Teachiog Connected Speech Rules to Japanese Speakers of English so ris to Avoid a Staccato Speech Rhythm
Marie A. MeIenca
A Thesis
in
The Centre
for
Teaching English as a Second Language
Presented in Partial Fulfilment of the Requirements for the Degree of Master of Arts at Concordia University Montréal (Québec) Canada
March 200 1
O Marie Melenca, 200 1 National Library Bibliothèque nationale 1*1 of Canada du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395. me Wellington Ottawa ON KiA ON4 Ottawa ON KI A ON4 Canada Canada Your Ne Votre rélërenCe
Our fi& Notre rë!drnce
The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant a la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sell reproduire, prêter, distribuer ou copies of this thesis in microfonn, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de reproduction sur papier ou sur format électronique.
The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation. Abstract
Teaching Connected Speech Rutes to Japanese Speakers of English so as to Avoid a Staccato Speech Rhythm
Marie Melenca
This study explores the effectiveness of teaching Japanese speakers of English how to connect speech so as to avoid a robotic speech rhythm. A control and an experimental group were each given three one-hour classes in English. The control group read poetry aloud without any explicit instructions. The experimental group were instructed in how to link words with three dserent sound boundaries: (1) consonant to vowel, (2) vowel to vowel, and, (3) consonant to consonant. The ability to link word pairs was rated for both the control and expenmentai groups in pre- and post-treatment situations, and this was compared to a native-speaker baseline of connected word pairs.
This research focussed on whether instruction in connected speech mles would assist in reducing the syncopated rhythm which ofien characterises the speech of
Japanese speakers of English. Although this study was an exploratory one, with a very small sample size, the results were heartening: directly teaching these rules seemed to heighten the awareness of the need to link word pairs. However, the results indicate several problem areas on which researchers and teachers need to focus in order to optimise time spent in pronunciation training. That is, a longer period of instruction with practice may enable a more successfil output. Furthemore, it would be necessary to control expenmental variables more carefùlly and use a larger sample size in order to obtain more definitive support for the clah that teaching comected speech deswould, indeed, generally improve the speech rhythm of Japanese leamers of English. Acknowledgements
This project was given its initial stimulus, and then received unswerving support eom David Levy, the English Director in the Department of Continuing Education at
McGilI University. The researcher i s indebt ed to Gregory Glidden and Marie-Claude
Barrette for their help in judging the speech samples for the study. The researcher is also gratefiil for the kind and helpful assistance of Elaine Shatenstein and RoseLyn SmalI for their assistance in proof-reading severd drafts of the thesis in progress.
The author is also thankfirl for the generosity of the Concordia English Language
Institute (CELI) of Concordia University for providing volunteer Japanese Learners of
English, as well as for Concordia University's TESL Centre for providing classroom space in which to conduct the experiment.
Dr. Beth Gatbonton and Carolyn Cronk were both excellent thesis readers and gave timely advice and encouragement; 1 am eternally groteful for their support.
1 am very gratefül to Dr. Palmer Acheson, who agreed to take on the position of thesis supervisor after the experiment had been conducted; his enthusiasrn for rny research brought a most welcome breath of fresh air. 1 am also grateful to Dr. Lori Moms who was the first thesis supervisor; her suggestions were helpful-
1 dedicate this research to my late father, Mikhail Melenca, whose unshakeable belief in me kept me to the task. TABLE OF CONTENTS
ACKNOWLEDGEMENTS ...... IV
LIST OF FIGURES ...... VIII
LIST OF TABLES ...... IX
LIST OF INTERNATIONAL PHONETIC ALPHABET SYMBOLS ...... X
CHAPTER 1: INTRODUCTION ...... 1
CHAPTER 2 : LITERATURE REVIEW ...... 3
Research on Linking ...... -3 Cotzsonan2 to Vowel - CV ...... -3 Vowel to ?/owel- VV ...... 4 Consottarat to Comonartr - CC...... ,...... 5 Resyllabrfication and Szmpl1~7cafion ...... 7 Terminology ...... 8 Research on Disconnected Speech ...... 11 Research on the Pronunciation Problems of JSsE ...... 13 Reasons for Teaching Linking ...... -19 Summary and Conclusions ...... 22 Research Question ...... -23
CHAPTER 3: DESIGN ...... 24
Overall Design ...... -24 Participants: Roles and Materials ...... 24 v Procedures and Measures ...... 28 Design of Pedzgogical Acrivity ...... ~...... -28 Recording mtd Anabsis of JSE Speech Samples...... 29
CHAPTER 4: RESULTS ...... 31
CHAPTER 5: SUMMARY. DISCUSSION. REC OMMENDAT1 ON S ...... 40
Research Questions ...... -40 Nalive Langriage Trmsfer ...... 41 hredm~guageStratem ...... 43 MaCommon Areas of Discotmecriori ...... 34 Do RernediaC Teachirlg Activities .He@?...... -45 WhaI lo Teach artd met?...... 45 Measuring Learner Performance ...... -46 Disamion of "ReadingA hd'' Resrrlts ...... 48 Discussion of Free-Speech Monologzre Remlts ...... 50 Transcription Discoveries ...... -52 Dismssion of Experimer~ialCot~di~ioris ...... 54 Dis~~tssiotrof Experimenter Bias ...... -54 Disncssiotz of Limitations ...... 5.5
Recommendations ...... -56
REFERENCE S ...... 59
APPENDICES ...... 64 Appendix A: Request for Teachers' Assistance in Recruiting Japanese Speakers of English...... 64 Appendix B: Participant Recruitment Letter ...... -65 Appendix C: Consent Form to Participate in Research (JSsE) ...... 66 Appendix D: Consent Form to Participate in Research (NSsE) ...... 68
Appendix E : Pre- and Post-Experiment Questions (JSsE) ...... 70 Appendix F: Pre-Experiment Questions CNSsE) ...... -71 Appendix G:Clear Speaking Test ...... -...... 72 Appendix H: Experimental Teaching Unit 1 - CV ...... 73 - Appendix 1: Experimental Teaching Unie 2 - VV ...... 73 Appendix J: Experimental Teaching Unit 3 - CC ...... 77 Appendix K: Control Activity : Poem 1 .. - ...... -79 Appendix L: Control Activity: Poem 2 ..-...... 80
Appendix M: Control Activity: Poem 3 .- ...... -81 Appendix N: Transcriptions of Speech Samples while Reading Moud ...... -82 Appendix O: Transcriptions of Free-Speech Samples ...... -129 List of Figures
Figure 1: Comparison of Linking by Native Speakers of English while
Reading Aloud versus Free Speech (%)
Figure 2: Comparison of Actual Versus PotentiaZ Linked Items Made
by Al1 Participants While Reading Aloud (%)
Figure 3 : Comparison of Actual Versus Potential Linked Items Made
by AI1 Participants in Free Speech (%)
... Vlll List of Tables
Table 1 : Types of CV environments across word and syllable boundaries 4
Table 2: Types of W environments across word and syllable boundaries 5
Table 3 : Types of CC environments across word and syllable boundaries 6
Table 4: Number of Linked Word Pairs (CV and W Environments):
Reading Aloud
Table 5: Number of Linked Word Pairs (CV and W Environments):
Speaking Freely List of International Phonetic Alphabet ~~rnbols'
l Excerpt fiom Trask (1996, p. 394) Chapter 1: Introduction
As a pronunciation teacher of English as a second and foreign language
(ESL/EFL), the researcher has noticed that Japanese speakers of English (JSsE) cm often attain near native-speaker proficiency in written production while falling short of native- speaker (NS) fluency in oral production, particularly in tems of prosodic features, especially linking. A high incidence of this phenornenon has been observed, and for this reason, a research focus on the staccato-like speech patterns of JSsE was chosen. In addition, a series of pedagogical activities to assist these learners in irnproving their
English speaking proficiency was designed in order to assist JSsE in improving their ability to connect speech in a more effective manner. The goal of this activity is to help learners of English to achieve fluidity of communication in a native-like manner, using comected speech as a central reflection of that redity.
A multi-level experiment was created to explore the following questions:
1- What are the most common types of linking problems across word- and syllable-
boundaries that contribute to the staccato-like speech of JSsE?
2. Can specific teaching strategies assist in improving the connected-speech patterns of
JSsE?
Before defining connected speech, otherwise known as linking and catenation, it is useh1 to outline what influences oral production. Ensuring the clarity of meaning in discourse can be provided in spoken langage by a variety of means, some of them socio- contextual, and others related to the actuai words spoken in the speech Stream. The rnechanisms include: (a) Grammatical cornpetence, @) semantic cornpetence, (c) 1 phonernic contrasts, (d) phonotactic rules which influence the occurrence of linking, stress, (e) intonation patterns, (0 psychological variables such as attitude and motivation, and, (g) sociolinguistic competence, which assumes the choice of an appropriate speech register, including the recognition of where it is appropriate to link or to separate phonemes.
The importance of each of these mechanisms may Vary from language to language. Native speakers of ~n~lish'attend to al1 eight mechanisms (in varying degrees) to engage in meaningfid and productive communication, as do speakers of other languages. The acoustic level in discourse includes prosodic eues, such as linking and timing.
This paper is primarily concemed with phonotactic rules, sociolinguistic competence, and the appropriate and competent production of both segmental and suprasegmental items. These influence the ability to appropnately link - or not - items in the speech Stream, thereby ensuring the clarity of meaning in discourse.
This thesis will atternpt to: (a) coethat JSsE produce disconnected speech in three environments: CV, W and CC, (b) describe and show the results of a series of three lessons designed to teach JSsE how to hkwithin these three environments, and, (c) contrast the oral output of the experimental and control groups in pre- and post-test conditions.
* This paper focuses on standard North American English (NAE).
2 Chapter 2: Literahire Mew
McNerney and Mendelsohn (1992) define linking, or connected speech, as having no pauses between words within a stream of speech.
In sentences where linking is required students must not pronounce words as separate entities, but make the words flow smoothly together. Linking should be introduced to students not only as a natural aspect of connected speech, but also as a necessary one for comprehensibility. There are times when a potentially ambiguous sentence can only be disambiguated when the appropnate linking .. . is used. (p. 194)
Celce-Murcia, Brinton and Goodwin (1996) have added that linking, which they cal1 connected speech, includes ". .. the comecting of the final sound of one word or syllable to the initial sound of the next" (p. i 58). The five environments of where linking occurs are discussed below: (a) Consonant to vowel (CV), (b) vowel to vowel (VV), (c) consonant to identical consonant (CC), (d) consonant to non-identical consonant (CICI), and (e) resyllabification (Avery & Ehrlich, 1992; Celce-Murcia et al., 1996; and Dauer,
Consonant to Vowel - C V
In word- and syllable-boundary environments, the final consonant (C) is pushed ont0 the following vowel (V). Table 1 shows the types of arîiculatory settings in word or syllable endings possible in CV boundaries. TABLE 1
Types of CV environments across word and syllable boundaries
-- - -- ~ ---~
Phonological Environment Examp les
C (labial) + V stop it, slotppy
C (dental) + V laugh at, flue
C (alveolar) + V played on, atom
C (palatal)+ V rage on, Cajun
C (velar) + V drag out, bakery
Vowel to Vowel - VV
Linking W across word- and syllable-boundaries involves glide insertion- This occurs when linking a word- or syllable-final tense vowel or a diphthong to a following vowel (Celce-Murcia et al., 1996). Vowels in the NAE inwentory, which End themselves in word- and syllable-final position, end in off-glides, /y/ and /w/~,and these make links easy to produce. In fact, open syllables in English contain only tense vowels which end in off-glides (Kreidier, 1989). Table 2 shows the types of articulatory settings in word or
Several vowels in NAE are tense and have off- glide endings - [iy], [ey], [uw], and [ow] - and these are sometimes reflected in NAE spelling, as is evident in some words, such as "t1-f' and "how. " Some phonetic inventories do not indicate off-glides in theu representations; instead, these inventories use two dots to indicate tension. That is, Ci:] represents [iy], [e:] represents [ey], [u:] represents [uw], and [O:] represents [ow]. Some of these systems may also avoid the use of off- glides for diphthongs, replacing toi] for [oy], [au] for [aw], and [ai] for [ay]. Whatever the system of notation used, the off- glide is heard in vowel-to-vowel Iinking. This thesis wiIl show tense vowels and diphthongs using off- glides for ease of notation. Furthemore, several of the resources used in this research (such as Avery & Ehrlich, 1992; Celce-Murcia et al., 1996; and Gilbert, 1993b3 also use off- glides in their notations of NAE. 4 syllable endings possible in W boundaries.
TABLE 2
Types of W environments across word and syllable boundaries
Phonological Environment Exarnples
/iy/ + V see it, create
/3y/ + V toy art, lawyer
/ey/ + V say it, crayon
layl + V try it, triangle
/UW/ + V true orange, congnious
/ow/ t V go on, noël
/aw/ + V how about, flour
Linking identical consonants causes a native speaker of English (NSE) to hold the consonant for a longer period than the same consonant in other environments, for example: (a) deep pot, (b) push Shirley, and (c) love very. Table 3 shows the types of articulatory settings in word or syllable endings possible in CC boundaries.
Linking between two non-identical consonants creates an interesting phenornenon in NSsE discourse: NSsE often do not release the first consonant before commencing the second. Acoustically, the CC transition is thus dorninated by the second consonant and it may be that NSsE use partly semantic or other contextuai cues to interpret what was 5 heard (Celce-Murcia et al., 1996). However, when a nonnative speaker (FINS) of English hem the NSE production of, "pet cat," they rnay hear "pekat," and so, may misunderstand.
TABLE 3
Types of CC environments across word and syllable boundaries
P honologicai Environment Examples
P+t stop trying, iaptop
lab technician
keep speaking, tip sy
need money
let down
bad judge
pet lizard, Atlantic
big boy, rugby
pet cat, Picton
big zoo
deck shoes, direction
g+k log cabin ResyIIab zjicat tion and SinzpliJicatzon
According to Celce-Murcia et al. (1996), resyllabification occurs when a consonant cluster in a word- or syllable-final position is followed by a vowel-initial word or syllable. The final consonant of the initial cluster is then pronounced as part of the following word or syllable. An example of resyllabification is: "find out" which sounds like "fine doubt," and "left OV~X-" sounds like "lef tover,"
These environments cause consonant clusters which are relatively difficult even for NSsE, and ways are found to simplify their pronunciation. Some examples of particularly problematic sequences include: (a) stopped near [st aptniy r], and, (b) robbed banks [rabdbægks].Unless one chooses to exaggerate each phoneme, thus sounding somewhat artificial, the final consonant of the first word is unreleased. When there are three heterogeneous stops, the difficulty increases, such as: (c) liked boys
[layktbci yz], (d) picked pockets [prktpaka tSI, and, (e) nagged people
[nægdpiypal].These consonant clusters are often problematic for many L2 leamers of
Engllsh as many speak languages that have simpler canonic structures (Prator &
Robinett, 1985).
It should be noted that in al1 these areas of comected speech, i. e. CV, W, CC,
CiCî, resyllabification and simplification, it is important for the interlocutor to respect thought-group boundaries.
The term thought grmp refers to a discrete stretch of speech that foms a semantically and grammatically coherent segment of discourse. When we think about where a speaker can logically pause in the stream of speech, we cm separate an utterance into thought groups. Although written discourse provides some markers for these divisions or pauses (i. e., 7 commas, semicolons, periods, dashes), in spoken discourse a speaker may pause at points where such punctuation does not always occur in a den transcription of the utterance. (Celce-Murcia et al., 1996, p. 175)
Linking usually does not occur between two clauses, or two distinct ideas. For example, in the sentence: "The elephant, of which 1 spoke, is gone." W linking occurs between "the" and "elephant," CV catenation would not occur between "elephant" and
"of," CV iinking occurs between ccwhich"and "1," but CV linking would not occur between "spoke" and "is." Note, however, that a NSE could in rapid speech overlook these conventions and link across thought-group boundaries.
The terminology of this particular phonological process proves to be varied, and it appears that several terms are not treated with equal significance. Since English has historically been strongly influenced by the French Ianguage, it is usehl to clari@ exactly what influence may have been received fiom this language, as well as to compare what is defined as connected speech. Lebel (1967)- and more recently, Garant-Viau (1994) give three types of linking: (a) l'enchaînement vocaIiqueI (b) I 'enchahementconsonantique, and (c) la liaison.
L'enchaînement vocalique resembles W linking, in that there is a smooth connection between the two vowels, with one exception: of€-glides are not compulsory in order to assist in the comection. Rather, each vowel must maintain its syllabic duration and fom, without any glottal stop inserted between the two vowels. For example, Noël is pronounced as: [na E 11 without glide nor glottal insertion (Garant-Viau, 1994). 8 L'enchaînement consonantique resembles a combination of CV and resyllabification. Garant-Viau (1994) states that the release of a consonant seems to be delayed rather than anticipated. In fact, it leaves the impression that the final consonant becomes the initial consonant of the following word. For example, il a can be represented as: [i-la]. As stated before, English has a similar phenomenon in consonant clusters, especiaily in respect to resyllabification: the final consonant of the initial cluster is then pronounced as part of the following word or syllable.
L'enchaînement consonantique shouid not be confiised with liaison. According to
Garant-Viau (1994), liaison is the survivor of an age when al1 consonants were pronounced. A large number of lexical units end in silent consonants which resurface when placed in front a vowel-initial item, thus se~ngas the consonant at the head of the next word. Furtherrnore, liaison is used more frequently in formal register, whereas in informa1 speech one tends not to use this aspect. Some examples of liaison with words that have silent final consonants in the wake of another consonant, and which resurfaces when a vowel-initiai word fiom the sarne rhythmic group follows: (a) the "s" of "les"
El&] resuffaces in les e~tfants[lezafa], (b) the "s" of ils [il] resurfaces in ils ont w V
[llzs], (c) the "s" of très RE] resuffaces in trèsw imporlanf [t~&zep=)~t&],(d) the
"t" of tout [tu] resurfaces in tourV à coup [tutaku],(e) the "s" ofÉtats [eta]
resurfices in Étatsw Unis [e tazuni] and, (f) the "s" of Champs [S S] resurfaces in
Champsu Élysées [S Cize lize]. Aithough English has inherited some aspects of l 'enchaîtzement vocalique and Z 'enchainement conso~tanripe,it does not have a parallel phenomenon to liaison. This is contrary to the claim that Gimson (1970) rnakes when he 9 claims that liaison refers to the linking of a consonant to a following vowel. Furthemore,
English demonstrates higher i2dences of linking in informa1 speech, which is the
opposite of French usage. Alt hough French is far fkom being a linguistic equivalent to
Japanese, it does, however, demonstrate an example of connected speech in an unrelated
language.
It may be that al1 Ianguages have some form of comected speech. Indeed, as
Pinker (1994) claims:
In the speech sound wave, one word mns into the next seamlessly; there are no little silences between spoken words the way there are white spaces between written words. We simply hallucinate word boundaries when we reach the edge of a stretch of sound that matches some entry in our mental dictionary. This becomes apparent when we listen to speech in a foreign language: it is impossible to tell where one word ends and the next begins. (pp. 159-160)
The rules for connected speech may Vary among languages; some attention needs to be paid to these rules in order to assist in the reception and production of discourse.
Temperley (198 7) suggests that ". .. [cl loser examination of Iinking shows its more profound effect on English pronunciation than is usually recognised, and that its neglect lads to misrepresentation and unnaturd expectations" (p. 65). Indeed, students should be made aware of the pervasiveness of linking which is related to rhythm, contractions, and sound changes (h40rleyy 1979). Browne and Huckin (1 987) also promote the
"...importance of learning how sounds are linked [and that it] is essential to good oral communication in English, because the listener needs these cues to interpret the meaning of the message" (p. 30). Research on Disconnected Speech
There are times in naturd NSE speech when linking across word- and syllable- boundaries Coes not occur. Not only do the phonemes become more carefully articulated, but also a glottal stop preceded by a short time lapse fills the gap between CV; for example, "get up7?[ger~p] becomes [get'. TAPI. It can also occur between W gaps; for instance, "go over" [gowowvar]becomes [gow. lowvar].
The following examples illustrate some environments where discomection rnay occur:
1. Glottal insertion may occur between CV environments in word- and syllable-
boundary position, which may be expressed as:
c (C)-V (C) + C(C) 2 V(C)
For example:
2. Glottal insertion rnay occur between W environments in word- or syllable-boundq
position, which may be expressed as:
C(C)V - VC(C) + C(C)V 2 VC(C)
For example:
Discomection may occur when a NSE wishes to be clearly understood; for example, in a classroom environment, a public speaking event, or in any situation when emphasis on information is required. An illustration of ernp hasis happens when asserting one's credibility, such as the emphatic staternent: "This is the right answer!" In this case, disconnection occurs between the first two and last two words. Also, attempts at
11 clarification cause norrnalIy linked environments to be separated, as is the case when the
NSE is emotionally highly agitated (Hatch, 1992) as in the angry expression, "Get out!"
[gst. ?awt].
A failure to link in English presents a problern in effective communication for the
learner of English. Indeed, the leamer needs to know the various prosodic features of the
target language which allow NSsE to comrnunicate effectively with each other. Prosody,
which is the suprasegmental system made up of stress, rhythm, linking, and intonation, is
used in detecting a speaker's emotions, sincerity and conviction, and this dlows
harmonious communication to continue (Hatch, 1992). When any of the elements of the
suprasegmental system are incorrectly used, this can cause confusion and
misinterpretation of a NNS output; NSsE can corne to view their interlocutor as being
arrogant or rude, and thus choose to close the doors of fi-iendly communication. Indeed,
non-native patterns in pronunciation and hesitation were found to be very strong contributors to listener distraction, irritation and annoyance (Fayer & Krasinski, 1987;
Gynan, 1984). Furthermore, when NNSs of English consistently disconnect the link between words, the expressive effect for the NS listener can be rather disconcerting, for the speaker is ofien viewed as aggressive and abrupt (Hatch, 1992; Anderson-Hsieh,
Riney & Koehler, 1994; S. Browne, persona1 communication, February 1, 1999).
Past research on the comprehensibility of NNS speech produdon shows that the tenn "pronunciation" has not always been clearly defined, and that non-professional NS judges are often unable to define exactly what element of "pronunciation" causes non- intelligibility (Hadden, 199 1; Ludwig, 1982; Varonis & Gass, 1982). Ludwig (1 982) reviewed the literature on error analysis conceming the measurement of 12 comprehensibility and irritation in NS of NNS attempts to comrnunicate. She fomd that
intelligible pronunciation, alongside accurate grammar and appropxiate vocabulary, are
necessary for the understanding of what contributes to comprehensibility and imtability.
When ESL teachers are asked to judge the intelligibility of NNS output, they tend to focus more on grammatical than oral correctness, since they have become
". .. accustomed to student's language, [and so,] teachers may no longer be able to determine which errors impede comprehension" (Hadden, 199 1, p. 3). With this in mind there are several pronunciation experts (Celce-Murcia et al., 1996;Dauer, 1993; Gilbert,
1993; Morley 1987; and Seidlhofer, 1994~)who have made it their mission to redefine and promote pronunciation with very specific components, such as: segmentais; word and sentence stress; rhythm; intonation; phrasing; pausing; blending; and linking. Each of these elements affects in varying degrees the intelligibility of NNS output, and the mastery of each element contributes to fluency (Celce-Murcia et al., 1996; Dauer, 1994).
Research on the Pronunczation Froblems of JSsE
Several variables might easily affect the speaking mannerisms of JSsE. For example, cultural traits may influence the characteristics of speech production: JSsE may prefer not to speak, but rather, allow their classrnates to initiate communication. Indeed, the most comfortable communicative interaction for JSsE is frequently perceived to be
4 These authors, among many others, are members of TESOL, which is an American professional organization that has a variety of interest sections. The TESOL Speech/Pronunciation Interest Section, or SPRIS, often shares information with the British equivalent, the IATEFL pronunciation special interest group. 13 none at al1 (Hinenoya & Gatbonton, 2000). This tendency to speak as little as possible leads to a lack of real-life oppomuiity to practice and refine speech patterns in a language being lemed. This is especially true for acquiring and applying the rules of comected speech.
Many pronunciation specialists (such as Celce-Murcia, et al., 1996; Morley, 1987,
1994; Gilbert, 19%b; Seidlhofer, 1994; Avery & Ehrlich, 1992) daim that pronunciation instruction has been overshadowed by a strong emphasis on reading and writing skills in many ESL courses, which they suggest hinders progress in oral production. Indeed, it is arguable t hat disconnected speech may derive firom strategies acquired during the iearning process associated with these types of Ieamïng. Gilbert (1993b) implies that L2 students - and this seems to be true for JSsE - may have learned to speak English from written material, and so, as there are white spaces between words, the L2 learner of
English may speak with an oral imitation of white space.
Many pronunciation specialists agree that JSsE seem to place more emphasis on grammatical and Iexical strategies than on phonologicai or phonetic strategies when leaniing the TL, and this tendency seems to stem £kom their language-training experience.
If one scans the most comrnonly used instructional textbooks in the ESL classroom, it is clear that the focus of language instruction is on grammar and vocabulary. For example, several colleges and universities in the Montreai area use Communicute What you Mean:
A Concise Advanced Grammar poil ock, 1997), EngIish on Line (Fournier, 1999),
RefinNzg Composition Skills (Smalley & Ruetten, 1999, and UndersfG171ding and Using
E@sh Grammar (Azar, 1989), al1 of which have a strong emphasis on grarnrnar and writing skills.
14 Courses at certain language institutions which teach English as a second language
(as opposed to teaching it as a foreign language), for example in Montréal (Québec), there is an emphasis on teaching al1 four skills of reading, writing, listening, and speaking. It is tme that conversational skills are taught and practised; however, there seems to be very little treatment of, indeed, even an avoidance of teaching prosodic features of English, especially of linking. The following three textbooks are among many used in the curriculum at McGill University and other language schools in Montréal.
These books deal scantily with word stress, sentence stress, and morphophonemic des, but nowhere can the rules of comected speech be found in these and similar textbooks:
Canadimt Cortcepts (Bens h & Thib audeau, 1993), A Canadian Comersation Book,
(Carver, Fotinos & Cooper, 1993), and Interchange (Richards, 1992).
The much heavier ernphasis on grarnmar and lexicon in their classroom expenence naturally leads the students to perceive this aspect of language Ieanllng as being most important to language proficiency. As a result, JSsE may unwittingly rely on a formulait and formalistic knowledge of English (Avery & Ehrlich, 1992; Celce-Murcia et al., 1996; Gilbert, 1993b; Morley, 1987, 2994; personal communications with ESL pronunciation professionals of TESOL's Speech/Pronunciation Interest Section).
From classroom and experimental observations, 1 have noticed that JSsE often fail to appropriately Iink in CV and W enWonments in conversational contexts. This disconnection is realised as an addition of a slight temporal pause, foliowed by a glottal stop. For example, "cup of' would be pronounced as: [ khph. ?av ] instead of
[knpav], and "go on" would be pronounced as: [go.161 iastead of: [gowan]. Even when JSsE are prolific interlocuters, their pronunciation contains a high frequency of 15 disconnection across word- and syllable-boundaries. Anderson-Hsieh et al. (1994) confum that JSsE ofien do not link between words as NSsE do. Indeed, they tend to pay noticeable attention to every word, so that eacb one sounds like a separate unit; thus, their speech sounds choppy, and their messages might be misconstrued (Morley, 1987, 1994;
Gilbert, 1993b).
In a study on the speech modifications of JSsE, Anderson-Hsieh, Riney, and
Koehler (1994) Iooked at how intermediate-proficiency (IP) and high-proficiency (HI?)
JSsE differ in their usage of connected speech fkom NSsE. They also looked at the effect of L l transfer on linking, specifically, rate of vowel reduction, consonant deletion, and the choice of strategy for simplieing consonant clusters. The researchers had three sets of subjects read from an abbreviated version of an oral test: One of the three groups consisted of five NSsE, the second consisted of five JSsE with an IP level, and the third five JSsE with an HP level. Six NSE evaluators with training in the administration of the
SPEAK test rated, among other variables, the pronunciation of JSE speech sarnples.
In general, Anderson-Hsieh et al. (1994) fiund that the NS group linked words together more than did the HP and JP groups. Indeed, they confirmed that the HP and IP groups were apt to break the link across word- and syllable-boundaries by inserting a g1ottal stop before the word-initial vowel in the second word. In addition, they claimed that foreign words with consonant clusters incorporated into the Japanese language were sirnplified by inserting vowels between any adjacent consonants. For example, the language learner of Japanese sees that "ice cream" is written in Romanized letters as aisukurimzi, "necktie" as nekutaiyand "text" as tekimto (f?om Jqanese for busy people,
1994, pp. 221, 222, 224) This allows the foreign syllable structure to conform to the
16 canonical syllable structure of Japanese, which is that of a predorninantly open and simple-phoneme syllable. Anderson-Hsieh et al. (1994) further suggest that L1 transfer affects comected speech and not just isolated units of speech.
Hancin-Bhatt & Rakesh (1997), in a study of the transfer effects of the phonotactics of Chinese onto English, " .. . have provided empincal evidence that transfer effects are a significant part of L2 syllable structure, and that they interact with developmental effects.. ." (p. 376). Chinese and Japanese tend towards a similar canonical structure, that is, that of open syllables with the possibility of a syllable ending in a nasal. Thus, the findings of Hancin-Bhatt and Rakesh (1997) may apply to JSsE in that transfer effects seem to contribute to the rough rhythm of output by JSsE.
What may prove problematic for JSsE, indeed, for many NNsE, is starting frorn a restricted phonemic inventory, both in number of sounds and in their distribution, which includes vowels without off-glides (Avery & Ehrlich, 1992; Thompson, 1987). Hence, the student may instead insert a temporal pause and a gloteal stop to compensate for what is missing in the phonological schema of his or her first language. Furthermore, JSsE often insert short unvoiced vowels, which also serve to 'round off final consonants. This epenthetic vowel appears not only in CC environrnents, but also in CV environments. For example, when a ISE produces "make a," a NSE may hear [meik. Pa] (where [v] represents the unrounded high back vowel used in Japanese) instead of beyka] as produced by native speakers (Thompson, 1987; Anderson-Hsieh et al., 1994).
The perceived epenthetic temporal pause may be caused by the phenomenon of vowel devoicing which is transferred fi-om Japanese to English. Indeed, Japanese high vowels are ". . . devoiced when preceded and foilowed by voiceless obstruents" (Frornkin 17 and Rodman, as cited in Vance, 1987, p. 48). However, nonhigh vowels are also devoiced
in the same environrnents, but with a much lower fiequency than high vowels (Vance,
1987). Hence, this tramfers into English as paragoge - the addition of one or more
segments to the end ofa word (Trask, p. 256). This appears to cause non-native-lilce
articulation, and may be realised as a temporal pause to the ears of NSsE.
Linking occurs in English and in French, and it probably occurs to sorne degree in
al1 languages of the world. Vance (1987) claims that historically,
[It] was once widely accepted that Izmson was completely regular in standard Japanese fiom about 1300 to about 1600. p]t may well be that liaison has always been sporadic and ordinarîly confined to particular words. Nonetheless, there is no doubt that it was once much more widespread than it is in the modem standard language. (p. 164)
What is particularly interesting is that modem Japanese demonstrates the use of a glottal stop [ 2 ] after short vowels. It can also occur in interjections and emphatic
speech, as it does in English speech, for exarnple, "an emphatic exclamation such as ouch! often begins with [ 1 ] " (Martin, as cited in Vance, 1987, p. 32). Furthemore,
McCawley (as cited in Vance, 1987) claims that a glottal precedes any syllable-initial vowel. In view of this, it appears that glottal insertion by Japanese speakers occurs in fiont of vowel-initial syllables, and it is likely they will transfer this effect onto their speech patterns in English.
Many pronunciation specialists often clah that the phonotactic structure of Asian languages pose a problem for second language leamers of English (Celce-Murcia et al.,
1996; Avery & Ehrlich, 1993; Gilbert, 19936; Morley, 1987, 1994). Indeed, most Asian languages are perceived as having an open-syllable structure, although not necessarily so. The syllable structure of Japanese is represented by Vance (1987, p. 64)as:
As cm be seen, the Japanese canonic shape is quite limited. It is based on the most
popular structure being CV. Only one glide (G) is allowed: /y/. A second short vowel is
permitted within the syllable boundarïes only if the first vowel is short, and this is
directly related to the mora-timing of the rhythm of Japanese. Merthe initial vowel, a
nasal may occur in this position, however, it is often phonernically represented as a nasalization of a preceding vowel (Vance, 1987). Finally, the "Qrepresents a glottal which occurs quite frequently in syllable-final position.
Al1 of the above findings suggest that JSsE may be unwittingly tramferring these patterns to English- This is detrimental to achieving oral fluency for Japanese learners of
English.
Reasons for Teaching Linking
The teaching of co~ectedspeech skills to JSsE, and indeed to dl L2 learners, is essential for effective cross-cultural communication. Several pedagogical pronunciation textbooks (Celce-Murcia, et ai., 1996; Gilbert, 1993; Dauer, 1993; Avery & Ehrlich,
1992) highly recommend the incorporation of linking activities. These guides indicate that it is necessary to teach students appropriate perceptual cues of English so that they are able to move fiom one word to the next, since the habitua1 - and largely uncûnscious
- speech patterns carried over from their L1 are not effective. Knowledge of when to link
19 across word- and syllable-boundaries rnay be difficult to acquire if not explicitly taught.
Morley (1987, 1994) points out that most ESL textbooks guide students to leam the lexical, sernantic and syntactic niles of the English language, yet give students little instruction in pronunciation.
Pronunciation should be viewed as important because there is always some potential to improve the ability of the NNS to be understood more clearly. It is important to note that an experienced ESL teacher may no longer be able to determine which errors irnpede comprehension. Lndeed, teachers rnay become, in a sense, "native listeners"
(Brodkey, 1972; Hadden, 1991). Knowing that teachers may have this ability to understand even the most unintelligible speech, could help teachers identify areas for concentrated effort in teaching pronunciation.
It is often argued that L2 learners of English do not need to leam the rules of linking. Some teachers claim that a learner's transitions between words will become more native-like as fluency develops, or that the lack of these transitions result in nothing worse than a foreign accent (see Morley, 1994). However, it has been shown that a learner can develop fluency without achieving an acceptable level of intelligible pronunciation, and that words become linked in ways not cornmon in English
(Temperley, 1987).
A robotic rhythm occurs as a result of disco~ectedspeech, and this may lead to a lack of cornprehension; in fact, Dauer (1993) claims that the speech of leamers of English sounds choppy when disconnected, and it is difficult for NSsE to know which words belong together in phrases. This leads to communication breakdown.
Additionally, in order to assist learners of English with listening comprehension
20 skills, it is important for them to uaderstand the phonological changes that occur in discourse. Brown (1977) suggests that in normal informal speech, NSsE usually concentrate on the content rather than on the precise articulation of what is being said.
NSsE tend to articulate in the most time-efficient marner, that is, they use linking, blending, reduction and simplification to assist in getting the message across in the least amount of time.
Indeed, Seidlhofer (1994) remarks that ". .,articulatory (over-) precision is a stylistic device. Lt is a conscious choice if we want to be insistent or threatening, but it may also make us sound "wearyingly precise and pedantic, even offensively so. .. . In human interaction, articulatory irnprecision is clearly the more natural and the more fùnctional option" (pp. 25-26). An illustration of a negative perception of choppy speech is found in Hatch (I992), where she describes a situation where she and her students perceived an Asian student as being angry. When codkonted, he was puuled and explained that he was not at al1 angry, so she realised that his ".. .staccato style [and use off syllable timing .. . [caused a] .. . defensive reaction from other students in the cIassroom.. ." (p. 279), and herself as well.
More serious examples are discussed on the SUPRAS Listserve, the officia1 name of the online group of pronunciation teachers and specialists f?om around the world.
Indeed, issues related to obtaining and retaining employment have been discussed, as has been reflected in articles in XESOL Matlers by members of the SpeecidPronunciation
Interest Section of TESOL. "Issues surrounding employees being fired or denied advancement because of foreign accents are now being argued in the courts" (Fried
Goetz, 2000, p. 14). This is an area that requires empiricai research fkom the phonological
21 and fiom the discourse analysis points of view.
Surnrnary and Conclusions
There is a general agreement in the literature that teaching of pronunciation to leamers of English should include teaching suprasegmentais, including linking, and that it should not focus primarily on minimal phonemic pairs. The relative importance of these different areas of nonnative pronunciation for intelligibility has been shown. It has been argued that the most critical area of pronunciation is prosody, which inciudes connected speech. Linking words together helps to reduce the possible negative effects of choppy speech, especially if JSsE, indeed al1 NNSsE, wish to conduct harmonious and successfiil discourse (Celce-Murcia et al., 1996; Gilbert, 1993b; Hatch, 1983 ; McNerney
& Mendelsohn, 1987; Morley, 1994; Seidihofer, 1994; Wong, 1987). Connected speech has an important effect on the progress in oral skills of not only JSsE but also many other
NNSs of English. A significant step towards more effective Ianguage teaching is to clari@ its place among the conversational skills JSsE must lem(Gilbert, 1993).
For the purpose of this study, the researcher has chosen to explore two environments where linking occurs across word- and syllable-boundaries: linking CV, and linking W. These environments were selected as a result of the researcher's observations that JSsE produce a high fiequency of errors in these areas, and that these errors appear to lead to miscornmunication and misunderstanding in NNS/NS interactions. Only two environments were chosen in order to keep the study as narrowly focussed as possible; this ensured optimal efficiency in data miiection and analysis. 22 Research Question
The question central to this study is whether explicit instruction wiil assist JSsE in reducing the frequency of discomected word pairs. Chapter 3: Design
Overail Design
The main aim of this thesis is to examine whether Japanese second language
leamers of English might improve their ability to connect word pairs in obligatory
contexts, thus attaining a smoother rhythm of speech. To attempt to answer this question,
a small exploratory study was created using a controi group (CG) and an experimental
group (EG). A group of NSsE provided the baseline of word pairs that were obligatory
contexts of comected speech. Word pairs were obtained fkom the text At the Travel
Agent's OfJice Gilbert (1993b); additional word pairs were derived fiom the free-speech
monologues obtained in response to an open-ended question. The controi and
expenmental groups were each given three one-hour lessons. The CG read poetry aloud
with a teacher who had received no training in pronunciation pedagogy. The EG were
taught how to connect word pairs in obligatory contexts by the researcher, who served in
the role of the EG teacher.
Participants: Roles and Materials
Nine JSsE were recruited fiom the wntinuing education department at Concordia
University, and the CG teacher and two judges were recruited nom the continuing education department at McGill University in Montréal. Permission was granted from the English language director at Concordia English Language Institute (CELI) to petition the assistance of CELI instructors to search for interested JSsE to participate in this experiment. The instnxctors were asked to provide an information sheet to those interested, who could contact the researcher at their convenience (see Appendix A). The participants consisted of the following groups: one group of five JSsE who made up the
EG, one group of four JSsE who made up the CG, four NSsE, two teachers, and two judges.
Japanese participants were chosen for this study because their speech usually contains a high fiequency of discomection while speaking English. Al1 participating
JSsE arrived in Montreal at least one year prior to the experiment, and they were between the ages of 20 and 29. Every participating JSsE had studied English since the age of 12 or
13 during their high school years. Six JSsE were registered in an intermediate or advanced level of English as a Second language, while another three were enrolled in
B.A. programs at Concordia University, thus, with fairly advanced language skills.
Al1 but one Japanese participant had never taken a pronunciation course prior to this expenment. The one JSE (participant number 14) who had taken a pronunciation course had only received cursory phonetic training over a one-year period in first-year university in Japan. She appeared to have no understanding of linking as an essential pronunciation component, judging by her lack of linking skills in the initial interview with her. This was later confimied in the experirnental treatment stage when she seemed pleasantly surprised that pronunciation practice did not solely mean the teaching of consonants and vowels.
The JSsE were assigned to either the CG or the EG. This was done by drawing
25 each participant nurnber out of a hat. Since there was ody one male participant among the JSsE, he was intentionally placed in the EG. It appeared reasonable to have an even number of female participants in both groups. Placing the male student in the EG allowed observation of any changes in his speech performance that may not have occurred if he was placed in the CG. An indicator of possible sex difference in the leaming of prosodic features was thus potentially available.
The two teachers in the pedagogical expenment were NSsE and were teaching
ESL at McGill University. Having a strong command of teaching pronunciation skills, especially linking, the researcher served as instructor for the EG, since no one else in her academic community taught pronunciation as a single ski11 course at the tirne. The CG instructor, who had Meif any understanding of how to teach pronunciation, was a volunteer recruited from among the researcher's' teaching colleagues and tau& advanced ESL writing courses at McGill University. Both teachers had at least five years of teaching expenence, with the EG teacher having had three more than the CG teacher.
A baseline of obligatory contexts for comected speech was required. Four unilingual NSsE were asked to provide speech sarnples. NSE 1, an Afro-American male participant, lived in New York City. The other three NSsE were al1 Canadian women who lived in the Montreal area. Al1 NSsE have university degrees, and the average age was slightly higher than those of the JSsE. They were al1 acquaintances of the researcher, selected for their lack of daily use of any other language than Englisb
The researcher transcribed word pairs taken fiom the discourse that the two groups of JSsE as well as the group of NSsE produced in eee-speech samples. Word pairs fiom the output of the Clear Speaking Test (Gilbert, 1993a; see Appendix G) were
26 also transcribed,
Once the transcripts were completed, the researcher identified points where
linking did and did not occur in the speech of these three groups. Two judges venfied the
accuracy of the researcher' s transcripts. Both judges have master's degrees, and they
were chosen for their experience in teaching connected speech. One of the judges was a
NSE who has French family members. At that tirne, he had taught ESL pronunciation for
more than three years. The second judge was a balanced bilingual speaker of French and
English; she had studied both English and French linguistics at French and English
universities. More specifically, she had an excellent understanding and naturd usage of
English phonoiogy.
In terms of inter-rater reliability, the researcher asked the judges (also referred to
here as the "raters") to listen to 10% of recorded speech samples of both speaking tasks to
determine whether or not the researcher accurately transcribed each word pair with the
appropriate comection or discomection. The researcher asked each rater to listen to the
cassettes, look at the transcriptions given in Appendixes N and 0, then indicate
agreement by marking a check mark next to each word pair with a CV or W
environment. This gave the raters a chance to point out any errors or omissions. One judge agreed with the transcriptions and connections 99% of the time, and the other judge
agreed 97% of the time. No omissions were found by the judges. The transcripts can thus
be considered highly reliable with regard to occurrence of linking in CV and W
contexts. Procedures and Measures
Design of Pedagogical Activity
Three 30-minute pedagogical units with an explicit focus on connected speech were created for the EG. These units were given over three consecutive days. Each teaching unit focused on the explicit linking strategies for one of the following three environments: (a) CV, (b) W, and, (c) CC. The first day involved students being taught how to link in a CV environment, for example, how to link "In an airplane" (see
Appendix H). On the second day, the EG teacher reviewed the previous lesson, and then the JSsE leamed the rule for linking in a W environment, for example, "May 1 ask" (see
Appendix 1). On the third day the first two deswere reviewed, and then the instructor taught the rule for linking CC word pairs, for example, "Bill loves pie" (see Appendix J).
Each lesson lasted one hour and followed the same format: presentation, elicited repetition, and pair practice. See appendices H, 1, and J, respectively. The three lessons occurred on consecutive days over a long weekend.
By the time the third lesson was completed, the researcher realised that a grasp of the more complex phonological niles associated with linking in a CC context was beyond what one could reasonably expect language students to acquire within such a short time.
Their performance as seen in post-treatment testing confirmed this judgement.
A senes of three 30-minute pedagogical activities were created for the CG as well. These activities did not focus on pronunciation, but rather required students to read poetry aloud once only in chorus with the teacher after she rad it aloud to them. The focus of their lessons was to discuss the meaning of each text aven. In the first class, the
CG teacher directed the cIass in a discussion of Uphill by Christina Rossetti (see
Appendix K), in the second, The RoadNot Taken by Robert Frost (see Appendix L), and the third, Taken Up by Charles Martin (see Appendix M). Each lesson followed the same format: a reading of the poem by the teacher, a single choral repetition, and then a discussion of the meaning of each poem directed by the teacher through the use of open questions. Each poem provided the content of one pedagogical unit. Each class Iasted one hour, and the classes were given over the same three consecutive days at the same times and in the same building as those for the EG.
Recording and Analyszs of JSE Speech Sumples
Before recording the JSsE, a baseline of obligatory contexts was required. The researcher recorded the four NSsE individually reading the text, "Ar the Travel Agent S
Oflce" shown in Appendix G. Upon completion, each was asked a question about what he or she did the previous evening in order to elicit a monologue (see Appendix F).
Every JSE was recorded to test their ability to pronounce the target items on the same day, exactly one week pnor to the block of teaching activities. One week after the block of teaching activities, al1 nine JSsE were recorded on the same day. Each JSsE was individually recorded reading the same dialogue read by the NSsE (see Appendix G).
Following the reading, each participant answered the same question as the JSsE (Le.
"What did you do last night?") intended to elicit a monologue. Before receiving the next participant, each monoiogue segment was timed so that each spoken piece analysed lasted approximately one minute. The researcher transcnbed al1 speech samples two weeks later when al1 the recordings were collected, the experiment was completed, and
the JSsE had been given free ESL pronunciation classes as remuneration for their
participation.
The researcher transcribed the 49 CV and W word pairs that are found in the dialogue used for reading aloud (see Appendixes G and N). She also transcribed a single page of descriptive monologues (see Appendixes E and 0)marking where linking did and did not occur in the CV and W environments. Chapter 4: Results
The text, "At the Travel Agent's Office7' (see Appendix G) was read aioud by al1 participants. The NSsE provided a baseline, JSsE provided samples for the pre-treatment condition, and dl JSsE read the same text aloud in post-treatment condition. There were exactly 49 pairs of words with a CV or W environment which could potentially be linked in the text provided. Narrow transcriptions of these pairs as pronounced by each participant are given in Appendix N. Table 4 (see page 32) provides raw data of the performance of al1 13 participants. TABLE 4
Number of Linked Word Pairs (CV and W Environments):
Reading Aloud
Group and Participant Pre-Test Post-Test
Baseline
NSE i 24
NSE 2 41
NSE 3 36
NSE 4 3 1
Control Group
JSE 6
JSE 7
JSE 8
JSE 9
Experïmental Group
JSE Il 14 17
JSE 12
JSE 13
JSE 14
JSE 15 26 30
Note: Total number of items to be potentially linked in the set text = 49 Al1 participants in the study were asked a question in order to elicit a sample of fiee speech. The NSsE and the pre-test JSsE were asked what they did the previous evening, and they were encouraged to elaborate as much as possible. The post-test JSsE were asked what they would do that evening, and they were asked to be specific and descriptive. Narrow transcriptions of the word pairs containing CV and W environments fiom each participant's sample are provided in Appendix O.
In the free-speech task there was quite a diverse range of number of word pairs that could potentially be linked, as well as great diversity in the number of word pairs that were actually linked as produced by each JSsE. Table 5 (see page 34) shows the raw data of the word pairs produced by each participant in pre-test and post-test conditions. What may appear to have been increase or decrease in the ability to link word pairs in terms of percentages in reality does not occur. The cornparison of percentages, which are obtained by dividing the actual number by the potential number of linked word pairs, is more readily seen in Figure 3 (see page 39). Narrow transcriptions of al1 word pairs are given in Appendix O.
There is a comrnon belief among my ESL colleagues that both first and second language speakers have a tendency to read aloud more carefully. Figure 1 (see page 35) shows the individual performances of each NSE while reading aloud and while speaking fieely. It is interesting to note that the average percentage of linking while reading a text is at 67%, and while speaking fieely is at 73%, which seerns to suggest that connection of word pain occurs with a high and approximately equal percentage under both conditions. TABLE 5
Number of Linked Word Pairs (CV and W Environments): Speaking Freely
Pre-Test Post-Test
Group and Participant Poten- Actuai % Poten- Actual YO tial tial
Baseline
NSE 1
NSE 2
NSE 3
NSE 4
Control Group
JSE 6
JSE 7
JSE 8
JSE 9
Experimental Group
JSE 11
JSE 12
JSE 13
JSE 14
JSE 15 FIGURE 1
Cornparison of Linking by Native Speakers of English while
Reading Aloud versus Free Speech (%)
Note: These are the percentages of actual versus potential linked items while reading aloud, as weil as the percentages of linked items while responding fieely to a question, for native speakers of English. These graph lines are shown again in Figures 2 and 3 for cornparison with Japanese speakers of English. Figure 2 (see page 3 7) shows that individual performances in pre- and post- treatment measures vary considerably. It is however noteworthy that the performance of al1 five EG participants either improved or remained relatively stable in linking ability when reading a set text. Among the four CG participants, three either stayed the same or showed a somewhat lower level. Atypically, one showed a higher percentage of linking for this condition after the experirnental period. FIGURE 2
Comparison of Actual Versus Potential Linked Items Made by All Participants While Reading Aloud (%)
Participants Native Speakers Control Group Experimental Group 1 2 3 4 6 7 8 9 11 12 13 14 15 Baseline 49 84 73 63 Pre-Test *a 35 53 43 41 29 16 31 45 53 Post-Test H 33 43 59 35 35 73 65 47 61 Figure 3 (see page 39) shows the results of connected word pairs of al1 participants while speaking eeely. Again, there is considerable difference in individual performance across the JSsE.
The results for the EG show that only two participants improved noticeably in their ability to connect word pairs: JSE 11 improved by 23% while JSE 12 improved by
24%. An unexpected result came fiom JSE 14 whose ability to cunnea word pairs decreased by 28%. Most surprising, however, is the observation that two of the CG participants also show significant increases in linking. It is important to keep in mind while looking at the results in Table 5 (see page 34) and Figure 3 (see page 39) that the individual performances of each participant, whichever group they were part of, show a great deal of variation in the number of word pairs that could be potentially linked as well as those that were actually Iinked. FIGURE 3
Comprrison of Actual Versus Potential Linked items Made by All Participants in Free Speech (%)
Participants Native Speakers Control Group Experiment al Group 1 2 3 4 6 7 8 9 11 12 13 14 15 Baseline 58 75 77 83
Pre-Test 0- *@ 27 42 17 28 13 14 19 57 34 Post-Test u 30 62 67 23 36 38 30 29 37 Chapter 5: Summary, Discussion, Recommendations
The sample size ofthis exploratory study is far too small to make any conclusive statement. Nevertheless, the differences between the pre- and post-test sarnples of the
JSsE reading aloud and speaking fieely seem to suggest that the explicit teaching of the rules of connected speech may improve the quality of the oral English output of JSsE. It seemed especially interesting to note that there was a difference for the EG between pre- and post-tests when the participants read aioud; however, there were several variables which may have affected the outcorne.
The initial purpose of this study was to explore the usefulness of a pedagogical technique for correcting a significant speech problem for JSsE and other leamers of
English, specifically that of producing connected speech. In this respect, the findings of this study offer a glimmer of hope as to what type of consciousness-raising activities rnight heIp students improve their ability to connect speech.
Research Questions
At the onset of this exploratory experiment in chapter 1, two questions were asked:
1. What are the most common types of linking problems across word- and syllable-
boundaries that contribute to the staccato-like speech of JSsE?
2. Can specific teaching strategies assist in improving the connected-speech patterns of
JSsE? 40 Native language transfer seemed to be a strong factor in the Japanese speakers'
pronunciation of English. It is clearly reflected in the way certain p honemes are
pronounced: For example, JSsE ofien change the suffix "er" to [el, such as in the word
ccdriver"which was pronounced by ISE 8 as [drpiv~s](line 174 in Appendk O).
According to Vance (1987), Japanese syllable structure does not allow [r] nor [ar] to be
in either syllable- or word-final position, and so, when the Japanese speak English, they
tend to drop the [r]. Furthemore, when a word ends with an "er" SUEX,the closest match
in Japanese is the above mentioned Ce]- -,-
When it came to determining the extent to which native laquage transfer affected
COM~C~~speech, one can only speculate. Looking at the results fiom this exploratory
study, the NSsE linked 67% of the time while reading aloud, and 73% of the time whiie
speaking fieely. The average of al1 nine JSsE prior to the experiment was 3 8% while
reading aloud, and 33% while speaking fieely (these amounts were calculated by adding
al1 nine percentages and dividing by nine in the pre-test conditions of both speaking
situations). So, it appears that JSsE generally speak with approximately 29% more
instances of discomected speech than NSsE while reading aloud.
It is especially difficult to draw any firm conclusions about the fiee speech of the
JSsE in cornparison to what the NSsE produced. In response to the experimental question used to elicit f?ee speech ("What did you do last night?") the NSsE spke with rnuch more sophisticated language, did not need any coaxing, and they almost never hesitated about what they were going to Say. Indeed, there was plenty of matenal with which to 41 work. In contrast, the JSsE generally used simpler language, needed to be coaxed (for
example, some paused for too long, so, 1 asked them to describe their dimer, if they told
me they made Cinner), and more often than not, became mute because they could not
express themselves.
While 1 transcribed their output, 1was quite fnistrated with the small arnounts of
word pairs with a CV or W environment that did not have hesitations or fillers such as
'curn,yy"uh," and "ah," which were produced by especially JSsE 6, 8, 9, 11, 12, 13, and 14
(see Table 5, page 34, and Appendix O). The most hstrating factor is that they were al1
quite ganulous when not under the investigative lamp of research. Larsen-Freeman and
Long (1990)Say it best when discussing the difficulty in studying the output of
participants:
.. . the mere presence of an observer is likefy to cause the subjects to pay more attention to their speech and thus result in unspontaneous performance. Moreover, even if completely spontaneous production data were available, there are certain drawbacks to relying solely on them for insights into the second language acquisition process. First of dl, without the imposition of constraints in terms of the range of possible responses a subject is likely to produce, it is impossible to study al1 aspects of a leamer's developing performance. Certain language features could not be studied because they do not occur fi-equently in normal conversation. A researcher would have to wait a long tirne, for example, for subjects to produce enough [of the target item being researched] for the researcher to be able to Say anything meaningful about their acquisition. (p. 26)
It is far easier to wntrol the quantity and quality of obligatory wntexts in a
controlled reading. It is also clear that even NSsE produce only certain amounts of
connected speech at certain times. Table 5 (see page 34) shows that in the free speech of the NSsE, there was a range of 16 to 30 potential word pairs to be linked, whereas the CG
JSsE produced a range of 12 to 52 potential word pairs to be linked, and the EG produced a range of 14 to 29 potential word pairs.
It seemed that the JSsE were suffering fiom the shame of not speaking "perfect
English," which is a behaviour 1 have observed fiom many Japanese students 1 have
taught over the years. Indeed, as Hinenoya & Gatbonton (2000) have pointed out: The
most cornfortable communicative interaction for JSsE is fkequently perceived to be none
at dl. Indeed, JSE 14 was quite exasperated just prior to the post-test recording of her
reading. She felt that she codd no longer speak because her pronunciation "so poor!" My
guess is that she became more aware of what she was not doing in her speaking habits,
and so, refrained fiom saying anything that might be problematic. This may explain the
sudden drop of her percentage rate of successfül linking in the pre- and post-experimental
recordings fkom 57 to 29 (see Figure 3, page 39).
Several pronunciation specialists (SUPRAS, an on-line discussion group, as
described in chapter 2) have recently discussed whether teaching linking is a necessary
component in ESL classes. dongthe range of answers provided, one correspondent
suggested that the tendency to separate each word is some sort of interlanguage
phenomenon in relation to the orthography, that is, where there is white space on a page,
students reflect this space in their speech (see also Gilbert, 1993b).
hother pronunciation specialist offered an exarnple which many can attest to: As
a beginner student of Spanish, a simple phrase such as su que hora es?"= frequently
unintelligible to him because he did not Iink the three words. He explains that his elementary level disconnected version of the phrase seriously distorted his expectations as to what native speakers were saying to hun, and so, communication breakdown
ocwed. Clearly, there is a strong need to teach this element in the ESL class. We could follow the example given in many French language leafning classrooms: teach liaison as part of the cuniculum. For example, in beginner-level French classes, McGill University uses the textbook: Le nouveau sansfionfieres(Dominique, Girardet, Verdelhan, &
Verdelhan, 1990); liaison is taught f?om lesson one. In the intennediate classes they use the textbook: Plaisir des Som (Kaneman-Pougatch & Pedoya-Guimbretière, 1991 ) where liaison lessons are continued.
Ten years of casual observations of my own students of Japanese ongin leads me to think that attempts at saying a sentence with the comect articulation, word and sentence stress in order to generate the required message sometimes fail because of articulation dEculties at word boundaries. The difficulties generate a pause, and this can destroy the structure of the intended utterance, especially since a glottal is usually inserted in fiont of words beginning with a vowel.
Most Cornmon Areas ofDisconnection
In Appendix N are the transcriptions of al1 the 49 word pairs that could potentially be produced by the four NSsE participants and the Nne JSE participants. The pairs that were discomected by ail four NSsE are: "unfortunately I've" (see lines 15, 64, 113, and
162)- "by early" (see lines 26, 75, 124, 173), and "early evening" (see lines 27, 76, 125,
174). In the first pair, it seems that a pause was necessary in order to emphasise
"unfortunately," thereby necessitating a disconnection. For the other two word pairs, it is possible that the sequence of vowels [ay] to [ar] and [iy] to [iy] required tw much rnuscular tension, and so, a disconnection may have provided temporary relief.
Al1 nine JSsE disconnected the same three word pairs with the exception of JSE
12 of the EG who was able to link the word pair after the experiment. Looking at Figure
3 (se page 39), it is clear that she airnost tripled her ability to link word pairs, with an increase of 24% (fi-om 14% to 3 8%).
Do Rem edzal Teachzng Actzvities Help?
Although the inconclusiveness of the results provided by this study are somewhat disappointing, there are some areas which provide confirmation that teaching linking can, indeed, improve the rhythm of the speech of Japanese leamers of English. This was especially evident for JSsE 11, 12, and 13 in both speaking tasks. Most certainiy, this needs to be fbrther researched with larger sample groups and more tightly controlled variables. These variables will be discussed in the recomrnendations below. mat to Teach and When
Teaching the EG how to connect word pairs with a CC environment proved to be far more complex than the JSsE could manage within the three days of the experiment, as they had had plenty to manage with linking in CV and W environments. Indeed, it became quite ciear that teaching them how to link two consonants was a far more co mpiicated teaching task than the researcher had original1y anticipateci. It became obvious that teaching the three different desof connected speech would have been better achieved over a minimum of four weeks. However, the students were due to ieave the country for the summer break within a month of recmitment, and the experiment needed to be completed within a reasonable amount of time. 45 In addition, teaching students CC envkonments would require a further division
of learning tasks: (1) How to Link consonants, whether identical or not, without vowel
insertion; this would most iikely require an understanding of syllable structures between
the L1 and L2, (2) how to link identical consonants, such as "deep pot" often pronounced
by NSsE as [ diypat ' 1, (3) how to link non-identical consonants, where each
consonant retains the original articulatory features, such as "like black," pronounced as
[ 1a yk 'blæk 1, (4) how to link non-identical consonants, where each consonant
changes somewhat fiom the original articulatory features, such as "but the," - where the
[ t ] is dentalized and [ 3 ] becomes more obstruent as a result of assimilation, and (5)
how to link consonants with glides where articulatory features are aected, such as "can't you" which is often pronounced as [kæntS a 1.
Two types of tasks were chosen to measure learner performance: reading aloud and elicited £iee-speech monologues. These two tasks allowed me to observe the habits of comected speech in two different contexts in order to determine how much linking occurs. The reading task was far easier to measure and control, whereas the elicited monologues proved to be disappointing in terms of the amount and quality of speech samples produced by each participant. That is, the NSsE produced usenil amounts of material, whereas the JSsE: (1) did not provide enough material, (2) did not provide enough word pairs with CV and W environrnents, (3) were repetitive, or (4) used an 46 excessive amount of fillers, such as: "uh," "um," and "ah."
Appendix O has the transcriptions of al1 NSsE and JSsE participants who
provided monologues in pre- and post-test conditions. As already mentioned,
there are large dserences in the amount of data useful to this study. Table 5 (see
page 34) shows the raw data dong with the percentages of word pair items that
wuld have been and actually were linked. NSE 2 provided the smallest arnount of
items to be hked (16), as well as those that actuaily were linked (12), whereas
NSE 3 provided the largest arnounts (3 0 and 23). However, if we look at the
percentages, it was NSE 1 who had the lowest percentage of realised linked items
(58%) in his monologue, and NSE 4 had the highest amount (83%). It was
expected that NSsE would link at least 50% of the time, so these percentages
exceeded my expectations.
The results for the JSsE, however, are sobering, especially if we look at the
individual performances. Figure 3 (see page 39) clearly displays the performance of each
participant. An unexpected improvement came f?om JSE 8, who was in the CG: she has
an improvement of 50% fkom 17% to 67%, which was dramatic. However, the raw data
in Table 5 (see page 34) shows that she links only 4 out of 24 word pair items in the pre-
test condition. It is also shown that in the post-test condition, she provided 12 word pairs
that could have been linked - half of what she provided in the pre-test - and 8 were
Iinked, giving a percentage higher than that in the pre-test condition. So, her
performance, based solely on the wmparison of the two percentages, is rather misleading.
In the EG, the performance of JSE 14 is even more disappointing if the pre- and
47 post-test percentages are compared. She drops fkom 57% to 29% in her ability to link word pairs. Looking at the raw data in Table 5 (see page 34), the ratio ofword pairs produced and the ones actually linked is 1433 in the pre-test condition. In the post-test condition, it is 21 5. So, she produces more word pairs but links Iess fiequently in the post-test condition. The raw data therefore shows a less dramatic difference. It should be noted, however, that this subject was particufarly self-deprecating. During the class, when cailed upon, she ducked her head, sucked in air between her teeth to indicate that she was unsure of herseIf, then shook her head in disappointment if she did not perform correctly.
When she did perform correctly, she shook her head in disbelief. As mentioned above, she admitted that her English was "so poor," and so she demonstrated clearly the self- fulfilling prophecy which plagues many a classroorn. As Norman Vincent Peale (1993) puts it: "What you think, you will become - good or bad, weak or strong, defeated or victorious - so practice being a positive thinker in a time like this." (p. 73) It would be interesting to see research on linking being done with a healthy dose of positive thinking included in the experimental design.
Discussion of "Reading Aloud" Results
While preparing for this study, 1 was advised by a former professor of Concordia
University's TESL Centre that the results for L1 and L2 redrsof texts would show that they enunciate more carefiilly they do when producing fiee speech. Indeed, the research noticed this phenornenon when listening to ESL students reading aloud during oral examinations. However, this appears to be true for the phonetic articulation of English speech. For example, the article "a" was pronounced as [ ey ] instead of the more regularly used schwa vowel [a ] ;see Appendix N, lines 28, 3 1,32, 37,44, and 45.
Reading aloud does uot produce larger arnounts of discomected speech as much
as articulation. The average percentage ofcomected pairs for the CG in pre- and post-
experimental recordings regressed by only 1%, rendering the results for the EG more
interesting because there was a 2 1% increase (see Table 4, page 32).
From my own observations with ESL students over the last decade, it appears that
that L2 readers hesitate, backtrack and enunciate more carefully while reading aloud. So
it is to be expected that there would be some difference between controlled readings and
uncontrolled monologues. It is to be expected that when a text is read, the tendency to
speak with more pauses and a higher rate of discomection would be present. However, the differences in my data are very small. The fkquency of comected speech in the
controlled reading and spontaneous speech for NSsE are show in Figure 1 (see page 35).
In spite of the fact that the number of NSE participants is very small (n = 4), the small
difference between the two conditions is a potential research area: 1s there really any
significant difference in the way one reads aloud in cornparison to spontaneous speech?
Perhaps many are under the impression that there is a big difference, but we need
empincal evidence to confhm or refute this claim.
Table 4 (see page 32) shows that there was a decrease in the average of comected
items CG JSsE produced while reading aloud. Individually speaking, al1 but one JSE worsened in the ability to link word pairs. JSE 8 increased by 16% fiom 43% to 59% that is, out of 49 potential word pairs that coufd be linked, JSE 8 was able to link 21 pairs in the pre-test and 29 in the post-test. This excellent improvement may be a result of several hidden factors. At first glance, it appeared that she was one of the undergraduate 49 sîudents, but she was in fact enrolled in the Concordia English Language htitute (CELI) where she was finishing her 9h week of study. Her strong performance may be related to her high level of motivation. Moreover, she had admitted to me that she enjoyed singing
English songs at Karaoke bars; rnany pronunciation specialists believe that music can improve speech rhythm. Although the EG were asked to promise to not share the information given to them during the experiment, it is also possible that one or more of the experimental group students taught JSE 8 how to connect words-
The most dramatic improvement in the EG was shown by JSE 12 who had a 57% increase irom 16% to 73%- That is, out of 49 word pairs that could potentially be linked, she was able to luik 8 pairs in the pre-test and 36 in the post-test (see Table 4, page 32)-
JSE 12 was enrolled in an intermediate-level of English at CELI. She seemed the most highly motivated out of al1 the JSsE to improve her speech rhythm. It is agreed among
ESL teachers that students with high motivation ofien acquire the target language in a relatively effortless fashion.
Dzscussion of Free-Speech Monologue Results
The results for the elicited monologue task of ~ee,unscripted speech are partly disconcerting. However, one must recollect that the sample size is quite small. There was a large amount of participant variability in what each JSsE produced in this task. Taking a look at Appendix O, some JSsE had very little to say, using a restricted vocabuIax-y (JSsE
2, 12, and 14), while others were more expressive, using slightly more sophisticated language (JSsE 7,9,1 1, and 13).
Each JSE who participated in this experiment possessed a different educational background, and each demonstrated a different capacity to learn and produce what was
leamed. Four JSsE participants were enrolled in undergraduate programs at Concordia
University; two of each were put into the CG (JSE 7 and 9) and EG (JSE 1 1 and 13). Five
students were intermediate-level students fiom CELI who were my prime target: two
were placed in the CG (JSE 6 and 81, and three were placed in the EG (JSE 12, 14 and
15).
The four JSsE - numbers 7,9,11, and 13 - were enrolied in bachelor degree
propams at Concordia University. It is hard to say whether or not their acadernic status
had any idluence on their performance, since the results for both pre- and post-treatment monologues (see Figure 3, page 39) show mixed results. JSE 7 and JSE 9 were both in the control group; JSE 7 showed a 20% increase of linked word pairs fkom 42% to 62% of linked pairs. In contrast, JSE 9 showed a decrease of 5% Born 28% down to 23%.
However, the raw data in Table 5 (see page 34) shows ratios of potential word pairs to be linked to that those actudly made were 52 to 22 for JSE 7, whereas JSE 9 had a ratio of
21 to 13.
Two of the five EG participants, JSE 1 1 and 13, both showed encouraging improvements of 23% and Il%, respectively (see Figure 3, page 39). Looking at the raw data in Table 5 (see page 34), one sees thepre-test ratios of potential word pairs to be linked to those actudly made were 23 to 3 for JSE 11, whereas JSE 13 had a ratio of 3 1 to 6. Table 5 (see page 34) shows thepost-lest ratios of potential word pairs to be linked to those actually made were 25 to 9 for JSE 11, whereas JSE 13 had a ratio of 27 to 8.
The remaining five JSsE were ail enrolled in CELI. Participant numbers 6 and 8 were in the control group, and numbers 12, 14 and 15 were in the EG. The most dramatic
5 1 decrease in the EG was shown by ISE 14 who had a 28% decrease fiom 57% to 29% of comected word pairs (see Figure 3, page 39). The raw data shows that the pre-test ratio was 14 to 8 and the post-test ratio was 21 to 6 (see Table 5, page 34).
During the post-treatrnent recording of the students in both contexts, the researcher becarne aware of acertain level of discornfort that may be related to the cultural features of Japanese learners of English. Teachers need to remember that
Japanese students, and perhaps many other leamers ofEnglish, may have a certain level of performance anxiety, and so not be abIe to produce what is anticipated. Corder (In
Larsen-Freeman and Long, 1990) confirms that ".. .leamers will place limitations on the data themselves." (p. 26) Larsen-Freeman and Long (1990) continue by saying that,
.. . learners will often not reveal to researchers their entire linguistic repertoire; rather, they will use only those aspects in which they have the most confidence. They will avoid the troublesome aspects through cirnimlocution or some other device. And it may be precisely the troublesorne aspects of the second language in which the researcher is most interested. Thus, if the occasion does not lend itself for a particular aspect of linguistic performance to be manifest, or if learners are adept at circumlocuting aspects of the language which cause them diEculty, researchers will not be able to adduce any sort of evidence. (p. 26)
Transcription Disco verzes
As the researcher transcribed the output produced by al1 the participants, there were no problems wit h transcriptions of NSE speech. However, several doubts arose when transcnbing JSE speech. Certain phonological phenomena arose that do not occur in English: (1) Meta-cognitive understanding of the researcher, that is, "'mat do 1 think I hear?'(2) Final [n] deletion with nasalization of preceding vowel. 52 Perhaps one drawback was my preconception of what the JSsE would produce.
The meta-cognitive influence of language comprehension cannot be ignored. There were moments when 1 thought 1heard aspiration, but in fact the speaker had "rounded off7 certain sounds. This rounding off, or "paragoge" was mentioned in chapter one, section
2.4. The Japanese language is often characterised as having voiceless obstruents followed by voiceless high vowels. What 1thought was aspiration of voiceless obstruents, turned out to be paragoge. It appeared that the students were overgeneraiizing the aspiration of voiceless obstruents in word initial position as NSsE would normally do. However, knowing what this does not necessarily mean understanding how this applies to the results as shown. Perhaps future research could look at this phenornenon and see to what degree paragoge affects linking.
Final [n] deletion and subsequent nasalization of preceding vowel occurred in a high number of cases in the speech of al1 JSsE. In Appendix N7 there are six word pairs which have /ni as the final consonant in the first word. They are: "can 1," "on
Wednesday," "return on," "and you711," "an adventure," and, "than 1." Except for the first pair, the JSsE used a nasalised vowel instead of producing an alveolar /n/.It appeared that this nasalised vowel was an allophone of ln/,however, there seems to no literature to confirm this finding. It is interesting that the JSE participants were able to produce an alveolar /n/ for "cm P' which lead me to think that it might be in fiee variation with the nasalised vowel. What was discovered was that the JSsE produced an unusual way of linking: a nasal vowel linked with a following vowel without using the /y/ nor /w/ glides.
This is one of the characteristics of their interlanguage which needs to be more carefully researched.
53 Discussion of Experimenta[ Conditions
It became clear that certain enWonments were very difflcult to hear.
Indeed, having a reel-to-reel audio recorder with a splicer would have been a usefùl tool in order to extract onIy the word pairs that were being used as a focus.
This would have been useful in eliminating some of the contextual cues that may have been interfixing. Also, it would be very usefùl to use a spectrogram in order to determine if certain sounds are indeed produced.
Discussion of Experzmenter Bias
There was a possibility that experimenter bias occurred as a result of the researcher perforrning the role of the EG teacher. 1t was unfortunate that the researcher was unable to find anyone able to teach this area of pronunciation who was available at the time ofthe experiment. She was therefore obliged to apply her own expertise. The
CG teacher had a very different area of expertise, with experience in teaching ESL reading and wrïting at a major university in Montréal (Québec).
Furthemore, the experimenter transcribed the data herself without the help of an assistant. Nevertheless, the reliability of the data was verified by two judges who each agreed 97% of the time or better with the accuracy of the transcriptions as well as where comection did or did not occur.
Although the two judges both had academic training and expenence in teaching
ESL pronunciation, they both are fluent bilingual speakers of English and French. Their bilingualism may have changed their level of awareness of suprasegmental phenornena like linking in English. There were several limitations in this necessarily exploratoly study, the most apparent being the sample size. More robust conclusions as to the usefidness of directly teaching prosodic feanires would be possible following the use of the exercises developed for this study with a larger group of students. Ideally, the researcher would analyse data gathered by instmctors other than herself and be assisted by judges with no significant exposure to any language other than EngIish. A pedagogical intervention of several weeks rather than the few days available for the represent study would allow tirne for more significant changes in pronunciation behaviour and thus provide a better indication of differences between control and experiment al groups. Student anxieties and ot her variables influencing motivation could also be addressed and better controiled.
There was also experimenter bias, difficult to avoid in the context where this study was conducted. ESL teachers who speciaIise in teaching pronunciation are rare in
Montréal (Québec). Finding unilingual anglophones with ESL experience and the appropriate linguistic knowledge to serve as judges of transcription reliability is also a sizeable challenge in the Montréal area. Recom rn endations
The importance of increased intelligibility needs particular attention. Including
the rules of connected speech in ESL pedagogy can go far in irnproving the listening and
speaking skills of al1 learners of English. HaWig taught large numbers of learners of
English various pronunciation techniques, 1 am wnvinced that learning suprasegmentals
wiIl benefit at least some of them. I believe that teaching these Iearners the
suprasegmentals requires less time and produces more effective results than teaching
simply how to articulate various consonants and vowels. 1 am, however, keenly aware that empirical measurements are required in order to support this hypothesis, as well as to bring about some changes in teacher training institutions.
This investigation into the teaching of the rules of connected speech indicates that the effectiveness of instruction may not depend solely on explicit teaching and eiicited repetition; indeed, perhaps a more communicative strategy needs to be implemented.
There are clearly some issues to be resolved in successfully eliciting spontaneous speech, and so the method of speech elicitation needs carefbl consideration. For example, in fùture expenmentation, students could describe a well-known fairy tale from their own culture. Another suggestion might be to have pairs of participants discussing the differences in two pictures, which are identical except for a restricted number of differences. This would hopefully alfow NNSs a chance to provide output that would more closely match that of NSs in quantity and quality.
A more discrete monitoring system needs to be impfemented. This systern should involve a soundproof roorn where no interruptions will occur. A high-quality recording 56 device should be used in a discrete way so that the participants are not so keenly aware of its presence, thus Ieading to a more artificial output.
In preparation for fùture experiments, the recruitment variables would need to be more carefully considered. For example, knowing exactly what the participant had studied in their university courses wodd help in creating a more homogeneous group.
This would help the researcher to avoid the surprise 1 encountered when 1 discovered that some of my participants had studied linguistics.
The most apparent need for change should be the number of participants recruited. The practicality of explicitly teaching connected speech following the use of the exercises developed for this study with a larger group of students would possibly derive more conclusive results. It is recomrnended, however, that linking words beginning with a vowel be taught in one type of experiment, that is, CV and W comection. Linking words that begin with a consonant, that is, CC connection, should be investigated in a separate study, since the phonological strategies in CC environments used by NSsE are far too complex to be used in the same shidy.
Pedagogical intervention of four weeks, or more, with at least one-week intervals would allow time for more significant changes in pronunciation behaviour. While explicitly teaching participants how to connect words beginning with a vowel, it would be important to include lessons on how to articulate potentialty difficult sounds in word- final positions such as [y], [w], [r] and [n] which would be linked with the fotlowing vowel. It would also me helpful to give an oveMew of al1 the vowels which end in off- glides in order to help the students understand that these are essential elements in assisting them to connect words.
57 Pronunciation pedagogy still has some fascinating areas to be fùrther investigated, and it is hoped that these and other areas of prosodic features wili receive more attention in second Ianguage acquisition research. References
Acton, W. R, Gilbert, J., & Wong, R (199 1, March). Integratingpromnciation into the second Zanguage cumimIum. Paper presented at the 25" Annual International TESOL Convention, New York, New York,
Anderson-Hsieh, J., Riney, T., & Koehler, K. (1994). Connecteci speech modifications in the English of Sapanese ESL leamers. IDEAL, 7,3 1-52.
Avery, P ., & Ehrlich, S. (1 992). Teaching American EngZish promnciatïonn Oxford: Oflord University Press.
Azar, B. S. (1989). Understanding mzd using English grarnrnar. (zd ed.) . Englewood Cliffs, NJ: Prentice Hall Regents.
Berish, L., & Thibaudeau, S. (1993). Cenadm Concepts. Scarborough, Ontario, Canada: Prentice-Hall-
Brodkey, D. (1972). Dictation as measure of mutual intelligibility: A pilot study. Language Lea.Imzg, 22, 203-220.
Brown, G. (1977). Listening to spoken English. In C. N. Candlin (Ed.), Applied Li~zgr~zsticsmd hzgunge SZU&. Essex, UK: Longman.
Browne, S- C., & Huckin, T. N. (1987). Pronunciation tutorials for nomative technical professionals: A program description. In H. J. Morley (Ed.), Czme~ztper~ectives on promrnciation (41-58). Washington, DC: Teachers of English to Speakers of Other Languages.
Boudreault, M. ( 1967). Prononciarion dufrançais par le rhythrne. Québec: Les Presses de 1' Université Laval.
Carver, T. K., Fotinos, S. D., & Cooper, C. (1 993). A CdianConversation Book: English in Everyday Lif . Scarborough, Ontario, Canada: Prent ice-Hall.
Celce-Murcia, M., Brinton, D. M., & Guodwin, J. M. (1996). Teachingpromnciation: A reference for teachers of EngZish to speakers of other Imguages. New York: Cambridge University Press.
Chan, M. (1 98 7). Phrase by phrase: Promnciafion and listening in Arnerican Englsh. Englewood Cli ffs, NJ: RegentdPrentice Hall. Dauer, R M. (1 993). Accurate English: A coqdete course in pronunciation. Englewood Cliffs, NJ: Regentflrentice Hall.
Dickerson, W. (1 989). Stress in the Stream of speech - The rhythm of spoken Eng(ish: Teacher 's manual. Urbana, IL: University of Illinois Press.
Fayer, J. M-, & Krasinski, E. (1 987). Native and nomative judgrnents of intelligibility and irritation. Language IRaming, 37, 3 2 3 -326.
Fournier, P. (1999). EngIish on Zine: introaUcing college Englsh SM..(2d ed.). Scarborough, ON: Prentice Hall.
Fned Go- B. (2000). Speech/Pronunciation consulting. ESOL Matlers, FebmaryMmch, 14.
Garant-Viau, C. (1994). La portée des sons. Québec: Les Presses de l'université Laval.
Gilbert, J. B. (1 984). Pronunciation and listening comprehension. In H. J. Morley (Ed.), Cz1rrent pervectives on promrnciation (29-3 9). Washington, DC : Teachers of English to Speakers of Other Languages.
Gilbert, J. B. ( 1 993 a). Clear qeech - pro~nrnciatio~~and lzsiening comprehension in North A mericm English: Sttîdent 's book. New York: Cambridge University fress.
Gilbert, J- B. (1993 b). Clear qeech - Prornzrnczafion and listening comprehension in North Arnerican English: Teacher 'sresarrce book. New York: Cambridge University Press.
Gimson, A. C. (1970). An introduction to the promrrtciution of English (2"* ed.). London: Edward Arnold.
Grant, L. ( 1993 ). Well wid: Advunced EngIish promnciation. Boston: Heinle & Heinle Publishers.
Gynan, S. N. (1984). Attitudes toward interlanguage: What is the object of study? Modem Lanpage JmrnaI, 68,315-3 2 1.
Hadden, B. L. (199 1). Teacher and nonteacher perceptions of second-language comunication. Language Lemming, 41. 1-24-
Halliday, M. A K. (1989). Spoken andwrilten Imguage. Oxford: Oxford University Press. Hancin-B hatt, B ., & Rakesh, M. B. (1 997). Optima1 L2 syllables: Interactions of transfer and developmental effects. Studies in Second Language Acquisition, 19,33 1-3 78.
Hatch, E. M. (1 992). Dzscolrrse and Ianguage educatzon. Cambridge: Cambridge University Press.
Hatch, E. M. (1 983). Psycholinguzstics: A second Ianguage perqecfive. Ro wley, Massachusetts: Newbury House.
Hinenoya, K. & Gatbonton, E. (2000).Ethnocentrism, cultural traits, beliefs, and English Proficiency: A Japanese Sample. Language Jmrnal, 84,225-240.
Jqanese for busy people. (2"d ed.). (1 994). Tokyo: Association for Japanese-Language Teaching
Jones, R. H. (1 997). Beyond "Listen and Repeat" : Pronunciation Teaching Materials and Theories of Second Language Acquisition. Sysfem, 25, 103-1 12.
Kaneman-Pougatch, M. & Pedoya-Guimbretière, E. (199 1). Plaisir des so~zs.Paris: Hatier.
Kreidler, C. W. ( 1989). fie promnciution of English: A course book in phonology. Oxford: Blackwell.
Lado, R. (1957). Linguistics across cultures: Applied linguisticsfor Imgaage teachers. Ann Arbor: University of Michigan Press.
Larsen-Freeman, D. & Long, M. H. (1990). An Introduction to Second Language Acquisition Research. In C.N. Candlin (Ed.), AppIzed Linguistics and Language Sfucjt. London: Longman.
Lightbown, P. ( I 98 5). Great Expectations: Second-Language Acquisition Research and Classroom Teaching. AppliedLinguistics, 6, 173-189.
Ludwig, J. (1982). Native-Speaker Judgments of Second-language Leamers' Efforts at Communication: A Review. Mdem Lcntguage Journui, 66,274-283.
McNerney, M., & Mendelsohn, D. (1992). Suprasegmentals in the pronunciation class: Setting priorities. In P. Avery & S. Ehrlich (Eds.), Teaching American English promnciation, 1 8 5- 196. Oxford: Mord University Press.
Miller, S. F. (2000). Targetingpromnciation: neintonation, sourds, and rhythm of American English. Boston: Houghton Mifflin.
Morley, J. (1 979). improving spoken English Ann Arbor : University of Michigan Press. Morley, J. (Ed.). (1 987). Current perspectives on promnciation: practices anchored NI îheory. Washington DC: Teachers of English to Speakers of Other Languages.
Morley, J. (Ed.). (1 994). Promnciation pedogogy and theory: New vious, new directions. Alexandria, VA: Teachers of English to Speakers of Other Languages.
Peaie, N. V. (1 993). Pontive thinking everyclqy: An Inrpiration for each &y of the yem. New York: Fireside. -
Pennington, M., & Richards, J. (1986). Pronunciation revisited. ESOL Quarterly, 20, 207-225.
Pinker, S. (1995). The lmpage instinct: How the mindcreutes lunguage. New York: HarperPerennial.
Pollock, C.W. (1997). Commmicate wharyou rnecm: A concise ahiancedgramrnar. (2nded., revised by S. Eckstut.). New Jersey: Prentice Hall Regents.
Prator, C . H., Jr., & Robinett, B. W. (198 5). MmaI of Americmz English promnciation. New York: Hoit, Rinehart & Winston.
Richards, J. C. (1992). Interchange. Cambridge: Cambridge University Press.
Seidlhofer, B. (1994). Pronunciation. In C. N. Candiin, & H. G. Widdowson (Eds.), Language Teaching: A Scheme for Teacher Education. Mord: Oxford University Press.
Smdley, R. L. & Ruetten, M. K. (1 995). Refi~zingcompositior~ skiZIs: Rhetoric and gramrnar. (4" ed.). Boston: Heinle & Heinie.
Temperley, M. S. (1987). Linking and deletion in final consonant ciusters. In H. J. Morley (Ed.), Currenfper.pectives on prornmciaîion (59-82). Washington, DC: Teachers of English to Speakers of Other Languages.
Thompson, 1. (1987). Japanese Speakers. In M. Swan & B. Smith (Eds.), Leamer Englsh: A teacher 's guide to inteMerence ami other problems (2 12-223). Cambndge: Cambridge University Press.
Trask, R L. (1 996). A dictionary of phonetics and phonology . London: Routledge.
Vance, T. 1. (1987). An introduction to Jipmese phonology. New York: State University of New York Press.
Varonis, E. M., & Gass, S. (1982). The comprehensibility of non-native speech Studies in Second Language Acquisition, 4, 114- 13 6. W mg, R. ( 1987). Teachz~~gprotnrnciatio~~: Focus on rhythm. Englewood Cliffs, New Jersey: Prentice-Hall.
Yule, G., & Macdonald, D. (1994). The effects of pronunciation teaching. In J. Morley (Ed.), Promnciation pehgogy mld Iheory: Nau views, new directions (109- 1 18). Alexandria, VA: Teachers of English to Speakers of Other Languages. Appendices
Appendzx A: Request for Teachers ' Assistance in Recruiting Jupanese
Speakers of English
Dear teacher,
1 am currently working on a master's thesis in the Applied Linguistics program of the TESL Centre at Concordia University in Montreal, Quebec. My research focus is the pronunciation of Japanese students of Engiish (JSE).
I have consulted with Mr- Charles Brown, and he hasr kindly aIIowed me to recmit JSE from your intensive English program as well as to corntact you for help.
PIease give them the attached information sheet and kindly ask them to contact me this week, if possible. Also, 1have posted additional she~son the bulletin boards in your department, in hopes that JSE will take note and wish to voiunteer. It is important for them to know that their participation is completely voiuntary, and that they will not be penalised if they do not wish to participate.
Thank you for your assistance.
Sincerely,
Marie Melenca Applied Linguist ics Graduate Concordia University TESL Centre 2070 MacKay Street, 2" floor (5 14) 768 3299 Appendix B: Participant Recruitment Letter
Dear student fiom Japan,
I am cwently working on a master's thesis in the Applied Linguistics depariment at Concordia University in Montreal (Quebec). My research focus is the pronunciation of Japanese students of English (JSsE). I am looking for JSE who are willing to participate in my teaching expenment between April8 to 18.1 would Iike to meet *dents who:
P are in intemediate or advanced levels of English > have lived out of Japan for less than two years P speak no other languages than Japanese and English R are between the ages of 1 8 and 3 5
In return for your assistance, I will give you a free pronunciation class equivalent to the time you spend participating in my research.
Ifyou are interested in participating, or if you would like to have more information, please contact me by leaving a message in the secretary's office of the TESL Centre, by phoning me, or by sending me an e-mail.
Thank you!
Marie Melenca TESL Centre 2070 MacKay Street, 2ndfloor Telephone: (5 14) 768 3299 [leave a message] E-mail: [email protected]~
Please cut dong the line, keep the above information for yourself and leave the lower part of this page with me.
Name:
Telephone number:
University:
Level : Appendix C: Consent Form to Participate in Research (JSsE)
1agree to participate in a program of research being conducted by Marie Melenca as part of her master's degree thesis under the supervision of Dr. Lori Moms of the Applied Linguistics program of the TESL Centre at Concordia University.
PURPOSE
1 have been informed that the purpose of the research is to better understand the pronunciation diffiiculties of Japanese speakers of English. PROCEDURES
The researcher will record my voice while 1 read a text and discuss a topic in English. 1 will participate in 3 lessons of 30 minutes each for over a period of 3 days.
When the information has been collected, studied and written, my identity will not be revealed to anyone other than the researcher.
CONDITIONS OF PARTICIPATION
1understand that I am fkee to lave at any time and discontinue my participation at any time without penalty. t7 1 understand that my participation in this study is CONFIDENTIAL (that is, the researcher will know, but will not disclose my identity in the menthesis.)
(7 I understand that information about my age, language skills and communication abilities will be kept CONFIDENTIAL.
1understand that the data fiom this study may be published but no information will permit me to be identified.
Cl 1 understand the purpose of this study and know that here is no hidden motive of which 1 have not been inforrned. Ei9= Please ~LKII to the back of the page CONSENT FORM TO PARTICIPATE IN RESEARCH (JSsE)
1 HAVE CAREFULLY STUDIED THE ABOVE AND UNDERSTAND THIS AGREEMENT. 1 FREELY CONSENT AND AGREE TO PARTICIPATE IN THIS STUDY.
NAME (PLEASE PRINT)
SIGNATURE
WITNESS SIGNATURE
DATE Appendix Dr Consent Fom to Participate in Research (1VSsE)
I agree to participate in a program of research being conducted by Marie Melenca as part of her master's degree thesis under the supervision of Dr. Lori Morris of the Applied Linguistics prograrn in the TESL Centre at Concordia University.
PURPOSE
I have been informed that the purpose of the research is to better understand the pronunciation difficulties of Japanese speakers of English. PROCEDURES
The researcher will record my voice while 1 read a text and discuss a topic in English. 1 will participate in 3 lessons of 30 minutes each for over a period of 3 days.
When the information has been collected, studied and written, my identity will not be revealed to anyone other than the researcher.
CONDITIONS OF PARTICIPATION
1 understand that 1 am fiee to leave at any time and discontinue my participation at any time without penalty.
O 1 understand that my participation in this study is CONFIDENTIAL (that is, the researcher will know, but will not disclose my identity in the written thesis.)
1 understand that information about my age, language skills and communication abilities wiI1 be kept CONFIDENTIAL.
0 I understand that the data from this study may be published but no information will permit me to be identifiai.
1 understand the purpose of this study and know that here is no hidden motive of which 1 have not been informeci.
Please tum to the back of the page EiF CONSENT FORM TO PARTICIPATE IN RESEARCH (NSsE)
I HAVE CAREFULLY STUDIED THE ABOVE AND UNDERSTAND THIS AGREEMENT. I FREELY CONSENT AND AGREE TO PARTICIPATE IN THIS STUDY.
NAME (PLEASE PRZNT)
SIGNATURE
WITNESS SIGNATURE
DATE Appendix E: Pre- and Post-Experiment @restions (JSsE)
PRE-TEST INSTRUCTIONS READ TO EACH JSE PARTICIPANT b The researcher, Marie Melenca, will record you reading the dialogue, "At the Travel Agent's 0ftice."*
B Once you have done that, the researcher will ask you to answer a general question so that she can record your speech patterns when you speak in a normal manner.
B The question:
'b What did you do last night? That is, what kind of activities did you do? Be specific and descriptive.
POST-TEST INSTRUCTION. b The researcher, Marie Melenca, will record you reading the dialogue, "At the Travel Agent's Offke. "
B Once you have done that, the researcher will ask you to answer a general question so that she can record your speech patterns when you speak in a normal manner. b The question: What will you do tonight? Be specific and descriptive.
Gilbert (1993b) Appendix F: Pre-Experzment Questions (NSsE)
PRE-TEST INSTRUCTIONS TO NSE PARTICIPANTS b The researcher, Marie Melenca, will record you reading the dialogue, "At the Travel Agent's a fi ce."^ b Once you have done that, the researcher will ask you to answer a general question so that she can record your speech patterns when you speak in a normal manner. b The question:
B What did you do last night? That is, what kind of activities did you do? Be specific and descriptive.
6 Judy B. Gilbert (1 993) Ciear Speech: Teacher 's Resource Book, Q Cambridge University Press.
71 Appendix G: CIear Speaking est^
At the Travel Agent's Office
A: ' Can 1 help you?
B: Yes, I want to fly to Chicago on Wednesday the seventh and retum on Friday the
nint h,
A: Of October?
B: No, November. How much is the fare?
A: Fares are cheaper if you stay over Saturday night.
B: Thanks, but unfortunately I've already ar-ranged some business here
that Fnday. 8 So 1'11 just have to pay the extra cost.
A: What time of day would you prefer? 'O Moming or aftemoon?
B: IL Moming, because 1 have to be there by early evening. l2 1s there a meal?
A: I3 Yes, they'll be serving breakfast; and you'll also see a movie.
B: l4 Which movie?
A: l5 In both directions they'll show a short feature on plamed communities. l6 Going
east, the major film is City SIickers. '' I think it7sa cowboy comedy. l8 The movie
going west is Big Joe. l9 That's an adventure story about a boy who raises a wolf.
B: " Sounds good, but what's the fare?
A: ZL Eight hundred and fifiy dollars round trip.
B : 22 Thatys more than I expected!
7 Gilbert (1993a' p. xii) Appendix H: Experzmental Teachzng Unit 1 - C V
As you read this page ofprinted English, notice the spaces between the words. It is easy to see where one word ends and the next word begins.
In spoken English, however, there are little or no spaces [or pauses] between words in sentences. Words seem to run together. Students of English are ofien not sure where one word ends and the next word begins when they listen to spoken English. The speaker links words together in groups to help the listener get the sense [or meaning] of the sentence- In short sentences, dl words are linked together, with no pauses. In this Iesson we will practice linking words-
In linking, the end of one word blends into the begiming of the next word with little or no pause. This mark " " is used to show linking. Repeat the following sentences. Link w the words.
Who are you? wu I'm ready for class. V wu My office is empty. V V V
B. ~inkin8means that words should be joined smoothly to each other without adding extra sounds or omitting final consonants. There are three types of comected environments. You will iearn one of three types of comected speech today.
C Linking A Final Consonant Tu An Initial Vowel [CVJ
If a word ends in a consonant and the next word begins with a vowel, use the consonant to begin the syllable of the following word: do NOT insert a glottal (do not stop the air or pause) between the two words.
1. In an airplane. V V 2. Live all alone. V V
D. hactice the follo~vingsentences with CV e~tvironments.
* Al1 of part one, with minor modifications, including the examples, is from Morley (1979). 9 B, C, D (1-10) and E are fiom Dauer (1993). 1. It's an apple. w w 2. Come in. V 3. It7s al1 over. V w
4. lgiveV up. 5. He made a mess. w
6. Cookw ameal. 7. He robbed a bank. w 8. She baked a cake. V 9. She loves himl0. w kr. 10. He likesu 11. They Save old bottles. II V
12. They believed weverything. 13. We live in an apartment. wuw 14. He had a big car. w
15. They lovew animals.
16. Many men shavedw every day.
17. Highway workers pave V our roads.
18. Some dogs retrieved w al1 balls.
19. Rude people shoveV into lines. 20. We approved al1 -ber work. V w
El Mark the links between words where ihere is ofinal consonant and initial vowel environnrent
In the moming the sun was up and the tent was starting to get hot. Nick crawled out under the mosquito netting stretched across the mouth of the tent to look at the morning. The grass was wet on his hands as he came out. He held his trousers and his shoes in his hands. The Sun was just up over the Ml. There was the meadow, the river and the swamp. There were birch tress in the green of the swamp on the other side of the river.
--- -- 'O Participants were informed that pronouns with an initial "h" were not pronounced within a sentence. l' D (1 1-20) and E are from Gilbert (1993a). Appendix 1: Experimenlal Teaching Unit 2 - W
In linking, the end of one word blends into the beginning of the next word with little or no pause. This mark " " is used to show linking. w lL Linking A Final VdTo An initiai Vmvel [F/r//
If a word ends in a vowel, as well as fwl and [y], and the next word begins with another vowel, go fiom one vowel right into the other without stopping your voice. Use a short [y] sound after front vo~els'~,a short [w] after back vowels. and [r] afier words ending in 'er' . The word 'the' is usually pronounced as [&y] before words beginning with vowels. This occurs not only between words, but also between syilables.
1. go out w 2. the answer is u V
3. triV angular incision [triangular] V
Practice the following sentences with W enviro~me~ts.
Let's go over it. w
I'm taking bi Vology [biology] and geV ology [geology]. May 1 ask? UV Do 1 know bal3? V V 1'11 Wear it. V Did you see kr? w They owe kmrnoney. V V 1 always tie it up. V V His blue eyes were open. w w 10. Where is the V
12 Front vowels are the following: [ iy 3 as in 'beat', [ I ] as in 'bit', [ ey ] as in 'bait', [LJ as in 'bet', [a]as in 'bat7. l3 Participants were informed that pronouns with an initial "h were not pronounced within a sentence. 14 Sentences C (3-10) are tiom Dauer (1993). 1 1. What's the answer? V 12. He asked me. V
13.1 hate to Sayu it.
14. "I think you should find another sport," 1w answered 15. From now on, I'll do better. w 16. Then she asked, "How are you?' V V
17- Wew agree on reV dity [reality].
18. Wew always study geu ography [geography] 15.
D. Mark the links between wurk wkethere is afrnal vowel and initial vowel environment
We often go the beach on the weekend. We always go by car, because it's fastest. My roommate will try anything. He loves to su* but he isn't a great surfer. He always falls off the surfboard.
Sometimes I go over to my fiend's house. The first thing she says is,
"How are you?" If 1 reaily tell her how I am, she doesn't listen. This happens so ofien that 1 decided to answer "fine" every time. But then 1 said, "You always ask, but you don't listen. " So she apologised and said, "My biology exam is freaking me out. I'rn sony- "
15 Sentences B, C (1 1-1 8), and D [excluding the Iast 2 sentences] are fiom Gilbert (1993a). Appendix J: Experzmental Teaching Unit 3 - CC
A. Linking A Fînal Consonant To An Initial Comona~t[CC/
Ifthe final sound is the same sound as the beginning of the next word, say the words as one word. Do not say the sound twice, say the sound once.
1. wash shells V 2. bus system w 3. tax savingI6 w
Ifa word ends in a consonant and the next word begins with another consonant, go directly fiorn one consonant to the next without releasing the first one (that is, without opening your mouth too wide, nor dropping your tongue) or adding a vowel sound like 'uh'. Furthermore, you should not omit the final consonant either. Try to say the two consonant sounds simultaneously- This also happens across syllables.
4. keepV taking 5. You Iaugh too much. V 6. Give good ad vice. 17 V w
Practice the follmving sentences with CC environ ments
Sam rnight go. V V The sun never sets. V Bill loves pie. w We both think it's good. w 1 wish she'd corne. w w
1 haveV very good ideas.
Pleasew stopgushing. Cook it in a deepjot. 18
w Where does the busw stop? 10.1 like black cats. w V 11. What time will you eat tonight? V w l6 Gilbert (1993a) l7 Dauer (1 993) l8 Gilbert (1993a) 12. His vacation was terrible. w w 13. Don't stop dnving. 19 V u
C Mark the links behveen WO& where tiiere is afinai Consonant and initial Consonant environment,
1. What's a "lab beaker?"
2. Put ten dollars in the box.
3. 1s this the right town?
4. Open the red door.
5- Our luck could change.
6. 1 took a cab back to town.
7. She has a black cat.
D. Mark the links between words where there is a#nal Consonant and initiai Consonant environment
The ~eci~e~'
A: How do 1 rnake it? B: You need an egg, an apple, an eggplant, an ounce of oatmeal, one onion, some apricots, some almonds, some orange juice, and some oil. A: Anything else? B: That's it! A: That sounds Iike an incredibIe recipe. B: It may be an inediblez2one. 1just made it up.
- l9 Dauer (1993) 20 J. B. Gilbert (1993a) 21 R. M. Dauer (1993) 22 Students were given the following footnote: Inedible = can't be eaten. Appendlx Kr Control Activzîy: Poem I
ENGLISH LESSON #I
Poetry is a usefil way to lem English. Please follow the instructions of the tacher.
Does the road wind uphill aI1 the way? Yes, to the very end. Will the day's joumey take the whole long day? From morn to night, my fi-iend.
But is there for the night a resting-place? A roof for when the slow dark hours begin. May not the darkness hide it fkom my face? You cannot miss that im-
ShaIl I meet other wayfârers at night? Those who have gone before. Then must 1 knock, or cal1 when just in sight? They will not keep you standing at that door.
Shall I find cornfort, travel-sore and weak? Of labor you shall find the sum. Will there be beds for me and al1 who seek? Yea, beds for dl who corne. Appendix L: Control Actzvity: Poem 2
Poetry is a useful way to lemEnglish. Please foliow the instructions of the tacher.
Robert Frost (1 874- 1963) The Road Not Taken 1916
Two roads diverged in a yellow wood, And sorry I could not travel both And be one traveler, long 1 stood And looked down one as far as 1 couid To where it bent in the undergrowth;
Then took the other, as just as fair, And having perhaps the better daim, Because it was grassy and wanted Wear; Though as for that the passing there Had worn them really about the same,
And both that morning equaIly Lay In leaves no step had trodden black. Oh, 1 kept the first for another day! Yet knowing how way leads on to way, I doubted if 1 should ever corne back.
1 shall be tellhg this with a sigh Somewhere ages and ages hence: Two roads diverged in a wood, and 1 - I took the one less travekd by, And that has made al1 the difference. AppendU: M: Control Activzty: Poem 3
ENGLISH LESSON #3
Poetry is a useful way to leam English. Please follow the instructions of the teacher.
Charles Martin (b. 1942) Taken Up 1978
Tired of earth, they dwindled on their Ml, Watching and waiting in the moonlight until The aspens' leaves quite suddenly grew still,
No longer quaking as the disc descended, That giowing wheel of lights whose coming ended Al1 waiting and watch. When it landed
The ones within it one by one came forth, Stalking out awkwardly upon the earth, And those who watched them were confirmed in faith:
Mystenous voyagers fiom outer space, Attenuated, golden - shreds of lace Spun into seeds of the sunflower's spinning face -
Light was their speech, spanning mind to mind: We corne here not believing what we find - Can il be ymr desire fo Zeme behirzd neearih, whzch those cdedungels bless, Exchanging amplitude for emptiness? And in a single voice they answered Yes,
Discord of human melodies al1 blent To the unearthly strain of their assent. Corne then, the Strangers said, and those that were taken, went. Appendk N: Tmscriptzons of Speech Samples while ~eadin~ loua'^
Due to the lengthy nature of this appendix (47 pages), the legend wiII preface the data in order to assist in the reading of the foilowing transcriptions. The numbers in columns seven and eight were used to facilitate Excel calculations. Some transcriptions in column four show a penod (" - ") in order to indicate a temporal pause where no sound has been produced. Also, there will be some transcriptions with a colon (" : ") in column four, this indicates that the vowel or consonant has been lengthened. These symbols are in keeping with IPA transcription conventions. A dash (" - ") is used to indicate that the item was not given by the speaker. Cotumn seven will show either "O," which means disconnected, or "1" which means connected. Column eight may sometimes show a "2," which means that a glottal was inserted during disconnection. Note that the transcriptions are given for four native speakers of English in order to give a baseline. The transcriptions of the nine Japanese speakers of English are provided for pre- and post-test conditions. In each case, there are 49 word pairs that were to be Iinked.
NSE Baseline Speech Samples
L - 1, N1 can I kanaY n aY 2. NI help you h~lpyuw P Y 3. Nl 1 want aywan YW 4. N1 Chicago on J~kagow.Pan ow a 5. N1 on Wednesday anwenzdey n w
6 - N 1 seventh and scvanû .2an 8 æ
23 Taken fkom CIear Speafing Test Gilbert (1993 a, p. xii)
82 1 ratarn. ?an l remnon
I N1 fares are f ~aza
N1 stay over steyowvar
N1 I've already ayhæv. 2dr~diy
N1 already mmged 3 1r cdiya r eynds2' N1 So 1'11 sow. 2ayl
N1 the extra aiyskstra N1 1 tirneof 1 taym. 23v N1 1 day would 1 dey. wud
N1 1 oraftemoon 1 oræftanuwn N1 1 because I 1 biykoz. Zay
NI by early bay. 2arliy
NI early evening arliy. 2iyvn1~
24 Please note this native speaker of English is of Afro-Hispanic origin and lives in New York. His speech moves fkeely between rhotic and non-rhotic varieties of EngIish. 25 Elided /d/ N1 and you'll anyd nyl NI you'll also yal . 2alsow la02 siyey
N1 feature on
NI going east N1 film is
NI think it's
izey
wcs. 21z
NI 1 an adventure
1 1 N1 1 about a (this pair was skipped) - - - -
bat1.wats ItlIw Io I NI 1 hundred and
NI Iexpected ayakspektad
"- - This participant skipped a iine. '' Speaker said this word without /U. N2 1 can I kanay I
~2 - 1want aywant I Y N2 Chicago on
N2 on Wednesday N2 1 seventh and I ratarnan n avaktowbar I
N2 fares are
N2 cheaper if rfya I N2 1 stay over N2 / but unfortunately N2 unfortunately I've
N2 I've already
N2 already arranged sowal ow
day would deywud 1 Y wodd you N2 1 moming or because 1 by early 1 bay. Parliy early evening there a 1 Bera and you'll 1 anyal you'li also 1 yalalsow see a siya show a 1 Sow. Pa feature on 1 fiyt Sar . ?an going east Piyst film is f 11mrz think it's B~gk~ts it's a 1 itsa going west 1 gowigwest west is wrst~z that's an aætsan an adventure story about stcriyabawt about a abawra raises a 1 awulf but what's 1 ba?. w~ts hundred and hndradan than 1 aanay 1 ayakspcktad N2 I I expected IY la I N3 cm1 kanay
N3 1 want aywant 1 lY IW N3 Chicago on
N3 on Wednesday N3 1 seventh and ratarnan avaktowbar
N3 ifyou
- N3 stay over steyowvar
N3 but unfortunately N3 1 unfortunately 1' ve N3 1 I've already 1 ayv. Palr~diy N3 1 already arranged al r e diy. 2a reynd3d N3 So 1'11 sow. Payl ow ay
N3 the extra &y. 3ekstra Y& N3 time of taymav m a N3 day wouid deywud YW N3 1 would you because I bahzay z by early bay. 2arliy Y early evening arliy. 2iyvn1g Y there a and you'll anya 1 n you'll also yalalsow 1 see a 1 siye I Y show a I going east 1 goxg. 2iyst l film is f11m1z m think it's ( ergk~ts I it's a itsa s going west gowigwrst D west is 1 westxz that's an ( Bstsan an adventure story about stxiy. Pabawt Y about a 1 abawra I a wolf awulf a but what's ba?. w~ts 3 hundred and hndradan d 147. N3 1 expected ayakspektad Y 1 kanay In 149. 1 N4 1 helpyou
N4 Chicago on rkagow . ?an ow
N4 on Wednesday anwrnzdey n
N4 seventh and sevenû.?æn 8 ratarnan N4 of October avaktowbar v
N4 fares are f erzar z
159. N4 if you ~fyuw f
160. N4 stay over steyowvar Y
163. 1 N4 1 I've already 1 ayv. Palrcdiy already arranged alr ediy . 2 areyn! 164- 1 N4 I I Y 16 N4 So 1'11 sow. ?al ow
166. N4 the extra aiy. 2ckstra Y 167. ( N4 1 time of 1 taymav 168. 1 N4 1 day would 1 deywud 169. 1 N4 1 would you 1 wudga 1 ar. Pæftarnuwn because 1 bahzay 1 bay. ~arliy early evening there a and you'll you'll also see a 1 siya show a feature on going east 1 gorg. liyst film is f 11rn1z think it's 1 ûrgk~ts it's a 1 itsa going West gowi~wcst that's an 1 Bætsan an adventure 1 anædvant~ar story about 1 storiy. labawt about a 1 abawra raises a reyz~za a wolf awulf but what's ba? . w~ts hundred and hndradan da1
196. N4 Iexpected ayaksp~ktad Y a
JSE Pre-Test Speech Samples
can 1
help you
1 want
Chicago on
on Wednesday seventh and sevens . ?e Is le rite. 2s lé lé of October
fares are fee. 38: e .e cheaper if
if you if". Piu
stay over ste. ?ove
-- but unfortunately bet. ?%fortlunetri
unfortunately I've
I've already
aiready arranged
so. ?silw the extra 216. J6 the extra dze, 2ekistr.i3 e '217. 56 time of teim. ?ovU m
218- J6 daywould de. 2ud e
219. J6 would you ud3u d 220. 56 morning or moni~- 2oe g 221. J6 or afternoon oe. 3eftenU e
222. 56 because 1 bikoz". PE~ z"
223, 56 by early b-ei. leli ei
224. 56 early evening eli.livmigs i 225. J6 there a zee e
226. J6 and you'll ene.3 iul ne
- 227. 56 you'll also yul. Pdso 1 228. 36 see a sie i
230. 56 feature on fiutle.26 e
23 1. 56 going east goi~.2ist KI
232, 56 film is firem. lis m
233. 56 think it's si0kU.lits k"
234. 56 it's a its-e s
235. J6 going West goi~.west Q
236. 56 west is west. 3idz" t
237. J6 that' s an dzets. 22 s
238, J6 an advenhire 6.?edUbe. Pedebent Se é
239. 56 story about storiebe"t i 16 1 I expected 1 e. lekspekted 57 1 can 1 1 kenei 17 1 help you 1 heopiu
J7 1 Chicago on 57 on Wednesday 6.Pwenzdei 57 seventh and sevens . 2.ena 57 ( return on
- 57 of ~ctober of. 2oktobar 57 much is metxi
57 cheaper if txipar. ?ifu
I J7 1 if you ifUiu 57 stay over ste. 2ovar
57 I've already ~iv.Polredi 57 the extra teimov de. Zut
J7 1 orafternoon or. 2-eftarn.ul r J7 1 because 1 bik~s.Pei s
- 57 eariy evening eli -2ivenX l 57 there a
57 and you'll
yul. Polso I sie i J7 show a
57 feature on
57 going east
57 film is film. 2is m J7 1 think it7s
J7 1 going west
J7 ( that's an J7 an adventure 288. J7 story about stori. 2~beut
289. 57 about a
290. 57 raises a
291. 37 a wolf
292. 57 but what's
293. J7 hundred and
dë. PE~ 1 295. 1 J7 1 1 expected 296. 58 can I kënei heopiu
298. 58 I want
299. J8 Chicago on
300. J8 on Wednesday 6.wënsde
1 303. 1 J8 1 of October
ferer 1 306- 1 J8 1 cheaper if 307. J8 if you ifiu 308. 58 stay over ste. ?ove 309. 58 but unfortunately
1 3 10. 1 J8 ( unfortunately I've 3 1 1. ( J8 1 I've already 1 3 12. 1 J8 1 already arranged I so. ?eilw 314. J8 the extra de. 2ekstre
1 3 16. ( J8 ( day would de. Pud3 1 3 17. ( ~8 1 would you
1 320. 1 J8 1 because 1 bikoze
- 321. J8 byearly bei. Zeli 322. J8 early evening eli. Piv~nZ
- -- 323. J8 there a dee 324. J8 and youYI1
325. 38 you'll also yu1 . Polso
t 326. 1 J8 see a 327. 38 show a 1 328. 1 J8 1 feature on 1 329. 1 18 ( going east fir~m.Pis
8i~k.Pits 1 332. 1 J8 1 it's a 1333. (J8 (goingwest goig . west west. ?iz 1 335. 1 18 ( that's an 1'an adventure 58 about a
J8 a woif 2. Puf e 58 1 but what's
kene i I J9 help you heopiu
J9 1 want eiwent J9 Chicago on
J9 on Wednesday 6.wenzdei O
J9 seventh and
J9 retum on
.Tg of October of. 2ok"tobar f
1 J9 1 fares are fers. 22 s
19 1 but unfortunately bet. PenfortJœnet"lwi t J9 1 unfortunately I've 360. J9 17vealready wu.Polredi voo
1 already arranged olr edi . ?-erents t ieO
So 1'11 so. 2~i1" O i 0
the extra de.2ekstre leeo time of t-eimov mol day would de. Tut eu0 would you utSu tS u 1
morning or moenig.Zo-e XI00
or afternoon oe. 3efteri.U eeO
because 1 bikosei s ei 1
by early b~.2eeWLi eO early evening e~~li.2iv~ni iio
1 1 1 1 there a 1 deo le la 11 and you711 myul nY1
you'll also pl.20lso 100
see a si-e i-el show a SOQ 0~21 feature on fitse. 20n 00
going east goi~.2ist io
film is film. 21s mi0
think it's zigku.lits kio
it's a its~ se0
going West goi~west Dw1 west is west. Piz tiO that's an detsen -I1 an adventure en. 2sd"bentSa n
- -- - story about storiebeud about a
but what's hundred and
dë. Pei ë 1 expected ei. 2ekspekted 1 oi
1 can 1 kenei n heIp you heopiu P I want eiwent 1 ei Chicago on on Wednesday seventh and
- return on of October fares are ferz. 2e l cheaper if tsipar. Pif if you stay over ste. ?=var e but unfortunately 408. J 11 unfortunately I've
I've already
aiready arranged
the extra
time of
day would de. Pud3
would you
rnorning or
or afternoon
because 1 bikez. Pei by early
early evening
there a
and you' 11 en. 3iuwi0 you'll also see a
show a
feature on going east
film is film. 3is think it's
it's a
going west goi~.west west is west .2ez that's an
JI 1 an adventure en. P~dventSar n-eO2
Jll story about stori. P~b~ut ieO2 Jl l about a (Speaker drops article) -
J11 lezez. 212 zeO2 Jll a wolf
but what's
Jll hundred and hsndred. 2~n
den. 2ei n eiO 2 1 expected ei.Pekspekted kenei help you heopiu 1 want eiwent Chicago on
on Wednesday
seventh and
of October
fares are fer. ?iz cheaper if
if you
stay over ste. Povar but unfortunately bet . lanfor t Juneit"li It le
Z've already W.2olredi v O so 1'11 1 so. 2eilw the extra dze.?ekstr~ e e time of teim. Pov m O day would de. 2ud e u would you 1 ud3u moming or mornir]. ?or 0 0 or afiemoon or. 2eftenii re
I because 1 bikos. 2ei s ~i by early 1 bei. 2erli early evening 1 erli.2ivUni~ there a ae~ e e and you'll ~n"iu1 nu i you'll also 1 iul. ldso see a 1 si.?e show a JO.Pe O e feature on futlar. ?on r O 1 going east goig. list D 1
I I film is film. 21s ii think it's 1 BigkU.Pits west. 2iz that's an dets. ?en an adventure en. ?eduventSar story about stori.3eb~ut about a
a wolf but what's hundred and hendred. 31x1
dë. lei 1 expected ei . Pekspekted kenei
- - help you heopiu
- - 1 want eiwent
Chicago on S ikego. ZG on Wednesday
. seventh and sevens. 3en return on of October tares are fers. ?sr cheaper if t 1ipa if you ifiu 1 ste. ?ovar l unfortunately I've 1 anf ort~uetwi. Peiv I've already 1 av. ~olredi I alreadymanged I 01redi.P-elents l So I'I1 ( so. loi0 l the extra de. 2ekstr~ e time of teimov rn day would 1 de. 2ud
- because 1 1 bikez. ?~i I = by early early evening eri. Piv-enï i
I there a ze~ e and you'll enyul n you'll also yul .Z)elso 1
1 see a si= i show a Soc O going east goig. 3ist D film is film. 9is m think it's 1 sigk. Pits I going west west. ?is It li that's an ze tsen s e an adventure en. 2~dYventSar n e story about storiebeut
-- - li le about a
a wolf but what' s bet. ?wez hundred and héred. ?i2 d a
dë. 2~i )ë lei 1 expected
help you heopiu P i
1 want
Chicago on on Wednesday
I seventh and sevens . Pende s e
of October ovoktobar v O
fares are fers. 3e cheaper if 552. if you ifU,2iu fi0 1 stay over steyovar YOl
I 1 I've already evoledi vol
so 1'11 so. Teil O ei 0
the extra ze. 2ekstre eeO time of teimov" mol day wouId de. 2ud3 eu0
would you ud3u dY1
moming or rnonirjor DO1 or aflernoon or~ftarnü re1
because 1 bikos". 3-ei s ei 0 by early bei. 3erli ei r O early evening erli. 2ivonï iio there a zee el and you711 md.Piul di0
youYI1also iul. Pelso le0 see a sie il
show a Joe OBI feature on fitSar.2on roO
going east goir]. 2ist qjio film is film. 3iz mi0 514 it's a itse s
J14 1 going west goig. Pwest 1 I 514 west is west. 2iz t 1 dzats. 1é I
J14 1 story about 1 storiebeut 1 ebeute
J14 1 awolf 1 ew-elf but what's
hundred and
aë. ~i ë 1 expected ~i.Pekspekted ~i
can 1 kenei n
help you
1 want
Chicago on
on Wednesday anwenzdei n seventh and sevenû . Pen ritarnen of October
J 1 5 1 €ares are cheaper if txipar. Zif rio2 if you stay over ste. Povar e O but unfomnately be t . PenfO r t S enet li
I1 It1 leI unfortunately I've 1 enf ortSenetli . Zeiv li lei I've already 1 eiv. Zolredi so 1'11 / so. 2eiw"l the extra 1 Be. 2ekstre time of 1 teimov day would 1 aewud would you morning or morni~.2or Q 0
- -- or afternoon 1 or. 2-eftarnun because 1 1 bihzei by early 1 bei. 2erli early evening erli.Pivni~ i i there a dere r e and you' 11 1 =Wu1 you'll also 1 yul. ?olso I1 see a sie i show a feature on fyut Sar. ?on I going east film is film. 2iz
think it's
it's a itse s going west goigwest west is west. 3iz
that's an an adventure enedvent S ar story about storiabeud
about a
raises a reizaza z
a wolf
but what's be? .Puet 2 hundred and
1 expected eyekspektad
JSE Post-Test Speech Samples 1 kenei 56 help you helUpiu 56 1want eimn
56 Chicago on Sikego. Po J6 1 on Wednesday J6 1 seventh and 1 sevens. 20
56 retm on rite. 213 of October
fares are cheaper if if you stay over ste. 2ov-e but unfortunately bet. PanfortSunetUli unfortunately I've anf ort Sunetuli. 2evU I've already already arranged
so. Peilw the extra time of teim. ?ovu day would de. Pud would you morning or because 1 bikozu.Pei bei. ?-el1 early evening there a and you'll you'll also see a sie J6 show a 10. Pe J6 feature on fyut SE. 26 002
56 going east goiq. Sist gi02
J6 film is firum. 3is mi02
J6 think it's Bigk". 2its kUiO 2
J6 it's a itse SEI
J6 going West goiqwest DWl
56 west is west. 2idzU tiO2
J6 that's an dzets. 2e se02 J6 an adventure 12.2ed~bentSE @eO2 56 stoxy about stori. 2ebe"t ieO2
56 about a ~b€?"t~ tI
J6 raises a leim ZB~
56 a wolf e. Puf ~~02
J6 but what's betU.Puets tUu O 2
56 hundred and h~nd~ed.%nu de02
56 than 1 86. le ë~o2
56 1 expected e.Pekspekted eeO2
57 can 1 kenei n ei 1 57 help you heopiu pi1
57 1 want eiw~nt -ei w 1
57 Chicagoon Sikego6 001
J7 on Wednesday 0wenzdei Io IW 1 I 1 , i 693. 57 seventh and sevens . 2enda se02 695. 1 J7 1 of October of. Poktobar 696, 57 much is
697. 37 fares are ferer 698. J7 cheaper if 699. ( J7 1 if you 700. 1 57 1 aay over ste. 2ovar e 70 1. 1 J7 1 but unfortunately bet. ?enfortSunetri unfortunately I've
~iv.Polredi
704. already arranged
so. 2eilw 00
708. 1 17 1 day would de. 2ut
1 709. J7 would you
I 710. 57 morning or
or. Peftarnfi 1 712. ( J7 because 1 bikes. 3ei aile /O eli. 3iv-enï
716. J7 andyou'll see a show a Io.?= O e feature on fyutxe. l8 O going east goiq. Pist Tl i film is film. 21s m i think it' s Bigkits k i it's a its-e se1 going West goig.Pwest QW" west is westU.Piz tu i O that's an aets~ s~l
1 an adventure enevent J G ne1 story about storiyebeut Y e about a ~b~ute te raises a lei22 z e a wolf e. Puf eu0 but what' s (Speaker drops this pair) - - i hundred and h~ndred.2211 d e than 1 dë. 2ei 6 '~31 1 expected 1 ei. Pekspekted I can I kenei n ~i
1 help you heopiu P i
1 want eiwent ei w
Chicago on sikeg6. 26 O O on Wednesday 6.wënsde OwO rite. 26 O 744. 1 J8 1 of October
( 746. 1 J8 ( fares are
1 747. 1 J8 1 cheaper if ifiu 1 749. ( 18 1 stay over ste. ?OVE le Io / 750. ( J8 1 but unfortunately bedenfort Sœnetli Id IB 1 75 1. 1 J8 ( unfortunately I've 1 752. 1 18 1 I've already 1 753- 1 58 1 already arranged
the extra
1 757. 1 J8 1 day would de. md3 758. 58 would you 1 759. ( J8 1 moming or oref tena because 1 bikoze bei. ?di 763. J8 early evening
764. 58 there a 1 765. ( 58 ( and you'll you'll also / yul. 301~0
I 767. J8 see a si^? 768- J8 show a 1owe 769. J8 feature on fitsar. 28
770. 58 going east goigist
772. 1 J8 1 think it's 773- J8 it's a itse
774. 58 going West goir]. west 775. 1 J8 1 west is 1 westiz 776. 1 J8 ( that's an 1 detsé 777. J8 an adventure O. lodventlar
778. 58 story about stori. lebeut
779. 58 about a ebeute 780. J8 faises a r-eizez~
1 781. JS a wolf e.Puf
I 782. 58 but what's b-e. 3mts 783. J8 hundred and hendred. ?en
1 1 784. J8 than 1 de. Pei
1 785. J8 1 expected eyekspektet 786. J9 can I kenei 787. J9 help you heopiu 788. J9 1 want eiw-ent
789. J9 Chicago on sik~go.?on 790. J9 ' on Wednesday on.wenzdei nw1
I seventh and sevenû . ?-en
1 of October of. ?okutobar
fares are fers. ?E
cheaper if
if you
stay over ste. Pov~r but unfortunately bet . Penfort SœnetUlWi unfortunately 1' ve enf ort Sœnet"lwi. levU I've already wu.Polredi already arranged olredi. 2~rentst
so. 2~11~
the extra
time of teim. 2ov
- day would de. Put would you
because 1
early evening eli.Pivenï
there a de. 2-e J9 you'll also pl.Polso I O
J9 see a sie i e
J9 show a 10.2e O e
J9 feature on fyutxe. ?on e O
r J9 film is film. Zis m i
J9 think it's 8igk".2its k i
1 J9 it's a its~ s 12 1 goigwest J9 west is west. 3iz
J9 that's an detsen J9 an adventure mebent 1-e n e
J9 story abotit stori. P~b~ud i .e 1 ebeude J9 I about a Id le J9 raises a lezzz~ z e J9 a wolf E. Puf e u J9 but what's bet"w~ts t w
J9 hundred and h~ndred.?en d e J9 IthanI 1 dé. Pei 18 lei J9 1 I expected 1 oi. Pekspekted lei (e JI1 Ica111 ken~i n ei
J1 I help you heopiu P i Ill 1 want ~iw~nt ei w
1 838. Jll Chicago on sikego. 30 0602
839. J11 on Wednesday Gwensde Owl
840. Jll seventh and seven0en Be1
841. JI1 retmon ritmi2 n1
842, JI1 ofoctober .ev.Pektebar veO2
843. JI1 much is m=tJ.?i txi 0 2 844- J11 fares are ferzar za1
845. Jll cheaper if t Siparif r i 1
846. JI 1if you ifyu fY1 847. J11 stay over ste. Povar eo1
848. JI1 butunfortunately b-eP.PenfortSunetli 2eO2
849. J11 unfortunately17ve enfortSunetli.?eivu gj. O 2
850. J1 1 I've already eiv". 3eledi v"eO 2
85 1. J11 already arranged oledi. PelenS t i~O2
852. Jl1 So I'll so. %ilw O eiO 2
853. JI1 the extra de. Zekstr-e eeO2
854. J11 time of teim. ?.ev me02 855. J11 daywould de. Pudg eu0 1 856. JI1 wouldyou uds;~ dY l 857. JI1 morningor rno~nig.20 gooz
858. Jll or afternoon o. Pæf tenfi 0~12
859. Jli because 1 bik~zei z ei 1 860. J11 by early bei. ?di eieO 2
86 1. JI I eariy evening ~li. ?iv-enï il02 there a
JI1 and you'll 1 en. 2iu
- - JI1 you'lÏ also iu. ?=Go
Jll see a
show a
Jll feature on
going east film is film. 2is I Jll think it's
Jll it's a
going West goig. west Jll west is west. 2iz
that's an
Jll an adventure
story about
about a (Drops article) -
- - a wolf
but what's hundred and hendred. ?en d
1 expected eikspekted
can I help you heopiu P Chicago on on Wednesday seventh and
of October ovoktobar faes are ferzar cheaper if if you
1 1 stay over steyovar Y 10 1 but unfortunately bedmfort Suneituli
I unfortunately 1' ve enfortSuneituli. Peiv i ( oi 0 2 I've already eivolredi vol already arranged olrediyereintJt Yel
so. 2~11" O eiO 2 the extra teimov day would de. ?ud eu02
- -- would you ud3u dY1
morni~.?or go02 or afternoon because 1 bikozei
I 1 beyerli YQ1 early evening there a aee e and you'll .enyu1 n you'll also yulelso 1 see a show a feature on fut Saron r going east goig. Pist Q film is filmis m
1 think it's Bigkits k it's a itse s going West goigwest rl westiz I t an adventure 1 enedvent Sar I story about 1 storiyebaut I Y about a 1 ebeute It raises a 1 reizize a wolf but what's bet. w~t It hundred and hendreden d than 1 demi n
1 expected eyekspekted Y can 1 kenei n help you 1 heopiu 1 want -eimnt Chicago on 1 5k~~ow.e on Wednesday ewenzdei
- seventh and -1 Fe=. ~sn of October 1 avektobar fares are 1 ferzer cheaper if txipar. ?if if you stay over ste. Povar
I but unfomnately 1 beenf ort Sue twi unfominately I've enfor t Sue twi . IZeiv I've already ( ev. Polredi
]de. Pekstre time of teimov day would 1 de. lud woufd you 1 ud3u or afternoon oref tenY because 1 bik-ezei 958. JI3 by early bel. 2eri
I 959.. JI3 early evening er iyivenl 960. 513 there a ze12 961. J13 andyou'll enyul 962. 513 you'll also yu1 . Zelso 1 sie 964. 1 J13 1 show a 965. 1 J13 1 feature on 966. 513 goingeast goig. Pist
967. J13 film is filmis
968, 513 think it's sigki ts 969. J13 it'sa its~ 970. 513 going West goigwest 971. 313 westis westis
972. J13 that'san z~t~n 973. 1 JI3 1 an adventure 1 ~nedventS ar 1 974. 1 J 13 1 story about ( storiebeut 1 975. 1 J13 1 about a 1 ebeut. 212 976. 1113 1 raisesa 1 reizezo I I I 977. 1 J13 1 a wolf 1 ewof 978. 513 but what's bet. 2wez
I I 1 979. 1 J13 1 hundred and 1 h0red. 2a
1 eyekspekted kenei In lei help you heopiu Ip li 1 want
- - Chicago on sikego. ?on 1.01 O on Wednesday
-- seventh and
I 1 rite. 28 10 10
of October ovoktobar v O
- . -.- fares are fers. 2.er Is le cheaper if
if you
stay over steyovar Y 0 but unfortunately
unfortunately I've
-. -- .. . . I've already evoledi alread y arranged
So 1'11 so.2~il O ei
t the extra ze. Zekstre E e time of teim. Pov day would de. Pud3 le lu would you
morning or because 1 bikos". ?~i s ~i02 by early b-ei. lerli ei sr 0 2 early evening erli. Piv-eni ii02 there a zee ee1 and you'll enyu1 nY1 you711also iul . Tels0 le02 see a si= ie1 show a 10~ 01 feature on fitSar.2011 roO2 going east (Speaker drops pair) - - - - film is (Speaker drops pair) - - - - think it's (Speaker drops pair) - - - - it's a (Speaker drops pair) - - - - going west goigwest r]W1 west is westiz ti1 that's an dzetsen se1 an adventure amduvent1 ar el story about storieb~ut ie1 about a eb~ute t-el raises a hiz~ ze1 a wolf ~?w~?lf w1 but what's b~d"w.et s du1 hundred and hendr eden da1 than I 8ë. ?ei ë eiO 2 1 expected can 1 help you I want
Chicago on on Wednesday seventh and sevenûan of October fares are ferz-er cheaper if if you stay over steyovar but unfortunately unfortunately I've eiv. 2olredi already arranged
so. ?eiwOl the extra time of day would would you - or afternoon 1 oreftarnun because 1 bikhz~i
by early bei. flerli
- early evening 1 prli. 2ivni~ there a I dere 1 and you'll .enyu1 you'll also see a 1 sie show a S0-e feature on fyutsar-Pon
going east goir]. 3ist film is filrniz think it's it's a 1 itse 1068. J15 going west 1 goigwest
I t-t-1069. J15 west is 1 west. liz an adventure 1 enedvent Sar story about storiab~ud about a abeuda
a wolf 1 awulf but what's hundred and 1 hmdreden 1079. JI5 1 expected ~iyekspektad Yel Appendù: O: Transcriptions of Free-Speech Samples
Due to the lengthy nature of this appendix (22 pages), the legend will preface the data in order to assist in the reading of the following transcriptions. The numbers in columns seven and eight were used to facilitate Excel cdculations. Some transcriptions in column four show a period (" . ") in order to indicate a temporal pause where no sound has been produced. Also, there will be some transcriptions with a colon (" : ") in column four; this indicates that the vowel or consonant has been lengthened. These symbois are in keeping with IPA transcription conventions. Column seven will show either "O," which means disconnected, or "1" which means connected- Column eight may sometimes show a "2," which means that a glottal was inserted during disconnection- Note that the transcriptions are given for four native speakers of English in order to give a baseline. The transcriptions of the nine Japanese speakers of English are provided for pre- and post-test conditions. The number of word pairs that were to be linked Vary for each speaker.
NSE Baseline Speech Samples
well on
on O?'
or about
'cause 1 did a drda
lot of lara
28 Please note this native speaker of English is of Mo-Hispanic origin and lives in New York. fis speech moves fieely between rhotic and non-rhotic varieties of English. 129 that 1 Baray t ay 1
IO. well um wel. 2avm la0
frrst 1 farst. 2ay t ay O
with a
and uh ceanda d a
1 uh afterwards a. Pæftawaadz a æ
afterwards uh æf taw~adza z a
went uh wmta ta
I I for a f~aa a a
couple of khpla 1 a dnnks at the 1 drrnksalBa be a biya YaI have uh hæva va1
uh a a. 2av a a 0
lounging area lawnd31geriya DE1
ok what 1 owkey. w~t' by bJ Io did 1 d~day 1
Well 1 wcl:ay 1 ay 1
and as ænd. 3æz dæo
as 1 æz. Zay z ay 0
class it klæs~t s 11 ( sow. ?ay 32- N2 outof awrav
34. N2 rain at reynæt
36. N2 with one w18w~n
37. N2 theuh aa.2a
38. N2 pharrnacy and farmasiyan 39- N2 spent extra spsnt . Pekstra 40. N3 couple of khplav
41. N3 because 1 biykazay
42- N3 homeuh hom. la 43. N3 uh earfier a. 2arliyar 1 fram. ?a 1 45. 1 N3 1 stay overseas 1 steyowvarsiyz 1 46. N3 went overseas wentowvarsiyz 47. N3 overseas uh owvarsiyza
48. N3 week of wiykav 1 49. 1 N3 1 was in 50. N3 for a f~ra
51. N3 week and wiykan
52. N3 then 1 &nay 153. IN3 Iwasin 1 wazrn 55, N3 weeksI s ay 1 I wiyksay 56. N3 fora fcma r a 1 57. N3 month in mnûrn 811
58- N3 and when anw~n nwl
59. N3 whenI wmay n ay 1
60. N3 home um howm. ?am m a02
61. N3 course 1 kxsay s ay 1
62, N3 had a hæda da1
63. N3 lot of larav r a 1
64. N3 catch up kwap tS a 1
65. N3 upon apan Pa1 66. N3 on and anan na1
67. N3 and uh ænda da1
68. N3 uh yesterday a.ycstardey aY"
69. N3 yesterday in ycstardey~n ey I 1
70. N4 night 1 nayray t ay 71. N4 gym and d31man m a 1
72. N4 takeout teykawt ' ka1 73. N4 out Indian awt' . P~niyan t1O2
74. N4 um ifs am. 21ts m 102
75. N4 it'sa rtsa s a 1
76. N4 pIace on pleysan s a 1
77. N4 Prince Arthur prrnsarûar s a 1
78. N4 Arthur and arûaran r a 1 79. N4 chicken with txrkan-w18 nwO
80. N4 yeah and yæ. 2ænd ææO2 81. N4 and1 ænday da1
82. N4 II ay. Zay ay ay O 2 83. N4 1 was aywaz ay w 1 84. N4 SOI soway w ay 1 85. N4 ate it eyr~t t11
86. N4 then1 acnay n ay 1
87. N4 lots of latsa s a 1
JSE Pre-Test Speech Samples
88. J6 school uh skul:-2.~2 89. J6 had a hede
90. J6 and uh
he. 2-ey , ' 92. 16 got a got. 2x2
93. J6 and uh 1 94. 1 J6 1 one hour 95. 56 hour and
zë. Pei no. 2~i 98. J6 it's uh
99. J7 yesterday I yestede. 2ei eiwë wente
1 dfiiends and frenz. Pen z ene n
57 uh we
57 we went
57 Nickels and zenwi n
J7 and uh
J7 one of eov I 57 fkiend uh f rende 57 uh is eiz l e 57 major in 17 ( in education 57 1 education and J7 and 1
57 1 am eim 121
J7 minor in meine:. 2ï e J7 in education
sowi talking about bet what's bet .wets t Japanese and and uh Canadian educational
. - systems and sistems. 2ex 7 and uh both of bosov
-7b oves v took a teacher and tit Je.?end she is her opinion opinion is is anyway iz. ?eniwe z strong and strog. ?en D and uh ene n have a lot of of influence off fluens f influence on
so.wi consider oh
ka opinions oke. 2~i e ei homework and and uh P-' a. Zei IP lei which was letterman and yes um yes . Pem: Is le showed ah le. Ped le le at the ede Id le program uh
1 went bed 1 bed. 2ei d ei ai. Pei ei ei
t 1 went eiwent ei w but uh they were showing a SoiIJ. Px? ID le driver uh 172. J8 not a note t I I teen agers tl.2cd3es ï driver uh dr~ives.2~n s J8 1 oke. ?ai 17 J9 1 went eiwent ei
177. 59 to a tue u
178. J9 concert uh konsart. Pe t l 179. J9 that was $etu.wez t 180. J9 was um wez~rn z
'181. J9 urn arabic ~m.2erebik m 182. 1 J9 1 concert and 1 konsart. Pend 183, J9 and um end~m d
184. J9 um ok em. 2oke m
185. J9 ok uh oke. i'e: e
186. J9 no actually no:.%kt Su~i O
187. J9 actually 1 ekt Suei. 2ei i 1 188. 1 19 ( forgot and
190. 1 J9 ( at uh Côte
1 191. J9 name of nemov m 192. J9 place was pleswez s
193. J9 so 1 soei O
194. J9 forgot also f o:get .?olso t it' s a 1 its~ 1 kind of 1 kenov think um 1 sink. lem I dun it's 1 deno. Pits it's a itse kind of k~inof space you 1 oke. Pei write assignment mireseZrnent
I for ESL fo.2iesel 1 so. ?el I write 1 aireit for about foebeut three hours 0ri. 2~2s hours on 1 eues. 16 on only lB.?dli only an 1 Bli. Pen
essay essay about 1 ese. lebeut
-- the al1 1 ze. 201 all immigrant 1 01. Zimiglant îhat is 1 zot. Piz It li Io then 1 zë. Pei
1wrote ~irot
that 1 zet - Pei Jll then 1 1 zë. ?ai 1 went 1 ei.went lei Iw 1 Jll night uh 1 neit. Pe: It le uh I e:.Pei B ~i
hornework is howmw-e:k. 2iz" k i 0 and uh ende d e
yes um yrs. 3e:m: s e um about e:m: . PebeutU m e
Kossovo and kosobo.P~nd O E and uh ende d .e
uh and e - ?-end G e and about endabeutu d~1 about Ottawa gbeut". Potew~
I Ottawa about 1 otowe. 1ebeut about uh zbouto
OK uh 1 oke. le: sins. Z2i
1 was having a today 1 tude. 2ei 1 was was working working on presentation 1
1 went
leiberi. 2e some information information up up and ep. ?end and uh
1 wanted eiwened but even though I
kerd. ?et
et. 2ov Mcgill it doesn't work because it's it's only its.?onli JI3 so 1 so. 2~i
J13 there uh aer . Pe J13 1 ten o'clock 1 ten~klek 313 o'clock or oklgk-30~
513 or eleven oe. 2ileven
51 3 eieven o'clock ilevenokl~k
514 did uh did. PE
J14 did uh dide: 31 4 1 things at 1 eto: 514 just 1 d3esei 514 maybeI mebiei 514 talk with tok. wi2
514 livewith livwis
J14 maybe 1 mebi. 2~i
J14 talked with toktewis J14 night oh neyto 1 ewat JI 5 night I neil. 2ei J15 madea medg J1S and1 ~nei 1 watched 1 eiwet 1 ei movie with muviwi8 i boyfnend and 1 boifrenen made uh 1 med-e: 1 e:.PP~ made a mede d thing it's Bigits D called uh stew ko1d.e. sityu d
-- -- stew and 1stW. Pend I and uh end-e d vegetables are vedzt~balzar z are inside erinseid r inside uh inseid. le d onion and 1 ni en. ?end l broccoli and 1 brekoli . Pend l yeah and 1 ie. Pen l and it's 1 anits I it's uh 1 itse I ha as 1 he. ?es 1 Q as you 1 ez~u creamy and 1 krimiend and white end. weit d and uh ende d JI5 uh it's e. 2its e i02
JSE Post-Test Speech Samples
56 tonight 1
-- J6 1 cook uh because
I 56 because uh bikoze e.Pei
so. 2~i
I J6 beef yeah maybe 1 mebiei
b -- eiwio
- - 56 yeah uh ye. 5%
1 56 hm yeah f 57 hope 1 hopei
- -- eiwil
17 1 because 1
-- J7 have a heve
57 lot of lotov
J7 yeah uh
57 uh yeah
t - J7 1 because 1 bikoz. 3-ei
57 just went a.est. went 1 wentoba It Io 57 so I 1 sooi 10 lei. 57 do it duit u i 57 it you ityu t Y 1 hev. Pei J7 mernorise everyt hi ng 1 betei
57 music I'm
J7 I'rn uh thinking
' 57 something else
57 because I'm
uh one
one of w-enof n O fnend uh
J8 1 and we endwi Id IW
18 have any heveni v e
J8 oh yes oyes OY 18 yes 1 yesei
T8 yeah it' s tonight 1
- - but uh maybe 1'11 mebieio 1 i and uh
homework w final exam but uh
-e. lits to uh next
so. 2eim
1 not in not. ?in t no only no. 26nli O
uh international there are part uh one part is short one iust 1 to explain tu. lekusupulen what is br example fonreksempo ut 382. 1Jl1 whatis mtiz t l -- - 383. JI1 global warming globolwemlg 1 - 2
2
2
387. 1 J11 ( choose a 1 tSuz.2~ 2 388. 1 J11 1 and I 1 end. lei I 389. 1 JI1 1 should explain I Sud. leksplë 390. 1 JI 1 1 explain about 1 eksplë. ?ebet
392. 1 JI 1 1 pages I 1 pedziz. 2oi
394- Jll so 1 so. 2ei O
395. Jll guess I ges. PQ~ s 396. 1 JI 1 1 1 will 1 eiwio 397. JI2 1will eiwio ei
398. 512 yes um ies. l~m s
400. 512 II -ei.P2i -ei 40 1. Ji2 some interviews semïnt~vius m
402. J 1 2 interviews on ïntevius.26 s
403. J12 Friday about f reide - 2ebeUt e
404. JI2 fitness and f itkes~nd s
405. 512 diet uh d~iet. ?G t JI2 1 something about J12 1 about about
J12 1 uh one
1 J 1 3 1 tonight uh
513 whoah fu. ?G u e
JI3 ahyeah t2.2~~ Y
JI3 doing a duirp DE
J 13 final wit h feinowi0 O w
513 anduh end^ d .rs J13 uh yeah e.ye GY 513 final in feineolin 1 i
113 1 Indo European
I 513 because uh bikoz~ z e J13 1 uh it's
t J13 uh1 e. 2ei e
513 read uh lots rid~ d 1 lotsof" J13 forit fodit d
J13 itand it. Zen t
523 anduh BnB n
513 uhyeah .e. 2iz e
513 doyou duiu u
J13 knowit nomi t O 1 so. 26 1 On l J14 1 Friday I 1 fY1eide.1~i l JI4 whatI wetei t
514 doon du. 28 u 514 Saturday and s~tzde.2~nd e
514 and 1 end. Pei d 114 1 made a list 1 medelist I l J14 and1 md. P-ei d JI4 have uh lot hevelot v
514 lot uh things 10t-esi~s t
29 Incomplete attempt in saying the word "hard." know what what uh w-et. 2e te uh what what I so 1 so. 2ei O ei finish up fini1.1~~ Se ha it's he. Zits e li still I'm sUtioeim O '~3 and uh 1 ende uh yeah B. ?ie E i well uh wele 1 a uh Sm e. ?eim B ei Perport to airport 1 du. lu le and I'm 1 end~im Id lei
I at um etem te at I'm 1 et. 2eim It lei 1 heus. 2aet s-six oh
t seven or sevenor n O yeah in yz. ?in e i 1 Bene uh I'm B. Peim e B
149 474. J15 I'm I'm ~imeirn
1 475. J15 toarrive tuereiv
476. J15 amivein ~reivin
477. JI5 Japan at d3ep~net
478, 515 atuh et=
479. J15 uh afternoon e. 2eft