noun 名詞だから/ meishi dakara (because it’s a noun) 4

204

The second and third most frequently occurring codes were also seen for other examples in the survey: firstly the opinion that it was not incorrect to write loanwords in rōmaji (12), and secondly the related code ‘freedom of the individual’, where respondents stated that people were free to choose whichever script they thought best (10).

Five respondents also gave a ‘Japanese feeling’ in their reasons for their choice of ‘appropriate’. This is interesting to compare with similar responses in the examples of loanwords in hiragana in the survey. While the use of hiragana, a script associated with Japanese-ness (18), could give a Japanese feeling to words like ‘れもん/remon’ (lemon) (11) ‘はぴねす/hapinesu’ (happiness) (5), and even ‘ふらい/furai (fly) (2), in this case the word and its pronunciation (hankachi) was felt to give a ‘Japanese feeling’ to rōmaji, a script which is commonly associated with foreign words and languages (75). One respondent specified an intended impression of ‘Japanese, but still modern’, 147 which suggests the existence of different kinds of ‘Japanese-ness’, for example, traditional and modern, as well as the feeling that ‘tradition’ is the default option, as evidenced by the connecting particle ‘が/ga (but)’. The role of pronunciation is interesting to compare with the comments relating to the text containing ‘ ふらい /furai’ (fly), and the broader discussion of English and Japanese pronunciation described in the discussion (section 7.3.2)

The remaining 52 respondents selected ‘inappropriate’ for the use of rōmaji to write the word ‘hankachi’. These responses were coded, and the reasons they gave amounted to 81 tokens. The ten most frequently given reasons given for this answer appear in the following table (5.25).

For the respondents that selected ‘inappropriate’, the most frequent reason given was that it was hard to read/understand (18), as one respondent explained ‘to be honest, in rōmaji it’s hard to read. It’s think it’s just because I’m Japanese, and am used to the Japanese language, but it just doesn’t cross over to being easy to understand.’148 This response suggests a sense of expectation that rōmaji would be easy to read, 149 and is explained by the respondents familiarity with Japanese, and perhaps also seeing ‘hankachi’ in Japanese (katakana). The number of people whose answers were coded as ‘hard to read’ (18) is twice the number of respondents who selected ‘appropriate’ giving the reason that it was understandable (9).

147和だがモダンな印象をあたえるため 148正直ローマ字だとわかりにくいです。これは私自身日本人であり日本語に慣れ親しんでいるからこそだと思いますが、わかりやすいに越したことはありません。 149 This was not, however, stated in their response to the questions of the associations of rōmaji. 205

Table 5.25: Reasons for inappropriacy of 'hankachi' in rōmaji

No. of Code Example tokens

1. Hard to read/understand わかりづらい/wakarizurai (hard to understand) 18

ハンカチは英語では handkerchief である/hankachi wa 2. Spelling is incorrect eigo de wa handkerchief de aru (‘hankachi’ in English is 14 ‘handkerchief’)

3. Unnecessary, pointless 必要が無い/ hitsuyō ga nai (not necessary) 9

外国人が見たときハンカチだと認識できない/ gaikokujin ga mita toki hankachi da to ninshiki dekinai 4. Unclear for foreigners 7 (when a foreign person looks at it, they won’t realize [it says] ‘handkerchief’)

Not English, not 英語でも日本語でもないから/ eigo demo nihongo demo 5. 6 Japanese nai kara (it’s not English and it’s not Japanese)

6. Weird 変 /hen (weird) 6

教養の低さを感じてしまう/ kyōyō no hikusa o 7. Negative image 5 kanjiteshimau (it ends up having an uneducated feeling)

見慣れないから / minarenai kara (not used to seeing it 8. Unusual, rare 3 [written like that])

正しい表記とは思わないから / tadashī hyōki to wa Katakana should be used 9. omowanaikara (I don’t believe that’s the correct way of 2 for loanwords writing it)

むしろはんかちと平仮名で書いて、その上に Kawaii とローマ字にした方が良かったかも / mushiro hankachi to Hiragana as preferred 10. hiragana de kaite, sono ue ni kawaii to rōmaji ni shita hō ga 2 alternative orthography yokatta kamo (it might have been better to have ‘hankachi’ in hiragana, and ‘kawaii’ in rōmaji instead)

The second most frequent response was that the spelling was incorrect (14), this may also be related to the fourth category ‘unclear for foreigners’ (7). This is interesting in light of the

206

number of respondents who thought that the motivation for the orthography was in fact to appeal to a foreign market. While the move from katakana to the roman script, one more widely understood by non-Japanese people, may have had intelligibility as its aim, this has in fact been lost or at least mitigated by the perceived difficulty of recovering the referent ‘handkerchief’ from the loanword ‘hankachi’. One respondent pointed out that ‘hankachi’ was not a particularly familiar (Japanese) word for foreigners, ‘“hankachi” isn’t a word that’s generally used in English like “sushi” is…’150

The claim that ‘hankachi’ was not Japanese, and yet not English (6) also relates to the unusual practice of putting a loanword in rōmaji script. One respondent specifically commented on the translingual nature of this case: ‘Because ‘handkerchief’ is a word that’s come from overseas, I don’t know why they’ve given it a Japanese reading and then purposely put in back into roman script. It’s puzzling.’151

Although in much smaller numbers, other participants commented on alternative orthographic options, such as katakana (2) and even hiragana (2). The suggestion of hiragana as a potential option may have been a task effect; this being the final sample text in the survey, respondents had already been presented with three examples of loanwords in hiragana, and were perhaps primed to see this as an alternative. ‘Hankachi’ was, however one of the more frequently occurring words in the corpus of hiragana loanwords (see section 4.1.3), and so these comments might also be related to being exposed to this word in hiragana previously.

The final question for this text asked for respondents’ views on the perceived motivation for the designer in writing this loanword in rōmaji. Like the previous examples, the results for those selecting ‘appropriate’ or ‘inappropriate’ were calculated separately, but no major differences were found. The results are presented in the following table (5.26).

The most common perceived reason for the use of rōmaji for the word ‘hankachi’ was that it was designed to appeal to non-Japanese people (55). This was reported by both respondents who thought the use appropriate and inappropriate, despite variation in the perceived success of this strategy.

150 ハンカチは Sushi などのように一般的に使われる英語ではないから… 151 ハンカチは外国から来たこ言葉なのに、それを日本語読みしてそれをわざわざまたローマ字読みする意味がわからない。ややこしい。 207

Table 5.26: Respondents’ perceived reasons for the use of rōmaji for 'hankachi'

No. of Code Example tokens For non-Japanese 1. 外人向け/gaijin muke (for foreigners) 55 people

お洒落さを出すため/osharesa o dasu tame (to be 2. Cool 17 fashionable)

日本らしさをコンセプトに/ nihonrashisa o konseputo 3. Japanese 11 ni (For the concept of ‘Japanese-ness’…)

表記にインパクトをもたせるため / hyōki ni inpakuto o 4. Eye catching 11 motaseru tame (to give the phrase more impact)

デザインを優先して / dezain o yūsen shite (giving 5. Design 11 priority to the design…)

かわいらしさを伝えよう/kawairashisa o tsutaeyō (to 6. Cute 9 express cuteness)

モダンな印象にするため / modanna inshō ni suru tame 7. Modern 8 (to give a modern impression)

8. Do not know わかりません/wakarimasen/ I don’t know 8

「日本のハンカチ」であることを強調するため/ Specify a Japanese 9. ‘nihon no hankachi’ de aru koto o kyōchō suru tame (to 6 handkerchief emphasise that it’s a ‘Japanese handkerchief’)

10. Hybridity 日本製の英語/nihonsei no eigo (English made in Japan) 5

Other reasons were given in much smaller numbers, for example that it made the product seem cool (17), or added connotations of Japanese-ness (11), or was simply designed to be eye catching (11). Those who indicated it was an aspect of the ‘design’ (11) may also be referring to one of these aims.

The idea that rōmaji could give a ‘modern’ impression (8) is interesting, since it was not given by respondents in the script association section (section 5.2.4), but may be an extension 208

of the ‘foreign’ or ‘international’ associations of rōmaji. A ‘modern’ feeling was, however, one of the connotations of rōmaji noted by Inoue (2005), see section 2.4.2.

5.4 Summary of layer 3: survey responses to scripts and sample texts

The results of the first section of the survey both reinforced commonly cited associations of the scripts in use in Japan, as well as revealed the complex and contrasting associations a script can have even for a single respondent. The reported connotations of the Japanese scripts highlight their diverse origins, flexible uses and long histories, and provide a crucial backdrop for interpreting opinions on the use of hiragana for loanwords in Japanese.

The scripts and their associations are, however, just one element of the multimodal social semiotic understanding of meaning described in chapter 2. Therefore, the meanings of particular loanword in hiragana cannot be understood when devoid of their context, as the case studies have demonstrated. The contextualized meanings of and motivations for loanwords in hiragana were therefore explored in the following section of the survey, which sought responses to authentic texts.

Overall, the use of hiragana for the loanwords in the texts presented was deemed appropriate by the majority of survey respondents, with an average of 73% across the three texts. That writing loanwords in hiragana was ‘not wrong’ was a common reason given for all three examples, as was that the freedom of the individual meant that writers were free to use the scripts as they saw fit. Another reason frequently given for the ‘ はぴねす/hapinesu’ (Happiness) clinic text and ‘蜂蜜れもん/hachimitsu remon (Honey Lemon) drink text was that the words in question were names or proper nouns. This was later mentioned by a participant in the interviews, when asked why the text ‘蜂蜜れもん/hachimitsu remon (Honey Lemon) received a higher level of appropriate answers than ‘I can ふらい/furai (fly)’, he stated simply because one was a product name, and one was not.152

For the respondents who selected ‘inappropriate’ (27%) for the examples of loanwords in hiragana, the fact that the word in question was English or a loanword ranked highly in all three examples. ‘はぴねす/hapinesu’ (Happiness) clinic and ‘蜂蜜れもん/hachimitsu remon’

152 The participant was Takuya, (G4). 209

(Honey Lemon) drink were also explicitly labelled as being incorrect by some respondents. A feeling of weirdness or unnaturalness was also not uncommon for the ‘はぴねす/hapinesu’ (Happiness) clinic and ‘ふらい/furai’ (fly) sticker examples; these two were also often said to be hard to read or hard to understand.

The responses to the questions regarding the perceived motivation for using hiragana script for loanwords are collated in the following table (5.27). The numbers in brackets besides the loanwords in each of the texts represent the number of respondents and the total number of reasons given.

Table 5.27: Summary of the perceived reasons for the use of hiragana for loanwords

Happiness (154/257) Lemon (144/236) Fly (146/223)

1. Familiar Made in Japan Cute

2. Soft Japanese-ness Suits character

3. Gentle, kind Eye-catching For children

4. Elderly are target Katakana is unsuitable Yurui

5. Easy to read/understand Old-fashioned, traditional Easy to read/understand

6. Cute Familiar Impressionistic/individual

7. Eye-catching Soft Soft

8. Yasashī (gentle/easy) Balance with other scripts Do not know

9. Healing Sweetness/not sour Humour/playful

10. Memorable Easy to read/understand Childlike/childish

Table 5.27 above highlights the similarities given for each text as well as the differences. There were three themes that were common to all three sample texts: ‘softness’, ‘ease of reading’, and an ‘attention-getting’ quality. ‘Softness’ and ‘ease of reading’ are both established connotations of the hiragana script, and it is therefore interesting that these have been carried over to the texts in which this script has been used. One of the most basic motivations for using hiragana for loanwords is perhaps to invoke some of these associations for words which are normally written in the angular, harsh katakana script. The

210

metafunctions at work here would seem to be different, however, as the feeling of ‘softness’ is ideational, saying something about the state of the referent, while the ‘ease of reading’ may be more interpersonal, for example, in the recognition of a particular imagined readership. This interpretation is supported by references to the imagined readership of the ‘はぴねす /hapinesu’ (Happiness) clinic and ‘ふらい/furai’ (fly) texts also appearing in the above table, being perceived to be for the elderly and for children respectively. Hiragana may be felt to be more suitable in each case for different reasons: for children who may not know katakana or kanji, or for the elderly for whom this script is said to be more familiar.

The ‘attention-getting’ or ‘eye-catching’ quality, also common to all three texts, relates to the fact that the use of hiragana for loanwords is a marked use of this script.153 However, many participants in the interviews commented that they had never noticed loanwords being written in hiragana before (see section 6.1), which at first seems difficult to reconcile with this ‘eye- catching’ aspect of its use. However, it is possible that while they are not immediately noticeable as part of a person’s surroundings, when asked to reflect on the choice of this script over katakana, the script is felt to have an eye-catching or impressionistic quality.

Other interpretations of the use of hiragana were specific to the text in which they were found. For example, as mentioned above, the elderly were often felt to be the target market for the clinic called ‘はぴねす/hapinesu’ (Happiness), and this was felt to be one of the motivations for the choice of hiragana in this example. Approachability also ranked much higher for this text than it did for ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) drink, and it was not a consideration for ‘ふらい/furai’ (fly) at all. The warm, approachable atmosphere of the clinic’s sign is also represented through other semiotic resources in the text, as discussed in case study 1 (section 4.2.1).

The associations of hiragana with Japanese-ness and Japanese culture154 were particularly strong for the example of the ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) drink, and the text contains many other references to Japanese culture, as described in case study 2 (section 4.2.2). The unsuitability of katakana was also specifically referred to for this example.

153 An ‘attention-getting quality’ was also noted to be one of the motivations for the non-standard spellings in the product and business names investigated by Davies (1987). 154 Davies (1987) also notes examples of non-standard spellings in product and business names that were used to connote a different time or place. 211

For the ‘ふらい/furai’ (fly) sticker, the cute and ‘yurui’ image of the character represented some of the most frequently cited reasons for the use of hiragana in this text. Humour and a connection to children were also mentioned for this text only. As above, other indexes of cuteness and humour were found in the text, as discussed in the case study 3, in section 4.2.3.

The perceived reasons for the use of hiragana in a text therefore draw upon both the established connotations of the script, as well as other text-specific reasons. Because of this, an understanding of both the common associations of the scripts, as investigated as part of the survey (section 5.2) as well as a detailed analysis of particular texts in the form of case studies (section 4.2) is necessary to gain a full understanding for how hiragana is being utilized as a semiotic resource within these texts.

212

Chapter 6: Findings from the interviews and focus groups

While the survey sought the opinions of a large number of respondents to a small number of texts, the interviews and focus groups involved a smaller number of persons (34) and obtained their interpretations of a wider range of texts (32). In small groups or individually, participants were asked for their impressions of texts representing the main genres of the survey, such as product names, business names, and social media posts. They were presented with colour photos of these texts, and asked what they believed the motivation of the writer was in using hiragana over katakana (and sometimes kanji), as well as how the nuance would change if the word did appear in an alternate script.

The interviews and focus groups allowed for a deeper understanding of specific participants’ interpretations of loanword use in hiragana, and in the case of the focus groups, also permitted participants to discuss and work through their interpretations with others. Like the survey, the findings were relevant to answering the third research question relating to the functions of loanwords in hiragana, but were also used to triangulate the findings relating to the first two research questions, namely, the typical characteristics of loanwords in hiragana, and the contexts in which they are usually found. This was achieved through general discussion concerning the examples produced in the course of the sessions, as well as participants’ reactions to specific examples.

In total, 15 interviews were conducted across different locations in Japan in October and November 2015. Seven different prefectures were chosen for the interviews in order to obtain views from participants from a variety of locations including urban and semi-urban, as well as ‘関東/kantō’ (eastern) and ‘関西/kansai’ (western) areas. Participants were not excluded from interviews that did not take place in their hometown, however, as the aim was not to compare the various groups with each other. When the participants’ hometowns were included, 16 different prefectures were represented.

As explained in the methodology chapter (section 3.4.4), one-to-one interviews as well as focus groups were conducted. In the following paragraphs, interviewees and focus group participants are referred to by their pseudonym and interview type, for example ‘G6’

213

represents focus group 6, and ‘S4’ represents (single) interview 4, as listed in tables 3.13 and 3.14 in chapter 3, ‘Research design and methodology’.

Many of the responses of the interviewees and participants reconfirmed and supported the findings of the online survey, for example that hiragana was used to add connotations such as Japanese-ness (Ayano G6, Kazu G1, Nagisa G3, Miho G3, Ken G3, Ryūichi S5, Yūta G4, Kei G4), or softness (Ryūichi S5, Mio G8, Hiro G1, Kazu G1, Ryō G2, Takuya G4) to a product or business, or the that the ease of reading this script (Natsumi G1, Daisuke G2, Yūta G4) meant that products or businesses were likely to be aimed at young children (Ayano G6, Yūsuke S6, Nagisa G3, Hiromasa G8, Yuka G1, Yui G2, Natsuki G2, Gensuke G4), or the elderly (Satomi G6, Miho G3, Hiromasa G8, Natsumi G1, Ryō G2, Yūta G4). Cuteness was another common association mentioned (Nagisa G3, Mio G8, Yūsuke S6, Yūta G4, Takuya G4), as well as the use of hiragana for loanwords having an eye-catching quality (Yūta G4, Miho G3, Gensuke G4). Takuya (G4) summed up several of these associations in a comment made towards the end of his interview, explaining that writing loanwords in hiragana can have a number of effects: making them look like Japanese words, or seem childlike, or soft, for example.155

For this reason, the following findings will focus on aspects of loanwords in hiragana not addressed by the survey, as well as differences of opinion that emerged within and between interviewees and focus group participants. This will allow these data sources to expand on the findings of the survey, and provide an in-depth understanding of how loanwords in hiragana are perceived by Japanese people. The following sections illustrate key findings of the interviews.

6.1 Participants’ general impressions

Some interviewees, such as Kazu (G1), were unsure of precisely what was meant by ‘loanwords in hiragana’ when the topic of the interview was initially explained, and the researcher asked if they had seen any such examples. This suggests that the script in which loanwords are written may not be especially salient for native Japanese speakers, with participants such as Yūta (G4) admitting he had honestly never noticed loanwords in hiragana

155 ひらがな書いてると、日本語っぽくなると言う面もあるし、ちょっと幼稚見える面もあるし、また何か柔らかさとかそう言うの引き出すところもあるし 214

before,156 and Natsuki (G2) and admitting she had never thought about this topic before that time.157

However, the overall response produced by most of the sample texts produced in the course of the interviews and focus groups was not one of surprise or disbelief as the texts containing product names, business names, social media examples, and even the hostess bar signs were produced, Kei (G4) noting in relation to the word ‘すなっく/sunakku’ (snack) for example, that he often sees it written like this.158 This suggests that the genres of text from which the samples for discussion in the focus groups and interviews were taken were representative of typical examples of loanwords in hiragana which people may be exposed to in their daily life. The few texts that did generate surprise and disbelief are discussed in section 6.5 below.

Some participants could suggest loanwords they had seen written in hiragana when asked at the beginning of the interview, for example Naoto (G4) suggested the comedy duo ‘くりーむしちゅー/kurīmu shichū’ (cream[y] stew),159 whose name is consistently written in hiragana; while Takuya (G4) suggested animal names might be common, for example ‘らいおん/raion’

(lion). Yui (G2) reported having seen the word ‘ちゅーりっぷ/chūrippu’160 (tulip) although she could not remember where, and Mio (G8) felt she had seen ‘かふぇおれ/kafe ore’(café au lait) in hiragana on a menu. Examples were also supplied in the course of the discussion, often when similar texts were produced. For example, after seeing the picture of the bottle of ‘ゆずれもん/yuzu remon’ (Yuzu Lemon) drink, Yūsuke (S6) remembered having seen the word ‘おれんじ/orenji’ (orange) in hiragana used as the name of a drink. At the very end of his focus group, Takuya reported feeling that he had seen the word ‘coffee’ in hiragana on coffee shop (喫茶店/kissaten) menus, specifically not ‘coffee’ with a bar (こーひー/kōhī), but with a ‘ふぃ/fi’ for the final vowel sound.161 One such example was found in the corpus, ‘こおふぃ/kōfi’ (coffee), although not in the location where Takuya’s focus group took place.

156正直に言うと気にしたことないんだよね 157 考えたことない 158 この「すなっく」とかめちゃ見ますもんね。 159 Another comedy team with a name featuring a loanword in hiragana, ‘さまぁーず/samāzu’ (Summers) also featured in the corpus. 160 It is not known how the vowels were elongated in the examples given by participants, but bars were more common (see section 4.1.1) in the layer 1 corpus. 161 「コーヒー」もひらがなで棒じゃなくて何か「ふぃ」、そういう表現の商品たまに喫茶店あるような気がする。 215

When shown the example of はぴば/hapiba (happy bi[rthday]) in hiragana, Mio (G8) remembered writing the complete phrase ‘Happy birthday to you’ in hiragana, as well as ‘happy valentine’, when she was younger, saying it was just something that teenage girls did. Yuka (G1) also reported having written ‘ここあ/kokoa’ (cocoa) in hiragana in a Facebook post, explaining that she felt it gave the word a cute, sweet, warm feeling. Satomi (G6) remembered having written ‘あいす食べたい’ (I wanna eat ice-cream) to a friend via social media, explaining that she sometimes used hiragana for loanwords to show she was joking (じゃれてる/jareteru) or messing around (ふざけてる/fuzaketeru). Ayano (G6) noted that the loanword ‘ファイト/faito’ (fight), used in Japanese to mean ‘go for it’ or ‘do your best’, even appeared in hiragana in the stamps available on the social networking platform ‘Line’ (see figure 6.1 below), and explained that it had a different feeling to the regular katakana version, being less strong or intense.

Figure 6.1: Social networking 'sticker’, ‘Fight!!’

6.2 Motivations for the use of hiragana for loanwords

As explained above, many of the perceived motivations for the use of hiragana for loanwords echoed the findings of the script association section of the survey (see section 5.2.2). For example, the bath toy ‘スプラッシュあいらんど/supurasshu airando’ (Splash Island) was felt to utilize this script as it was a product made for children by Yuka (G1), Natsuki (G2), and Ryūichi (S5), while hiragana gave a gentle feeling to the aged care facility ‘らいふアシスト馬車道/raifu ashisuto bashamichi’ (Life Assist Bashamichi) as reported by Natsumi (G1), and Ryō (G2). Ease of reading was also mentioned in relation to texts which appeared to be aimed at children or the elderly, and was also raised in a joking manner when the hostess bar texts were produced, which Takuya (G4) and Chisato (G6) noted might be designed to be easy to read for their drunken customers.

216

The following aspects emerged as alternative motivations of loanwords in hiragana, or explored the themes generated in the survey in more detail.

6.2.1 Unfamiliarity with English spelling

Nine participants suggested that the reason for the use of hiragana was that the writer did not know how to spell the English word, for example this was Yuka’s (G1) initial reaction in relation to the café called ‘れふぁいん/refain’ (Refine). This was particularly common when the English spelling of a word was perceived to be difficult, and was mentioned in relation to the word ‘ばれーぼーる/barēbōru’ (volleyball),162 which appeared on a print club image, by Chisato (G6) and Jun (G3), the disjunct between the English and Japanese rendering of the first syllables of the word being noted by Jun in particular. Natsuki (G2) made a similar argument in the case of the word ‘ふらい/furai’ (fly), which appears in the sticker featured in the survey, suggesting the designer may not know the spelling because the Japanese transliteration (furai) is ambiguous in referring to both /l/and /r/ sounds in English. In each case, a fellow member pointed out that they could still write the word in katakana, which was usually acknowledged by the initial speaker, and therefore eliminated a potential reason for the use of hiragana for loanwords in these focus groups.

In two discussions of spelling in relation to the print club texts, however, the initial speaker persevered with the idea that the problem of spelling was the initial motivating factor, suggesting the possible thought process of the writer which led to the loanword in hiragana. Hiromasa (G8) suggested that the writers first intended to write something in English, but not knowing the spelling considered writing the word in katakana. But this, he reasoned, would reveal the fact they did not know the spelling, so by using hiragana, it made it seem that it was on purpose. Similarly, Ryō’s (G2) line of thought was that English would have been the first choice, but without knowing how the spell the word(s), they were faced with the option of katakana, which would have seemed a bit stupid, so they used hiragana, which gave an additional connotation of cuteness. These explanations suggest an awareness of the orthographic options available to a writer, and their respective nuances and connotations.

162 In Japanese, /v/ and /l/ sounds do not have direct equivalents, and are rendered /b/ and /r/ respectively. 217

The frequency with which considerations of English spelling in the Roman alphabet came up as the initial explanation for loanwords in hiragana is of interest, as it suggests that writing words and phrases in English is another semiotic resource available to Japanese. This also relates to the comments from the survey in relation to the clinic called ‘はぴねす/hapinesu (Happiness) included in the survey (layer 2), for example, which stated that the original English form would have been hard to understand for some people (see section 5.2.3). English as a further option for loanwords is further explored in the discussion section 7.2.

6.2.2 Accidental use of hiragana for loanwords

While the current research concerns the meaning associated with loanwords in hiragana, the unmotivated or accidental use of this script was also mentioned by some participants, and represents two broad scenarios.

The first reported case of accidental use of hiragana for loanwords would commonly be called a typographical miss or ‘typo’, and results from the way Japanese is input via a keyboard (see section 1.1.1). In G4, Takuya, Naoto, Yūta all had the experience of accidentally writing loanwords in hiragana, either by forgetting to press the space bar to convert the word to katakana, or by pressing it twice to return the selected option to hiragana. This kind of error was commonly referred to in the case of inconsistencies, described in section 6.5 below. The role of word processing software in relation to loanwords in hiragana is also discussed in section 7.5.

The second type of accidental usage concerned the etymology of the word itself, as some participants believed that the origins of some commonly used loanwords may not be known to all native speakers. For example, Natsuki (G2) suggested that many Japanese people may not be aware that ‘カステラ/kasutera’ (castella) and ‘キャベツ/kyabetsu’ (cabbage) are loanwords, and she herself made an exclamation of surprise when the word ‘れもん/remon’ (lemon) was produced as an example of a ‘loanword’ in hiragana. Hiromasa (G8), on the strength of his personal conviction that the word ‘きゃべつ/kyabetsu’ (cabbage) seemed acceptable in hiragana, started to question his own intuition that it was a loanword partway through the interview, asking the researcher ‘What is “cabbage”, (kyabetsu) anyway? Is it originally an English word? Would you call this here (points at the photo prompt) a ‘cabbage’

218

in Australia?’ When the researcher replied in the affirmative, he replied, ‘Well, so it’s not a Japanese word then…’163

Hiromasa’s confusion suggests the strength with which the script can affect the interpretation of a word, including its origins, to a potential reader; the seemingly natural hiragana transcription leading him to question whether it might be a Japanese word after all. Writing particular loanwords in hiragana may also have a self-generating effect: the more a certain word appears in hiragana, the more its origins are obscured, and the more likely it is to be written that way by another person. The ‘weirdness’ of this particular word may gradually wear off from frequently being encountered, with the result that words such as ‘cabbage’ and ‘lemon’ are sometimes felt to be Japanese. This sedimentation of language forms is further explored in the discussion (section 7.5).

6.2.3 Balance and layout of text

Three respondents mentioned the choice of script in relation to the text as a whole, especially to the words coming directly before and after the loanword in hiragana, and the balance created between them. For example, in the label on a supermarket shelf ‘サラダ用きゃべつ /sarada yō kyabetsu’ (Cabbage for salads), 164 Ryō (G2) explained that despite being inconsistent with the product packaging, he actually preferred this version, saying that the version on the package felt weird (気持ち悪い/kimochi warui), because it had too much katakana. The use of hiragana for the loanword ‘cabbage’ therefore separates this phrase into three units, each in one of the Japanese scripts, hiragana (cabbage), kanji (for), and katakana (salads). This was also noted by Tarō (G4) in relation to ‘らいふアシスト馬車道/raifu ashisuto bashamichi’ (Life Assist Bashamichi), who commented that the three words were each written in one of the three Japanese scripts.

Two participants, Ayano (G6) and, Nagisa (G3) talked about the balance of scripts in the bath toy ‘スプラッシュあいらんど/supurasshu airando’ (Splash Island) which is preceded by

163 「きゃべつ」って何もんな？「きゃべつ」って元々は英語なの？…オーストラリアでこれは「きゃべつ」なんですか？ …じゃ、日本語ではないんだよね。 164 The product name is ‘サラダ用キャベツ’ with ‘cabbage’ in katakana, while the label on the shelf says ‘サラダ用きゃべつ’ with cabbage in hiragana. This image, as well as the others used in the interviews and focus groups, can be found in appendix 4. 219

the onomatopoeic expression ‘ころころ/koro’ (roly-poly) on the product packaging, as the toy contains balls that roll down a small slide into the bath. The use of hiragana for ‘island’ created a vertical stack where a katakana word (‘splash’) is sandwiched between two hiragana words (‘roly-poly’ and ‘island’),165 Since onomatopoeia are frequently found in both hiragana and katakana scripts, Nagisa went on to explain that the producers probably considered all the options for which words appeared in which script, and judged the alternation of hiragana-katakana-hiragana to be the best combination. The balance of scripts in a particular text was also mentioned by the survey respondents, for example in relation to ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon), described in section 5.3.3.

In examples of two-loanword phrases where only one of the words was written in hiragana, such as the vending machine drink ‘あいすココア/aisu kokoa’ (Ice Cocoa) and the product ‘たおるマフラー/taoru mafurā’ (Towel Muffler), as well as ‘スプラッシュあいらんど /supurasshu airando’ (Splash Island) mentioned above, participants sometimes commented on how the message would be different if the script used for each word was reversed. ‘Towel muffler’, was felt to be a suitable arrangement in five separate sessions (G3, G4, G8, S5, S6), but for different reasons; Nagisa (G3), Tarō (G4) and Ryūichi (S5) each expressed a belief that while the adjective or noun modifier could be in hiragana, the noun or actual product should be in katakana. Mio (G8) expanded on this idea, stating that in the reversed order, she would question the quality of the product, while Yūsuke (S6) argued that rather the ‘eye- catching’ function of hiragana would be lost, and the product would not stand out so much.

Despite conforming to the idea that noun modifiers should be the word which appeared in hiragana, the use of script in the example of ‘あいすココア/aisu kokoa’ (Ice Cocoa) was questioned by three participants, with some respondents feeling that the reverse order may have been more appropriate, in other words, ‘アイスここあ/aisu kokoa’ (Ice Cocoa). As Yūsuke (S6) explained, ‘cocoa’ has a cuter image than ‘ice’, and would therefore be the more fitting of the two words to appear in hiragana. Coincidentally, ‘cocoa’ was the word Yuka (G1) mentioned having written in hiragana on a Facebook post, its ‘cute’ image also being a motivating factor for her.

Other respondents mentioned the layout or arrangement of words on the page as considerations for the script used for particular words. For example, when shown the

165 ころころスプラッシュあいらんど/koro koro supurasshu airando (Roly-poly Splash Island) 220

alternative variations for the ‘I can ふらい/furai (fly)!!’ example, Miho (G3) commented that if the phrase appeared all in kana, the number of characters required would not fit in the text.166 Takuya (G4) noted that the smaller amount of space kana takes up compared to English might be a consideration in the hostess bar names, possibly because the signs they are frequently displayed on are so small.

Similar comments were also made by three participants in relation to the text from the survey which used kanji for the word ‘coffee’ (珈琲/kōhī), who judged it appropriate because ‘it’s better space-wise, using two characters (珈琲/kōhī) rather than four (コーヒー/kōhī)’,167 for example.

6.2.4 Obscuring word boundaries

Many participants noted that in some cases, the use of hiragana for loanwords had the effect of erasing or obscuring the boundary between two or more words. As Walker (2000, p.11) notes, the ‘articulation’ or format of a text can express either connection or separation between the various elements, in other words, it is an expression of the textual metafunction.

Jun (G3) commented on the layout in terms of the word boundaries and ease of reading, noting that successive hiragana loanwords on a single line were harder to read than those appearing on separate lines. He illustrated this point with the name of the drink, ‘みっくちゅじゅーちゅさわー/mikkuchu jyūchu sawā’( Mix Juice Sour), which he found easy to read as the words appear vertically in three separate columns. However, he predicted that if the aged care facility name, ‘ らいふアシスト馬車道 /raifu ashisuto bashamichi’ (Life Assist Bashamichi) which appears on a single line in another text, were to appear with both loanwords in hiragana, (eg. らいふあしすと/raifu ashisuto), it would be harder to read. Gensuke (G4) also noted that this phrase would be hard to read if both words appeared in hiragana. Similar ideas were noted by Hiromasa (G8), stating that ‘いろいろすーぷ/iroiro sūpu’ (Various soups) was hard to read as it appears on a single line, with the use of hiragana for the word ‘soup’ erasing the word boundary between these words. Yūta (G4) also

166 ‘I can’ in English/rōmaji requires four characters while the kana versions eg. ‘アイキャン/aikyan’ requires five. Japanese characters also require more space than rōmaji characters. 167 4 文字より 2 文字の方がスペース的に良かった。 221

commented on the difficulty in understanding this phrase, noting that without the picture, he would be unsure of the meaning. In G6, when Chisato made a similar remark about the absence of script to mark word boundaries in this text, Ayano (G6) agreed, noting that although the word boundary in ‘ゆずれもん/yuzu remon’ (Yuzu lemon) was also erased by the use of hiragana, the fact that these words appeared on different lines made this distinction for the reader. Another way of distinguishing word boundaries is by the insertion of spaces between words; despite this practice not being used in Japanese writing, the author of the social media post ‘いえすあいあむ/iesu ai amu’ (yes I am), appears to have done this. However, Kazu (G1) still reported that this example was hard to read.

Gensuke (G4) perceived the phrases ‘あいらびゅ/airabyu’ (I love you) and ‘ふぁっきゅー /fakkyū’ (fuck you) as being used as phrases or single elements by their writers, and Yūsuke (S6), also noted that most of these examples of consecutive words in hiragana were common combinations, including ‘I love you’ and ‘yes I am’, as mentioned above, or set phrases such as ‘Happy Halloween’ (はっぴーはろういん/happī harōin), and this predictability helped to ease the cognitive burden. He felt that if a complicated or unfamiliar phrase appeared in hiragana, such as ‘あいうっどらいく/ai uddo raiku’ (I would like…), it would be harder to understand.

The use of hiragana loanwords in social media was felt to be an attention-getting device by Ayano (G6), ‘It’s to make people interested to read it; because it’s hard to read, it’s like saying “Read it carefully!”’168 Satomi (G6) agreed, explaining that the reader is forced to spell the words out before understanding them ‘It’s like “N-E-W Z-E …” ohhh it’s “New Zealand”…You read it all and you know what they’re saying. You think “huh?” and then you read it and you get it.’169 This suggests that rather than any difficulty with the script itself, it is the unexpected combination of loanwords in hiragana that slows readers’ understanding of these words.

The difficulty in reading consecutive words in hiragana was felt to be intended in the case of product names for three participants. Satomi (G6) and Hiromasa (G8) thought this may be the case in ‘いろいろすーぷ/iroiro sūpu’ (Various soups), and Yuka (G1) with ‘みっくちゅじゅーちゅさわー/mikkuchu jyūchu sawā’ (Mix Juice Sour). Nagisa (G3) also felt that this

168 興味もって読んでもらうため。読み難くて、しっかり読む！って 169 “にゅーじー” あっ、ニュージーランドか。全部読んで意味がわかる。ん？って読んで、あっ！っという感じ 222

may have been a consideration in the print club image which said ‘ふぁっきゅー/fakkyū (fuck you)’, where the difficulty in deciphering the words served to obscure the meaning, and thus downgrade the offense of the message. This particular example further blends the word boundaries by linking the sounds of the two words with the single Japanese mora ‘きゅ/kyu’. Ryūichi (S5) suggested that rather than consecutive words in hiragana being difficult to read, they simply represent an ‘unnaturalness’ (違和感/iwakan), and that this may be a marketing strategy, to catch a potential consumer’s attention. This ‘unnaturalness’ as a strategy to make the product stand out was also mentioned by Yūsuke (S6) in relation to ‘野菜りっぷくりーむ/yasai rippu kurīmu’ (vegetable lip cream).

The kinds of responses described above suggest that besides the nuances and potential connotations of a particular script, there are also factors such as the arrangement of words within the text which are considered by authors, whether they are advertising executives designing product packaging, or high school students taking print club photos. In each case, the choice of script and the assembly of the other semiotic resources within a text demonstrate the textual metafunction a word can play within its context.

6.2.5 ‘Un-realness’

The use of hiragana for loanwords was said to give an air of ‘un-realness’ to the message, or a feeling that the person was joking, or did not really mean what they said, by seven of the participants. Hiro (G1) explained that in the cases of the print club images containing the words ‘あいらびゅ/airabyu’ (I love you) and ‘ふぁっきゅー/fakkyū’ (fuck you), the use of hiragana suggested that they did not really mean those words. Mio (G8) echoed these comments on ‘I love you’, saying that writing it in English would seem ‘real’ (本気/honki, 本当/hontō), but in hiragana it just seemed like a logo or design you might find on an item of clothing.170 The ‘fuck you’ example was unanimously agreed across all interview groups to have a stronger meaning and/or be more offensive if it was written in katakana, for example ‘In English or katakana it would be more real, like have a much more real feeling’ 171 as Nagisa (G3) explained, with the hiragana version described by Ayano G6, and Jun G3 as

170 Although she did not mention the ‘unrealness’ of the texts, Ayano (G6) also noted that the print-club examples seemed like motifs (模/moyō) or were used to decorate (飾/kazaru) the images. 171 英語で書いてあったり、カタカナで書いてあると本気っというか、ガチ感がすごいかも 223

merely ‘ふざけてる/fuzaketeru’ (messing around), for example. In G4, unseriousness was also spoken of in relation to the social media post containing the word ‘ふぁみりぃ/famirī’, (family) with Yūta jokingly wondering what the writer was trying to hide by writing the term in such a fashion, with Kei following up on this comment by wondering why they avoided using the Japanese word, ‘家族/kazoku’ (family).

This is evidence of the interpersonal function of the use of script, as the ‘seriousness’ with which the message is to be taken by the reader is influenced by the choice of script.

The playful connotation of words or phrases in hiragana was also brought up by Miho (G3)172 in relation to the ‘I can fly!!’ sticker, who, like seven of the survey respondents, explained that the use of hiragana in this phrase was appropriate because the character was not actually flying, he was ‘just messing around’ ( ふざけてる /fuzaketeru). This idea was also commented on by Nagisa (G3), explaining that the use of hiragana showed that the character was kidding around (なんちゃって/nanchatte). The idea of katakana representing the ‘real’ was also brought up in the ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) text used in the survey (see section 5.3.3), where the use of hiragana was felt to distinguish the word ‘lemon’ in the product name from a reference to real ‘lemons’ as an actual fruit ingredient elsewhere in the text by nine respondents.

Jun (G3) extended the playful aspect of phrases in hiragana to the facebook post ‘いえすあいあむ/iesu ai amu’ (yes I am), which responded to a compliment given to the writer, explaining that despite the affirmation expressed in the words, the script suggested a kind of modesty or restraint (遠慮/enryo) in taking the compliment, expressed through the purposeful ‘ふざけてる/fuzaketeru’ (messing around) with the choice of script. Kei (G4) also noted this example as containing a hint of embarrassment at the compliment, and the hiragana acceptance indicating the writer was not serious (真面目じゃなく/majimejanaku). Chisato (G6) expressed a similar interpretation, saying that English would have a seriousness (真剣さ /shinkensa) to it, while, hiragana showed she was just joking.

The un-realness of loanwords in hiragana was also applied by Jun and Nagisa (G3) to the names of hostess bars containing the words ‘らうんじ/raunji’ (lounge) or ‘すなっく

172 Miho did not participate in the online survey, and this was therefore the first time she had seen the text. 224

/sunakku’ (snack), 173 with Jun suggesting that when written in katakana, they suggest referents other than hostess bars, such as the ‘lounge room’ you might find in a house, or ‘snacks’ you can eat such as chips. Nagisa agreed, stating that ‘regular’ lounges such the ones in hotels are always written in standard katakana. With much laughter, they gave voice to the innuendo hinted at by the hiragana versions of these words, suggesting the intended message was ‘We’re not really a normal lounge…’(Nagisa), ‘We’re not a proper one…’ (Jun) or ‘we’re not that kind of lounge…’ (Nagisa).174

The ability of the script to suggest ‘un-realness’ suggests how script can perform the referential metafunction, as the difference between katakana and hiragana can suggest whether a comment such as ‘I love you’ or ‘Yes I am (cute)’ should be taken seriously or not, and whether a ‘lounge’ denotes a general seating area or a hostess bar.

6.2.6 Expressing Japanese-ness

Like ‘はぴねす/hapinesu’ (Happiness) clinic and ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) in the survey (see section 5.3), some respondents felt that hiragana had been used to give a feeling of ‘Japanese-ness’ to a referent, as mentioned by Ryūichi (S5), Ayano (G6), and Kazu (G1). In G3 it came up in relation to a number of texts, with Nagisa stating the towel muffler had a ‘和風/wafū’ (Japanese-y) feeling, Miho explaining that hiragana made the 野菜りっぷくりーむ/yasai rippu kurīmu (vegetable lip cream) more Japan-like,175 and Jun suggesting that the use of hiragana on the box of とまと/tomato (tomatoes) was to appeal to the fact they were domestically grown.176 In G2, Ryō also talked about the motivation for the inconsistent sign for the Uji green tea cake,177 ‘Because it’s Uji green-tea from Kyoto, the shopkeeper thought the images of “Kyoto”, “Japan”, “old” would suit, so they put it in hiragana. They thought it would look more Japanese-y.’178 In G8, Mio hinted at the difference in preferred script based on the cultural associations of the referent, as she mused ‘“Bread”

173 ‘らうんじちあき’ and ‘すなっく新’ were the examples produced in the survey. 174 「本当の普通のラウンジじゃない」「ちゃんとしていない」and「違うラウンジだよ…」 175 日本っぽくなります、すごく… 176 国産アピール 177 Uji is a district in Kyoto, famous for the green tea produced there. 178 このかすてらのひらがなは宇治抹茶。その京都の宇治のほうで作ってて。たぶん、店の人たちも京都は日本。その古いイメージがあるから、ここはひらがなにしたほうがいい、和風になるみたいからあるかな…これを作った人はそっちのほうがいい。和風に見えるから書いた。 225

(パン/pan) should be in katakana. But “an-pan”179 (あんぱん/anpan) would be in hiragana.

Ah, but “cream-pan” (クリームパン/kurīmupan) would certainly be in katakana…’180 Kei, Yūta and Gensuke (G4) also agreed that in the case of ‘an-pan’, the loanword ‘pan’ should appear in hiragana. Kei further indicated that ‘cream-pan’ would be best in katakana. The group laughed at the seeming inconsistency of their intuitions of script use for these examples.

Other respondents such as Kazu (G1), and Ryūichi (S5) and Hiromasa (G8) extended the Japan-ising influence of hiragana to suggest that items referred to with loanwords in this script item had been in Japan for a long time. Ryūichi (S5) also had the impression that hiragana made it seem that the product was suitable for Japanese people. This was particularly the case with items that were exogenous to Japan, for example, the instant curry brand ‘ すぱいす/supaisu’ (Spice). Using hiragana, the script strongly associated with Japanese culture, for the brand name of a product originally associated with the Indian subcontinent, suggested to Ryūichi that ‘it looks like the product is aimed at Japanese people, that the blend (of spices) is adjusted for them...they’re saying clearly “this is not a foreign product”, it suits the delicate palate of the Japanese.’181 This was also mentioned by survey respondents in relation to the brand of instant coffee which featured the word ‘coffee’ in kanji (珈琲/kōhī) rather than katakana (see section 5.3.5).

6.2.7 Expressing familiarity

Besides hiragana giving a ‘Japanese’ feeling to a referent, the more general feeling of ‘familiarity’ or closeness was also found to be able to be expressed through the hiragana script. This is quite different to the joking, ‘unreal’ associations described earlier, and shows the broad range of interpretations of loanwords in hiragana.

Giving a feeling of familiarity was brought up as a motivating factor for businesses to use hiragana for loanwords in their names, such as ‘らいふアシスト馬車道/raifu ashisuto bashamichi’ (Life Assist Bashamichi) by Ryō, (G2) and ‘はぐみぃー/hagu mīi’ (Hug me) by

179 ‘あんぱん/an-pan’ is a soft bread roll filled with traditional sweet red bean paste, and therefore associated with Japanese culture. 180「パン」はカタカナで。「あんぱん」はひらがなの感じ、でも「クリームパン」は絶対ひらがなじゃないことない？ Like an-pan, ‘cream-pan’ (cream-bread) is a roll filled with custard or cream. It is also a Japanese invention, but the dairy-based filling means that it is more strongly associated with western culture. 181日本人向けに…その混ぜ方も調整してるか…日本人皆持ってる繊細な舌にあうの作って。 226

Yūsuke (S6). Familiarity was also mentioned in relation to posts on social media, with Daisuke (G2) suggesting that it gave the messages a familiar feeling, however Yūsuke (S6), Ryō (G2) and Ayano (G6) all mentioned that loanwords in hiragana were only used in messages to particularly close friends. Naoto (G4) also noted that the social media examples of loanwords in hiragana had a sense of closeness (親しみさ/shitashimisa) to them.

In the case of the word ‘にゅーずぃーらんどぅ/nyūzīrandou’(New Zealand), Kazu’s (G1) impression was that it was a place that the writer knew well, and using hiragana expressed the place had a sense of being ‘home’ for the writer. Yūsuke (S6) noted similar connotations of familiarity with phrases such as ‘ はっぴーばれんたいん /happī barentain’ (Happy Valentine) and ‘はっぴーはろういん/happī harōin’ (Happy Halloween), suggesting that as foreign customs, they seem more authentic when written in English, but that writing them in hiragana gave them a feeling of familiarity (親しやすい/shitashimi yasui); this seems to suggest that hiragana can have the effect of bringing something closer towards the writer.

Yuka (G1) also referred to the potential of hiragana to suggest closeness in the example of the print club with the words ‘ふぁっきゅー/fakkyū’ (fuck you) written on it, suggesting that this image represents more than just mimicry of or aspiration to western culture (憧れ /akogare), but an appropriation of it, making it ‘closer to themselves’.182 By writing this obscene English phrase in hiragana, they not only downplay the seriousness of the offense as described in the previous section, they make it their own through writing it in the script most associated with Japanese culture, and one that is frequently cited as being ‘familiar’ or ‘close’ to the national consciousness. When this print club image was brought out in another interview group, Miho (G3) admitted that she had seen this phrase before on a Japanese friends’ facebook page, one who was a ‘returnee’ (帰国子女/kikokushijo) or person who has spent part of their childhood living overseas. While details of this particular post are not known, it is interesting to consider the possibility of the ‘familiarising’ effect of hiragana suggested by Yuka (G1) in relation to this second writer, who was known to have experience living abroad.

The idea of familiarity was also brought up in relation to the referents of hiragana loanwords. Kazu (G1) noted that writing ‘すーぷ/sūpu’ (soup) in hiragana gave the impression that it had been in Japan for a long time, and Hiromasa (also G8) noted that in the case of

182 もっと自分に身近にして 227

vegetables, ‘tomato’ and ‘cabbage’ were ingredients that were familiar to Japanese culture (身近の食材/mijikai no shokuzai). In the two word compound ‘たおるマフラー/taoru mafurā’ (towel muffler), Kei (G4) explained that having ‘towel’ in hiragana was acceptable because it has a familiarity (身近/mijika), while having the scripts reversed, and the word ‘muffler’ in hiragana would be unnatural because it does not have this familiar feeling.

Finally, the use of hiragana in the names of hostess bars was also felt to create a feeling of closeness or familiarity, with terms translating to ‘easy to get along with’ or ‘approachable’ such as ‘親しみやすい/shitashimi yasui’ by Yui (G2) or ‘親近感/shinkinkan’ (Natsuki G2) being used to describe the atmosphere of these establishments. Yūta (G4) noted that bars such as ‘らうんじちあき/raunji chiaki’ (Lounge Chiaki) had a feeling of being easy to enter (入り安さ/hairi yasusa) to them. The degree of ‘intimacy’ of the services provided was seen to be most extreme in the bar called ‘あいどる/aidoru’ (Idol) which was described by Taro (G4) as seeming to be ‘not a place for drinking’, 183 suggesting that other more explicit services might be the main trade.

Kei (G4) also commented on the use of hiragana for the names of hostess bars, referring to an earlier comment he made, that these words seemed to have slightly weak people such as children and the elderly as their targets. He went on to explain that the hostess bars with hiragana names seemed to be places where customers could return to that vulnerable feeling, or a feeling of being children.184 This interpretation links a number of functions of hiragana together, including familiarity, gentleness, and childhood.

The ‘familiarising’ effect of words appearing in hiragana script and their referents can therefore be seen across a number of very different examples from the interviews. This is felt to be related to the ‘Japan-ising’ function of the script, but rather than making connections on on a national cultural level, this function is active on a more personal, individual level, thereby suggesting the interpersonal function of script choice in Japanese.

183飲むところじゃない 184 さっき、あまり好きじゃないっていったじゃ、子供っぽいとか若い人、何ていうの。。。老人ターゲットみたいなか弱い人にターゲットしてる半面…夜の場所じゃない？だからなんとなくそういう気持ちに帰りたいんだとか、何かちょっと自分が子供に帰りたいと言う思せる場所でいうことでこういうかわいく待ってるよみたいなそういうなんだろう。普通の日常だとこれ見ても何も感じない、夜にこう入るみたいな… 228

6.2.8 Representing voice quality and pronunciation

The use of hiragana for loaned words and phrases was said to add a particular voice quality to the written words by six participants. As Joyce et al. (2012) note, many orthographic variants are attempts to represent a particular pronunciation in a given situation, and indeed, many of these examples also involve non-standard transliterations such as the greeting ‘ぐっもーにん /gummōnin’ (g’mornin’) described in case study 4, section 4.2.4, and/or combinations of characters outside of official language policy (see section 2.5.2).

Yuka (G1) explained that the attempt to more accurately represent speech is particularly important on social media, where the tone of a statement need to be communicated through written text, rather than through other semiotic resources available in face-to-face communication such as voice quality or facial expression. She made these comments in relation to the Facebook post ‘にゅーずぃーらんどぅ!/nyūzīrandou!’ (New Zealand!), which she said communicated something more like ‘NEW ZEALAAAAND!!!’, raising her voice and throwing both hands up in the air as she did so. Hiro (G1) agreed, saying that if it was in English or katakana, you would need to add something like ‘I’m having a great time!’ to the message, but by putting the country name in marked hiragana script it seemed complete as it was, he argued, already containing the ‘hands in the air’ feeling that Yuka suggested. Takuya and Gensuke (G4) also noted this feeling for this particular example, explaining that it gave an exited feeling (テンション上げて/tenshon agete),185 with Takuya further noting ‘you wouldn’t write like this if you were feeling down.’186

This example is also interesting because besides the use of hiragana, it contains two transliteration variants not found in the standard katakana rendition of ‘New Zealand’ (ニュージーランド/nyū jīrando). Firstly, it contains an approximation of a ‘zi’ sound (ずぃ), rather than the conventional ‘ji’ (ジ). Secondly, a small ‘う/u’ sound has been added at the end of ‘Zealand’. Nagisa (G3) wondered if the intended effect of these changes was to mimic a native English speaker’s pronunciation, or perhaps their teacher’s pronunciation. Yūsuke (S6) made a similar comment, suggesting that the writer had made it sound like English and then written it in hiragana. He wondered if, like many Japanese learners of English, the writer

185 ‘テンション/tenshon’ is a frequently used loanword used to denote the degree of excitement experienced by a person. 186 落ち込んでいる時にそれをしない 229

was embarrassed of speaking ‘katakana English’, and was keen to exhibit knowledge of a more native-like pronunciation.187

Another Facebook post, the message ‘はぴばーすでい/hapi bāsudei’ (happy birthday) also contained a transliteration that can be seen as more native-like than the standard ‘ハッピーバースデー/happi bāsdē’. Although the degree of native-ness was not mentioned specifically, it was thought to be motivated by a particular pronunciation by Yūta (G4).188 In the same group, Kei (G4) felt the use of hiragana in this post served to give the message an excited voice quality. When asked by the researcher if he had ever written this phrase in hiragana, he laughed and replied that his ‘テンション/tenshon’ (‘tension’, see footnote 192) had never been so high as to motivate a shift to hiragana. His group member Takuya (G4) agreed, quipping ‘To get the tension that high, alcohol might be involved…’189 While this comment appeared to be a joke and was met with laughter by his group members, it is noteworthy in suggesting that loanwords in hiragana represent a degree of excitement that a business person (as opposed to a teenage girl, perhaps) would not normally express, and may even express a ludic, drunken voice quality if such a business person were to use this marked orthography.

The name of the children’s rhythmics classroom ‘はぐみぃー/hagu mīi’(Hug me) also contained an unusual transliteration device, the vowel in ‘me’ being elongated by both a (half-size) hiragana ‘い/i’ as well as a bar. As explained in section 4.1.1, the use of both vowel-lengthening features in a single word was unusual within the examples in the corpus, occurring in only five of the 162 long vowels in the corpus. Nagisa (G3) noticed this unusual transliteration and suggested that the aim may have been to mimic a child’s pronunciation of the word, with a particularly long ‘meeeeee!’ sound. This echoes Miyake’s (2007) comments on the functions of half-size characters affecting the prosody and rhythm of a word, as well as adding a touch of cuteness.

187 While the ‘ずぃ/zi’ transliteration seems to be an attempt to represent the English pronunciation of ‘Zea-’, the motivation for ‘どぅ/dou’ is less clear. It may be at attempt to recreate the ‘d’ stop in the native English pronunciation of ‘-land’, and/or an attempt to overwrite the ‘o’ sound that forms part of the character ‘ど/do’. Another explanation is its use as an emphatic or lengthening device, similar to its use in manga and telop for phrases such as ‘だめだぞぅ/dame da zou’ (‘Don’t!’) in male speech. This transliteration is discussed in relation to gender in section 6.3 below. 188 発音ぽっくみたいな 189 ここに行くには多分、アルコールが… 230

The above examples have shown that the use of hiragana can be seen as adding a particular voice quality to written words, therefore blurring the boundary between speech and writing. The addition of non-standard transliteration in many of these examples is noteworthy, although as noted in the analysis of the ‘ぐっもーにん/gummōnin’ (g’mornin’) tweet (Case study 4, 4.2.4), no concrete conclusions can be drawn without investigating whether non- standard transliteration also occurs in loanwords in katakana. However, the effect of the combination of hiragana script with a more accurate transliteration can be seen as simultaneously marking a pronunciation shift closer to an external native-like norm, as well as using the ‘closeness’ and ‘familiarity’ of hiragana as described in the previous section to bring a native-like pronunciation of English within the range of a Japanese speaker of English. The notion of hiragana focusing attention on the sound of a word or phrase is further discussed in section 7.3.2.

6.2.9 Highlighting language play

The business name ‘はぐみぃー/hagu mīi’(Hug me) was initially misread by Kazu (G1) and Chisato (G6) as ‘はぐくみ/hagukumi’, the meaning ‘the raising of a child’, and Ayano (G6) noted that this word was used, in kanji, in the explanation of the business in the text. She therefore argued that this paronym was the motivation for the use of hiragana in this particular case. Paronyms and homonyms were some of the instances of language play uncovered in the corpus analysis (see section 4.1.6).

In a discussion on manga, Nagisa (G3) reported a feeling that manga titles for ‘おたく/otaku’ (geeks) in particular were common sites of loanwords in hiragana, with Miho (G3) supplying the title ‘K-on!’, transliterated as ‘けいおん/keion’, as a representative example, the cover of which is shown in figure 6.2 below. On later investigation by the researcher, the title was found to represent the clipped form of ‘軽音楽/kei ongaku’, or light/pop music, as the subject of the manga was a light music club at a Japanese senior high school. The anglicized transliteration ‘K-on!’ was perhaps intended to invoke English phrases such as ‘Keep on!’ or ‘Carry on!’, for example.190

190 The author’s pen-name, ‘かきふらい/kakifurai’ is also of interest, and is described in the discussion, section 7.3.2. 231

Figure 6.2: Manga cover, 'K-on!!'191

6.3 Perceptions of age and gender

In the interviews, six participants, both male and female, identified hiragana as being feminine and associated with women’s writing. As Ryō (G2) explained, because hiragana was once the only script used by women, it has associations of softness and femininity today, even if people are unaware of the historical reasons for this. Kei (G4) also noted that many of the examples did not seem to be aimed at men (男対象/otoko taishō), but rather at women, children and older people. Miho (G3) suggested that the gentle, feminine influence may have been the reason it was chosen for businesses such as ‘はぐみぃー/hagu mīi’ (Hug me) and ‘らいふアシスト馬車道/raifu ashisuto bashamichi’ (Life assist Bashamichi), which care for children and the elderly respectively. Ryūichi (S5), in discussing loanwords in hiragana in product names, reported that if he had to decide either way, he would suggest these products were aimed at women rather than men.

Other genres, such as the hostess bars, rather than being aimed at women, were perceived to be using the feminine influence of hiragana to draw attention to the presence of women inside these establishments, as Satomi (G6) explained, through the round, gentle hiragana script. The feminising influence of hiragana script was so strong that it led Nagisa (G3) to reflect that if the names of the bars called ‘ゆにこーん/yunikōn’ (Unicorn) or ‘すぷりんぐ

191 http://k-on.wikia.com/wiki/K-ON!_(Manga) 232

/supuringu’ (Spring) were written in katakana, rather than hiragana, they would suggest host bars rather than a hostess bars. Miho (G3) agreed with this hypothesis, adding that a host club with a hiragana loanword name would be weird (おかしい/okashī), to which Nagisa added ‘gross’ (気持ち悪い/kimochi warui), unless, she reasoned, it was a host club in Shinjuku 2- chome, a well-known gay district in Tokyo. This conversation highlights the power of script to suggest femininity, or a stereotypically feminized version of male homosexuality, for some interviewees. It also reflects the perceptions of ‘feminized speech’ reported by Okamoto & Shibamoto-Smith (2016), where heterosexual men using women’s speech (女言葉/onna kotoba) were reported as seeming effeminate (女々しい/memeshī) or ‘like a drag queen’ (おかまっぽい/okamappoi).

Participants such as Naoto and Yūta (G4) noted that the hiragana loanwords in the social media examples were probably written by women, and participants were often asked if they could guess whether the authors were men or women. While ‘all women’ was the most frequent and immediate guess, made by seven individuals, on reflection, many men such as Jun (G3), Yūta (G4), Takuya (G4), Kazu (G1), and Hiro (G1), correctly identified the text containing ‘New Zealand’ (にゅーずぃーらんどぅ!/nyūzīrandou!) as being written by a male. Hiromasa (G8) also guessed that this example was written by a male, but when asked why, he thought that it was perhaps more to do with the content of the message than the script. Since the post was the reply to the question ‘Where are you?’ as had been explained by the researcher, he felt a female writer would perhaps expand their reply to by giving more details of their visit besides the name of the country, while the simple, straightforward answer to the question as given above struck him as the way in which a male would respond. Gensuke (G4) explained his reason for identifying this text as being written by a male as relating to the marked ‘どぅ/du’ sound at the end of the phrase, suggesting that a female writer would be more likely to extend the sound as ‘どー/dō’ (see also footnote 187 above).

While Yui (G2) expressed a belief that while girls must be more numerous in using loanwords in hiragana on social media, she said it was possible that ‘guys who liked anime’192 (アニメ好きの男の子/animezuki no otoko no ko) might use it, with Natsuki (G2) agreeing, supplying the term ‘geeks’ (おたく/otaku) for Yui, and explaining that ‘normal boys’ would

192 Anime is the clipped form of the loanword アニメーション/animēshon (animation), and is used in Japan, as it is in the west, to refer to animated television series and movies specifically produced in Japan. 233

be unlikely to use hiragana in that way. Nagisa (G3) also expressed a link between hiragana loanwords and geek culture, including manga, anime, and maid cafes. These kinds of comments, like the ones referring to gay host bars above, point to the feminising influence of hiragana being extended to particular kinds of males, including ‘geeks/otaku’ who may not fit a stereotypically masculine image.

The conversations between Yui and Natsuki (G2) concerned ‘boys’ (男の子/otoko no ko) rather than ‘men’ (男の人/otoko no hito), and other participants expressed a belief that age might be a factor in the use of hiragana loanwords by males; Jun (G3) had the impression of high school or first or second year university students, for example. When the researcher revealed that the message ‘はぴばーすでい/hapi bāsudei’ (happy birthday) was from a male to another male, Mio, (G8) aged in her 40s, exclaimed ‘From a guy to a guy?! Whoa…. well, I guess times are changing...’193 implying that in the past, her male peers would not have used hiragana in this way. The participants of G6, all women in their 30s, also expressed surprise that the birthday greetings were sent from one male to another.

Age was also brought up in relation to the hiragana loanwords used as names for businesses. In G6, Ayano suggested that the hostess bars with hiragana names presented were for ‘old men’ (おじさん達/ojisantachi), with Chisato adding that people in their 50s and 60s were the target market for these types of bars. Satomi agreed, suggesting that a bar aimed at young people would be more likely to use English for their name. Nagisa (G3) made a similar comment in discussing the florist called ‘ふろーりあ/furōria’ (Floria), reporting that it seemed to be for ‘old people’ (お年寄り/otoshiyori), as a shop for young people would be more likely to use katakana or English. Yūta (G4), wondering out loud why English was not used for the word ‘life’ in ‘らいふアシスト馬車道/raifu ashisuto bashamichi’ (Life Assist Bashamichi), supplied his own answer, suggesting older people might not be able to read it, and wonder what ‘リフェ/rife’ (ri-feh) was, mocking the pronunciation of the English word ‘life’ by someone unfamiliar with the language.

In G3, Miho and Nagisa thought that the use of hiragana made the hostess bars themselves seem old, Nagisa suggesting from the 1980s. In G8, Mio suggested an old-fashioned (昔の感

じ /mukashi no kanji) or Showa-period 194 ( 昭和っぽい/shōwappoi) feeling, and also

193 男性から男性へですか。へー…そっか、何か時代は変わってんなー。 194 The Shōwa period was from 1926 – 1989. 234

suggested that these bars were aimed at older people (年配の方/nenpai no kata). Hiromasa, citing the ‘generation gap’ (ジェネレーションギャップ/jenerēshon gyappu) between himself, aged in his 50s, and Mio in her 40s, disagreed with this interpretation, saying that the use of hiragana merely gave a soft (柔らか/ yawaraka), informal feeling (砕けた感じ /kudaketa kanji) to the business, marking it as a place to relax. Yūta (G4), also in his 50s, conversely suggested that writing words like ‘snack’ and ‘lounge’ in katakana had an outdated feeling (古臭い/furukusai), while hiragana gave it a rougher, edgier feeling. In G4, Takuya initially noted that the youthful image of hiragana suggested the staff might be young women, but later commented that ‘らうんじちあき/raunji chiaki’ (Lounge Chiaki) would probably have a more mature proprietress. Kei disagreed with this statement, with Tarō eventually supplying ‘late twenties’ as the approximate age.

Age was sometimes considered a factor for the use of hiragana loanwords on social media, with Chisato (G6), who is in her 30s, explaining that it was fine for young people to mess around with these kinds of relaxed writing styles195 but working adults (社会人/shakaijin) would not use them, except in private messages to their closest friends. Daisuke and Ryō (G1) also noted that the use of hiragana for loanwords in emails or messages would be limited to people you were close to. Age was also commented on by Jun and Nagisa (G3), who interpreted the Facebook post ‘I went to see the cherry blossoms with my family…’ (昨日ふぁみりぃでお花見 /kinō famirī de ohanami) as referring to an outing with the writer’s parents, until it was revealed by the researcher that the word ‘family’ referred to the writer’s husband and young child. Jun expressed surprise that someone in their twenties would use this phrase, stretching it at most to someone in their early twenties, stating a belief that people his age (30s) would not write in this manner. On reflection, Nagisa, also in her 30s, disagreed with Jun, suggesting that it was possible, and that the use of hiragana in this post might be a kind of emphasis, interpreting it as ‘my lovely husband and cute baby. ’ She added, however, that this choice of script made the message seem slightly ‘girlish’ (ギャルっぽい). The word ‘girl/gyaru’ used here refers to a particular girl sub-culture, and the letters (ギャル文字 /gyaru moji) associated with it, popular in the 1990s (see section 2.6). She remarked later that hiragana in general, and an avoidance of kanji, are also traits of ‘gyaru’ culture.

195 ふざけて、崩した文字 235

While Mio (G8) did not mention ‘ギャル文字/gyaru-moji’ (girl-letters) specifically, she was familiar with styles of girls writing popular in Japan in the 1980s and 90s, and frequently drew parallels between these and the texts presented in the focus group. She remarked that the use of superfluous small characters (小文字/ko-moji) in the social media examples such as ‘にゅーずぃーらんどぅ/nyūzīrandou’ (New Zealand) and ‘ふぁみりぃ /famirī’ (family) reminded her of the craze for using small characters, for example to start common phrases such as ‘thank you!’ (eg.‘ぁりがとう!/arigatō!’), and noted the handwriting in the print club was reminiscent of ‘round letters’ (丸文字/marumoji) which are still common in handwritten texts by teenage girls and younger women. These comments link loanwords in hiragana to the evolving canon of female writing styles, from the handwritten maru-moji beginning in the 1970s to the creative gyaru-moji of the cell-phone era of the 1990s and early 2000s, suggesting that these styles continue to be drawn upon and reinterpreted by succeeding generations of not only women, as demonstrated above, but men too.

The associations of the hiragana script with ‘femininity’ as reported in the survey (section 4.2.2), as well as previous research (section 2.4), therefore suggests only part of the whole picture of the use of loanwords in hiragana. Of course, some examples presented in the survey and interviews are simply reported to be utilising the ‘feminising’ influence of this script, but this can be applied to either the consumers or the producers, for example a hiragana loanword may indicate a particular product may be ‘aimed at women’, or that a business with men as their target market such as a hostess bar ‘has women inside’. Furthermore, the feminising effect is felt by some to be limited to particular ages of women, namely young women, or conversely to be extended to young men, or gay men. Questions of age are also interesting in light of fashions of girl writing from past decades such as maru- moji and gyaru-moji; originally these styles were associated with teenagers and women in their early 20s, and may still be seen so by the generations that pioneered and used them, who are now in their 30s and 40s.

For today’s teenagers, however, these styles may seem outdated and uncool. Original users and disseminators of these early styles may also use them into adulthood, such as to describe their ‘family’ consisting of husband and child as in the social media post described above. While the association of the hiragana script with ‘femininity’ and girl culture is undoubtedly one factor behind its use for loanwords in some cases, the interviews and focus groups, as well as the case studies, reveal the breadth of functions performed by loanwords in hiragana.

236

6.4 Influence of typography

Within texts such as those presented in the interviews and focus groups, script choice was just one semiotic resource deployed, and some interviewees commented how the choice of typeface also affected their interpretations. For example, Nagisa (G3) reported that the old- style café feeling given by the use of hiragana in the name of the vending machine drink ‘あ

いすココア/aisu kokoa’ (Ice cocoa) shown in figure 6.3196 was supported by the typeface used, drawing attention to the way the strokes in the character ‘い/i’ are connected in the middle. In the case of the box of tomatoes (とまと/tomato) pictured in figure 6.4, Miho (G3) suggested that the typeface chosen was ‘Japanese-y’ ( 日本っぽい /nihonppoi), and in combination with the hiragana script was perhaps trying to appeal to potential customers as a domestically-grown product, which was also mentioned by Jun (G3) previously.

Figure 6.3: Typeface used for 'Ice cocoa' Figure 6.4: Typeface used for 'Tomatoes'

Mio (G8) commented on the ‘flowery’ typeface chosen for the florist called ‘Floria’, (ふろーりあ/furōria) describing how it added to the cuteness created by the hiragana script, an effect which, she argued, would be lessened if the word appeared in the more conventional script used for the aged care facility ‘Life Assist’ which was also presented at that stage of the interview. Cuteness was also the perceived motivation of the use of hiragana for the handwritten sign ‘Happy Valentine’, (はっぴーばれんたいん/happī barentain) which Yūsuke (S6) explained was emphasized by the slightly misshapen characters and oversized and misplaced diacritics (濁点/dakuten and 半濁点/handakuten), features reminiscent of the ‘round letters’ (丸文字/maru-moji) style described in section 2.6. In G4, Kei also mentioned maru-moji when the print-club examples were introduced, with Yūta explaining that the

196 The images used in this section (5.3.4) have been cropped to highlight the particular typeface being discussed. The actual images used in the interviews and focus groups can be found in the research design and methodology section (3.4.4) and in appendix 4. 237

round shapes and curves made it easy to write in a cute fashion, noting that it would be hard to write katakana characters in this cute way.

Figure 6.5: Lettering used for 'Happy Valentine' Figure 6.6: Typeface used for 'Floria'

Chisato (G6) explained how the use of font showed the ‘type’ (-系/-kei) of bar, labelling the examples in the photo prompts ‘beautiful type’ (きれい系/kirei-kei) for ‘Lounge Chiaki’ (らうんじちあき/raunji chiaki); ‘cute type’ (かわいい系/kawaii-kei) for ‘Unicorn’ (ゆにこーん/yunikōn); and ‘normal type’ (普通系/futsū-kei) for ‘Spring’ (すぷりんぐ/supuringu). Kei (G4) also mentioned the cute font used for ‘Unicorn’, which he believed suggested young women would be working there, an interpretation also supported by the English words ‘idol197 Cafe & BAR’ on the sign.

Figure 6.7: Typeface used for 'Lounge Chiaki' Figure 6.8:Typeface used for 'Unicorn'

Figure 6.9: Typeface used for 'Spring'

197 An ‘アイドル/aidoru’ (idol) is a common loanword which refers to a TV star or singer in their teens or early twenties, usually with a cute, girlish image. This bar is likely to have young hostesses rather than actual celebrities. 238

These comments on typeface remind us that when viewers encounter loanwords in hiragana in authentic contexts, and their impressions of the script are mediated through other semiotic resources used in the text as well. It also suggests how the choice of hiragana works as part of a multimodal whole, with themes such as ‘nostalgia’ or ‘cuteness’ being communicated through combinations of resources such as script, font, colour, and direction, for example, as highlighted in the case studies (section 4.2)

6.5 Reactions to inconsistencies

Different kinds of ‘inconsistencies’ which had been noted in the corpus texts were also presented in the interviews and focus groups. The first set contained two examples where the brand or name of the product contained a loanword in hiragana, for example ‘すぱいす /supaisu’ (Spice) instant curry, and the same word (spice[s]198) occurred elsewhere in the text in standard katakana. In this text, it is used in the explanation of the product, ‘with the refreshing echoes of bouillon and spices’,199 and is pictured in figure 6.10 below.

Figure 6.10: Inconsistent use of script in same text, ‘Spice’ curry

These examples were not considered problematic by any of the interviewees or participants, in each case they stated that it was acceptable for the name of a product to use a loanword in hiragana, but in the explanation of the product, the loanword should be written in standard katakana, as Hiromasa (G8) labelled it, in ‘normal Japanese’ (通常日本語/tsūjō nihongo). When asked how the image of the product would change if the explanation of the product also contained a loanword in hiragana, common responses were that the product would seem

198 Japanese does not use a plural form for words such as ‘spice’. 199ブイヨンとスパイスの爽快な余韻 239

untrustworthy (信用できない/shinyō dekinai) by Chisato (G6), like the company was messing around (ふざけてる/fuzaketeru) by Hiro (G1), ‘weird’ (おかしい/okashī) by Jun (G3), ‘making fun’ (舐める/nameru) by Hiromasa (G8), or having a ‘playful heart’ (遊び心 /asobigokoro) by Ryō (G2). As Ryūichi (S5) stated ‘You might ask yourself, “Is this product for babies…?”’200 suggesting the immature or infantile nuance the use of script could give. In each case, therefore, it was felt that while there was flexibility in the script used for the product name, the use of language for the explanation of the product should conform to standard conventions.

The second set of inconsistent examples contained products with hiragana loanwords in their names, but the signs created by the shops to display the prices had been written in katakana, such as the tomatos pictured in figure 6.11 below. In these cases, all participants also felt that it was unproblematic, that it was fine for the shop to have used a different script to the original packaging, although Kei (G4) laughingly expressed a sense of sympathy with the product’s designers, saying ‘the supermarket staff really didn’t understand the feeling of the product…’201

Figure 6.11: Inconsistent use of script in related texts, tomato display

Three participants argued that, in the case of the display of tomatoes at the supermarket shown in figure 6.11 above, the seeming inconsistency was because the sign referred not to the product name, but to the general class of foodstuff. Yui (G2) suggested that the same sign was used for tomatoes from many different producers, explaining that while this time they

200 「赤ちゃんむけですか」と思うかもしれない。 201 スーパーの人とか…この商品の気持ちを汲み取ってない… 240

were ‘とまと/tomato’ (in hiragana), sometimes they might be ‘トマト/tomato’ (in katakana), and that was why in this case the lettering was incongruous. Takuya (G4) also suggested that the same sign might be used for many types of tomatoes. The first and second kinds of inconsistency described so far were relatively common in the corpus (see section 4.1.10), which may explain why they were not considered errors or mistakes by participants.

The third set of examples illustrated the reverse situation, where the product names were in katakana, while the price labels used the marked hiragana, such as those pictured in figures 6.12 and 6.13 below. These examples were much rarer, and generally generated reactions of surprise from participants, such as ‘めっちゃ意外/meccha igai’ (really unexpected) from Kei (G4).

Figure 6.12: Inconsistent signage, ‘Cabbage for salads’ Figure 6.13: Inconsistent signage, ‘Uji green-tea castella’

One example, the package of ‘サラダ用きゃべつ/sarada yō kyabetsu’ (cabbage for salads) shown in figure 6.12 above was generally felt to be an error or mistake, with five people citing a ‘conversion error’ (変換ミス/henkan misu), as the source. Yūsuke (S6) backed up this belief by referring to his experience working in a convenience store, where the ‘pops’ or price labels had to precisely reflect the name of the product so customers could identify the prices. Mio’s (G8) experience supported this statement, she explained that at the supermarket, if the product name is in hiragana, the sign usually is too. Takuya (G4) also mentioned this, joking that a customer might question if the packed cabbage labelled in katakana was indeed 94 yen as the sign containing hiragana said, or it had been mixed up with another packet which was more expensive. However, Yūsuke (S6) also mentioned a playful manager as a possible source of this inconsistency, one who was ‘mischievous’ (お茶目/ochame), and

241

Kazu (G1) laughingly suggested that the desired impression of the inconsistent sign was that ‘a really hip guy makes these!’202

For the second example, ‘宇治抹茶かすてら/uji matcha kasutera’, (Uji Green-tea Castella) (figure 6.13), however, three people also judged to be a strategic choice by the sign maker, either to appeal to a sense of Japanese-ness (Nagisa G3), being made from traditional green tea from a famous tea district in Kyoto, as well as being easier to read (Mio G8) or to be more eye-catching (Yūsuke S6). Takuya’s (G4) interpretation was more detailed, he explained that as the product appeared to something suitable for a gift, the label containing the word ‘castella’ in hiragana (かすてら/kasutera) was designed to catch the buyers eye, while for the person who received the gift, the katakana orthography of ‘castella’ (カステラ/kasutera) gave a feeling of being written properly or carefully (ちゃんとしていた/chanto shiteita). Others such as Hiromasa (G8) and Gensuke (G4) thought that, like the ‘cabbage for salads’ example described above, it was also just an accident.

The slight difference in interpretation of these two examples is interesting, because on the script level, they both feature products with loanword names in katakana, displayed with price labels showing the same word in marked hiragana. It seems that the nature of the product (one associated with Japanese tea), as well as the other semiotic resources within the text, for example the brush like-font and traditional imagery of the castella packaging, made it more likely for viewers to imagine this example to be a motivated choice rather than an oversight or error. The examples in the third set are also interesting to compare with the inconsistencies in the second set; while the second, featuring hiragana product names with katakana labels, were not perceived to be unusual or generate much surprise, the reverse situation, with katakana product names and hiragana labels did, and was often felt to be a mistake. Like the inconsistencies in the first set, some texts or parts of texts were more strictly held to the conventions of standard Japanese writing than others.

The text presented in the survey which generated the most intense expressions of surprise and bewilderment also surrounded the word ‘cabbage’, where half and quarter cabbages where displayed on two shelves one above the other, with the price label for the quarter cabbages being written in katakana (キャベツ/kyabetsu), while the half cabbages were written in hiragana (きゃべつ/kyabetsu), as shown in figure 6.14 below.

202 めっちゃポップなやつを作ってるんですよ！って 242

Figure 6.14: Inconsistent signage on cabbage display

Laughter accompanied exclamations from Hiromasa (G8) ‘there’s no deep meaning to this!’,203 and Kei (G4) ‘this is surely a mistake!’204 when this image was produced. The most frequent explanation, given by five participants, was that the signs were made by different people, with some people suggesting the sloppiness or inexperience of a part-time staff member was responsible for the offending hiragana sign. Yuka (G1) and Nagisa (G3) also pointed out that the hiragana sign was not made as neatly as the katakana sign, appearing to be made from paper rather than plastic, and not being centred vertically (Yuka) and being bent and not attached properly (Nagisa). However two respondents, Mio (G8) and Yūsuke (S6), admitted the possibility that the signs were made this way on purpose to distinguish the two different products (half and quarter cabbages).

Cabbage was one of the loanwords which many people felt to be acceptable in hiragana, (discussed further in section 6.7), and in this case of the display of cabbages Jun (G3) felt that while either hiragana or katakana was acceptable for writing ‘cabbage’, a sense of weirdness (違和感/iwakan) resulted from seeing both options side by side. Other participants such as Mio (G8) admitted they would not have noticed this inconsistency. One final point of note on this example was raised by Yuka (G1), who pointed out that some of the characters used for the hiragana and katakana versions of the word ‘cabbage’ were quite similar, for example the characters for ‘ya’ (や/ヤ), and those for ‘be’ (ベ/べ) are almost identical. It is possible,

203 深い意味がない！ 204 これはミスしかないっすか？！ 243

therefore, that an inconsistency in script with this particular word was more likely to go unnoticed than a word which brought out the often cited aesthetics of hiragana being ‘round’ and katakana being ‘angular’.

6.6 Alternative transcriptions of ‘I can fly’

An image of the sticker used in the survey (layer 3) was shown to the interview participants, and many agreed that the use of hiragana for the word ‘ふらい/furai’ (fly) in an English sentence ‘I can fly’ seemed unnatural or inappropriate. The following seven options were then presented on a card, and interviewees were invited to reflect on transcriptions they would be likely or unlikely to use if they were the designer. The options aimed to investigate what made this particular example the least ‘appropriate’ of the samples presented in the survey, and what precisely was the problem of ‘mixing’ that was frequently given as a reason for this answer (see section 5.3.4). The comments of interviewees are summarized beneath each of the options.

1. あいきゃんふらい!! (All hiragana: ai kyan furai!!) This example was felt to be cute by Ryō (G2) and Yūsuke (S6). Jun (G3) reported that it seemed childlike (幼い/osanai) and therefore suited the character, Kei (G4) also mentioned its suitability for the character. Miho (G3) felt it had potential, but the extra characters required would mean it would not fit in the corresponding space on the sticker. Hiromasa and Mio (G8) both reported it was hard to read, and Nagisa (G3) felt that the meaning would only be clear by referring to the picture. Gensuke (G4) felt that with the addition of spaces to separate the words, it would be suitable.

2. アイキャンフライ!! (All katakana: ai kyan furai!!) This example reminded some interviewees of other words: Jun (G3) and Ryō (G2) saw the word ‘ice candy’ (アイスキャンデー/aisukyandē)205 as the two phrases shared a number of characters, while having ‘fly’ in katakana reminded Nagisa (G3) of fried food (フライ/furai). Hiromasa (G8) felt it was too harsh (かたい/katai) to suit the character, while Mio (G8) reported finding it hard to read. Chisato (G6)

205 This is a relatively familiar loanword, referring to an icy-pole (AusEng) or ice-lolly (BrEng). 244

explained that it appeared as one word, with no word boundaries differentiated by contrasting scripts, and Natsuki (G2) simply said it ‘wasn’t right’ (違う/chigau).

3. I can fly!! (English) This version was one of the most popular, reported to be cool by Nagisa (G3), and good by Kei (G4), Daisuke and Natsuki (G2), and the ‘only option’ by Gensuke (G4). While this version was Hiromasa’s (G8) first choice, he found the word ‘fly’ in particular hard to read in English, and Kazu (G1) argued that the strength of ‘fly’ in hiragana disappeared if it appeared in English. Miho (G3) also felt it was too serious for a character who was just ‘mucking around’ (ふざけてる/fuzaketeru). Chisato (G6) felt that elementary school students would not understand the meaning.

4. I can フライ!! (Combination of English and katakana: I can furai!!) Natsuki (G2) felt this was fitting, but Ryō (G2) and Ayano (G6) again mentioned its association with fried food. Hiromasa (G8) thought this was a viable option, Mio (G8) agreed it was easy to read but felt it was a bit harsh (かたい/katai) for the character.

5. アイキャンふらい!! (Combination of katakana and hiragana: ai kyan furai!!) This version had the same associations of ‘ice candy’ for Jun (G3), although Hiromasa (G8) reported this was a version he might use.

6. I can 飛ぶ!! (Combination of English and Japanese: I can tobu [fly]!!) This option drew the most mixed reviews throughout the interviews and focus groups. It was liked by Takuya and Kei (G4), and Yuka (G1); however it was disliked by Nagisa (G3), and reported to be weird by Yui and Daisuke (G2), and not right (違う /chigau) by Mio (G8). Miho (G3) explained that this option would not work with the text as it makes it seem like the character is actually flying, whereas Jun’s (G3) dislike for this option had a phonetic basis. He explained that the vowel sound of the word ‘fly’ could stretch (伸びる/nobiru), giving a feeling of ‘flyyyyy!!’, whereas the Japanese word for fly (飛ぶ/tobu) stopped too abruptly.

245

7. Fly できる!! (Combination of English and Japanese: Fly dekiru [can]!!) Like the above version, Jun (G3) felt the final sound in this version also did not stretch the way the original transcription did, and Miho (G3) again felt this would only be appropriate if the hamster could actually fly. It was also said to be weird by Daisuke and Yui (G2), and unsuitable by Takuya (G4).

The discussions on the potential alternatives for the sticker featuring the words ‘I can ふらい /furai (fly)’, which was labelled inappropriate by around half of the survey respondents (see section 5.3.4), highlighted the range of considerations open to these individuals in their imagined role of designer. These included the way the word sounded, the space it would take up within the text, the action of flying/not flying taking place in the accompanying image, as well as words which, on first glance, might be misread such as ‘ice-candy’.

Some respondents knew very quickly which versions they were drawn towards, while others considered each of the examples in turn. Hiromasa (G8) traced the train of thought of the actual designer in the following way, imagining them writing the phrase in English first (3), then deciding it was hard to read and putting the word ‘fly’ in katakana instead (4), but then considering the harsh (かたい/katai) image of this version unsuitable for the cute character, they chose to put this word in hiragana as it appears on the actual sticker.

No single alternative example was felt to be a better choice than the original version, although the example containing only English (I can fly!!) was one of the more popular choices. Like the survey respondents, many of the difficulties mentioned by interviewees with this example seemed to be specific to this particular loanword, being a homonym with the Japanese ‘fried food’ (フライ/furai), and difficult to spell because of the collapse of /l/ and /r/ sounds in Japanese. The low number of judgements of ‘appropriate’ in the survey may also be a result of this text’s unusualness in terms of the results of the corpus study described in 4.1; being a verb rather than a noun, an unusual genre, the speech ascribed to an entity, and part of an English sentence. Interestingly, despite many interviewees initially reporting that the use of hiragana in the original sticker was strange, after considering the alternatives, many decided that it might be the most suitable version after all. As Ayano (G6) explained, it may be because of the perceived similarities in age and English level of the designer and their

246

customers: ‘the person who made it and the people who are buying it are our age, so the original is most fitting. For kids, all Japanese would be better.’206

6.7 Suitability of particular loanwords in hiragana

Many interviewees reported an intuition about which loanwords seemed acceptable in hiragana and which did not, but the rationale behind these feelings often differed. Many participants such as Yuka (G1), Yūta (G4) and Natsuki (G2) expressed the belief that the word ‘れもん/remon’ (lemon) seemed acceptable in hiragana because it had a kanji, for example. This reason was also frequently given in the survey for the appropriacy of using hiragana for the word ‘lemon’ in the product name ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon). The presence of a kanji was also used as evidence of a word’s establishment or long history in the Japanese language, meaning that it seemed equally suitable in hiragana or katakana, as Kei (G4) explained. He also suggested the word ‘らーめん/rāmen’ (ramen)207 as word with Chinese origins that had a kanji and seemed equally suitable in hiragana; his group member Tarō (G4) agreed, noting it had an image of being a kind of Japanese dish (日本食のイメージ/nihonshoku no imēji). There is a certain irony in the presence of a (Sino-Japanese) kanji being used to justify the Japanese-ness of a dish of Chinese origins, however, it serves to highlight the role of a kanji variant in establishing a non-native word such as ‘ramen’ or ‘lemon’ as a Japanese word, and the translingual flows between China and Japan in general.

When participants suggested that the existence of a kanji was responsible for a word’s naturalness in hiragana, the words ‘珈琲/kōhī’ (coffee) and ‘硝子/garasu’ (glass) were suggested by the researcher as a potential counterexamples, because, like ‘lemon’, they can be written in kanji, but do not appear frequently within the corpus of hiragana loanwords collected for the study. These words were often felt to be unnatural in hiragana, ‘coffee’ was reported to seem weird (変/hen) in hiragana by Gensuke (G4), and more specifically the problem was reported by Yuka (G1) and Daisuke (G2) to be with the long vowel sounds, of which this word has two. As described in section 3.4.1, the method of elongating vowel

206 これを作ってるような年齢の人と、買うような年齢の人と、私にはあってるから、多分、これがしっくり来るだと思う。もしかしたら子供…本当の子供とかだったら、全部日本語ほうがいいっと言うかもしれない。 207 As described in the introduction, loanwords of Chinese origin were excluded from this study precisely because words such as ‘ramen’ are difficult to distinguish from ‘漢語/kango’ (words of Chinese origin). 247

sounds differs between words written in hiragana and katakana, meaning that the presence of two katakana-style bars in a hiragana words can seem unnatural to a reader (こーひー/kōhī), although this was used in two of the three cases of ‘coffee’ in hiragana in the corpus.208

In the case of ‘glass’, however, there was no elongated vowel, and most participants were at a loss to explain why it seemed unnatural in hiragana. Daisuke (G2), however, suggested that it was due to the innate properties of the referent, being a hard substance, while hiragana was often associated with the word ‘soft’, (柔らかい/yawarakai), an intuition backed up by the survey results (see chapter 5, section 5.2.1). Comments such as this illustrate the multiple factors at work in people’s judgements of the suitability of particular words in hiragana, and that while the presence of an alternative kanji orthography makes some loanwords seem suitable in hiragana, it does not hold for all cases.

Furthermore, a number of words which did not have kanji were also cited as seeming equally suitable in hiragana. ‘Cabbage’ was frequently cited as a loanword which seemed acceptable in hiragana, and indeed, it was also one of the most frequently occurring words in the corpus (13 times), and also accounted for five of the examples of inconsistency, such as the ‘サラダ用きゃべつ/sarada you kyabetsu’(cabbage for salads) pack and the display of cabbages noted in section 6.5 above. Cabbage was also said to be a vegetable that was familiar to Japanese culture as mentioned by Hiromasa (G8) above (section 6.2.7).

Two members of one particular focus group, Jun and Nagisa (G3), had an unusual logic for the acceptability of writing ‘cabbage’ in hiragana. The pronunciation of this word, ‘kyabetsu’, was felt to have differed so far from the native English pronunciation that it had become a ‘Japanese word’, and for this reason, it did not seem unnatural in hiragana. A word such as ‘coffee/kōhī’, they argued, would be understood by a monolingual English speaker, and therefore was a loanword, and should be written in katakana. As Jun explained, ‘If the pronunciation is similar, I feel like it’s just as it was in the foreign language, but if it’s different like “cabbage (kyabetsu)”, it’s Japanese– well, it’s a loanword but it’s not strange to call it “Japanese.”’209 Nagisa agreed, wondering whether the loanword used for ‘bread’ in

208 The third example took an unusual approach in transliterating the word ‘こおふぃ/kofi’. These changes appear to be a move towards a more native-like pronunciation of the word as described in section 6.2.8 above, although a similar spelling has a historical precedent dated 1928 as noted by Hida & Sato (2002). 209 発音して似てたら、外国語のままな気がするし「きゃべつ」とかみたいに変わったら、僕なんかするともう外国、外来- まぁ外来語だけど日本語と言ってもおかしくないのかなぁ 248

Japanese, ‘pan’, would be understood in its native Portugal. This information, it seemed, was necessary for her to decide whether it was a loanword or a Japanese word, and therefore the appropriate script with which to write it. When the researcher commented on the unusual role pronunciation played for these participants in distinguishing loanwords from nativized ‘Japanese words’, Nagisa suggested it may be due to their experience with the English language, as both she and Jun teach it and are proficient speakers.

‘Tomato’ was mentioned by Ryūichi (S5) and Nagisa (G3) as another word which seemed suitable in either hiragana (とまと) or katakana (トマト), and also appeared frequently in the corpus. This word, however, seems to provide a counter example to the argument put forward by Nagisa and Jun (G3) as explained in the previous paragraph, as the loanword ‘tomato’ is phonetically quite similar to its English counterpart.

While the topic of words that would not seem suitable in hiragana was not a topic of the interviews and focus groups, participants sometimes volunteered these examples as they discussed the samples presented, and worked through various hypotheses to explain their intuitions. As a counterexample to ‘きゃべつ/kyabetsu’ (cabbage), ‘れたす/retasu’ (lettuce) was often suggested as a word which would seem weird in hiragana, for example by Ryūichi (S5), Mio and Hiromasa (G8), and indeed it did not appear in the corpus at all. Natsuki (G2) also mentioned that ‘さらだ/sarada’ (salad) would appear strange in hiragana, and Mio (G8) felt the same way about the word ‘ぽてと/poteto’ (potato), explaining this was because it had a strong sense of being an ‘English’ word, although it did appear in the corpus five times. ‘Potato’ is unusual amongst the other fruit and vegetable words mentioned as there is a similar Japanese term, ‘じゃがいも/jagaimo’, as mentioned in section 4.1.1, so the presence of this further alternative may have strengthened the identification of ‘potato/poteto’ as a foreign word for her.

These intuitions about the acceptability of some loanwords in hiragana over others also came to the fore when participants were asked if and how they would correct a 10 year old child,210 or a non-native speaker of Japanese, whose written work contained loanwords in hiragana. Some noted particular words they would or would not correct, for example Hiromasa (G8) stated he would not correct the word ‘きゃべつ/kyabetsu’ (cabbage) in hiragana, but in the

210 By this age, a child would be expected to know the rules and conventions of use for the Japanese scripts. 249

case of ‘れたす/retasu’ (lettuce) would explain to the writer that it was ‘a little weird’.211 Yui (G1) explained she would not correct the word ‘れもん/remon (lemon), but would insist on corrections to both ‘こーひー/kōhī (coffee) and ‘あいらんど/airando’ (island) if they appeared in hiragana, exclaiming with mock severity, ‘I’d say “write that in katakana!”’212

These kinds of comments reflect a kind of ‘scale of correctness’ of loanwords in hiragana, as Natsuki (G2) explained, ‘れもん/remon’ (lemon) is one of the loanwords that is ‘closer to being correct’213 in hiragana. It is therefore possible that other loanwords might represent the middle and ‘incorrect’ end of the proposed scale respectively. The results of this study provide some tentative support for this, including the frequency counts of the corpus in layer one (see chapter 4, section 4.1.1) as well as the responses to the survey, and the focus groups and interviews. Some words such as ‘lemon’ and ‘cabbage’ seem to be more widely accepted, ‘tomato’ produces varied responses, and ‘coffee’ is often said to seem strange in hiragana.214 However, no concrete conclusions of the acceptability of particular loanwords in hiragana can be drawn at this stage, and it is therefore put forward as a suggestion for further research in the conclusion chapter, section 8.3. While some words do seem to be more acceptable in hiragana than others, the very fluidity of the Japanese writing system and its relation to particular lexical items, expanded on in the following section, was one of the main findings of this study.

6.8 Fluidity of script use

In the survey, many respondents noted that the examples of loanwords in hiragana presented were ‘not wrong’, and this sentiment was also expressed by participants in the interviews and focus groups. Their comments suggested that the relationship between script and origin of a word can be quite fluid, with both hiragana and katakana being options for both loanwords and Japanese words. As Natsuki (G2) explained, ‘people might not be conscious that they’re loanwords, they’re just like words that are only ever written in kanji…so with all these kinds of original Japanese words, it’s possible to want to use hiragana or katakana, they’re thought

211 ちょっとおかしいよ 212 「コーヒー」をカタカナにしなさい！ 213 正しいより近い 214 ‘Coffee’ does, however, appear in the corpus of hiragana loanwords. The loanwords that are least acceptable in hiragana would be unlikely to appear in a corpus of hiragana loanwords at all. 250

of as being the same.’215 This suggests a more flexible relationship between script and lexical item than is suggested in many descriptions of the writing system as explained in the introduction (section 1.1.1).

When loanwords appeared in marked hiragana script rather than the usual katakana, they were not felt by any of the participants or interviewees to be a particularly serious transgression of linguistic rules, even in the case of the label of the ‘サラダ用きゃべつ /sarada you kyabetsu’ (cabbage for salads), which some felt to be an error (see section 6.2.2 above). Nagisa (G3) argued that while she believed it was a ‘conversion mistake’ (変換ミス /henkan misu), it was ‘within the acceptable limits’ (許容範囲/kyoyōhani) of the language. Ryūichi (S5) made a similar comment on this example, saying ‘It’s not the kind of mistake that people would say “That’s wrong, fix it!”, most people would just overlook it, it’s just a weird feeling.’216

In the course of the interviews, many participants admitting having never noticed loanwords in hiragana before, Hiromasa (G8), for example, said that in general Japanese people do not notice them, but on being shown them, he felt they were indeed strange (おかしい/okashī). Jun (G3) made a similar comment about the display of half and quarter cabbages with labels in two different scripts described in section 6.5 above, theorising that while they were each acceptable individually, the fact that they were presented in such close proximity was weird (変/hen). Mio (G8) did not even feel this arrangement was problematic, explaining that ‘if you just glance at it, it simply says “cabbage (きゃべつ)” (gesturing at an imagined top shelf)… and “cabbage (キャベツ)” (gesturing at an imagined bottom shelf).’217 The fact that these ‘marked’ transliterations were not noticed by many of the participants also supports the idea that it is ‘not wrong’ to write loanwords in hiragana in some contexts.

Besides being not necessarily incorrect, loanwords in hiragana were also felt to be potentially meaningful linguistic choices, even in the writing of less-expert users of language such as children and non-native speakers. When participants were asked whether they would correct or draw attention to the use of hiragana for a loanword in the writing of a 10 year old child,

215 外来語として意識していないのか…もう、日本語として、外来語というか意識しずに一つの単語も、漢字しかないものと同じような感覚で、日本古来の日本語の単語でも、もしかしたらそのひらがなを使ったり、カタカナを使ったりというもあるかもしれない。それと同じ感覚かもしれない。 216 これ間違っているから直さなくちゃって皆が言うほどのミスでもないから、皆適当に見逃してるんだと思いますけど。何かこれはちょっと不思議な感じする。 217 ぱっと見たら、きゃべつ…キャベツだ。 251

for example, many people such as Jun, Miho, Nagisa (G3), Yūta (G4) and Hiromasa (G8) all reported that if they thought there was a reason (理由/riyū) or aim (狙い/nerai) for this choice of script, they would not. Miho (G3) added that young people might use hiragana for loanwords to be cute (かわいい/kawaii) or to mess around (ふざけてる/fuzaketeru), for example. It is interesting that ‘meaningful choice’ seems to be the first interpretation the use of hiragana for loanwords for these participants, suggesting that it would only be in cases were no reason or aim was discerned, that the possibility of an error was considered.

Even when the non-native writer appeared to be unaware of the conventions of script use, Jun (G3)’s imagined explanation fell short of actually labelling loanwords in hiragana incorrect: ‘I’d say “there’s this way (katakana) of writing too”, I wouldn’t say it was the correct way, I’d just say it was the general way (of writing loanwords).’218 His choice of the word ‘general’ (一般的/ippanteki) over other words such as ‘correct’ or ‘appropriate’ reflects a descriptivist rather than a prescriptivist approach to script use, and one that echoes an understanding that the ‘rules’ of language are the result of the sedimentation of particular ways of script use.

One participant, Ryūichi (S5), noted that the flexibility of script use was one of the special features of the Japanese language:

One good thing about Japanese is, well, to think of it as being able to change the impression of a word by writing it in hiragana or katakana, that’s not a bad thing. So everyone learns “these words are really in katakana”, for example, but then “there’s also this way of writing, too.” To understand the original way, and then accept this (variation) on top of it, is fine.219

Ryūichi’s (S5) statement above refers to the importance of learning the ‘standard’ or unmarked way of writing loanwords, for example, in school, but also the usefulness of understanding the flexibility of the Japanese writing system and the possibilities for variation. Like the other examples of ‘non-standard’ orthography described in section 2.6, it is only

218 明らかに分かっていないでやっていたら、こういう書き方 [カタカナ]もあるよとか、こういう場面では、この使い方のほうが、まぁ正しいってまで言わないけど、一般的だよと言う風に言うかもしれない 219 「表現の一つ」と言う風にまあ… 日本語のいいところだと思えば、別に、同じ言葉でもカタカナによって、ひらがなによって印象を変えることができるというのは、表現力がある言葉だって考えれば別に悪くはないし。だから、「本当はカタカナなんです」っていうのをまず、たとえば、皆が勉強して、でも、「こういう風に使ったりすることもあるよ」っという風に。もともとのことを分かった上でみんなが受け入れていけば別にいいと思う 252

through a knowledge of the ‘standard’ that these transgressions become meaningful, funny, or cool, for example.

Comments on the context in which these written words occurred were also made which mediated the general feelings of acceptability of loanwords in hiragana. For example, Hiromasa (G8) suggested they would not be appropriate in school essays in the case of children. Nagisa (G3) initially stated that she would not draw attention to loanwords in hiragana in the writing of children, but then added she might if she was the Japanese (国語 /kokugo) teacher at a Japanese school. This comment in itself is telling, as it suggests that marked forms might be acceptable in another class, such as science or social studies, but not in the class dedicated to the study of the national language. In the case of adult non-native speakers of Japanese, context was also raised, for example by Natsuki (G2), and Yui (G2), who noted loanwords in hiragana would not be suitable for official (オフィシャル/ofisharu) written work, and Yūsuke (S6) also reported they would not be appropriate for documents in the workplace or ‘normal society’ (普通の社会/futsū no shakai). His proposed method of explaining this distinction showed a nuanced understanding of the context in which the writing appeared, suggesting ‘I would explain that the meaning is the same, but katakana might be more appropriate for the situation...but it’s case by case, it needs to suit both the referent and the circumstances.’220 Yūsuke’s noting not only the context but also the referent of the word is significant, and has parallels to the discussion of the loanword for bread, ‘pan’ in section 6.2.6 above, where hiragana was said to suit the Japanese varieties of bread such as ‘あんぱん/anpan’ (sweet-bean bread), but not those such as ‘クリームパン/kurīmu pan’ (cream-bread).

Participants such as Yūta (G4) also drew attention to the relevance of context of the examples shown in the focus group, noting that many were ‘more like logos than words’221 as a number of these examples were product and business names, suggesting a different frame of reference for these items. Later in this group, Gensuke (G4) also took up this idea, explaining ‘If you saw (loanwords in hiragana) in a document it would be weird, but as a logo or this kind of product name, well, you feel it’s kind of ok.’222 Ryūichi (S5) also expressed reservations about the appropriate contexts for loanwords in hiragana, suggesting that they would be

220 意味は一緒だけどカタカナの方が適切… 場面とか、ケースバイケースって言うんですよね。その状況とかそういう物にあわせますけど 221 言葉よりロゴ 222 文書として見たら変。ロゴとかそういう商品名としたらまぁ別にいいと思うという感じですよね 253

unlikely to be found in genres such as newspaper articles, or television captions (テロップ /teroppu).

Besides a consideration of the context and the referent, a personal preference for a particular script over another was also mentioned as an influential factor in orthography, and further underscores the fluidity of script use in Japan. As Kei (G4) explained, ‘for me personally, I don’t really like seeing Japanese written like this (loanwords in hiragana)… It feels a bit like they’re treating me as stupid.’ 223 In reply to this comment, his group member Gensuke suggested that this dislike might stem from Kei’s advanced English ability. While Kei denied this was an influencing factor, rephrasing his belief that he was being made fun of (なめてる /nameteru) by being addressed with texts containing loanwords in hiragana, it is interesting that for Gensuke at least, proficiency in the source language, and perhaps knowledge of the original/rōmaji spelling, was thought to predispose someone to dislike this orthographic variant.

Natsuki (G2) also mentioned the role of ‘personal difference’ (個人差/kojinsa) in preference in script use, explaining:

There’s also room for personal preference with this. Words like ‘cabbage’…some people would not be ok with anything other than katakana, and then there’s people who are fine with hiragana. I think there’s some words with room for individual difference, and there’s others that the majority of people would agree are not fitting in anything but katakana.224

This comment ties together influences from both the word in question, as described in section 6.7 above, with the individual preferences of the writer or reader of the word. Context, as an additional factor, was also described in this section. Therefore, the use of hiragana for a loanword is subject to many considerations, all of which result in a certain fluidity of script use for users of the language.

223 個人的に僕もあまり日本語のこういう表記が好きじゃない… 何か馬鹿にされたというか… 224 個人差もあると思うんですよね。「キャベツ」が絶対カタカナじゃないといやの人もいるだろうし、ひらがなでもオッケー人もいる。その個人差があるものがあれば、ほとんどの人カタカナじゃないとしゅっくりこないものもあるだろうし。何なんだろうその差が… 254

6.9 Summary of layer 4: interpretations of loanwords in hiragana

The findings from the interviews and focus groups serve to provide an in-depth understanding of how participants interpreted and created meanings from representations of loanwords in hiragana, and therefore related to the third research question. They were also used to triangulate the results of the other layers of the research project, most directly they expanded the results of the online survey (layer 3), but they were also useful in gauging whether the typical traits identified from the texts collected for the corpus (layer 1) were representative of the kinds of texts participants encountered in their daily lives. The comments made by the participants were also relevant to the case studies (layer 2), as many drew attention to other semiotic resources in the text such as font or colour. The interview and focus group findings built on those of the previous layers by offering a deeper understanding of how loanwords in hiragana are perceived by native speakers of the language.

Script was found to be a powerful semiotic resource, having the power to suggest different referential meanings, for example whether a ‘lounge’ is a ‘normal’ lounge or a hostess bar (or even a host bar); or affect the interpersonal message of a piece of writing, making statements such as ‘I love you’ or ‘fuck you’ playful rather than serious, or express a sense of familiarity with a foreign country such as ‘New Zealand’. The choice of hiragana script was also reported to be capable of making a foreign dish such as curry more ‘Japanese’, and yet the careful transliteration of some loaned words and phrases in hiragana suggests a move toward authentic ‘foreign’ speech, away from ‘katakana English’ in both script and pronunciation. The choice of script was also found to be more closely linked to personal intuition and beliefs than it was to rules or conventions concerning the origins of particular loanwords. However, context was also felt to be important, with many participants mediating their comments on the orthographic options available to the writer by discussing them in terms of specific genres, including product packaging, business names, and social media.

255

Chapter 7: Discussion

In the Routledge Handbook of Applied Linguistics, van Leeuwen (2011) suggests that future research in multimodality should take a more interdisciplinary approach, engaging with fields such as art and design theory or functional linguistics. He also argues that studies should explore cultures outside the western writing tradition in order to promote cultural diversity within the field. The current research project has strived to do both, engaging with a range of theoretical bases as well as engaging with a semiotic resource specific to the Japanese language. In this way, it hopes to play a small part in the development of the field of multimodal studies by describing findings from a non-western cultural tradition, namely, how script can be used as a semiotic resource within Japanese written texts.

The discussion chapter will locate the findings of this study within its broader theoretical context by linking them to key theoretical concepts and debates such as translingualism and the status of English in Japan, as well as discussing why and how semiotic resources are deployed in multimodal texts. Loanwords in hiragana can be seen as performing the metafunctions described by Halliday (1978) and others in the field of SFL such as Eggins (2004), however, these metafunctions should be understood as overlapping rather than discrete, even if one function is foregrounded in the interpretation of a particular text.225

This chapter will firstly discuss the often reported belief that it was ‘not wrong’ to write loanwords in hiragana, linking this idea to the translingual and trans-orthographical place of these loanwords within the Japanese language. The functions of loanwords in hiragana will then be discussed, first in relation to the socio-historical associations of the hiragana script, and secondly with a focus on how meanings beyond these associations are revealed by examining loanwords in hiragana in context. Next, how script use in Japanese writing can be considered a mode within a multimodal text will be explained. The chapter will end with some final thoughts on the role of technology in the sedimentation of language, and how it can form the basis for future research on the topic of loanwords in hiragana. This echoes van Leeuwen’s third suggestion for establishing the field of multimodal studies, engaging with technology.

225 This approach was also taken by Zappavigna (2015, p. 6) in her analysis of the metafunctions of hashtags in twitter posts, where she explained ‘while it is useful to isolate these functions for the purpose of analysis, they are in fact enacted simultaneously in any linguistic performance, and are not mutually exclusive.’ 256

7.1 ‘Not wrong’: the acceptability of loanwords in hiragana

The link between loanwords and katakana script is established to the degree that these foreign words are also known as ‘カタカナ語/katakana go’ (katakana words) in Japan, and therefore an unexpected finding of the current research project was the general degree of acceptability reported for these examples of atypical script use in the contexts in which they were found.

The survey was the main method used to investigate perceived appropriateness of loanwords in hiragana, and across the three sample texts used in the survey, the average number of people who judged them to be appropriate was 73%. One of the most frequently given reasons for appropriacy across all three texts was that writing a loanword in hiragana was ‘not wrong’, although significantly, not one respondent went as far to say it was ‘right’. In the case of the text containing ‘lemon’ in hiragana, 13 respondents noted that it was not uncommon to see the word written with this script. These findings suggest that the use of particular scripts for particular lexical items may not be as rigid as the term ‘katakana words’ suggests.

The texts used in the survey were all taken from the corpus collated in the first stage of the research project, and therefore constituted authentic contexts of loanwords in hiragana. The appropriacy is therefore related to the kinds of texts the loanwords appeared in, and the referents the hiragana loanwords related to. For example, the corpus analysis demonstrated that product packaging and business signage were frequent sites of hiragana loanwords, and it was examples such as these which were labelled ‘appropriate’ by respondents to the survey. The view that writing loanwords is ‘not wrong’, therefore, refers to words within these contexts, and may not necessarily be extended to others, such as business correspondence, for example. The interviews and focus groups used a wider variety of texts than the survey, and reinforced the finding of general feelings of acceptability of loanwords in hiragana revealed in the survey, with participants not exhibiting reactions of surprise, disbelief or dislike of the majority of examples. Of the 32 texts produced in these sessions, only two examples,226 both representing cases of inconsistency, were met with much surprise, and were felt to be unintentional by some participants.

226 These were the price tag for ‘サラダ用きゃべつ/sarada yō kyabetsu’ (cabbage for salads) and the display of half and quarter cabbages with the word ‘cabbage’ represented in hiragana on one sign, and katakana on the other. 257

The view that it was not necessarily wrong to write loanwords in hiragana was also supported by the number of words collected for the corpus. It contained a total of 628 words across 497 texts, the majority of which were collected by the researcher over 28 days of fieldwork immersed in the linguistic landscape of Japan. While loanwords written in hiragana are still no doubt a very small proportion of the total number of loanwords in written language, the fact that this amount was collected in a broad sweep of relatively short duration suggests that a more detailed analysis of particular domains may uncover many more, both in areas identified by the corpus as sites of loanwords in hiragana (eg. product packaging), as well as potential sites suggested by the interview participants (for example, manga, and pop-song lyrics).

The corpus collected did contain a wide variety of text genres: besides the product packaging and business signage mentioned above there were examples from social media; publication titles including books, manga, magazines; informal written texts such as print club images; and the names of characters and comedians. This also supports the finding of the acceptability of loanwords in hiragana, in other words, they may be ‘not wrong’ for a range of different domains. It is felt that one of the strengths of the research design was to demonstrate the use of loanwords in hiragana in such a wide variety of genres, which were created under a variety of circumstances, from hand-written messages on print club images to computer-mediated input, and from those created by professional marketing teams to those created by individuals on social media.

The findings of the case studies (section 4.2), also supported the idea that it was ‘not wrong’ to write loanwords in hiragana by illustrating that script could be deployed skilfully and meaningfully as part of a multimodal message encoded in a text. Hiragana script has a range of semiotic functions, and the particular function foregrounded was shown to be linked with other semiotic resources in the text, such as colour, shape, typeface, and text direction. Cues such as these led to two very different interpretations of similar types of texts, both were examples of product labels containing loanwords in hiragana, despite the fact that the actual product name was written using standard katakana: ‘宇治抹茶かすてら/uji matcha kasutera’ (Uji green-tea castella) and ‘サラダ用きゃべつ/sarada yō kyabetsu’ (cabbage for salads). The first case was sometimes judged to be purposeful, drawing on the Japanese connotations of hiragana and the product’s relation to the Japanese district famous for traditional green tea, while the second case was often judged to be a ‘typo’ or ‘ 変換ミス/henkan misu’

258

(conversion error), there being no other semiotic resources felt to support the use of non- standard hiragana as a meaningful choice. In other words, readers engaged with these texts multimodally, leading them to distinguish a case where hiragana script was used as an acceptable, meaningful choice for a loanword, and a case where it was a suspected error, despite the similar orthographic context of these texts.

However, within the general feelings of acceptability cited so far, the findings suggested that some words were ‘less wrong’ in hiragana than others. As one of the interview participants explained, there were some loanwords which had room for individual difference when it came to the appropriateness of writing them in hiragana, while there were other words that the majority of people would agree should be in katakana. This suggests the idea of a ‘scale of correctness’ for loanwords in hiragana, with words such as ‘きゃべつ/kyabetsu’ (cabbage) and ‘れもん/remon’ (lemon) consistently said to be unproblematic in hiragana, appearing higher up on the scale. The relative frequency of seeing the word ‘lemon’ in hiragana was also reported in the survey, with 13 respondents giving this as a reason for its appropriacy. ‘Lemon’ and ‘cabbage’ were also two of the most frequently occurring in the corpus, this finding is discussed in section 6.7.

The fact that ‘lemon’ could also be written in kanji (檸檬/remon) was also frequently given as a reason for the word’s appropriacy in hiragana across both the survey and interview findings; the logic seemed suggest that words which had a kanji are ‘Japanese’, and ‘Japanese’ words can be written in hiragana. This type of ‘folk-linguistics’ is also noted by Hosokawa (2015), who suggests that factors outside a word’s etymological origins such as script may affect the identification of a word as ‘Japanese’ or ‘foreign’. The presence of a kanji, however, was not the only factor affecting a word’s position on the cline of acceptability in hiragana, ‘珈琲/kōhī’ (coffee), for example, can also be written in kanji, and was even stated to be a word which was ‘becoming Japanese’ by some survey respondents, but was not often felt to be acceptable in hiragana, and appeared only three times in the corpus. Some interview participants suggested this was because of the two long vowels, which, as described in previous sections, necessitate a choice of katakana-style bars or hiragana-style extra vowel characters, and the three times the word appeared in the corpus did show some variety in this respect. However, the corpus demonstrated that the presence of a long vowel did not necessarily preclude a word’s appearance in hiragana, with 26% of the words containing long

259

vowels, although the variation in the ways these sounds were transcribed may suggests some discrepancy in how such sounds should be written in hiragana loanwords.

Interestingly, while ‘れもん/remon’ (lemon) was one of the most frequently occurring loanwords in hiragana, it was never used to refer to the actual fruit, for example in the fresh produce section of a supermarket as the ‘きゃべつ/kyabetsu’ (cabbage) and ‘とまと/tomato (tomato) examples often did, but only to lemon-flavoured drinks and lollies. Similarly, the word ‘ぽてと/poteto’ (potato) was only used to refer to processed food or menu items, rather than the raw vegetable, for which the Japanese term ‘じゃがいも/jagaimo’ was consistently used. The presence or absence of words in the corpus is therefore only one lens through which to consider the phenomenon of loanwords in hiragana, the referential meanings of these words as well as the genres in which they occur should also be considered. The cases of ‘lemon’ and ‘potato’ not being used to describe the raw ingredients may also relate to the suggestion of ‘un-realness’ related to the use of hiragana for loanwords, which is further explored in a later section (7.3.2).

Overall, the fact that loanwords in hiragana were frequently judged to be ‘not wrong’ as part of the survey suggests the blurring of the boundaries between Japanese words in hiragana script and loanwords in katakana script. This finding was supported by the general reactions and specific comments from the interview and focus group participants, as well as the breadth and diversity of the corpus. The case studies also showed how loanwords in hiragana as semiotic resources were skilfully deployed amongst other resources in the texts. The blurring of lexicon/script boundaries will be explored in the following section, which discusses the role orthography plays in translingualism.

7.2 Trans-orthography

While the second research question sought to ascertain in what contexts loanwords in hiragana occurred, the answer lies not only in the genres of texts represented in the corpus, but in the broader cultural context in which these words are arising. Loanwords in Japanese, prime examples of translingual lexical items, can also be said to be ‘trans-orthographic’ when they appear in hiragana, problematising the role of script in demarcating ‘Japanese words’ and ‘foreign words’, and underscoring the flexibility of the Japanese writing system.

260

As early as 1995, Honna argued that loanwords exhibited a ‘drive for Japanization’:

…what is conspicuous in these borrowing patterns is the drive for Japanization. Although English loans are superficially recognized as they are due to their katakana representation, they are structurally and semantically treated as Japanese words. And (sic) they are incorporated into Japanese. They have become part of the Japanese lexicon and grammar…’

Honna (1995, p. 52)

As Honna notes, the source word may undergo changes in phonology as well as semantic shift/narrowing/expansion when it becomes used as a Japanese loanword, as explained in section 2.5, and are therefore ‘semantically’ and ‘structurally’ part of the Japanese language. However, the use of the katakana still gives these words an ambivalent status between a Japanese and a foreign word. On the one hand, they appear in Japanese characters, as opposed to the Roman (or other) alphabet, marking their transition into the Japanese language. On the other, the katakana script still marks these words as ‘foreign’, as non-Japanese words, with Kay (1995) claiming that the hiragana and katakana scripts divide the Japanese lexicon into foreign and native words as neatly as rooms in a Japanese house and their furnishings are divided into ‘Western’ and ‘Japanese’ style.

For this reason, the presence of loanwords in hiragana, the topic of the current study, is of importance, as these words question this clear-cut distinction between loanwords and Japanese words, as they represent cases where this orthographic marker of ‘foreignness’ is removed. When a particular loanword is different to the source word in both pronunciation and meaning, the katakana script may be the last marker of foreign-ness. This shift from (less) foreign to (more) native is only apparent through the written form, suggesting that ‘trans-orthography’ might more accurately represent the phenomenon described.

The translingual and trans-orthographic status of loanwords in hiragana suggests an paradoxical position between ‘native’ and ‘foreign’, although the boundaries between these terms should be understood as relational rather than absolute. Hosokawa (2015) draws on these relational boundaries to describe loanwords as an ‘internal Other’ to a Japanese ‘self’ (2015, p. 51). An established loanword in katakana, being ‘internalized’ into the language is therefore ‘more Japanese’ than a foreign word in the roman alphabet, for example, but its

261

status as ‘Other’ means that it is simultaneously less Japanese than a Sino-Japanese word. These ‘漢語/kango’ in turn might be considered less Japanese than the lexicon untouched by Chinese influence, the ‘大和言葉/yamato-kotoba’. Similarly, relational boundaries can be seen between loanwords in hiragana and those in katakana, the hiragana words being further along the cline towards ‘Japanese words’, or Hosokawa’s ‘Japanese “self”’.

Evidence for this shift towards the ‘Japanese self’ can be seen in the associations of ‘Japanese-ness’ of the hiragana script reported in the survey (section 5.2.1) as well as previous studies (see section 2.4), and furthermore in the feelings of closeness or familiarity reported across many examples of loanwords in hiragana. The relational boundaries of the acceptability of hiragana loanwords are also suggested by the ‘scale of correctness’ described in the previous section. Loanwords which can also be written in kanji, such as ‘lemon’, exploit their trans-orthographic nature to the fullest, the script used being able to connote foreign (English speaking) culture (LEMON or レモン/remon), Japanese culture (れもん/remon) or even Chinese culture (檸檬/remon) through the socio-historical associations of the script.

The kanji characters themselves are a further example of trans-orthography, and although more frequently associated with Chinese culture than Japanese culture (see section 5.2.3), the presence of a kanji alternative was also cited as a marker of ‘Japanese-ness’ in the case of loanwords. From the survey responses to the texts containing loanwords that could also be written in kanji, ‘檸檬/remon’ (lemon) and ‘珈琲/kōhī’ (coffee), two different kinds of opinions were noted, the first stating that these were Japanese words, and the second stating that they were becoming Japanese words. For example, the following comments were made about the word ‘lemon’: ‘It feels like a word which is transitioning from a loanword to a Japanese word’,227 and ‘It’s already treated like a Japanese word, not a loanword.’228 This difference in nuance, although outside the scope of this study to pursue in more detail, is considered significant in a discussion of the translingual, trans-orthographic status of loanwords described in this chapter, as it captures the changing position of these words within the Japanese language. The kanji characters which formed the first writing system of Japan also set a precedent for a culture of borrowing, absorption, and indigenization of new terms, as noted by Hoffer (1980), and many of the respondents cited by Hosokawa (2015).

227外来語から日本語へと移行した言葉だと感じる 228 すでに外来語ではなく日本語として扱われているから 262

Although these authors focussed on loanwords in standard katakana, it is argued that those in hiragana provide a particularly apt example of the continuation of this practice.

Loanwords written in hiragana are particularly important in the context of the number of studies on the role of English/rōmaji in countries such as Japan, where its decorative and emblematic functions have been the main focus (eg. Blommaert, 2012; Hyde, 2002; James, 2014; Stanlaw, 1987). Evidence of the consideration of English/rōmaji as an alternative option also appeared in this study, for example the survey respondents who cited the English version of the word ‘Happiness’ being difficult as a motivation for the use of hiragana this word appeared in the name of the clinic (see section 5.3.2), and the frequency with which ‘not knowing the English spelling’ was given as a reason for the use of hiragana for a loanword in the interviews and focus groups (section 6.2.1). While loanwords in this study are indeed part of the ‘polylingual blend’ described by Blommaert (2012), they do not appear in rōmaji/English like the other emblematic tokens he describes, nor do they exploit ‘anglography’ or ‘English as a visual language’ as described by James (2014). Inoue (2005) argues that the orthographic shift from katakana to rōmaji/English is part of the process of the ‘modernization of Japanese’ (2005, p. 157), however loanwords in hiragana in fact represent movement in the opposite direction, it is in fact an orthographical shift towards Japanese-ness, towards a script strongly linked to traditional Japanese culture and language. While Hosokawa (2015) calls loanwords ‘the outsiders within’ the Japanese language, the move to the hiragana script suggests a move towards assimilation with other Japanese lexical items.

Despite the focus of this research not being on words in rōmaji or katakana, the findings do contribute to our understanding of the role played by English in Japan, as the majority of the loanwords in the corpus were of English origin (82%, see section 4.1.5). This is consistent with the large number of loanwords from English, in other words, the ratio of English hiragana loanwords in the corpus in this study appears to be proportionate to the ratio of English loanwords used in Japanese in total.

However, an alternate interpretation would ask whether loanwords from English are actually more likely to appear in hiragana than words from other languages. For example, since English is considered to be a ‘global language’, and is more widely studied in Japan both in the school system and through private classes for adults (英会話/eikaiwa), it is likely that the Japanese are more familiar with English words than those from other languages. Since the burden of comprehension is slightly higher with a hiragana loanword than its unmarked

263

katakana counterpart, a less familiar loanword may be considered too confusing in its hiragana form. These less familiar loanwords might be more likely to appear in the standard script for foreign words, katakana, or in rōmaji, to more clearly mark them as non-Japanese words. A second factor in considering the ‘global’ status of English is that it may be more open to ownership and the indigenising effects of the hiragana script; English lexical items might be more available to be ‘owned’ by the Japanese while those from French or Italian, for example, continue to be ‘loaned’.

With an awareness of language as a local practice (Pennycook, 2010), and a growing feeling of familiarity and therefore ownership of these once ‘alien’ words, the ‘foreignness’ of these terms may no longer need to be marked through the script, suggesting a relaxation of the established link between loanwords and katakana. The implications of this are further discussed in the conclusion.

7.3 The functions of loanwords in hiragana

The third research question was designed to gain insight into why loanwords appear in hiragana, and chapters 5 and 6, the findings from the survey and interviews/focus groups, outlined how the functions of hiragana emerged through the analyses of each layer of the research design. The survey, and interviews and focus groups were of most relevance in investigating the functions performed by loanwords in hiragana, however, these results were also supported and triangulated through the corpus analysis and the case studies.

The variety of meanings reported to be associated with the hiragana script represent the social semiotic nature of meaning making, these are ‘motivated signs’ where individuals bring together form and meaning in particular social contexts. The functions revealed by the current study show how the use of script, in this particular case for loanwords, is capable of fulfilling all three of the metafunctions proposed by Halliday (1978) in relation to language, which have been extended to other semiotic resources such as image (Kress & van Leeuwen, 2006) and typography (Stöckl, 2005).

These functions can roughly be divided into two groups: functions associated with the socio- historical associations of the hiragana script, and functions which were related to the

264

communicative context in which these words were found. These will be described in the following sections.

7.3.1 Socio-historical associations of hiragana

The survey investigated the associations of the hiragana script, with words such as soft, gentle, easy, children, cute, Japan, and feminine appearing frequently in the responses (findings section 5.2.1), these answers are consistent with previous studies on the associations of the script, such as Iwahara et. al (2003), which are summarized in section 2.4. The functions of loanwords in hiragana were often reported to be linked to these associations, in other words, by putting a loanword in hiragana, certain connotations such as ‘Japanese-ness’ or ‘gentleness’ could be layered on to the original form. This parallels the ‘special katakana words’ (特殊カタカナ語/tokushu katakana-go) described by Akizuki (2005), which are words other than loanwords written in katakana, such as ‘スミマセン/sumimasen’ (sorry). Akizuki argues that ‘special kakakana words’ allow the author to both avoid the connotations of hiragana or kanji, and draw upon those of katakana.

At first glance, this interpretation seems more closely aligned with a structuralist, Saussurean view of semiotics, namely that these signs are made up of a signifier (hiragana script) with a variety of signifieds (gentle, familiar, and so on) rather than the social semiotic approach to meaning making that forms the basis for this research project. However, to ask how the signifier ‘hiragana’ can be reported to mean ‘Japanese-like’ in one text and ‘aimed at children’ in another, or to account for differences in opinion between individuals, reinforces how meanings are created in particular contexts by particular individuals, rather than through any intrinsic properties of the semiotic resource itself.

The following sections will discuss some of the common associations of hiragana script, Japanese-ness, femininity, and familiarity; and how they relate to the metafunctions.

Japanese-ness

The hiragana script has strong connections to traditional Japanese culture through its associations with ‘大和言葉/yamato-kotoba’, the language of Japan which pre-dated contact with the Chinese (Loveday, 1996), as well as its relational position with the other scripts,

265

katakana being linked to foreign words (loanwords) and kanji having historical connections to Chinese culture. The case study of the drink ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) (section 4.2.2) drew on the connotations of Japanese culture through the hiragana script as well as a variety of other resources, the brush-like calligraphic typeface, and the vertical script, for example. These and other similar cues were found in many other texts in the corpus, for example, those associated with locally grown fresh produce, omiyage (souvenir) products, and texts associated traditional Japanese customs such as the Girls’ Festival (ひな祭り/hina matsuri) or the summer gift giving season (お中元/ochūgen). A particular type of Japanese-ness, one connected with long standing traditions and culture, is therefore also suggested, and is interesting to contrast with the survey item which contained the loanword ‘hankachi’ in rōmaji, for example. This was described by one respondent as ‘Japanese, but still modern’ suggesting that ‘Japanese’ is ‘traditional’ by default.

These examples suggest how hiragana can perform the ideational function, telling the reader something about the origins or cultural alignment of the referent. Furthermore, as a marker of Japanese-ness to a Japanese audience, evidence of the interpersonal function can also be seen; the use of this script establishes the relationship between the writer and reader as part of a shared Japanese history and culture. This, however, should also be problematised from a social semiotic perspective, as the claim to a ‘shared culture’ could be questioned by those from social groups often excluded from this shared Japanese-ness: Japanese of mixed racial heritage, or ‘returnees’ (帰国子女/kikokushijo), for example.

Within the general association of ‘Japanese-ness’, the way in which it was applied to a referent did show some variety. For example, it could make a referent seem ‘more Japanese’, with terms such as ‘和風/wafū’ (Japanese style), and ‘日本っぽい /nihonppoi’ (Japanese-y) being commonly reported in the survey and interviews, especially in products associated with Japanese culture, such as ‘Green-tea pudding’ (抹茶ぷりん/ matcha purin). It was also interpreted with a change of emphasis as ‘made in Japan’ (日本製/nihonsei).229

Other respondents and interview participants felt that the effect of hiragana was to show that a foreign product was suitable for Japanese people, or had been adjusted to their tastes; this was reported for the instant curry ‘すぱいす/supaisu’ (Spice), and also for the clinic

229 This is not necessarily ‘Japanese-y’, many cars or electronic good made in Japan (日本製/nihonsei) would not be considered to be ‘和風/wafū’ (Japanese-y), for example. 266

‘Happiness’ (はぴねす/hapinesu) being perceived to provide ‘Japanese-style service’. The association of hiragana with ‘Japanese-ness’, therefore, shows slight variations of nuance in its application, findings that could only be generated by engaging with a range of texts and research approaches as were used in this study.

The sense of ‘Japanese-ness’ described above suggests the cultural appropriation which Androutsopoulos (2000) noted in the use of ‘Germanized’ spellings of English words. However, he also noted a secondary and quite different function, the same style of Germanized spelling being used to mark the speech of a non-initiate of the subculture, unfamiliar with or unable to replicate the pronunciation of the borrowed English term. This function was also found in one particular example of a loanword in hiragana, the sticker with the phrase ‘I can ふらい (furai/fly)’. Some respondents identified this as a ‘Japanese accent’,230 and even criticized the designers of the sticker for seeming to endorse ‘katakana English’. 231 This derisive use of localized orthography (Germanised spelling or hiragana script) to mark disfluent or non-native line pronunciation is yet another facet of the use of hiragana as a marker of ‘Japanese-ness’.

Femininity

Femininity and a historical connection to women’s writing is a commonly cited association of hiragana (Iwahara et al., 2003; Miyake, 2007; Smith & Schmidt, 1996). However, this trait also showed variation in the way it was interpreted by the survey respondents and interview/focus group participants. While the texts in the corpus could not reliably be coded as being aimed at women or men, some items such as cosmetics and toiletries, or personal accessories such as hand-towels and handkerchiefs, did seem to be appealing to women more than men, and used conventional semiotic resources associated with products aimed at women, such as pastel colours, puffy typefaces or ‘maru-moji’, and motifs such as hearts and flowers, for example. The word ‘cute’ (かわいい/kawaii) was often mentioned in relation to hiragana in these cases. The use of hiragana to suggest femininity also suggests the ideational function of hiragana script, giving the message that the referent is ‘designed for women’, for example.

230日本人的な発音のイメージ 231For example, 英語の発音をひらがなで当て字のように当てはめてしまうと、日本人は日本語訛りの英語しか話せなくなる. (If hiragana is used to spell out English words, Japanese people will only ever be able to speak English with a Japanese accent.) 267

However, like the use of hiragana as a marker of ‘Japanese-ness’ described above, the script was used to connote femininity in different ways. Rather than appealing to women as a target market, in the case of hostess bars, the feminising influence of hiragana was felt to refer to the fact that there were women inside these establishments. What these businesses were selling, in other words, was not just the alcoholic drinks but the attention of a female companion; this was how they differentiated themselves from other bars, as one interviewee explained. Building on the feminising influence of hiragana script, one interviewee also reasoned that a host bar with a name in hiragana script would suggest a gay host bar, thereby suggesting a stereotypically emasculated image of male homosexuality.

As with the previous example, the precise interpretation of the often cited association of hiragana as ‘feminine’ or connected with women was shown to have different facets depending on the text and the person interpreting it. This reinforces the role of social semiotics, as this approach highlights the fact that meaning is made by individual members of society, rather than being intrinsic to the properties of particular scripts or other semiotic resources.

Familiarity

While ‘familiarity’ was not a commonly cited association of hiragana in the literature, it was described as ‘intimate’ by Sherry and Camargo (1987), and ‘intimate’ and ‘private’ by Robertson (2015). ‘Familiarity’ was a code present in the survey findings for the associations of hiragana, albeit only six times, and may also overlap with a number of other codes which appeared in larger numbers, for example ‘Japanese-ness’ (18), ‘foundational/basic’ (8), and also ‘child’ (45), and ‘beginner’ (5). The use of hiragana to connote ‘familiarity’ foregrounds the interpersonal function, suggesting a long-standing and intimate relationship, real or imagined, between the referent and the reader.

While ‘familiarity’ was not frequently given as an association of the hiragana script, it was often mentioned as an effect when a loanwords was written in hiragana. Creating a sense of (pseudo-) familiarity with a product or business was a commonly given as a reason for the use of hiragana in two of the survey texts, ‘はぴねす/hapinesu’ (Happiness) clinic and ‘蜂蜜れもん/hachimitsu remon’ (Honey Lemon) drink. The use of hiragana as a marker of familiarity was also evident in the interview responses, for example many participants noted it gave a friendly or familiar feeling to a business name written in this script. This was often mentioned in relation to the large number of hostess bars that used hiragana loanwords as part 268

of their name, where it was suggested that this sense of familiarity made these establishments seem ‘easy to enter’ and/or a place to relax and be oneself.

In the case of online communication, hiragana was reported to be especially used with close friends, therefore also reflecting a sense of familiarity within the relationship. Loanwords in hiragana were also said to express a sense of intimacy with the referent, for example, that the country ‘にゅーずぃーらんどぅ/nyū zīrandou’ (New Zealand) was a place the writer knew well, or that the obscene phrase written in hiragana across the print club photo was an act of bringing the expression ‘closer to themselves 232 ’ rather than just mimicking a phrase associated with western culture (see section 6.2.7).

Applying the associations of hiragana such as familiarity, femininity, or Japanese-ness to a word normally in katakana is one aspect of why loanwords are sometimes written in hiragana. However, to limit the interpretation of this phenomenon to only these associations would be to overlook the variety of hereto undescribed functions that these words are performing, many of which are based in the contexts in which they occur. The following section will discuss these additional functions.

7.3.2 Interpreting loanwords in hiragana in context

In the following sections, the functions of hiragana when used for loanwords beyond the conventional associations of the script will be discussed. Many of these examples rely on the contexts in which they occur, providing a deeper understanding of how script plays a role in multimodal texts.

Emphasis

Hiragana was reported to have an emphasising effect, making product or business names stand out amongst their competitors, and was also mentioned in relation to posts on social media, where it was felt to make a particular post stand out amongst other similar texts.

Product and business names are well represented in the corpus, and business names are often sites of non-standard spelling in English contexts, for example ‘Krispy Kreme Donuts’; Davies (1987) notes a number of examples of this kind. Loanwords in hiragana in particular

232 もっと自分に身近にして 269

present a kind of language puzzle, readers may experience a sense of achievement for decoding the marked hiragana loanword. This theory was suggested by Rebuck (2002) in discussing unusual loanwords in Japanese, arguing that deciphering the meaning of a word such as ‘オートクチュール/ōtokuchūru’ (haute couture) may give the reader intellectual satisfaction, in a similar way to solving a crossword puzzle. This seems to be even more true of a loanword in hiragana, especially if it involves an element of language play, as described in a later section.

However, the technique of using loanwords in hiragana for product names was reported to be overused and hackneyed by one of the survey respondents, who explained ‘because now an exceedingly large number of companies are using this type of expression (loanwords in hiragana), the truth is that the effectiveness of this technique is wearing thin.’ 233 The predominance of product and business names in the corpus supports this perception of the popularity, and perhaps saturation point, of this marketing technique.

The use of hiragana for loanwords as an emphatic device is not only apparent from their use as brand names, but their comparative salience to other elements in the text due to their size or colour, for example, marking them as important, which is an expression of the ideational metafunction (Kress & van Leeuwen, 2006). This was demonstrated in the case study for the drink ‘蜂蜜れもん/hachimitsu remon’, where the word ‘れもん/remon’ (lemon) appears in a typeface twice the size of the next largest word in that text.

The use of hiragana for emphasis is interesting because this a function normally ascribed to katakana. Katakana’s emphatic effect is sometimes compared to how italics are used in English; Walker (2000, p. 12) notes that italics are used for ‘a number of communicative functions, including distinction (when used for a book title), differentiation (when used to denote a foreign word), and emphasis (to draw attention to a particular word).’ Indeed, katakana is often used in these ways to emphasize a word which normally appears in kanji or hiragana: as a general emphasising device, to highlight foreign (loan) words, and to mark onomatopoeia, for example. The problem therefore arises of how to emphasize a word when its default option is ‘emphasized’, as in the case of a conventional loanword in katakana. The widespread use of loanwords in Japanese has meant that the emphatic function of katakana in the context of words of foreign origin is unlikely to be as strong as when these words first entered common use.

233今は非常に多くの会社がこの表現を用いるので、そのデザイン効果がうすれていっているのが事実 270

The answer of how to emphasize a katakana loanword is, possibly, to transcribe it in hiragana, despite this script being described as ‘basic’ or ‘normal’ by the survey respondents (see section 5.2.1). The ‘emphasising’ effect comes not from the script itself, but the unexpected combination of the hiragana script with a foreign loanword. It is the context, in this case, the word being written, which allows the reader to interpret the reason for the use of hiragana as being an emphatic device. This interpretation may be backed up by the word’s salience within the text as described above, for example, or by being placed first in many of the social media examples, such as that described in case study 4.

Unseriousness and Un-realness

Like the emphatic function described above, the ‘un-real’ or playful effect of hiragana was reported across a variety of texts in the corpus. Both the variety of genres and contexts where this aspect was mentioned as well as the relative frequency of these interpretations make this function of hiragana one of the most significant in the study. These examples represent two types of ‘un-realness’, the first a joking, kidding-around quality, and the second a non- conventional or non-literal interpretation of the word. Examples of each will be described in the following sections.

The ‘joking’ or ‘messing around’ quality was most frequently expressed by the Japanese word ‘ふざける/fuzaketeru’. The use of hiragana for loanwords was reported to make the obscene phrase ‘fuck you’ (ふぁっきゅー/fakkyū) seem like a joke; the declaration ‘I love you’ (あいらびゅ/airabyu) to be less meaningful than its katakana or English versions; and the phrase ‘yes I am’ (いえすあいあむ/iesu ai amu) to appear modest when accepting a compliment. All of these examples from the corpus were found in genres associated with informal communication between peers, such as social media posts and print club images. In online communication, loanwords in hiragana may therefore be used to represent absent verbal cues such as facial expressions or a ludic voice quality; they are represented orthographically, in this case, though a marked use of script. This has parallels to how Danesi (2016) argues that emoji play a role in replacing cues such as voice quality or facial expression missing from online communication. For example a smiley face () can indicate that the preceding statement is a joke, or not meant to be taken seriously, while in face-to- face communication this ‘joking’ or ‘unserious’ quality could be represented by eye-rolling or laughing. In the examples above, it is not hard to imagine these phrases being accompanied by emoji in English-speaking contexts to represent the same joking quality, for

271

example ‘Yes I am ;)’. The idea of ‘un-realness’, or the seriousness with which a message should be taken, relates to modality and therefore the interpersonal function, in other words modality is represented orthographically in these examples. Hiragana loanwords are used in the above examples to establish the tenor of the speech event and the relationship between the persons involved.

The second aspect of ‘un-realness’ foregrounds the ideational function, although, as mentioned in the introduction to this chapter, these categories are not mutually exclusive and overlap can occur. In these cases, the hiragana and katakana versions of a word were reported to have different referents, in Saussure’s terms, they are different signifiers with different signifieds.

In the case of the drink ‘蜂蜜れもん/hachimitsu remon’, nine survey respondents suggested that the designer had used hiragana for the product name because the katakana word referred to the real fruit, appearing in the phrase ‘国産レモン使用/kokusan remon shiyō’ (made with domestic lemons) elsewhere in the text. A similar distinction was made by participants in the interviews between the hiragana and katakana versions of the word ‘spice’ which appeared on the packet of curry; the hiragana word (すぱいす/supaisu) was reported to refer to the product name,234 while the katakana word (スパイス/supaisu) referred to the spices used in producing it, such as turmeric and cardamom.

This notion of a non-literal or non-conventional meaning as a function of loanwords in hiragana can be seen in other business names appearing in the corpus, the hair salon ‘みるく /miruku’ (Milk) may be emphasising that they do not actually sell milk; and the dog-trimming salon called ‘ほっと DOG/hotto DOG’ (hot dog), it is hoped, has no connection to the actual sausage product. Hostess bars were another example of this ‘un-realness’, those with the word ‘らうんじ/raunji’ (lounge) in hiragana on their business signage were described as ‘not that kind of lounge…’, for example.

While not mentioned by any of the participants, the product line ‘みっくちゅじゅーちゅ /mikkuchu jyūchu’ (Mix Juice) which produces the variant ‘Mix Juice Sour’ present in the corpus also revealed an interesting aspect of un-realness. It appears the company originally wanted to call the product ‘ミックスジュース/mikkusu jyūsu’ (Mix Juice), in standard

234 Some participants also pointed out that this was the name of the Indian restaurant which licensed the curry. 272

katakana and without the baby-like pronunciation. However, as the definition of ‘juice’ was 100% (fruit) juice, and ‘Mix Juice’ contained only 20%, they decided on the current orthography of non-standard hiragana and affected pronunciation235.

In the case of the non-flying hamster, either of these two described aspects of ‘un-realness’ can be seen, depending on the reader’s interpretation of the image. If we accept that the hamster is only pretending to fly, clumsily squashing the squirrel in the process, the interpersonal function of joking or messing around is foregrounded. However, if we interpret the scene as the hamster attempting to fly by jumping on top of the flying squirrel, the ideational function, describing the state of things, is foregrounded, as some survey respondents and interviewees explained; if the hamster claimed ‘I can フライ /furai!!’ (katakana) or ‘I can fly!!’ (English), it would suggest he really could fly, and the punchline of the joke would be lost. In other words, ‘ふらい/furai’ (fly) was felt to have a different meaning to the equivalent word in English or katakana.

Another set of cases in which both metafunctions are realized is the word でーと/dēto ‘date’, which appeared in social media posts and print club images relating to social outings between two female friends. While these might, of course, be same-sex couples on dates, it is also possible to see this hiragana loanword function as a marker of un-realness, either interpersonally, with a joking tone, suggesting a message such as ‘neither of us have boyfriends, so we have to go on dates with each other…!’ or in terms of the ideational function, ‘this is not an actual, literal date (デート/dēto), it’s just a ‘date’ with my friend (でーと/dēto)’

These examples illustrate that the metafunctions are not mutually exclusive, and texts can be interpreted in similar ways while foregrounding different metafunctions. The examples above all contain a sense of un-realness, but this can be represented through the interpersonal function connected to modality, the seriousness with which a statement should be taken, or through the ideational function, where hiragana and katakana versions of the same word differ in the authenticity of their referents.

235 See http://www.sangaria.co.jp/mikkuchu/index.html; and https://ja.wikipedia.org/wiki/みっくちゅじゅーちゅ 273

Creating or obscuring word boundaries

Placing a loanword in hiragana can either create or obscure a word boundary, depending on the script used for the surrounding words. This is evidence of the textual metafunction as it contributes to the cohesion of the text as a whole, and relates to the number of references to a ‘sense of balance with other scripts’ reported across the examples used in the survey. The use of hiragana for loanwords was also noted to create what van Leeuwen (2005, p. 9) would call a visual ‘rhyme’ when its use resulted in the alternation of scripts, for example the words ‘ころころスプラッシュあいらんど/koro supurasshu airando’ (Roly-poly Splash Island) stacked vertically on the toy’s packaging.

Writing a loanword in hiragana can create a word boundary where none would normally exist when two loanwords occur consecutively.236 Hiragana loanwords in combination with those in standard katakana often appeared in the corpus, such as ‘たおるマフラー/taoru mafurā’ (towel muffler), which in standard transcription would result in a long string of katakana characters, ‘タオルマフラー/taoru mafurā’. Thus, having one of these words in hiragana was reported by interviewees to mark the boundaries of the words, making it easier to read. The arrangement of having the first loanword in hiragana was reported to be preferable to the second, the hypothetical transcription ‘タオルまふらー/taoru mafurā’ was described as less effective and even less reliable as a product.

In the case of the product ‘いろいろすーぷ/iroiro sūpu’ (Various Soups), writing the word ‘soup’ in hiragana obscures the word boundary between a Japanese word in hiragana and a loanword in katakana, as written conventionally this phrase would appear as ‘いろいろスープ/iroiro sūpu’. This lack of a clear boundary between the two words in the actual product name was felt to be an attention-seeking strategy by some participants, the slight difficulty or unnaturalness of reading a long set of consecutive hiragana characters forcing the reader to unpack this word more slowly and carefully. In this way it also functions as an emphatic device as described above.

At other times the word boundary, while obscured by the choice of hiragana for a loanword, was defined through other semiotic resources such as colour. For example the sign for the

236 However, sometimes a dot or interpunct (・), called a ‘中点/nakaten’ is used for this very purpose, and is commonly used in the names of non-Japanese people, such as ‘ウィリアム・シェイクスピア/uiriamu・ sheikusupia’ (William Shakespeare). 274

game centre ‘Blazing Land’, (らんらんらんど/ranran rando), used three different colours to separate the phrase into three elements (ran-ran-rando). The combination of marked script use and colour therefore highlights the alliteration between the onomatopoeic ‘らんらん/ran’ (blazing) and the first syllable of the loanword ‘らんど/rando’ (land), while maintaining intelligibility, an effective use of the modes of colour and script.

The functions of marking or obscuring word boundaries are used most creatively in cases of lexical blends, often combinations of loanwords and Japanese words. These lexical blends and other instances of word play are described in the following section.

Marking language play

The large number of homonyms in Japanese means that puns and language play have a long history in Japan (see section 2.7). The findings of the current research have described the role loanwords in hiragana play in bilingual puns, with 38 word/phrase units in the corpus containing an instance of language play. Knospe (2015) illustrates how the non-standard use of orthographical devices such as hyphens and capital letters create ‘problem spots’ which alert the reader to the presence of an instance of word play. It is argued that loanwords in hiragana represent similar ‘problem spots’ which prompt the Japanese reader to uncover the language play encoded.

The corpus contained two broad types of word play; firstly, homonyms and paronyms, for example the name of a brand of heart-shaped cookies with a dove motif on the packaging called ‘はーと/hāto’ (heart), which played on the paronymic ‘はと/hato’ [dove]. The second type consisted of lexical blends or ‘portmanteau words’, for example the brand of rice-chips called ‘こめっぷす/komeppusu’, made from the Japanese ‘米/こめ/kome’ (rice) and the front-clipped loanword ‘ちっぷす/chippusu’ (chips). This example is typical of the lexical blends in being a combination of a loanword and a Japanese word, and therefore, having both katakana and hiragana/kanji in their source words. Because a combination of hiragana and katakana in a single word is highly marked, one script was chosen for the resulting blend, resulting in a (partial) loanword in hiragana. As mentioned above, this is also evidence of the function of hiragana loanwords to obscure word boundaries. The creation of word boundaries where none exist in the loanword for humorous effect can be seen in other examples of word play, such as the homonymic ‘サイだぁー/saidā’ (cider/it’s a rhinoceros).

275

Because lexical blends of Japanese words and loanwords in katakana were not the focus of this study, it is not possible to say to what extent katakana is chosen for the resulting blend. However, examples do occur, such an the 2007 campaign slogan for Roots, a brand of canned coffee, which ran ‘飲スピレーション/insupirēshon’, with the first character of the loanword ‘インスピレーション/insupirēshon’ (inspiration) replaced with the kanji ‘飲/in’ (to drink), making it similar to the concealed words in this study, only identifiable in the written form. It is hypothesized, however, that katakana blends such as these would be less likely to occur, as there is a chance that the katakana segment would simply being interpreted as an unfamiliar loanword, in the above example, ‘スピレーション/supirēshon’ (‘spiration’). In other words, they would be less likely to be identified as a ‘problem spot’, allowing access to the instance of word play as Knospe (2015) suggests.

Knospe (2015) discusses the bilingual English/German puns in German newspapers in the context of an ‘emerging’ bilingualism in Germany. The Japanese situation shows interesting similarities and differences to the German context. Firstly, while the general English proficiency in Germany is ‘emerging’, in Japan it is still reported to be quite low (Aoki, 2017). For this reason, the Japanese instances of word play cannot strictly be called ‘bilingual’, as the non-Japanese elements are not foreign words, but ‘loanwords’, which, as discussed previously, lie at the border of ‘Japanese’ and ‘foreign language’. In other words, these foreign lexical items are already in use in the Japanese language, appearing in loanword dictionaries and so on, and are not new or ‘nonce’ borrowings from a foreign language. However, three of the puns identified in the corpus involved English/English homophones, which presuppose a higher level of English competence than Japanese/English ones.237

Despite these differences, the instances of wordplay described by Knospe (2015) and those that appear in the corpus show a number of similarities: the humour resulting from combinations of native and foreign elements, and the broad audience to whom these jokes seek to appeal. Knospe’s examples are from newspapers, and the examples from the corpus are from genres such as commercial product names, marking these not as ‘in-jokes’ for a small number of bilingual elites, but as a brand of humour felt to be understood and appreciated by the general public. Most relevant to the current study are the ‘additional

237 The homophonic puns with two English words were the maid café ‘まーめいど/māmeido’ (Mermaid); the band name ‘ふぉーゆー/fō yū’, which plays on the English ‘for [you]’ and the fact that the band consisted of ‘four [people named] Yū’); and the dog trimming salon ‘ほっと DOG/hotto DOG’ (hot dog), referencing the takeaway food as well as a fashionable canine. 276

typographical means’ used to highlight an instance of wordplay described by Knospe (2015, p. 177), such as the use of hyphens or capitalization where none would usually occur. ‘Additional typographical means’ also seems to be an apt description of the use of hiragana for marking language play involving loanwords in the Japanese examples described in the previous paragraph.

Language play was also mentioned in the interviews, for example three participants felt that hiragana was used for the children’s rhythmics classroom ‘はぐみぃー/hagu mīi’ (Hug me) in order to reference the near-homonym, ‘はぐくみ/hagukumi’ (raising [a child]). Tangential references to language play were also made by one interviewee in particular, who mentioned seeing the word ‘island’, which appeared in one of the interview prompts, partially in hiragana with the kanji for ‘love’ representing the first syllable (愛らんど/airando).238 The manga series ‘K-on!’, was also mentioned as an example of a loanword in hiragana, and was found to be an instance of language play, referencing the clipped form of ‘軽音楽/keiongaku’, or light (pop) music. To add another layer of language play to this particular example, the pen-name of the author of this series is ‘kakifly’, transliterated as ‘かきふらい/kakifurai’ (fried oysters) and appears in both rōmaji and hiragana on the manga cover pictured in figure 6.2. ‘Kakifly/かきふらい’ therefore also plays on the sound of ‘fly’ and ‘fry’ often mentioned by survey respondents and interviewees in relation to the sticker featuring the words ‘I can ふらい/furai (fly)!!’

Pronunciation and script choice

Considering the role of script in pronunciation problematises the traditional binary concepts of ‘written’ and ‘spoken’ language. In the survey, two associations of the hiragana script related to pronunciation, although these were reported in smaller numbers than the more general connotations described in section 7.3.1. Hiragana was noted to represent ‘(only) sounds’, perhaps in contrast to the meaningful kanji ideograms, and the role of hiragana ‘for reading’ in representing the pronunciation of difficult kanji was also mentioned by some respondents.

In the interviews, the use of hiragana for loanwords was sometimes thought to be used to highlight the sounds of the words themselves, especially in cases where the transliteration of the loanword was also marked or non-standard. This included the name of the children’s

238 This word can therefore be understood as ‘island’ or ‘love-land’, giving the combined message ‘love-island’. 277

rhythmics classroom ‘はぐみぃー/hagu mīi’ (Hug me), which uses two vowel-lengthening devices, a bar and an (half-size) ‘い/i’, and was interpreted in one case as representing the speech of a small child, something like ‘hug meeeeee’ in English.

Some of the examples of loanwords in hiragana represented more accurate transliterations of established loanwords similar to those identified by Akizuki (2005), for example the approximated /z/ sound in ‘にゅーずぃーらんどぅ!/nyū zīrandou!’ (New Zealand!) created by the illicit combination239 of a ‘ず/zu’ and a half-size ‘ぃ/i’, replacing the standard ‘じ/ji’. The case study on the greeting ぐっもーにん/gummōnin (g’mornin’) incorporated a linked sound between the two words ‘good’ and ‘morning’, accomplished with an illicit combination of characters, ‘っも/mmo’, rather than the standard ‘んも/mmo’, as well as removing the final mora ‘グ/gu’ found at the end of the standard transliteration of the word (モーニング/mōningu). Two phrases from print club images also contained linked sounds at the word boundaries, ‘あいらびゅ/airabyu’ (I love you), and ‘ふぁっきゅー/fakkyū’. In both of these cases the linking of the two words has removed the need for an extra mora/vowel sound at the boundary of the two words, and are thus a more accurate rendering of connected speech.

These kinds of examples are interesting in the context of ‘katakana English’; if a writer is keen to show that they do not speak this stigmatized variety, they can do so by firstly not writing in katakana, and secondly, by displaying a knowledge of a more accurate English representation of the word. In the survey, katakana was associated with terms such as ‘robotic’, ‘mechanical’ and ‘disfluent’, and is often used in manga and ‘telop’ (television captions) for representing the speech of non-Japanese speakers of the language, to represent their perceived ‘foreign’ accent.

In the cases of ‘New Zealand’ and ‘g’mornin’’, the hiragana versions encode a more fluent and native like pronunciation of these loaned terms compared to their standard katakana equivalents. This may be evidence of the ideational function, marking the speaker as a having

239 Since the conventions and permitted sound combinations for transcribing loanwords in Japanese in the Hepburn system (ヘボン式/hebon shiki), and ‘kun-rei system’ (訓令式/kunrei shiki) relate to katakana, it could be argued that such combinations in hiragana are outside of the scope of these systems and therefore not ‘illicit’. However, discussions with native-Japanese speaking researchers suggest that since combinations such as ‘ずぃ /zi’ are also not found in other hiragana/Japanese words, they still feel slightly unusual or awkward. 278

native-like pronunciation, or be seen in terms of the interpersonal function, involving claims to authenticity, for example.

However, in one case a loanword in hiragana was felt to represent a less-fluent pronunciation of a word, in the sticker containing the phrase ‘I can fly!! (ふらい/furai)’. As mentioned above, it was interpreted by some survey respondents and interviewees as representing the character’s inability to accurately pronounce the final word, the unconventional script choice representing an unconventional (Japanized) pronunciation. This may be motivated by the fact that this word contains an ‘l’ sound, the distinction between ‘l’ and ‘r’ being notoriously difficult for Japanese learners of English. In addition, there is no way to approximate a more native-like English ‘l’ sound in Japanese as the writer of ‘New Zealand’ did for the ‘z’ sound described above. In this case, English/rōmaji or even katakana would represent the model of standard pronunciation. However, like the examples described above, this text utilizes hiragana to highlight the pronunciation of a particular word, regardless of whether that pronunciation is more or less native-like.

The use of hiragana in these cases, it is argued, has the function of drawing the readers’ attention to the sounds represented, and thus shares some ground with the emphatic function described in section 7.3.2. Hiragana’s reported associations of representing ‘sounds’ suggest parallels to the use of phonemic script, foregrounding how certain words are pronounced or heard. The idea of a certain authenticity of pronunciation stands in contrast to the ‘joking’ quality or un-realness associated with the script described in previous sections. However, rather than representing contradictory findings, it highlights the role context and the other semiotic resources used in multimodal texts play in interpreting meaning. The findings on the role of hiragana in marking (non)native-like pronunciation also ride the translingual flows described in section 7.2, representing the English ability of the Japanese, ranging from the stigmatized ‘katakana English’ to high levels of proficiency; and the pronunciation of both loanwords and foreign words by the Japanese. Loanwords in hiragana are therefore not just a phenomenon of written language, but also provide a window into perceptions of spoken language in the Japanese context.

279

7.4 Script as a mode in a multimodal text

While the previous sections have explored the main functions of loanwords in hiragana, they have also demonstrated how script use in Japanese can be seen as performing as a mode in a multimodal text.

As described in section 2.3, the three metafunctions described by Halliday (1978), the ideational, interpersonal, and textual, were argued to be the three kinds of meaning that language is designed to make. These functions have been extended to other semiotic resources, for example visual texts such as images and graphs (Kress & van Leeuwen, 2006); advertising images (Oyama,1998), and typefaces (Stöckl, 2005). Kress (2011) argues that the fulfilment of these functions is a requirement of a particular semiotic resource being considered a mode within a social semiotic theory of communication. Script choice in Japanese, it is argued, is also a mode, as it meets the criteria set by Kress (2011), the findings of this study demonstrating how it can be shown to express each of Halliday’s metafunctions.

The ideational function is used to represent the world around us, the state of things, for example. Hiragana has been shown to differentiate referents in the same text, for example a brand name and a generic item, as in the case of the words ‘れもん/remon’ (lemon), or and ‘すぱいす/supaisu’ (spice). In both these cases, the hiragana version was reported to refer to the brand or product name, while the katakana version referred to an ingredient of the aforementioned product. The use of hiragana for a loanword can also add an extra layer of meaning to the referent by taking advantage of the connotations of hiragana such as Japanese, traditional, or cute, for example. This illustrates that the use of script can perform the ideational function, making a statement about the world by giving information about a particular referent.

The interpersonal function concerns relationships, in the case of written language, between the writer and reader of a text. Hiragana script was described as giving a sense of intimacy or familiarity to the referents it was used for, in the cases of products and businesses, and on social media, was said to be used particularly with close friends. The interpersonal function can also be used to communicate modality, or the seriousness with which a message should be taken, leading to comments of the ‘un-real’ or joking nature of some of the texts.

280

The textual function, which creates cohesion between the elements of a text into a meaningful whole, is most evident in the use of hiragana for loanwords to both obscure and create word boundaries. This was used to humorous ends in the case of lexical blends created from a Japanese word and a loanword. This technique can also result in visual ‘rhymes’, connecting words in particular scripts to other words in that script within a text.

As Kress suggests, particular modes can be made to carry a greater or lesser ‘informational weight (functional load)’ (2011, p. 60) within a multimodal text. Taking script as a particular mode, it is easy to see how the ‘functional load’ of this semiotic resource increases when it is used in a marked context, namely, for a foreign loanword. Script is carrying more ‘informational weight’ in the case of loanwords in hiragana than for loanwords in katakana, or Japanese words in hiragana, for example.

This demonstration of how script can function as a mode within a multimodal text is an important contribution to multimodal studies, and reinforces van Leeuwen’s (2011) call for the discipline to engage with languages and texts outside the western cultural tradition. The alternative scripts used for different words within even a short phrase appears to be unique to the Japanese language, and the current study has shown how even one specific aspect of script use, the use of hiragana for loanwords, can be used meaningfully, effectively and flexibly.

7.5 The sedimentation of script choice

Aspects of the current topic of loanwords in hiragana can be seen as demonstrations of the ever-evolving nature of language, but also as examples of how these changes can influence and often reinforce each other, and become sedimented over time. It is no coincidence that two of the most frequently occurring words in the corpus, ‘れもん/remon’ (lemon) and ‘きゃべつ/kyabetsu’ (cabbage), were also the two words most frequently reported to seem acceptable in hiragana in the interviews. Asking whether the relative frequency of these words means that they are felt to be more acceptable, or whether the acceptability of these words leads to their frequent occurrence is not the point, rather, these factors mutually reinforce each other over periods of use, becoming sedimented in the language. Eventually, this may reduce the ‘script familiarity effect’ reported by Tamaoka (1998, p.1), meaning that particular loanwords will be processed as easily in either kana script. This also has 281

implications for the testing of Japanese as a foreign language (and perhaps a native language too) as suggested in the following chapter, section 8.1. Technology plays a major role in innovative language practices in a number of ways, and these will be explored in the following paragraphs.

Firstly, as explained in section 1.1.1, when words are entered into Japanese word processing software, they initially appear in hiragana, and a drop-down menu presents the orthographic options for that series of morae including hiragana, katakana, and sometimes one or more kanji. Therefore, in order to write the word ‘レモン/remon’ (lemon), when the characters R- E-M-O-N are entered, the word appears as ‘れもん/remon’, with katakana (レモン) and kanji (檸檬) appearing as options, along with hiragana, as shown in figure 7.1 below.

Figure 7.1: Script options for 'lemon'

The writer then selects the katakana option appearing as the first option in figure 7.1, in order to have the word appear as ‘レモン/remon’ in the document. However, this exposure to loanwords in hiragana, even if, in the majority of cases, they are not selected by the user, may play a role in their dissemination. Hiragana appears as a choice for these foreign terms, the menu of orthographic options suggesting a SFL network diagram, however constrained the choice may be from the context in which the word will appear. As one interviewee reported, it was possible that the choice to use hiragana was made at the time of entering the text and seeing the orthographic options on the screen in front of them, ‘I wonder if they’re looking at it while they’re typing…when they see (the hiragana option), they purposefully think about the possibility of ‘joking around’, and decide to make a playful comment.’240

240書き込む時を見ているからかな。見ていると、こういうふざけたやつ明らかに、ふざけたコメントしてる。(Yūta, G4) 282

The selection of hiragana for a loanword could also be made unintentionally, and some interviewees admitted to making ‘typos’ by accidently selecting hiragana for loanword they had entered. Such errors were also thought to be the cause of some of the inconsistent examples in the corpus. An erroneous input of the word ‘サンキュー/sankyū’ (thank you) was also responsible for the string of humorous homonyms described by Sunakawa (2008).

Whether the selection of hiragana for a loanword is made in any of these three ways: planned and intentionally; intentionally at the moment of typing; or unintentionally; an interesting change occurs in the software of programs such as Microsoft Word and Google Japanese keyboard. The order in which the orthographic options produced by the software occurs adapts itself to the most recent orthographic preference of the user. While the first instance of entering the letters ‘R-E-M-O-N’ gives the options in the order shown in figure 7.1 above, if hiragana is selected in the first instance, the second time those same letters are entered, the options appear as follows (figure 7.2)

Figure 7.2: Adjusted script options for 'lemon'

In this case, the marked hiragana option ‘れもん/remon’ (lemon), is presented first in the hierarchy of options offered, rather than the standard katakana ‘レモン/remon’ (lemon) which appears first in figure 7.1. The sedimentation of language practices is therefore being reflected even in the tools used to mediate written language.

Drawing conclusions about the effect of technology on the occurrence of loanwords in hiragana would require, among other things, a diachronic investigation of the topic, rather than the synchronic approach taken in this study. However, the potential effects of tools used to mediate written text is noted here, and further investigation of this matter suggested as topic of future study in the following chapter, section 8.2.

283

Besides the mechanics of entering text via Japanese word-processing software, there is another important way in which the use of technology contributes to the spread and sedimentation of language practices. These novel forms can be disseminated much more widely and quickly than they could in the past through social media and other online platforms. A demonstration of how social media has contributed to the uptake of a particular loanword in hiragana is related to the Japanese model nicknamed ‘Chīpopo’ (Chisato Yoshiki), whose tweet containing the phrase ‘ぐっもーにん/gummōnin’ (G’mornin’), was analysed in case study 4. As noted in section 4.2.4, Yoshiki’s tweets often contain loanwords and phrases in hiragana, suggesting this may be a trademark of her social media presence. Another of her trademarks is a pose in which she points to her cheeks, pulling a slightly pouting face, and this, in turn has been dubbed the ‘ちいぽぽーず/chīpopōzu’ (Chīpo-pose). The term itself a lexical blend of a Japanese word and loanword, like the others described in section 4.1.6, and uses hiragana to obscure the word boundary via the shared character ‘ぽ /po’. It is therefore both typical of the other instances of word play described in the corpus, as well as a clever and appropriate coinage for a word connected to a person whose own language frequently utilizes loanwords in hiragana.

The ‘Chīpo-pose’ has been mimicked by girls producing print-club images, (see figure 7.3 below), and each instance is captioned as such, being labelled with the words ‘ちいぽぽーず /chīpopōzu’ (Chīpo-pose).

Figure 7.3: Print club images containing the 'Chīpopose'

284

As noted in section 3.4.1, besides the stickers produced by the print club booth itself, the images can be sent to the user’s email account, as well as automatically uploaded to online galleries.241 Since these images are available online, they are likely to have been seen by many others, as well as shared by the individuals themselves, further disseminating the new term. The last image, it will be noted, is of a group of boys attempting the ‘Chīpo-pose’; interestingly, their caption does not contain the blended term, but the two words ‘ちいぽぽポーズ/chīpopo pōzu’ (Chīpopo Pose), and furthermore, the loanword ‘pose’ is written in standard katakana. The cleverly coined term, in this case, has missed its linguistic mark, reinforcing the status of these boys as outsiders to the girl-culture world of print club, fashion models, and the language associated with these practices.

The craze for doing the ‘Chīpo-pose’ has in turn generated its own hashtag (#ちいぽぽーず) further establishing its place on social media. Users uploading these images to twitter or Facebook can therefore tag their own renditions of the pose, and also search for others who have created similar representations. This, in turn further disseminates and sediments the term in the sub-culture.

This brief discussion of the role of social media in spreading a particular loanword in hiragana is an illustrative example of the cyclic nature of these language practices: terms encountered on social media, for example, by reading a celebrity’s twitter feed, can be taken up by individuals, loaned and even owned when re-used in the creation of new texts such as the print club images in figure 7.3 above. These images are then fed back into social media platforms to be further disseminated. Like the discussion of text-input software, however, these comments do not attempt to provide the final word on the role of social media in the spread or sedimentation of loanwords in hiragana, but rather to note its role in this process and to put the issue forward as a topic for further investigation.

7.6 Summary of the discussion chapter

This chapter has brought together the findings of the research project in order to provide a nuanced understanding of the practice of writing loanwords in hiragana gained from the four

241 Only three tokens of the word ‘ちいぽぽーず/chīpopōzu’ (Chīpopose) were included in the corpus, however, to avoid over representation of this term as described in 3.4.1. 285

layers of analysis: the descriptive statistics of the corpus, the case studies, the online survey and the interviews and focus group discussions. By doing this, it has highlighted the general acceptability of loanwords in hiragana, establishing this practice in the larger context of the translingual and trans-orthographical properties of the Japanese language and writing system.

By integrating the findings across all layers of analysis a variety of functions of loanwords in hiragana have been described, which can relate to the socio-historical associations of the script, and/or the contexts in which they occur. The context-based interpretations of these words and phrases highlight the social-semiotic nature of meaning making, reinforcing how forms and meanings do not stand in pre-existing, arbitrary relationships, but are brought together by the sign-maker in specific contexts.

The findings of this study also demonstrate that, like other semiotic resources such as typography or layout, script should be considered a mode within a multimodal text according to the test proposed by Kress (2011). The current research therefore contributes to an understanding of how multimodal texts are used to create meaning in contexts outside the western cultural tradition.

Finally, this chapter has highlighted the role technology plays in disseminating and sedimenting novel language practices such as loanwords in hiragana. While it has been outside the scope of this study to explore this in more detail, these thoughts are noted as suggestions for further research in the following chapter.

286

Chapter 8: Conclusion

The current research set out to understand the role of loanwords written in the marked hiragana script. This phenomenon is worthy of investigation as it problematises the association of these foreign borrowings with the katakana script, and the claim that ‘loanword’ is synonymous with ‘カタカナ語/katakana-go’ (katakana word).

Understanding the role of hiragana loanwords lead to the following three research questions:

1. How can loanwords in hiragana be characterised?

2. In what contexts can loanwords in hiragana be found?

3. Why do loanwords appear in hiragana?

The four layers of analysis, a combination of both quantitative and qualitative data which arose from both analyses of the texts themselves (the descriptive statistics of the corpus and the case studies), as well as the opinions of native Japanese speakers in Japan (the survey and interviews), led to the following findings.

Loanwords in hiragana exhibit some typical features such as usually having English as their source language, while languages which have had historical contact with Japan, such as Portuguese, were the source of some of the most frequently occurring words, such as ‘ぱん /pan’(bread) and ‘かすてら/kasutera’ (castella). This is not surprising as these words entered the Japanese language when the conventions of script use were not as defined as they are today. Loanwords in hiragana are usually nouns, with adjectives found in fewer numbers and verbs being very rare. Some of these loanwords contain long vowels, and in most cases, they are lengthened with the bar associated with katakana words rather than the extra vowel typical of hiragana words. In some cases, both methods are used. In terms of referential meaning, around half of the words in the corpus refer to the names of products and retail stores.

Given this referential use of loanwords in hiragana, it is not surprising that product packaging and business signage were common genres found in the corpus. Other genres found to be sites of these words included social media posts, publication titles, and print-club images. It is unclear whether the use of hiragana for loanwords may remain in these genres or spread to

287

other contexts in the future. In terms of the appearance and arrangement of loanwords within a text, they were usually found singularly, although two consecutive words were not uncommon, and some examples of three consecutive words were also found. The majority of texts contained only one word/phrase unit in hiragana. Various examples of inconsistency were also apparent, where the same loanword was transcribed in an alternative script either within the text, or between associated texts.

The question of why loanwords occur in hiragana was the most complex to answer, and via the four layers of analysis, a number of functions performed by hiragana loanwords have been described. The findings have demonstrated that loanwords in marked hiragana script are capable of fulfilling the metafunctions proposed by Halliday (1978), thus acting as a ‘mode’ within a multimodal text.

Of the functions seen to be performed by these words, some relate to the socio-historical associations of the hiragana script itself, for example adding connotations of ‘Japanese-ness’ or familiarity to a referent. These connotations demonstrate the ideational and interpersonal functions respectively. Other functions relate to the multimodal contexts in which they were found, a central tenet of meaning-making within a social semiotics context. For example, hiragana was shown to indicate a sense of ‘un-realness’ (ideational metafunction) or a suggestion that a statement should not be taken seriously (interpersonal metafunction). The textual function was most apparent in the use of hiragana loanwords to either define word boundaries or conversely to obscure them. This was often used to humorous effect in the significant number of examples of language play within the corpus, perhaps representing a new aspect of Japan’s long history of homonymic puns. Another important function of loanwords in hiragana was to mark more native-like pronunciation, which, it is argued, may be a reaction to the stigmatism associated with ‘katakana-English’.

Script is thus shown to be an effective and flexible resource capable of adding layers of meaning to a written word, using loanwords in hiragana as a specific case of a marked or unconventional use of script. The findings also provide evidence of the translingual/trans- orthographic nature of loanwords by analysing the shift from katakana, a script strongly associated with foreign words and culture, to hiragana, one representative of native Japanese words. That many respondents in both the survey and interviews noted that it was ‘not wrong’ to write loanwords in hiragana indicates a sense of ownership of these ‘foreign’ words, and while these comments were limited to particular genres of text which the participants were

288

presented with, parallels can be seen with the adoption of the Chinese writing system and associated lexicon in previous centuries.

8.1 Theoretical implications

The findings of this study make several important contributions to a diverse range of fields.

Firstly, the findings contribute to the understanding of loanwords in Japanese, namely, that they are sometimes written in hiragana. While this may sound like an insignificant claim, it is important because seminal texts describing Japanese language contact such as Loveday (1996) and Stanlaw (2004) do not mention this practice being used today. 242 It also problematises the use of the term ‘カタカナ語/katakana-go’ (katakana-words), as being synonymous to ‘外来語/gairaigo’ (loanwords). The findings of the current project were therefore vital in establishing the phenomenon of ‘loanwords in hiragana’ as an object of study.

Secondly, the findings highlight the complexities of the choice of script for loanwords, as some of the examples of loanwords in hiragana were consistently reported to be ‘more Japanese’ than others. While the presence of a kanji played a role in the process of indigenization for loanwords such as ‘れもん/檸檬/remon’ (lemon), it was not the deciding factor, as the word ‘きゃべつ/kyabetsu’ (cabbage), for example, was also consistently reported to be acceptable in hiragana.

The points raised above have implications for pedagogy, and the teaching of Japanese as a second or additional language. In language classes, for example, katakana is routinely described as the script for foreign words, and students’ understanding of the conventions surrounding the use of hiragana and katakana is often tested by asking for the appropriate script for simple words such as ‘kuruma’ (car), which would normally be written in hiragana or kanji (くるま/車); and ‘basu’ (bus), which, as a loanword, would conventionally be transcribed in katakana (バス). It would raise the validity of such tests to avoid using loanwords such as ‘れもん/remon’ (lemon), and ‘きゃべつ/kyabetsu’ (cabbage), which were

242 Historical references to periods where katakana was not used consistently for loanwords, such as ‘たばこ /tabako’ (tobacco, cigarettes) do feature. 289

judged by a number of participants in this study as being equally suitable in hiragana in some contexts of use, and as the corpus shows, are also some of the most frequently occurring loanwords in hiragana. Furthermore, it may benefit more advanced learners of Japanese to be made aware of these atypical script variations, and the kinds of communicative metafunctions they can be used to perform. Becoming competent in an additional language involves being aware of the nuances of similar forms, and orthography, it is argued, is an important aspect of this.

An understanding of the complexities of script use in Japanese leads to another important contribution to theory, namely, how the trans-orthography described in the current research relates to the established field of translingualism. Not only are the loanwords in this study no longer ‘branded’ as foreign words through the katakana script, they are transcribed in a script strongly associated with native Japanese words (和語/wago). This can be seen as an act of ownership of these words, a blurring of the boundaries of the lexicon of ‘self’ and ‘other’. While some scholars have drawn parallels between the absorption of loanwords into Japanese with that of Sino-Japanese terms in the past (e.g. Hoffer, 1980; Hosokawa, 2015; Stanlaw, 2004), the appearance of loanwords in hiragana script appears to represent a further step towards the assimilation of these ‘foreign’ words. Furthermore, these lexical items are repurposed in local contexts of use, being used to perform a variety of functions, including marking bilingual puns, representing more accurate phonetic transcriptions, and creating and obscuring word boundaries. These functions reflect the localized uses that the marked script choice for loanwords have been used to perform, reinforcing the idea of ‘repetition that is something else, sameness that is difference’ in global flows of language (Pennycook, 2010, p. 37).

The final contribution relates to the field of multimodality, as the findings of this research demonstrate how script can function as a semiotic resource in a multimodal whole. The social semiotic ‘meanings’ of loanwords in hiragana are not just the products of the socio-historical associations of the script, but are context-based, thus the same orthographical resource (hiragana) can communicate a sense of Japanese traditional culture in one text and a play a role in the seriousness of a particular statement in another. These diverse meanings are supported by other semiotic resources within the text, as demonstrated by the case studies as well as the comments of the survey respondents and interviewees. Context is central to a social semiotics, and therefore also an important element of SFL.

290

This contribution echoes van Leeuwen’s (2011) call for multimodal studies to look outside the western cultural tradition in order to develop the field as a branch of applied linguistics. While this study is based in a non-western culture, it is also significant in investigating a semiotic resource unique to that country, namely, an element of Japan’s multi-script writing system. While studies of more general semiotic resources such as image or typography in various cultural contexts also benefit the field of multimodal studies, descriptions and analyses of semiotic resources unique to a language or culture are of great interest. It is hoped that the findings of the present research promote the continued expansion of the field of multimodality in non-western cultures, and the description of further semiotic resources not present in western language contexts.

8.2 Critical reflections and suggestions for further research

In this closing section, the researcher’s experience of conducting this study is reflected upon, and comments on the research design and methodology are offered. The section concludes with suggestions for further research.

The status of the researcher as a non-native speaker of the language under investigation was reflected upon at various times throughout the project. While this necessitated cross- referencing of interpretations at many stages of the project with a range of native-speaking Japanese researchers, the non-native status of the researcher was felt to have worked as an advantage when conducting the interviews in particular. Knowing they could not rely on a shared cultural background, participants seemed to understand the need to be explicit in describing the connotations or meanings that loanwords in hiragana held for them and why, for example, referring to the historical association of the script with women’s writing which linked hiragana with femininity, or pointing out particular strokes or loops of individual characters that gave an impression of cuteness. If the researcher had shared the cultural background of the interviewees, it is possible that the responses would have been less explicit, relying on a tacit understanding of common pool of cultural knowledge.

The research topic itself was also felt to be conducive to investigation by a non-native speaker. While participants often admitted they had never considered the meanings behind loanwords in hiragana before, the act of a non-native speaker asking a native speaker the

291

meaning of an aspect of the language is a relatively commonplace experience. While the interviews and focus groups were conducted in a professional manner, the conversations often felt less like research interviews, and more like an expert/native speaker attempting to explain a complex aspect of their language to a non-expert user. This helped create a relaxed atmosphere during these sessions, and allowed the participants to take the role of a sharer of knowledge, rather than an interviewee.

An aspect of the study that was both a strength and a weakness was the broad range of genres from which the corpus examples were taken. While limiting the focus of the study to a particular genre such as product names or social media posts would have allowed more engagement with the topic of loanwords in hiragana relating to a specific context, this was also felt to reduce the impact of the findings. In other words, the breadth of contexts in which loanwords in hiragana were found was felt to be a significant finding of the research project. The study of loanwords in hiragana within particular contexts will add to our understanding of script use, however, and are suggested as a topic for further research. By limiting the scope of such as study to the functions of loanwords in hiragana in computer-mediated communication (CMC), for example, more detailed context-based findings of this practice can be investigated. Another genre not included in the current research design, manga, was suggested as a potential source of loanwords in hiragana by one of the interviewees, and is considered worthy of investigation.

Another aspect of the research design that had the effect of limiting the findings was the focus on the receivers/readers of the texts, rather than the producers. While the interpretations of the audience are vital in understanding the meanings created by the use of hiragana for loanwords, they could be supplemented by investigating the intended message of the producers of these texts. This is also offered as a suggestion for further research, and could focus on computer-mediated communication (CMC) as described above, or take a marketing perspective and investigate the design briefs for various products or businesses, and how script plays a role in the texts produced by graphic designers and marketing departments.

Some final suggestions for further research arose directly from the findings of this study. Firstly, the idea that some loanwords seemed more appropriate in hiragana than others is worthy of exploration, in order to establish a ‘scale of correctness’ for a wider range of loanwords, and to see if this scale varies across different demographics of society. As noted in section 5.1, the survey sample did not compare responses across age and gender, for

292

example. Secondly, since older loanwords such as those from Portuguese, and English words dating from the Meiji period were some of the most frequently occurring in the corpus, taking a diachronic perspective of loanword orthography would add to the findings of this study. Incorporating loanwords of Chinese origin, such as ‘ぎょうざ/gyōza’ (gyoza) would also be of interest due to their close relationship with Sino-Japanese words. Variation could be traced across periods of major loanword uptake such as at the opening of Japan after the years of national isolation (鎖国/sakoku), and that following the Second World War, as well as periods of nationalism and the linguistic cleansing associated with them. It would also allow for an investigation of the effects of computer-mediated script input as described in the previous chapter, in order to examine whether exposure to loanwords in hiragana through their being presented as an ‘option’ in word processing software coincides with a rise in their use. Following van Leeuwen’s (2011) call for studies in multimodality to engage further with technology, an analysis of how Japanese software handles non-standard orthographical choices, including selection of script, as well as combinations of characters would also be of interest.

8.3 Epilogue: A return to the Japanese house

In the introduction, it was noted that Kay (1995) used the metaphor of rooms in a Japanese house to describe the compartmentalization of the ‘native’ and ‘foreign’ in Japan. The rooms in a Japanese house, she argues, are clearly defined as ‘Japanese style’ or ‘western’ style by their structure and furnishings, and the language used to describe them: Japanese rooms have ‘障子/shōji’ (paper sliding screens) while western rooms have ‘カーテン/kāten’ (curtains), for example, and the lexical items mirror this distinction: Japanese terms for the Japanese- style room, and loanwords in katakana for the western-style rooms.

Katakana, Kay explains, is vital in maintaining this compartmentalization of language, allowing for the introduction of loanwords as new lexical items while yet ‘protecting’ native words from their influence by branding them as foreign items, resulting in the desired ‘incomplete linguistic assimilation’ of loanwords.

A loanword in hiragana, to continue the house metaphor, is the linguistic equivalent of a bed (or more precisely, a ‘べっど/beddo’) in a Japanese-style room, rather than the traditional ‘布

293

団 /futon’ (futon), and a brief online search suggests that this is not an inconceivable arrangement. 243 Japanese houses, like the Japanese language, are not so neatly compartmentalized as Kay’s metaphor suggests; styles shift and flow with the occupants lives, needs, and preferences. Kitchens contain both cutlery and chopsticks, and either are options for a ‘パスタ/pasuta’ (pasta) or a cubed ‘ステーキ/sutēki’ (steak) at the western-style dining table. After dinner, European-style baked goods from traditional tea houses in Kyoto can be consumed in the Japanese-style room, as the J-pop artists on the television pepper their lyrics with English words. Occupants take a ‘シャワー/shawā’ (shower) or an ‘お風呂/ofuro’ (bath), usually both, before retiring to the sleeping arrangement of their choice, which may include a bed in a Japanese-style room (figure 8.1), or a futon in a western-style room.244

Loanwords in hiragana represent a similar fusion of language and culture, and are just one aspect of the practices of translingualism and ‘glocalization’ apparent in Japan as well as other contexts of cultural contact around the world. Without the katakana script branding these words as instances of ‘incomplete linguistic assimilation’, they are open to being utilized for a range of metafunctions as described in the preceding chapters. It is hoped that future researchers continue to engage with this complex and beautiful language and its writing system.

Figure 8.1: Japanese style room with bed245

243 In fact, after typing ‘和室/washitsu’ (Japanese-style room) into Japanese Google, ‘和室ベッド/washitsu beddo’ (Japanese-style room bed) and ‘和室カーテン/washitsu kāten’ (Japanese-style room curtains) were two suggested search options. 244 ‘洋室布団/yōshitsu futon’ (western-style room futon) was also a suggested search option for the term ‘洋室 /yōshitsu’ (western style room). 245 http://stylish-room.com/japanese-style-bed-coordinate/ 294

References

Androutsopoulos, J. K. (2000). Non-standard spellings in media texts: The case of German fanzines. Journal of Sociolinguistics, 4(4), 514–533.

Akizuki, K. (2005). Arienai nihongo [Unbelievable Japanese]. Tokyo: Chikuma Shinsho

Aoki, M. (2017, April 6). Japan’s latest English proficiency scores disappoint. The Japan Times. Retrieved 4 March 2015 from http://www.japantimes.co.jp/news/2017/04/06/national/japans-latest-english- proficiency-scores-disappoint/#.WYq_ClUjHRY

Arksey, H., & Knight, P. (1999). Interviewing for social scientists. London: Sage.

Backhaus, P. (2007). Alphabet ante portas: How English text invades Japanese public space. Visable Language, 41(1), 70–87.

Backhaus, P. (2012, October 22) Politicians may 'ru' the day their names became verbs. The Japan Times. Retrieved 20 December 2017 from https://www.japantimes.co.jp/life/2012/10/22/language/politicians-may-ru-the-day-their- names-became-verbs/#.WjmEOminGyI

Bartal, O. (2013). Text as image in Japanese advertising typography design. Design Issues, 29(1), 51–66. http://dx.doi.org/10.1162/DESI_a_00196

Bestor, T., Steinhoff, P., & Bestor, V. (Eds.). (2003). Doing fieldwork in Japan. Honolulu: University of Hawai'i Press.

Blommaert, J. (2012). Sociolinguistics & English language studies. Urban Language & Literacies, 22 (June), 1–26.

Blommaert, J. (2013). Writing as a sociolinguistic object. Journal of Sociolinguistics, 17(4), 440–459. http://dx.doi.org/10.1111/josl.12042

Bloomberg, L. D., & Volpe, M. F. (2015). Completing your qualitative dissertation : A roadmap from beginning to end. California: Sage

Brown, J. (2011). Re-framing 'kawaii': Interrogating global anxieties surrounding the aesthetic of 'cute' in Japanese art and consumer products. The International Journal of 295

the Image. 1(2) 1-10.

Borggreen, G. (2011). Cute and cool in contemporary Japanese visual arts. The Copenhagen Journal of Asian Studies, 29(1), 39-60.

Buck, J. H. (1969). Japanese characters in written Japanese. Paper presented at the Southeastern Conference on Linguistics, Florida State University, Tallahassee, March 28-30, 1969.

Canagarajah, S. (2013). Translingual practice: Global Englishes and cosmopolitan relations. New York: Routledge.

Cheng, W. (2012). Exploring corpus linguistics. London: Routledge.

Coupland, N., & Garret, P. (2010). Linguistic landscapes, discursive frames and metacultural performance: The case of Welsh Patagonia. International Journal of the Sociology of Language, 205, 7–36.

Crystal, D. (1998a). Language play. Chicago: University of Chicago Press.

Crystal, D. (1998b). Toward a typographical linguistics: Keynote address for the ATypI 1997 Conference, University of Reading, 13 September 1997. Type, 2, 7–23.

Dale, P. N. (1986). The myth of Japanese uniqueness. Kent: Croom Helm.

Danesi, M. (2016). The semiotics of emoji. London: Bloomsbury Publishing.

Daulton, F. E. (2004). The creation and comprehension of English loanwords in the Japanese media. Journal of Multilingual and Multicultural Development, 25(4), 285–296. http://doi.org/10.1080/01434630408666533

Davies, E. E. (1987). Eyeplay: On some uses of nonstandard spelling. Language and Communication, 7(1), 47–58.

Dickinson, J. A. (2015). Plastic letters: Alphabet mixing and ideologies of print in Ukrainian shop signs. Pragmatics, 25(4), 517–534.

Dovchin, S., Sultana, S., & Pennycook, A. (2016). Unequal translingual Englishes in the Asian peripheries. Asian Englishes, (May), 1–17. http://dx.doi.org/10.1080/13488678.2016.1171673

296

Doyle, J. R., & Bottomley, P. A. (2006). Dressed for the occasion: Font-product congruity in the perception of logotype., 16(2), 112–123. http://dx.doi.org/10.1207/s15327663jcp1602_2

Doyle, J. R., & Bottomley, P. A. (2009). The massage in the medium: Transfer of connotative meaning from typeface to names and products. Applied Cognitive Psychology, 23, 396– 409.

Eggins, S. (2004). An introduction to systemic functional linguistics (2nd ed.). London: Continuum International Publishing Group.

Flick, U. (2006). An introduction to qualitative research (3rd ed.). London: Sage.

Flood, J. L. (1996). Humanism, 'nationalism', and the semiology of typography. In M. Tavoni (Ed.), Italia ed Europa nella linguistica del Rinascimento: confronti e relazioni. Atti del convegno internazionale Ferrara, Palazzo Paradiso, 20–24 marzo 1991. Vol. 2: L’Italia e l’Europa non romanza. Le lingue orientali. [Italy in Eurpoe in the Linguistics of the Renaissance:Comparisons and Relationships. Proeceedings of the international conference] (pp. 179–196). Panini: Ferrara.

Fukuda, R. (2006). Gairaigo to gendai shakai. [Loanwords and today's society] Kokuritsu kokugo kenkyūjō [National Institute for Japanese Language and Linguistics (NINJAL)]

Gottlieb, N. (1995). Kanji politics. Padstow: T J Press.

Gottlieb, N. (2000). Word-processing technology in Japan: Kanji and the keyboard (1st ed.). Richmond, Surrey, UK: Routledge.

Gottlieb, N. (2010a). Playing with language in e-Japan: Old wine in new bottles. Japanese Studies, 30(3), 393–407. http://dx.doi.org/10.1080/10371397.2010.518600

Gottlieb, N. (2010b). Technology and the writing system in Japan. In P. Heinrich & C. Galan (Eds.), Language life in Japan: Transformations and prospects. Hoboken: Taylor and Francis.

Gray, D. E. (2009). Doing research in the real world (2nd ed.). London: Sage.

Haarman, H. (1986). Verbal strategies in Japanese fashion magazines: A study in impersonal bilingualism and ethnosymbolism. International Journal of the Sociology of Language,

297

58, 107–121.

Halliday, M. A. K. (1978). Language as a social semiotic. Maryland: University Park Press.

Hida, Y., & Sato, T. (2002). Gendai nihongo kōza 6: moji, hyōki [Course in contemporary Japanese 6: letters and expression].Tokyo: Meiji shoin.

Hodge, R., & Kress, G. (1988). Social semiotics. Cambridge: Polity Press.

Hoffer, B. (1980). English loanwords in Japanese: Some cultural implications. Language Sciences, 12(1), 1–21. http://dx.doi.org/10.1016/0388-0001(90)90020-H

Honna, N. (1995). English in Japanese society: Language within language. Journal of Multilingual and Multicultural Development, 16(1–2), 45–62.

Hosokawa, N. (2015). Nationalism and linguistic purism in contemporary Japan: National sentiment expressed through public attitudes towards foreignisms. Studies in Ethnicity and Nationalism, 15(1), 48–65. http://dx.doi.org/10.1111/sena.12135

Huang, C. W., & Archer, A. (2014). Fluidity of modes in the translation of manga: The case of Kishimoto’s Naruto. Visual Communication, 13(4), 471–486.

Hyde, B. (2002). Japan’s emblematic English. English Today, 18(3).

Igarashi, Y. (2007). The changing role of katakana in the Japanese writing system: Processing and pedagogical dimensions for native speakers and foreign learners (Doctoral dissertation). Retrieved 18 March 2015 from DSpace Dissertation Database. (2007-08-16T01:05:03Z)

Inoue, F. (2005). Econolinguistic aspects of multilingual signs in Japan. International Journal of the Sociology of Language, 176, 157–177.

Irwin, M. (2011). Loanwords in Japanese. Philadelphia: John Benjamins Publishing Company.

Ishii, K. (2013). Taishōki no 'Fujin Kōron' ni okeru gairaigo hyōki no hensen [Notational shift of loanwords in 'Fujin Koron' published in the Taisho period]. Ningen bunka sōsei kagaku rongyō [Dissertation repository of the schools of human culture and creative science, Ochanomizu University], March. Retreived 11 June 2015 from http://teapot.lib.ocha.ac.jp/ocha/handle/10083/52686

298

Ishikawa, K. (2007). Hiragana no bigaku [Aesthetics of hiragana]. Tokyo: Shinchosha.

Ishikawa, K., & Rubrecht, B. G. (2008). English loanword use on Japanese television. In JALT2007 Conference Proceedings (pp. 307–316). Tokyo.

Iwahara, A., Hatta, T., & Maehara, A. (2003). The effects of a sense of compatibility between type of script and word in written Japanese. Reading and Writing: An Interdisciplinary Journal, 16, 377–397.

James, A. (2014). English as a visual language: Theorising the social semiotics of anglography in polylingual texts. English Text Construction, 7(1), 18–52. http://dx.doi.org/10.1075/etc.7.1.02jam

Japan Technical Communicators Association. (2015). Guidelines for writing loanwords (katakana), 3rd edition. Retrieved 29 October 2015 from https://www.jtca.org/standardization/katakana_guide_3_20150929.pdf

Jaworski, A. (2015). Globalese: A new visual-linguistic register. Social Semiotics, 25(2), 217–235. http://dx.doi.org/10.1080/10350330.2015.1010317

Jenkins, J. (2003). World Englishes: A resource book for students. New York: Routledge.

Jewitt, C. (Ed.). (2011). The Routledge handbook of multimodal analysis. London: Routledge.

Jewitt, C., Bezemer, J., & O’Halloran, K. (2016). Introduing multimodality. New York: Routledge.

Jones, G. M., & Schieffelin, B. B. (2009). Talking text and talking back: 'My BFF Jill' from boob tube to youtube. Journal of Computer-Mediated Communication, 14, 1050–1079. http://dx.doi.org/10.1111/j.1083-6101.2009.01481.x

Joyce, T., Hodošček, B., & Nishina, K. (2012). Orthographic representation and variation within the Japanese writing system: Some corpus-based observations. Written Language and Literacy, 15(2), 254–278. http://dx.doi.org/10.1075/wll.15.2.01rob

Kachru, B. B. (1986). The alchemy of English: The spread, functions and models of non- native Englishes. Oxford: Pergamon Press.

Kallen, J. L., & Ni Dhonnacha, E. (2014). Language and inter-language in urban Irish and Japanese linguistic landscapes. In E. Shohamy, E. Ben-Rafael, & M. Barni (Eds.),

299

Linguistic landscape in the city (pp. 19–36). Bristol: Multilingual Matters.

Kataoka, K. (1997). Affect and letter-writing: Unconventional conventions in casual writing by young Japanese women. Language in Society, 26(1), 103–136.

Kataoka, K. (2003). Form and function of emotive pictorial signs in casual letter writing. Written Language & Literacy, 6(1), 1–29. http://dx.doi.org/10.1075/wll.6.1.02kat

Kay, G. (1995). English loanwords in Japanese. World Englishes, 14(1), 67–76.

Kess, J. F., & Miyamoto, T. (1999). The Japanese mental lexicon: Psycholinguistic studies of kana and kanji processing. Amsterdam: John Benjamins.

Kikuchi, S. (1994). Daigakusei no gairaigo ishiki: imēji, hyōki, goruīshiki no kensa kara [University students’ consciousness of western loanwords: from a study of image, transcription, and lexical awareness]. Iwate daigaku, kyōiku gakubu, fuzoku kyōiku jissen kyōiku shidō sentā kenykyū kiyō [Iwate University Department of Education, Educational Practice, and Research Guidance Centre Bulletin], 4, 61–73.

Knospe, S. (2015). A congnitive model for bilingual puns. In E. Z. Winter-Froemel (Ed.), Wordplay and metalinguistic/metadiscursive reflection. Berlin: De Gruyter Mouton

Kress, G. (2010). Multimodality: A social semiotic approach to contemporary communication. New York: Routledge.

Kress, G. (2011). What is a mode? In C. Jewitt (Ed.), The Routledge handbook of multimodal analysis. London: Routledge.

Kress, G., & Leeuwen, T. Van. (2006). Reading images: The grammar of visual design (2nd ed.). New York: Routledge.

Loveday, L. J. (1996). Language contact in Japan. New York: Oxford University Press.

Maynard, S.K. (2008). Maruchijanru danwa ron [Multigenre Discourse Theory]. Tokyo: Kuroshio

Miller, L. (2003). Graffiti photos: Expressive art in Japanese girls’ culture. Harvard Asia Quarterly, (Summer).

Miller, L. (2004). Those naughty teenage girls: Japanese kogals, slang, and media

300

assessments. Journal of Linguistic Anthropology, 14(2), 225–247. http://dx.doi.org/10.1525/jlin.2004.14.2.225

Miller, L. (2011). Subversive script and novel graphs in Japanese girls’ culture. Language & Communication, 31(1), 16–26. http://dx.doi.org/10.1016/j.langcom.2010.11.003

Miller, R. A. (1982). Japan’s modern myth. New York: Weatherhill.

Ministry of Education Culture Sports Science and Technology Japan. (1991). Gairaigo no hyōki [The transcription of loanwords]. Retrieved 14 February 2017, from http://www.mext.go.jp/b_menu/hakusho/nc/t19910207001/t19910207001.html

Miyake, K. (2007). How young Japanese express their emotions visually in mobile phone messages: A sociolinguistic analysis. Japanese Studies, 27(1), 53–72. http://dx.doi.org/10.1080/10371390701268646

Miyoshi, H. (2013). Nihon no raberu: Meiji, Taishō, Shōwa [Japanese Labels: Meiji, Taisho, and Showa periods] (2nd ed.). Kyoto: Shikosha.

Moriarty, M. (2013). Contesting language ideologies in the linguistic landscape of an Irish tourist town. International Journal of Bilingualism, (April 2013). http://dx.doi.org/10.1177/1367006913484209

Nagashima, H. (2006). Sha-re: A widely accepted form of Japanese wordplay. In J. M. Davis (Ed.), Understanding humor in Japan. Detroit: Wayne State University Press.

Nakamura, K. (2012, December 8). Gyaru-moji, ba o wakimae ['Gyaru-moji' should be determined by context]. Asahi Shimbun, p. 14.

Nathan, J. (2012, August 2). Ice cream’s Jewish innovators. Tablet. Retrieved 9 April 2016 from http://www.tabletmag.com/jewish-life-and-religion/108106/ice-creams-jewish- innovators

National Institute for Japanese Language and Linguistics (2006). 'Gairaigo' īkae teian: wakarinikui gairaigo o wakariyasuku sure tame no kotobazukai no kufū ['Loanword' alternative expressions:A scheme of terms for making hard to understand loanwords easy to understand. Retrieved 5 August 2015 from http://pj.ninjal.ac.jp/gairaigo/Teian1_4/iikae_teian1_4.pdf

301

Nittono, H. (2016). The two-layer model of 'kawaii': A behavioural science framework for understanding kawaii and cuteness. The East Asian Journal of Popular Culture. 2(1) 79- 95.

Okada, M. (2012). Wordplay as a selling strategy in advertisements and sales promotion. In J. Chovanec & I. Ermida (Eds.), Language and humour in the media. Newcastle upon Tyne: Cambridge Scholars Publishing.

Okamoto, S., Shibamoto-Smith, J.S. (2016) The social life of the Japanese language: Cultural discourse and situated practice. Cambridge: Cambridge University Press

Olah, B. (2007). English loanwords in Japanese: Effects, attitudes and usage as a means of improving spoken English ability. Bunkyō gakuin daigaku ningen gakubu kenkyū kiyō[Bunkyō Gakuin University, School of humanities research bulletin], 9(1), 177–188.

Olivo, W. (2001). Phat lines: Spelling conventions in rap music. Written Language & Literacy, 4(1), 67–85. http://dx.doi.org/10.1075/wll.4.1.05oli

Otake, T. (2013, June 27). Gifu man, 71, sues NHK for distress over its excess use of foreign words. The Japan Times. Retrieved 26 February 2015 from https://www.japantimes.co.jp/news/2013/06/27/national/gifu-man-71-sues-nhk-for- distress-over-its-excess-use-of-foreign-words/#.WabL3MgjFIr

Otake, T. (2007, September 23). Japanese: A language in a state of flux. The Japan Times. Retrieved 6 August 2015 from http://www.japantimes.co.jp/life/2007/09/23/to-be- sorted/japanese-a-language-in-a-state-of-flux/#.WXqGk4SGOUl

Otsuji, E., & Pennycook, A. (2010). Metrolingualism: Fixity, fluidity and language in flux. International Journal of Multilingualism, 7(3), 240–254. http://dx.doi.org/10.1080/14790710903414331

Oyama, R. (1998). Visual Semiotics: A study of images in Japanese advertisements. (Doctoral thesis, University of London, England). Retrieved 9 June 2015 from http://eprints.ioe.ac.uk/19140/7/298112_Redacted.pdf

Pennycook, A. (2010). Language as a local practice. Hoboken: Routledge.

Piller, I. (2001). Identity constructions in multilingual advertising. Language in Society, 30(2), 153–186. http://dx.doi.org/10.1017/S0047404501002019

302

Piller, I. (2003). Advertising as a site of language contact. Annual Review of Applied Linguisitics, 23, 170–183.

Randall, N. (2002). Lingo online: A report on the language of the keyboard generation. Retrieved April 9 2015 from http://www.arts.uwaterloo.ca/~nrandall/LingoOnline- finalreport.pdf

Rebuck, M. (2002). The function of English loanwords in Japanese. Nagoya University of Commerce and Business Administration Journal of Language, Culture and Communication, 4(1), 53–64.

Renner, V. (2015). Lexical blending as wordplay. In E. Z. Winter-Froemel (Ed.), Wordplay and metalinguistic/metadiscursive reflection. Berlin: De Gruyter Mouton

Ross, N. (2006). Writing in the information age. English Today, 22(3), 39–45. http://dx.doi.org/doi:10.1017/S0266078406003063

Sanseidō Henshūjo. (2010). Konsaisu katakanago jiten [Concise Loanword Dictionary] (9th ed.). Tokyo: Sanseidō.

Sasahara, H. (2002). Keitai mēru ni okeru moji hyōki no tokochō to sono eikyō [The features and effects of the transcript of characters in mobile-phone messages]. The Japanese Journal of Language in Society, 5(1), 105–116.

Sase, Y. (2009). Forms and functions of Englishisms in Japanese womens' fashion magazines. (Masters thesis, San Jose State University, U.S.A). Retrieved 9 December 2015 from http://scholarworks.sjsu.edu/cgi/viewcontent.cgi?article=4698&context=etd_theses

Saussure, F. (2015). Course in general linguistics. (R. Harris, Ed.). London: Bloomsbury.

Schorn, R., Brunner-Sperdin, A., & Ploner, J. (2014). The influence of color, shape, and font formatting on consumers’ perception of online drugstores, Advances in Consumer Research, 42, 357–361.

Seargeant, P., & Tagg, C. (2011). English on the internet and a 'post-varieties' approach to language. World Englishes, 30(4), 496–514. http://dx.doi.org/10.1111/j.1467- 971X.2011.01730.x

Sebba, M. (2012). Orthography as social action: Scripts, spelling, identity and power. In Jaffe,

303

A., Androutsopoulos, J., Sebba, M., Johnson, S., (Eds.) Orthography as social action. Berlin: De Gruyter.

Sekine, K. (2014, November 5). Gyoza wa gairaigo ['Gyoza' is a loanword]. Yomiuri Shimbun [Yomiuri News].

Sharma, B. K. (2012). Beyond social networking: Performing global Englishes in facebook by college youth in Nepal. Journal of Sociolinguistics, 16(4), 483–509.

Shelton, B., & Okayama, E. (2006). Between script and pictures in Japan. Visible Language, 40(2), 157–176.

Sherry, J. F., & Camargo, E. G. (1987). May your life be marvelous: English language labelling and the semiotics of Japanese promotion. The Journal of Consumer Research, 14(2), 174–188.

Shibatani, M. (1990). The languages of Japan. Cambridge: University Press.

Smith, J. S., & Schmidt, D. L. (1996). Variability in written Japanese: Towards a sociolinguistics of script choice. Visible Language, 30(1), 46–71.

Spitzmüller, J. (2012). Floating ideologies: Metamorphoses of graphic 'Germanness'. In A. Jaffe, J. K. Androutsopoulos, & M. Sebba (Eds.), Orthography as social action. Berlin: De Gruyter.

Spitzmüller, J. (2015). Graphic variation and graphic ideologies: A metapragmatic approach. Social Semiotics, 25(2), 126–141. http://dx.doi.org/10.1080/10350330.2015.1010323

Stanlaw, J. (1987). Japanese and English: Borrowing and contact. World Englishes, 6(2), 93– 109.

Stanlaw, J. (2004). Japanese English: Language and culture contact. Hong Kong: Hong Kong University Press.

Stöckl, H. (2005). Typography: Body and dress of a text – a signing mode between language and image. Visual Communication, 4(2), 204–214.

Sunakawa, C. (2008). Script choice and its indexical meanings: The usage of instant messaging among Japanese students in a U.S. city. Studies in English and American Literature, 43(March), 139–165.

304

Tamaoka, K., Hatsuzuka, M., Kess, J. F., & Bogdan, D. R. (1998). Hiragana vs. Katakana: The script familiarity effect on the processing of words and pseudo-words. The Science of Reading, 42(1).

Tomoda, T. (1999). The impact of loan-words on modern Japanese. Japan Forum, 11(1).

Tranter, N. (2008). Nonconventional script choice in Japan. International Journal of the Sociology of Language, 192, 133–151.

'Tsubuyaku' kuma shustubō jyōhō ['Tweeted' bear warning information] (2014, September 4). Yomiuri Shimbun [Yomiuri News], p. 30. van Leeuwen, T. (2005). Introducing social semiotics. New York: Routledge. van Leeuwen, T. (2006). Towards a semiotics of typography. Information Design Journal, 14(2), 139–155. http://dx.doi.org/10.1075/idj.14.2.06lee van Leeuwen, T. (2011). Multimodality. In Simpson, J. (Ed.) Routledge handbook of applied linguistics (pp. 668–682). Florence: Taylor and Francis.

Walker, S. (2000). Typography and language in everyday life. Florence: Taylor and Francis.

Woodside, A. G. (2010). Case study research: Theory, methods and practice. Bingley: Emerald Group Publishing Limited.

Yamada, S. (2013). Yoku aru 'kotoba' no shitsumon [Common questions about 'language']. Retrieved 28 May 2015 from http://pj.ninjal.ac.jp/QandA/topic/post-44/

Yeemar. (2005). Kotoba Kaigishitsu [Language Meeting-room]. Retrieved 5 September 2016 from http://kotobakai.seesaa.net/archives/200512-1.html

Zappavigna, M. (2015). Searchable talk: The linguistic functions of hashtags in tweets about Schapelle Corby. Global Media Journal, 9(1).

305

Appendices

Appendix 1: Full list of codes used in corpus analysis

Appendix 2: Survey (Japanese with English translation)

Appendix 3: Interview/focus group administration documents

Appendix 4: Interview/focus group questions and picture prompts

Appendix 5: Sample interview transcript 2

Appendix 1: Full list of codes used in corpus analysis

Coding at grapheme level (N=161 words with long vowels)

Vowel lengthening grapheme o Bar (eg.くりーむ) o Additional vowel character (eg. くりいむ)

Coding at word level (N=628 words)

Word Form o Noun o Verb o Adjective/Noun modifier o Other word form

Coding at word/phrase unit level (N= 516 units)

Referential Meaning o Blog title or blog post title o Business description o Business name . Goods . Services . Hostess bar . Restaurant/café . Other business name o CD or DVD title o Concrete noun o Event name o Game title o Menu item o Message, wish or greeting o Person or group name o Place name o Product description o Product name . Cosmetics/toiletries . Fresh produce . Omiyage food . Personal accessories . Processed food . Stationery . Toys and collectibles . Other product name 4

o Publication title . book . magazine . pamphlet . Community paper o Service name o Relationship or person description o Other referential meaning

Source Language o Dutch o English o French o German o Italian o Latin o Portuguese o Other language

Language Play o Homonym o Paronym o Lexical blend o Concealed word

Consecutive number of hiragana loanword(s) per unit o Partial word o Single word o Two words in a row o Three words in a row

Coding at text level (N=497 texts)

Number of units of hiragana loanword(s) per text o Single unit o Two units o Three units o Repetition

Genre o Advertisement o Business sign 5

o CD or DVD o Directional sign o Information sign o Menu . Shop window display . Printed menu . Online menu . Vending machine option o Print-club images o Product o Product label or sign o Product packaging o Publications . Book . Brochure/pamphlet . Community paper . Magazine . Newspaper . Manga o Website . Commercial website . Social networking site o Other genre

Coding within and between texts (N=93 cases of inconsistency)

Inconsistency o Alternative script for a particular loanword within a text o Alternative script for a particular loanword between texts o Vowel inconsistency Appendix 2: Survey (Japanese with English translation) 外来語の表記に関する調査

調査の概要

現在、私は「外来語の表記」に関する調査を行っています。日本語において、外来語（外国から来た言葉）は、様々な文字で表現されています。本アンケートでは、そうした外来語からどのような印象を受けるかについて、皆さまのご意見を伺いたいと思います。

本アンケートには匿名でお答えいただきます。そのため、アンケート結果から個人が特定されることはありません。所要時間は約10分です。ご協力どうぞよろしくお願いいたします。

* 1. 私は、「外来語の表記」についての調査に参加することに同意します。

はい

いいえ

1 外来語の表記に関する調査

あなたに関する質問です

* 2. 自分は日本語の母語話者であると認識していますか。

はい

いいえ

3. 性別

女

男

* 4. 年齢

<18

18-20

21 - 30

31 - 40

41 - 50

51 - 60

> 61

5. 最終学歴

小学校

中学校

高等学校

大学

大学院

2 6. この調査の後で、グループ・インタビュー（フォーカス・グループ）を行う予定です。後日、このグループ・インタビューへの参加協力のお誘いメールを送ってもよろしいでしょうか。

はい

いいえ

7. 質問６で、「はい」とお答えくださった方、お手数ですがメールアドレスをご記入ください。「いいえ」の方は、次の質問におすすみください。

8. 質問６で「はい」と答えた方にお尋ねします。グループ・インタビューに参加していただく場合、場所としてはどの都市が一番ご都合がよろしいでしょうか？

東京

名古屋

大阪

神戸

その他の都市

3 外来語の表記に関する調査

第１部文字の印象

9. 「ひらがな」からあなたはどんな印象を抱きますか？ひらがなを見たときに抱くあなたの印象を、名詞か形容詞を使って、下記の空欄にできるだけ多く書いてください。

4 外来語の表記に関する調査

第１部文字の印象

10. 「カタカナ」からあなたはどんな印象を抱きますか？カタカナを見たときに抱くあなたの印象を、名詞か形容詞を使って、下記の空欄にできるだけ多く書いてください。

5 外来語の表記に関する調査

第１部文字の印象

11. 「漢字」からあなたはどんな印象を抱きますか？漢字を見たときに抱くあなたの印象を、名詞か形容詞を使って、下記の空欄にできるだけ多く書いてください。

6 外来語の表記に関する調査

第１部文字の印象

12. 「ローマ字」からあなたはどんな印象を抱きますか？ローマ字を見たときに抱くあなたの印象を、名詞か形容詞を使って、下記の空欄にできるだけ多く書いてください。

7 外来語の表記に関する調査

第２部表記の実例

写真を見て、下記の質問にお答えください。

13. このお店の名前に、「はぴねす」という文字が入っています。ハピネス(happiness/幸せ）をひらがなで書くのは…

適切です。

不適切です。

14. その理由をお答えください。

15. このお店が、ひらがなを使ったのはなぜだと思いますか？

8 外来語の表記に関する調査

第２部表記の実例

写真を見て、下記の質問にお答えください。

16. この商品（ジュース）の名前には「れもん」という文字が使われています。レモンをひらがなで書くのは…

適切です。

不適切です。

17. その理由をお答えください。

18. この会社が、ひらがなを使ったのはなぜだと思いますか？

9 外来語の表記に関する調査

第２部表記の実例

写真を見て、下記の質問にお答えください。

19. この商品（シール）には「ふらい」という文字が使われています。フライ(fly/飛ぶ）をひらがなで書くのは…

適切です。

不適切です。

20. その理由をお答えください。

21. このデザイナーが、ひらがなを使ったのはなぜだと思いますか？

11 外来語の表記に関する調査

第２部表記の実例

写真を見て、下記の質問にお答えください。

22. この商品（インスタントコーヒー）には「珈琲」という文字が使われています。コーヒーを漢字で書くのは…

適切です。

不適切です。

23. その理由をお答えください。

24. この会社が、漢字を使ったのはなぜだと思いますか？

12 外来語の表記に関する調査

第２部表記の実例

写真を見て、下記の質問にお答えください。

25. この商品（ハンカチ）には「hankachi」という言葉が書いてあります。ハンカチをローマ字で書くのは…

適切です。

不適切です。

26. その理由をお答えください。

27. このデザイナーが、ローマ字を使ったのはなぜだと思いますか？

13 Initial page I am currently researching how loanwords are written. In Japanese, loanwords (words that have come from foreign countries) can be written in different ways. This survey seeks to find what impressions these different ways of writing have on people, so I’d be grateful for your opinion.

This survey will collect some personal details from you as well as your answers to the survey. However, your individual details will not be used for any purposes other than the research, or published in any research findings.

The survey will take around 10 minutes to complete. Please consider taking part.

1. Do you wish to participate in “How loanwords are written in Japanese”? Yes No

Questions about you

Please select the options that best apply to you.

2. Do you consider yourself to be a native speaker of Japanese? Yes No 3. Sex: Male / Female 4. Age: Under 20 21-30 31-40 41-50 51-60 61+ 5. Last stage of education: Primary / Middle / High school / Graduate / Postgraduate 6. Are you willing to take part in a follow-up focus group? Yes No 8. If yes, please enter your email address: If yes, which city would be most convenient for you? Tokyo Nagoya Osaka Kobe Other (Please specify)

Section 1: Impressions of the scripts

9. What kind of image does hiragana have? Please enter as many words (nouns, adjectives) below.

10. What kind of image does katakana have? Please enter as many words (nouns, adjectives) below.

11. What kind of image does kanji have? Please enter as many words (nouns, adjectives) below.

12. What kind of image does romaji have? Please enter as many words (nouns, adjectives) below.

Section 2: Examples of writing

Look at the photo below and answer the following questions.

This shop’s name contains ‘happiness’.

13. Using hiragana for ‘happiness’ is appropriate / inappropriate.

14. Please tell us the reason for your answer.

15. Why do you think the shop used hiragana?

Look at the photo below and answer the following questions.

This product name (a drink) contains the word ‘lemon’.

16. Using hiragana for ‘lemon’ is appropriate / inappropriate.

17. Please tell us the reason for your answer.

18. Why do you think the company used hiragana?

Look at the photo below and answer the following questions.

This product (a sticker) contains the word ‘fly’.

19. Using hiragana for ‘fly’ is appropriate / inappropriate.

20. Please tell us the reason for your answer

21. Why do you think the designer used hiragana? Look at the photo below and answer the following questions.

This product (instant coffee) contains the word ‘coffee’.

22. Using kanji for ‘coffee’ is appropriate / inappropriate.

23. Please tell us the reason for your answer.

24. Why do you think the company used kanji?

Look at the photo below and answer the following questions.

This product (a handkerchief) has the word ‘hankachi’ written on it.

25. Using romaji for ‘hankachi’ is appropriate / inappropriate.

26. Please tell us the reason for your answer.

27. Why do you think the designer used hiragana?

Thank you for your participation in this survey! Appendix 3: Interview and focus group documents

This appendix contains the following documents used in the interviews and focus groups:

Plain language statement

Consent form

Participant background questionnaire

研究の概要: グループインタビュー (フォーカスグループ)

研究題目：借用から所有へ：外来語表記の変化について

研究者：トムソン・シーリア（言語学部）

中根育子（アジア・センター）

クナート・ハナ（博士課程、言語学部）

この研究では、英語の外来語が、日本語において、どのように書かれているかについて調査します。特に本研究では、異なる方法で表記される外来語を、人々がどのように解釈し、どのように考えているかについて調査します。

本調査への参加に同意していただける場合は、以下の形式で調査に参加していただきます。

１）フォーカスグループで 2 人から５人ぐらいの人たち(日本語母語話者)と外来語について話し合います。

２）その会話の模様をビデオで録画します。

法の許す範囲で、匿名を使用し、あなたの（フォーカスグループにおいて）記録されたデータ（映像、音声等）について秘密を厳守いたします。あなたの名前と連絡先は、パスワードで守られたコンピューターのファイルに保存されます。それらは、記録されたデータとは別に保存されます。本調査で記録されたデータ（映像、音声等）が、他の調査状況もしくは学術的発表の場（会議、論文、報告書、書籍等）で、あなたについて言及する必要が出た場合、仮名を使います。第三者があなたの身元について推測できるような個人的情報は公になりません。

研究のために収集したデータは、言語学部研究所に５年間保管され、論文に使用された後、第三者の手に渡らないよう破棄されます。

この研究への参加協力は、強制ではありません。従って、ある時点で協力の取りやめを希望される場合、また提供した未処理のデータを撤回したい場合には、不利益を被らずに自由に取り消すことができます。本調査への参加や不参加、調査途中での中断をすることで、あなたが学生である場合学業成績には一切影響はありません。また，メルボルン大学との学術的なつながりにも、一切影響はありません。

ご協力いただける場合は、添付の同意書に署名をし、研究調査者に返却してください。これによって、この書面を読み、そこに書かれた内容を理解したということが確認され、研究者からインタビューの予定を決めるための連絡があります。この研究を元にした論文発表が終わったら、ご希望があれば発表された論文をお送りします。

さらに詳しい情報を必要とされる場合、また何か懸案事項がある場合には、研究調査者にご連絡ください。

トムソン・シーリア: + 61 3 8344 5488; [email protected]

中根育子: +61 3 8344 8893; [email protected]

クナート・ハナ： [email protected]

本研究は、メルボルン大学倫理委員会（the Human Research Ethics Committee）によって承認されております。本調査の実施についてご意見や懸念がありましたら、メルボルン大学倫理委員会委員長までご連絡ください。

（電話 +61 3 8344 2073, ファックス +61 3 9347 6739）

メルボルン大学言語学部 Consent form for Focus Group participants

PROJECT TITLE: From loaning to owning: A study of changes in English loanword orthography in Japan 外来語の借用から所有へ：外来語表記の変化について

Name of participant: email: Names of investigators: トムソン・シーリア、中根育子、クナート・ハナ

1. 私は上記の研究調査に参加すること、またそのためにフォーカス・グループの内容について説明を受けたということに同意します。また、その詳細を書面にしたものを保管用に受領しました。

2. 私は本コンセント・フォームへ署名し提出すること、ならびに研究者がそれを保管することに同意します。

3. 私は、フォーカスグループで３人から５人の人たちと外来語について話し合うこと、またフォーカスグループ内での会話がビデオテープで録画されることに同意します。PLAIN LANGUAGE STATEMENTに書いているとおり、研究者がこの録画データを使うことに同意します。

私は、以下のことを承認します:

(a) 私は、フォーカス・グループの結果がもたらしうる影響について、十分な説明を受けた

(b) 私は、どの時点でも、説明や不利益を被ることなしに自分の意志で参加を拒否することができるということ、また、提供したデータで処理されていないものを撤回することができるという説明を受けた

(d) 私は、何らかの法的必要性が生じた場合、私が提供した情報は５年間守秘が約束されるという通知を受けた

(f) この研究について発表されるすべての論文において仮名で言及されるということ

(g) 研究調査者を通し、この研究に関連した出版物を入手することができるという通知を受けた

この説明を受けた研究に参加することに同意します。 □ はい □ いいえ (チェックしてくだいさい)

この研究に関連した出版物を入手したいです。 □ はい □ いいえ (チェックしてくだいさい)

参加者署名年月日＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿

あたなに関する情報

１．お名前

------

２．当てはまる年齢群に○をつけてください

18 歳未満

19 – 30 歳

31 – 40 歳

41- 50 歳

51- 60 歳

61 – 70 歳

70 歳以上

３．誕生地（都市の名前をお答えください：例, 東京，名古屋）

------

４．６ヶ月以上住んだことなる都市の名前（海外も含む）

都市名: 合計滞在年数:

５．あなたの使う言語を教えて下さい。またそのレベルについてあてはまるレベルに○をつ

けてください:

言語名: 挨拶/日常会話 /上級 / ネイティブ

６．あなたの現在の職業を教えてください

------

７．あなたの趣味を教えて下さい

------

８．あなたの日常生活（仕事，趣味，勉強）において，日本語話者以外の話者や英語話者，英語の本・Email，英語のウェブサイトなどに触れる機会はありますか？もし触れる機会がある場合，それはどのような状況でどのぐらいの頻度でしょうか？詳細を教えて下さい（下のカッコの中に記入してください）。

質問は以上です。ご協力，大変ありがとうございました。

Appendix 4: Interview/Focus Group questions and picture prompts

Introductory questions 今日は、お忙しいところ集まって、ありがとうございます。私はハナといいます。オーストラリアのメルボルン大学の博士課程で言語学を勉強しています。

今日は、これから、日本語で外来語がどうやって表されているか、についての質問をしたいと思います。外来語というのは皆さんご存知のように外国から来た言葉ですね。例えばテレビとテーブルですね。

大丈夫ですね？はい、それでは最初の質問をしますね。

外来語は、普通はカタカナで書かれているとおもいますが、それでは、ひらがなで書かれている外来語を見たことがありますか？

それはどのようなものですか。例えば、商品や店の名前、映画のタイトルなど。

今から，このようなものについて話します。

(例を示す：1. れふぁいん、2. ゆずれもん)

Thank you for making time to participate in this project. My name is Hannah, I’m a PhD candidate in the Linguistics department of the University of Melbourne.

Today, I’d like to ask you some questions about how loanwords are written in Japanese. As you know, loanwords are words that have come from other countries, such as ‘television’ and ‘table’.

Is that ok? Right, I’m going to ask you the first question.

★ Have you noticed loanwords being written in hiragana before?

What kinds of things? Product or shop names, movie titles, for example.

Here are some examples of the kinds of texts we’re going to talk about today.

(Produce examples: 1. Refine, 2.Yuzu Lemon)

1. れふぁいん/refain (Refine)

2. ゆずれもん/yuzu remon (Yuzu Lemon)

Topic 1 – Product names

商品の名前についての質問をしたいと思います。たまに、企業はひらがなを使って外来語の商品を表記することがありますね。思い浮かびますか？例えば「サクサクぱんだ」みたいなものです。

この商品名の中には、ひらがなで書かれている外来語があります。

（例を示す：3. いろいろすーぷ、4. たおるマフラー、5. スプラッシュあいらんどう）

★なぜそのようにするのだと思いますか？

外来語をあえてひらがなで書くことには、何か特別な意味があると思いますか？

ひらがなで書かれている外来語をみたとき、そこからどんな印象を受けますか？

もし、その商品がカタカナで書かれていたら、ひらがなで書かれていた場合と比べて何か違う印象を受けると思いますか？

Sometimes, businesses use hiragana to write their product names, even though they’re loanwords. Can you think of any? For example, like ‘Saku-saku Panda’?

Here are some examples of products with hiragana loanwords in their names.

(Produce examples: 3. Various soups, 4. Towel Muffler, 5. Splash Island)

★Why do you think companies use loanwords in hiragana in their product names?

Do you think it sends a particular message?

Does it create a certain image?

Would the product seem different if it was written in katakana?

3. いろいろすーぷ/iroiro sūpu (Various Soups) 4.たおるマフラー/taoru mafurā (Towel Muffler)

5. スプラッシュあいらんど/supurasshu airando (Splash Island)

Topic 2 – Business names 次に、企業やお店の名前についての質問をしたいと思います。外来語を使っているのに、ひらがなを使って表記している会社やお店の名前がありますよね。思い浮かびますか？例えば、ファミリーレストランの「すかいらーく」です。

これは、外来語をひらがなで書いている企業名の一例です。

（例を示す：6. らいふアシスト馬車道、7. はっぐみぃー、8.ふろーりあ）

★その会社やお店は、なぜそのようにするのだと思いますか？

外来語をあえてひらがなで書くことには、何か特別な意味があると思いますか？

ひらがなで書かれている外来語から、どんな印象を受けますか？

もし、そのお店や会社の名前がカタカナで書かれていたら、ひらがなで書かれていた場合と比べて、何か違う印象を受けると思いますか？

ちょっとだけ町を歩いて、私なりに調べてみたんですね。そうしたら、ひらがなを使っているお店は、特にスナックやホステス・バーで多かったんです。

（例を示す：9.すなっく新、10. ゆにこーん、11. らうんじちあき、

12. すぷりんぐ、13. ぺんぎん、14. あいどる）

★でもそれは、なぜだと思いますか？

Next, I’d like to ask you about business names. Some business and shop names are written in hiragana, even though they’re loanwords, right? Can you think of any? Like the family restaurant chain, ‘Skylark’, for example?

Here are some businesses that use hiragana loanwords in their names.

(Produce examples: 6. Life Assist Bashamichi, 7. Hug Me, 8. Floria)

★Why do you think companies use loanwords in hiragana in their names?

Do you think it sends a particular message?

Does it create a certain impression?

Would the business seem different if it was written in katakana?

When I was walking around town, I investigated these business names. Of all the business that used hiragana loanwords for their names, snack and hostess bars were some of the most common.

(Produce examples: 9. Snack Shin, 10. Unicorn, 11. Lounge Chiaki, 12. Spring,

13. Penguin, 14. Idol)

★Why do you think that is?

6. らいふアシスト馬車道/ raifu ashisuto bashamichi (Life Assist Bashamichi)

7. はっぐみぃー/haggu mii (Hug me)

8. ふろーりあ/furōria (Floria)

9. すなっく新 / sunakku shin (Snack Shin) 10. ゆにこーん/yunikōn (Unicorn)

11. らうんじちあき/raunji chiaki (Lounge Chiaki) 12. すぷりんぐ/supuringu (Spring)

13. ぺんぎん/pengin (Penguin) 14. あいどる/aidoru (Idol)

Topic 3 – Print club

今度は，プリクラについて話題をうつします。プリクラって，とったあとに，いろいろなメッセージを写真に書き込みますよね。このまえ，外来語はあえてひらがなでかいたメッセージのあるプリクラを発見しました。

（例を示す：15. おーばーおーるまん、16. あいらびゅ、17. ばれーぼーる、

18. ふぁっきゅー）

★なぜそうだと思いますか？

それは、ひらがなには、可愛いイメージがあるからですかね？

それとも、女っぽい感じがするからでしょうか？

それとも、何かへの反抗や抵抗ですかね？

もし、このプリクラのメッセージがカタカナで書かれていたら、違うイメージしますか？

Next, I’d like to talk to you about ‘print club’. These stickers are made by having your picture taken, and then writing messages and other things on the photo. I found some print club images with loanwords in hiragana on them…

(Produce examples: 15. Overall Man, 16. I love you, 17. Volleyball, 18. Fuck you)

★Why do you think this is?

Is this because hiragana is cute?

Or feminine?

Is it an act of ‘rebellion’ against something?

How would these print club images be different if the writing was in katakana?

15. おーばーおーるまん/ōbāōrūman (overall-man) 16. あいらびゅ/airabyu (I love you)

17. ばれーぼーる/barēbōru (volleyball) 18. ふぁっきゅー/fakkyū (fuck you)

Topic 4 – Social Media

次は，ソーシャルメディアについてお伺いします。みなさんは、ソーシャル・メディアをつかっていますか？たとえば、フェースブック、ツイッター、ミクシーなどです。その時，外来語をひらがなで書いたことありますか。またそういう書き込みを見かけたことがありますか。その言葉って覚えますか？

とうこうこれは、フェースブックにアップされた、ひらがな外来語の投稿です。

（例を示す：19. ふぁみりー、20. にゅーずぃーらんどぅ、21. いえすあいあむ、

22. はぴばーすでい)

★この人たちは、なぜひらがなを使ったと思いますか？

こうした書き込みが、もしカタカナで書かれていたら、イメージは違いますか？

これを書いた人が、男性だったか女性だったかわかりますか？

実は、一つのサンプルは男性から別の男性の友達に向けて書かれたものです。これを聞いてびっくりしますか？

Next, I’m going to ask about social media. Do you use social media apps like ‘Facebook’, ‘Twitter’, or ‘Mixi’? Have you ever written loanwords in hiragana? Or have you noticed messages like that? Do you remember what word(s) they were?

Here are some examples of hiragana loanwords from facebook.

(Produce examples: 19. family, 20. New Zealand, 21. yes I am, 22. happy birthday)

★Why do you think they decided to write in hiragana?

Would the post be different if the message were in katakana?

Can you tell if the person who wrote it was a male or a female?

Are you surprised that one post was a message from one male to another?

19. ふぁみりぃ/famirī (family)

20. にゅーずぃーらんどぅ/nyuuziirandu (New Zealand)

21. いえすあいあむ/iesu ai amu (Yes I am)

22. はぴばーすでい/ hapi bāsudei (Happy birthday)

Topic 5 – Consecutive words

今まで集めたサンプルの中で、例をおみせしたように、ひらがな外来語は、一つの言葉で書かれている場合が多かったです。例えば「すーぷ」。だけど、たまに、二つや三つの言葉で書かれている場合もあります。

(例を示す: 23. はっぴーばれんたいん、 24. みっくしゅじゅーちゅさわー)

私のサンプルの中ではこれは、珍しかったです。

★なぜ、こうした例が珍しいと思いますか？

In the samples I’ve collected, like the ones I’ve shown you today, examples of one hiragana loanword are more common, such as the word ‘soup’. But sometimes I’ve found two and three consecutive loanwords in hiragana, for example…

(Produce examples: 23. happy valentine, 24. Mix juice sour)

★These examples were rarer. Why do you think this is?

23. はっぴーばれんたいん/happī barentain (Happy valentine)

24. みっくちゅじゅーちゅさわー/mikkuchu jyūchu sawā (Mix juice sour)

Topic 6 - Inconsistencies

たまに、外来語が書かている文脈で、カタカナとひらがなが一貫して使われてないときがあります。一貫性がないというのは、商品名と、商品の説明で、違った外来語表記がされている、という意味です。例をおみせしますね。

（例を示す：25. すぱいす、26. あいすココア）

商品の名前にはひらがなで「すぱいす」と書かれています。でも、説明の中ではカタカナで書かれています。そこで、質問をしたいのですが、

★なぜ、このデザイナーは、外来語の表記を一貫させなかったのだと思いますか？

もしも、全部がひらがなの場合にどんなイメージしますか？

逆に、全部はカタカナの場合は？

Sometimes, inconsistencies were found in how a loanword was written. This means that these texts contain different ways of writing loanwords in the product name and the product explanation. I’ll show you some examples

(Produce examples: 25. Spice, 26. Ice Cocoa)

So the product name is ‘Spice’, in this example, but in the explanation this same word is written in katakana. So I wanted to ask you,

★ Why do you think the designer didn’t use the script consistently?

How would the image change if it were all in hiragana?

Or conversely, all in katakana?

25. すぱいす/supaisu (Spice)

26. あいすココア/aisu kokoa (Ice Cocoa)

Inconsistencies cont.

一貫性がない例は、お店の中でもみられました。例えば、商品の名前とお店の商品のラベルにも見つけました。例えば、こんな感じです。

(例を示す：27. とまと、28. りっぷくりーむ)

★なぜこの店は、商品ラベルを、商品名のようにせずに、カタカナで表記したのでしょうか？

逆の例もみられました。商品名はカタカナで書かれているのに，店が作ったラベルでは，同じ商品名がひらがなで書かれていました。

（例を示す：29. サラダようきゃべつ、30. 宇治抹茶かすてら）

★なぜこの店は、商品ラベルを、商品名のようにせずに、ひらがなで表記したのでしょうか？

★ もう一つの「きゃべつ」の例を見せたいんですが、これ、どう思いますか？

（例を示す：31. キャベツ台）

Inconsistencies were also found inside shops. For example, the product name and the label made by the shop, like these examples

(Produce examples - 27. tomato, 28. lip cream)

★Why do you think the shops decided to write these words in katakana on the labels, rather than how they appeared on the actual products?

I also found some of the reverse situations. The product name written in katakana, but the label that the shop made contains the same word in hiragana.

(Produce examples: 29. cabbage for salads, 30. Uji Maccha Castella)

★Why do you think these shops decided to use hiragana for the labels?

★ I’d like to show you one more ‘cabbage’ example; what do you make of this..?

(Produce examples: 31. cabbage display)

27. とまと/tomato (tomato)

28. 野菜りっぷくりーむ/yasai rippu kuriimu (vegetable lip cream)

29. 宇治抹茶かすてら/uji maccha kasutera (Uji green tea castella)

30. サラダ用きゃべつ/sarada you kyabetsu (Cabbage for salads)

31. キャベツ台/kyabetsu dai (cabbage display)

Topic 7 – Appropriate/inappropriate samples from survey

以前におこなったアンケート調査の中で、ひらがな外来語が、適切か不適切かを聞きました。いまからお見せする例は、適切さに関して、一番大きな差がみられた例です。

（例を示す：32. 蜂蜜れもん、33. I can ふらい!!）

これは一番適切と評価されたのはこれで（れもん）と一番不適切だったのはこれ（ふらい）です。

★なぜこうした差がでたのだとおもいますか？

それは、「レモン」は漢字があるけど、「ふらい」には漢字がないから

でしょうか？

「れもん」は商品名だからもっと，表現の自由度が高いからですかね？

逆に、「ふらい」の場合、英語の文の中にあるから不適切だと思いますか？

★「ふらい」という言葉は、どの文字で書かれたらもっと適切になると

思いますか？

（例を示す：代替の書き方）

In a survey I conducted, I asked people whether they thought hiragana loanwords were appropriate or not. The examples I’m going to show you generated the greatest difference in opinion:

(Produce examples: 32. Honey Lemon, 33. I can fly!!)

These images were voted most appropriate (lemon) and least appropriate (fly).

★Why do you think this is?

Is it because ‘lemon’ also has a kanji and ‘fly’ doesn’t?

Is it because there is more freedom in product names/proper nouns?

Is it because ‘fly’ is part of an English sentence example?

★ How could the ‘fly’ example be made more appropriate?

(Produce 34. Alternatives to ‘I can fly!!’)

32. 蜂蜜れもん/hachimitsu remon (Honey Lemon)

33. I can ふらい/furai (fly)!!

1. あいきゃんふらい!!

2. アイキャンフライ!!

3. I can fly!!

4. I can フライ!!

5. アイキャンふらい!!

6. I can 飛ぶ!!

7. Fly できる!!

34. Alternatives to ‘I can ふらい!!’

Concluding questions

★ひらがなで書かれている外来語は、正しい日本語だと思いますか？

もし、１０歳の子供が、「れもん」という言葉をひらがなで書いたら、あなたは何か言いますか？

もし、あなたの外国人の友達が、ひらがな外来語を書いたら、あなたはしてき

指摘 EA しますか？

★「レモン」や「キャベツ」という言葉は、日本語だと思いますか。

それはなぜですか？

「たばこ」や「ぱん」という言葉はどうですか？

「コーヒー」や「ガラス」という言葉はどうですか？

★ Do you think writing loanwords in hiragana is ‘correct’ Japanese?

How would you respond to a 10-year-old child who wrote ‘lemon’ in hiragana?

What about a foreign friend writing ‘lemon’ in hiragana?

★ Do you consider words like ‘lemon’ and ‘cabbage’ to be Japanese? Why (not)?

What about ‘tabako’ (cigarettes) and ‘pan’ (bread)?

What about ‘koohii’ (coffee) and ‘garasu’ (glass)?

Appendix 5:

Sample interview transcript

Interview transcript excerpt: contains interview S3 (Yūsuke) on product names in hiragana (Towel Muffler, Various Soups) and business names in hiragana (Floria, Hug Me).

H: たまに企業は、ひらがなを使って、外来語の商品を表記することがありますね。先の例みたいなものですが、他に思い浮かびますか？

Sometimes companies use hiragana for their product names, right? Like the examples I showed you a minute ago. Can you think of any others like that?

Y: たまに「おれんじ」とかひらがなで表記されてるんじゃないかなと思います。

I think sometimes ‘orange’ is written in hiragana...

H: そうですか？その「おれんじ」は、ジュースとかですか？

Is that right? Like a juice or something?

Y: そうそう、ジュースですね。

Yeah, yeah, like a juice.

H: なるほど、ありがとうございます！では、もう少し例をおみせしますね。

You don’t say - thanks for that! Ok, I’m going to show you a few more examples.

(H produces 3.‘Various Soups’, 4.‘Towel Muffler’, and 5.‘Splash Island’)

この会社は、なぜこのように表記したのだと思いますか？

Why do you think these companies decided to use hiragana?

Y: そうね…たとえばその「たおる」なんですけど、他の会社や企業も、同じようなタオルとかタオルマフラーを出している場合があるじゃないですか、自分の食品を見せるとか…おしゃれに見せなきゃいけない。ということで「たおる」という文字をひらがなで使って、「マフラー」というのをカタカナにしてるのではないかなと思います。また、逆もしかりで、また「タオル」をカタカナで、「まふらー」がひらがなというのも、ぜんぜんそういうの…目立つためにやるのではないかと思う。逆に、このように「たおるマフラー」っていうひらがなとカタカナの混合…なんていうか「あれ？なんか違和感が …」という風になると思うので…この違和感を逆手に取って、逆に違和感を注目させるというイメージだと思います。狙いがあるんだと思います。

So, for example with the ‘towel muffler’, if other companies are also making some kind of towels or towel mufflers, they want to show off their own product, they have

to make their own product fashionable, so perhaps they decided to make the ‘towel’ part hiragana and the ‘muffler’ part katakana. The other way around would also be good. Because if ‘towel’ was in katakana and ‘muffler’ was in hiragana, it would still totally (be fine), it would still stand out. Conversely, it’s the hiragana-katakana combination, you think ‘huh, there’s something unnatural here…’. The unnaturalness is put to use; conversely the unnaturalness makes you notice it, that’s the image I get. That’s the aim, I think.

H: あぁ、わかりました。これ、いろいろすーぷの場合に、全部ひらがなになっちゃったら、それもなんか…

Ah, I got it, thanks. So in the case of ‘Various Soups’ being all in hiragana…

Y: そうですね。これも、たとえばコーンスープもこの「ぽっか札幌」という会社が出してるんですけど、別の会社さんが出した場合も、やはりその同じようにこの「いろいろスープ」がカタカナでかぶってたら、どっちも同じものに見えてしまう。なのでこっちのほうが名前が知れているとか、知名度で選べてしまって、自分の商品が売れなくなってしまうということがあるんではないのかと思いますね。「すーぷ」という言葉をひらがなにすることで、ちょっと違和感が出て、「あっ」、ちょっと初めて見る、チャレンジして買ってみようかなという好奇心から、あの「あっ、おいしいじゃん、これにしよう」につなげたい、そういう狙いがあるのだと思います。

Yeah so for example if there’s a company that makes corn soup- Well it’s Pokka Sapporo making the soup, and in the case where there’s other companies also producing ‘various’ soup, it might look the same if they are both in katakana. So writing it this way, the name will be remembered, and products are chosen by how popular their names are, so maybe your product wouldn’t sell as well (if it were in katakana), that’s what I think they’re thinking. So putting soup in hiragana makes it a little unnatural, so the first time you see it you think “I might try that”, and you try it because you are curious, and it’s delicious, so you buy it again, and that’s the aim I think.

H: 分かりました、ありがとうございます。次は、お店と会社の名前についてお伺いします。外来語を使っているのに，ひらがなを使って表記している会社やお店の名前がありますよね。先の「れふぁいん」みたいに、この例も準備しました。

I see, thanks for that. Next I’d like to ask you about company and shop names. There are shop and company names which are written in hiragana, even though they’re loanwords, right? Like the ‘Refine’ café I showed you before. I’ve prepared some other examples.

(H produces 6.‘Life assist’, 7.‘Hug me’ and 8.‘Floria’)

この会社やお店は，なぜそのようにするのだと思いますか？

Why do you think these companies used hiragana in their business names?

Y：そうですね、やはりあの先ほどとやはり一緒だと思うんですけど、このフラワーショップっていうお店日本中にいっぱいありますし、その中でも自分の商品を買わせるために、どうしてもあの…ほかの企業とすこし違うところを見せなければいけないんですね。それこそあの皆は皆は同じことやってて、でそれで値段、最後は値段になってしまうんですけど、そこで何かその外観の雰囲気とかもありますし、この名前からでも読みとれる雰囲気っていうのがあるので、特にこの花ってやはり綺麗であったり、可愛いいであったり、色々あるので、少しこういうフォローも何かこの「ろ」に近い形にてんてん少し可愛い感じ崩して、あの崩してかかれているので、そういう狙いがあるのだと思います。先ほどのあの、この、何でしたっけ…

Yeah I think the mood is kind of the same (as the previous examples). So for example there’s heaps of florists in Japan, so somehow in order to make people buy your product, so you have to show how you’re a bit different from the other companies. Because everyone is doing the same thing; the price, of course the price might be the final factor, but then there’s the style of the exterior, and you also get the atmosphere (of the shop) from the name. And so with flowers they’re pretty and cute, and so on, so they want to follow this feeling, make it a cute shape with the ‘ro’「ろ」 character and the diacritics being a bit clumsy, this is the aim I think. It’s the same as the one before, what was it...?

H: 「たおるマフラー」

“Towel Muffler”

Y: 「たおる」があったり、この「すーぷ」であったりと。やはり他の企業と少し違うところを見せて、あの、自分達の独自性というものを出さなければいけないと思うので、そのような狙いかなぁと思いますね。

...with the ‘towel’, and the ‘soup’, they want to show something a little different to the other companies, they have to show their originality, that’s the aim I think.

H: んんん、でも、もしもカタカナで書かれたら、イメージは何か違いますか？

So if it was in katakana, would the image be different?

Y: 普通の…あのこういう、普通のお店、その「普通」って説明しにくいんですけど、あの、その…何ていうんですかね…難しいですけど、ちょっとこの可愛いイメージ…が読み取れない…ようなお店なのではないかなと。ただの花があって、この普通のしっかりとした、すごい真面目なお店っていうんです

かね。少しあの、こういう崩すことによって出してる可愛さが、あの、感じられないような感じがあるのだと思います

It would just be a normal – a normal shop, well, it’s hard to explain what ‘normal’ is but…what can I say…it’s hard… this kind of cute image couldn’t be gotten from the shop (if it was in katakana). It would just have flowers and be a normal proper and really serious shop. It’s a little- you just wouldn’t feel the clumsy cuteness of this one (points at ‘Floria’ picture).

H: なるほど。この「はぐみぃー」と同じように、ちょっと可愛いイメージを…

Oh, ok. So is this ‘Hug me’ the same in giving a cute image?

Y: そう…ですよね。これは写真を見ると、やはり「託児付きの地域塾」って書いてあるのでやはり子供向けというのもありまして、あのへたにあの英語使ってしまったりカタカナ使うのであれば-あの園児募集と書いてあるので、子供、あのその小さいお子様をあの、とりたいというので、は-、親しみやすいように、ひらがなを使っているのだと思います。なので、このようにキャチュコ-キャチュフレーズだけはひらがなにして、あとはもうこの「子供の感性は世界より大きくて広い」だとか、「説明会随時開催中」などの言葉は、あの、保護者さん向けに書かれているものなので、なんとなくニュアンス的に「あっ、何かいいかも」というようなものを、子供にあの植え付けたいのだと思います。

Yeah...I think so. By looking at the photo you can see it’s a classroom-with-childcare, so it’s aimed at children, and if they used English or katakana... And “recruiting kindergarteners” is written there, they want to get small children in, it’s easy for children to feel a familiarity if it’s written in hiragana. That’s why only the catch co- (py), the catch-phrase is in hiragana. And to get the carers’ attention they’ve written things like ‘Children have a sensitivity that’s bigger than the whole world’, and ‘Information sessions available on demand’, so the whole nuance is ‘ah, this might be good!”, that’s the thought they want to plant in the minds of the children.

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Kunert, Hannah R.

Title: From loaning to owning: Japanese loanwords in hiragana

Date: 2017

Persistent Link: http://hdl.handle.net/11343/198270

File Description: From loaning to owning: Japanese loanwords in hiragana

Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.