Negation in Kata Kolok Grammaticalization Throughout Three Generations of Signers

UNIVERSITEIT VAN AMSTERDAM Graduate School for Humanities

Negation in Kata Kolok Grammaticalization throughout three generations of signers

Master’s Thesis

Hannah Lutzenberger Student number: 10852875

Supervised by: Dr. Roland Pfau Dr. Vadim Kimmelman Dr. Connie de Vos

Amsterdam 2017 Abstract (250 words)

Although all natural languages have ways of expressing negation, the linguistic realization is subject to typological variation (Dahl 2010; Payne 1985). Signed languages combine manual signs and non-manual elements. This leads to an intriguing dichotomy: While non-manual marker(s) alone are sufficient for negating a proposition in some signed languages (non- manual dominant system), the use of a negative manual sign is required in others (manual dominant system) (Zeshan 2004, 2006). Kata Kolok (KK), a young signing variety used in a Balinese village with a high incidence of congenital deafness (de Vos 2012; Winata et al. 1995), had previously been classified as an extreme example of the latter type: the manual sign NEG functions as the main negator and a negative headshake remains largely unused (Marsaja 2008). Adopting a corpus-based approach, the present study reevaluates this claim. The analysis of intergenerational data of six deaf native KK signers from the KK Corpus (de Vos 2016) reveals that the classification of KK negation is not as straightforward as formerly suggested. Although KK signers make extensive use of NEG, a negative headshake is widespread as well. Furthermore, signers from different generations show disparate tendencies in the use of specific markers. Specifically, the involvement of the manual negator slightly increases over time, and the headshake begins to spread within the youngest generation of signers. Thus, it is suggested that KK negation is a case in point for diachronic language change. This study contributes both theoretical and methodological insights to sign language typology and the study of diachronic language change.

Keywords sign language negation; village sign language; Kata Kolok; grammaticalization Acknowledgments

I am grateful to my wonderful supervisors, who gave me a guiding hand throughout the entire process of exploring KK negation. Their feedback and critical comments were as invaluable as their encouragement to take it yet a step further. In particular, I would like to thank Roland for keeping me on track and supporting me in letting go.

Further, I wish to thank Connie for encouraging me to dive into linguistics. I am incredibly grateful that she took the risk to take me to the field in 2015, where my fire for this language and community was light. I am overly excited to continue on this route. This project would not have been possible without the community. I would like to thank Connie for sharing her data, and especially, to all the deaf and hearing community members in the field for sharing their language and for warmly welcoming me. In particular, I wish to thank Ni Made Dadiastini, who became a dear friend and a lifesaver for any community- or language-related question. Thank you!

Lastly, I am grateful for everyone who helped me think, develop and articulate my ideas, listened to me, and provided feedback on previous drafts. I am indebted to Elisabet García González for her patience and her sharp mind, Frederik Harder for bike-rides full of negation and statistics, Jelke Bloem and Sara Iacozza for their help with linear mixed effect models, Freya Watkins for her superpower English, and Nick Palfreyman for sharing his expertise on Indonesian signing varieties with me. Lastly, I would like to express my gratitude towards my parents Susanna and Bernhard who never get tired of listening to yet another challenge in the project, and whose comments helped me making the figures to shine!

Glossary

Abbreviations of sign languages ASL American Sign Language

CSL Chinese Sign Language 中国手语 DGS German Sign Language Deutsche Gebärdensprache

HKSL Hong Kong Sign Language 香港手語 KK Kata Kolok Libras Brazilian Sign Language Língua Brasileira de Sinais LIS Italian Sign Language Lingua dei Segni Italiana LIU Jordanian Sign Language Lughat il-Ishaarah LSC Catalan Sign Language Llengua de Signes Catalana NGT Sign Language of the Netherlands Nederlands Gebarentaal NZSL New Zealand Sign Language TID Turkish Sign Language Türk İşaret Dili VSL Flemish Sign Language Vlaamse Gebarentaal

Abbreviations of non-manuals hs headshake neg negation nfe negative facial expression tp tongue protrusion

Transcription Conventions

GLOSS a manual form

GLOSS-GLOSS a single sign needing more than one word to explain its meaning

GLOSS#GLOSS two signs are articulated at the same time (fused)

GLOSS^GLOSS clitic

GLOSS | GLOSS clause boundary _____ duration of non-manual marking Contents

1 Introduction ...... 1

2 Negation: A Typological Overview ...... 3 2.1 Spoken Languages ...... 3 2.2 Sign Languages ...... 5 2.2.1 Manual and Non-Manual Marking ...... 5 2.2.2 A Typological Dichotomy ...... 7 2.3 Negative Concord ...... 9

3 Kata Kolok (KK) ...... 12 3.1 Linguistic and Demographic Overview ...... 12 3.2 Negation in KK ...... 13 3.3 Present Study ...... 14

4 Methodology ...... 16 4.1 Data ...... 16 4.2 Coding ...... 17 4.3 Procedure ...... 19

5 Results ...... 20 5.1 Overview of Results ...... 20 5.2 Manual Marking ...... 21 5.3 Non-Manual Marking ...... 24 5.3.1 Scope of the Headshake ...... 25 5.3.2 Tongue Protrusion ...... 26 5.3.3 Choice of Non-Manual Marking ...... 27 5.4 Statistical Trends ...... 29 5.5 Additional Observations ...... 31

6 Discussion ...... 34 6.1 KK in Typological Perspective ...... 34 6.2 Aspects of Grammaticalization ...... 36

7 Concluding Remarks ...... 40

References ...... 41

Appendix A ...... 45

1 Introduction

More than 6,000 different languages are used in various communities around the world. Crucially, both spoken and signed forms of communication exist. Contrary to popular misconception, signed languages are not a visuo-gestural equivalent of Esperanto, a constructed international auxiliary language. Instead, signed languages differ across geographical borders and sometimes even across cultural groups. Rural signing varieties represent a special socio-linguistic case. This type of signed language emerges rapidly in mostly isolated, rural enclaves often as a result of a high incidence of hereditary deafness. Compared to standardized spoken or signed languages, sign languages emerging in such contexts are relatively young. Moreover, they represent a communicative tool often shared by the deaf and a large number of hearing community members. There is a growing body of research investigating the diversity among rural signing varieties, comparing them with signed languages used in urban settings. These studies represent an invaluable contribution to sign language typology. The present study aims to contribute to this literature, focusing on negation in the sign language isolate Kata Kolok (KK), used in a village in Bali, Indonesia. Although negation is considered a linguistic universal, languages show a great deal of variation as to how negation is expressed (Dahl 2010; Payne 1985; Zeshan 2004, 2006). Moreover, negation is one of the few phenomena that has received a considerable amount of attention from the perspective of sign language typology (Zeshan 2004, 2006; Oomen & Pfau 2017). Across all signed languages studied, negation reveals a typological dichotomy. Nevertheless, KK does not seem to fit in the established classifications (Marsaja 2008). This study offers a review of the pattern of standard negation used in KK. Thus, the overarching research question of whether KK negation follows a manual dominant system with no or minimal use of non-manual elements is addressed. With this study, I contribute novel insights to the established typological categorization of negation. In addition, I raise both theoretical as well as methodological issues that are at the heart of the established classifications of sign language negation. Furthermore, KK is a relatively young signed language and may change more rapidly than established spoken or signed languages because of various factors such as its youth, the size of the community, and the setting of use. By taking an intergenerational approach, this study deals with the effects of diachronic language change over the course of three generations of signers.

The paper is structured as follows: I provide an overview of the typology of negation in spoken (Section 2.1) and signed (Section 2.2) languages and introduce the language of investigation, KK (Section 3). Section 4 explains the methodology of the present study. I adopt a corpus-driven approach and use intergenerational data to identify combinatorial patterns of different markers involved in standard negation and their syntactic position. Results are presented separately for the specific manual and non-manual markers (Section 5). The discussion in Section 6 is twofold: I evaluate the theoretical and practical implications of the observed patterns from a typological perspective (Section 6.1) and I offer speculations on potential pathways of diachronic language change (Section 6.2). This paper closes with some remarks on the limitations of the present study and suggestions for future research directions (Section 7).

2 Negation: A Typological Overview

Miestamo (2013) defines standard negation as “the basic way (or ways) a language has for negating declarative verbal main clauses.” Thus, negation denotes the most minimal technique to convert a sentence (S1) into a semantically opposite sentence (S2) so that S1 is true whenever S2 is false and vice versa (Dahl 2010). For instance, the negative polarity of the basic clause negator not affects the entire sentence in (1b). Notably, this does not apply to (1c) where the negative particle un- has scope locally over a single constituent, namely the adjective happy.

(1) a. John is happy. b. John is not happy. c. John is unhappy. [English]

Despite some variation in the theoretical approach, classifications of negation are often based on the nature of the negator (Dahl 2010; Payne 1985). Diagnostic tests target the scope of the negative element and thus, serve to distinguish sentential negation from constituent negation. For instance, in English, only the former can be followed by a positive tag question, e.g.‘is he?’ or continuations with neither e.g. ‘and neither am I’ (Payne 1985). Standard negation is by definition sentential. Hence, constituent negation as well as other specialized forms of negation such as negative derivational prefixes, negative quantifiers and adverbials, n-words (e.g. nothing), and negative existentials go beyond the scope of this study. Following Miestamo’s definition, I also exclude negative interrogatives, imperatives, and subordinate clauses. Lastly, this study is concerned with grammatically, rather than semantically negative contexts (e.g. wrong, reject, deny). The following sections provide a brief overview of different ways to express standard negation in spoken (Section 2.1) and signed languages (Section 2.2).

2.1 Spoken Languages All languages have ways to change the polarity of an affirmative into a negative clause. Nevertheless, the linguistic realization of such changes is subject to considerable cross- linguistic variation (Dahl 2010; Miestamo 2005, 2007). Typological studies of a large number of unrelated (spoken) languages show that morphological modifications, negative

particles, negative verbs, or negative nouns are all used to express negation (Dahl 2010; Payne 1985). Morphology: Languages like Turkish insert a negative, bound morpheme in the inflectional marking of the verb. Unlike (1c), this results in a negative interpretation of the entire sentence. In Turkish, the negative suffix -mV attaches to the verbal root, which yields a negative reading. As illustrated in (2b), adding -mu to the verbal stem oku ‘to read’ results in the negative sentence ‘I am not reading’. Across languages that form negation by means of morphology, negative suffixes are more common than prefixes or infixes. Occasionally, languages feature prefixation, e.g. Latvian. Infixation of negative morphemes, however, is rare, e.g. Dolakha Newar, a Tibeto-Burman language used in Nepal (Genetti 2007). In tonal languages, morphological changes can be realized by reduplicating or modifying tones. In Mbembe (used in Nigeria, part of the Niger-Congo family), for instance, the change from high (3a) to low (3b) tone on the tense prefix results in a negative reading. In Dahl’s (2010) sample, morphological negation appears as a typologically unmarked strategy.

(2) a. Oku-yor-um read-PROG-1SG ‘I am reading.’ b. Oku-mu-yor-um read-NEG-PROG-1SG ‘I am not reading.’ [Turkish | Dahl 2010: 14]

(3) a. mɔ́ -tá 3.FUT-go ‘He will go.’ b. mɔ̀-tá 3.NEG-go ‘He won’t go.’ [Mbembe | Dahl 2010: 17]

Negative particle: Particles are free standing and uninflected morphemes. For example, in Indonesian, the negative particle tidak signals a change from affirmative (4a) to negative (4b) polarity in a declarative sentence. At times, negation can consist of two independent, uninflected negators, e.g. in the case of ne...pas in standard French. This phenomenon will be discussed in more detail in Section 2.3.

(4) a. Saya tidur I asleep ‘I am asleep.’ b. Saya tidak tidur I NEG asleep ‘I am not asleep.’ [Indonesian | Dahl 2010: 19]

Negative verbs: Negative verbs (higher negative verbs and auxiliaries) represent a highly marked strategy for expressing negation. Higher negative verbs, such as in Tongan, are complement-taking matrix verbs. Negative auxiliaries take inflectional marking, whilst the lexical verb typically occurs in a nonfinite form. Finnish is a language that uses negative auxiliaries. As illustrated in (5), the negative auxiliary ei is inflected for person and number and the main verb lue remains uninflected.

(5) a. Pekka lukee P. read.PRS.3SG ‘Pekka is reading.’ b. Pekka ei lue P. NEG.3SG read ‘Pekka is not reading.’ [Finnish | Dahl 2010: 21]

2.2 Sign Languages Standard negation has been investigated in a substantial number of signed languages. It is argued that signed languages, for the most part, employ negative particles (Zeshan 2004, 2006). Nevertheless, there exists a proposal that negation may be analyzed as morphological construction in some signed languages, e.g. German Sign Language (DGS; Pfau 2008). Expressing negation by means of negative verbs is, at this point, unattested in signed languages.

2.2.1 Manual and Non-Manual Marking The visual-spatial modality gives rise to a great deal of simultaneity. Although the two hands represent separate articulators that may be used at the same time but independently, this is not very frequent. Nonetheless, the simultaneous combination of manual and non-manual elements, i.e. activity of the face and the body, is extremely common (Pfau & Quer 2010). Negation is a case in point where a manual sign, i.e. a negative particle, is used alongside a

non-manual element. Typologically speaking, the manual basic clause negator found most frequently is a side-to-side handwave, produced with either a 1-handshape (Figure 1a) or a 5- handshape (Figure 1b). In DGS (Figure 1c), the basic negator has no repetition but comprises a single sideward movement.

a. Catalan Sign Language b. Hong Kong Sign Language c. German Sign Language Figure 1: Form of the basic clause negator in Catalan Sign Language (a), Hong Kong Sign Language (b), and German Sign Language (c). Images a/c retrieved from Pfau & Quer (2007: 130), image (b) retrieved from Tang (2006: 217).

In contrast to spoken languages, bodily and facial behaviors can function not only as gestural, but also as phonological, morphological, syntactic, or pragmatic markers in signed languages (Pfau & Quer 2010). Syntactic non-manuals in particular have been well explored in the context of negation. The most widespread non-manual marker used is a negative headshake. Signers from Mediterranean countries, such as Greece, Turkey, and Jordan, also employ a backwards head tilt (Antzakas 2006; Hendriks 2008; Zeshan 2004, 2006). Most sign languages feature negative facial expressions including, but not limited to, squinted eyes and furrowed eyebrows. This array of non-manual elements make up the ‘not face’, a cross- linguistically and cross-culturally valid facial expression occurring in negative contexts among both gesturers and signers (Benitez-Quiros et al. 2016). Nevertheless, a negative facial expression is generally insufficient to negate a phrase. Chinese Sign Language (CSL), however, represents an exception: according to Yang & Fischer (2002), the negative facial expression functions as the main negator and a headshake alone does not yield a negative interpretation.

2.2.2 A Typological Dichotomy The use of manual and non-manual elements in sign language negation leads to a series of different combinatorial possibilities (see Table 1). Note that the two parts of Table 1 consider two partly overlapping features, namely (i) the obligatory presence of the manual element and (ii) the spreading of the non-manual element.

Table 1. Combinations of manual and non-manual marker. + indicates the obligatory presence of the marker; - indicates optionality; ø indicates that the marker is not used at all. i) manual element + + - ø

non-manual element ø - + +

manual manual non-manual non-manual only dominant dominant only

ii) obligatory presence of manual element + + - - spreading of non-manual + - + - element

manual non-manual ? ? dominant dominant

Table 1 lists eight combinatorial options that arise from the two aforementioned conditions (i) and (ii). Nevertheless, only two types have been attested in different signed languages to this point (Zeshan 2004, 2006; Oomen & Pfau 2017). Thus, sign language negation shows an intriguing dichotomy: some sign languages require the manual element to negate a proposition, while it is optional in others (Zeshan 2004, 2006). In so-called manual dominant sign languages, the use of a manual negator is obligatory in any negative context. A headshake may be – and often is – co-articulated with the manual negator. Nonetheless, the headshake by itself is insufficient to change the polarity of an affirmative sentence. Irrespective of the scope of the headshake, the omission of the negative particle leads to ungrammaticality, as illustrated in (6b) for Italian Sign Language (LIS). Whenever the negative headshake (hs) is present, it is lexically bound to the clause-final manual negator

NON, and does not spread over adjacent signs (6a).

(6) a. ___hs PAOLO CONTRACT SIGN NON b. * (____(______(___(____hs) PAOLO CONTRACT SIGN ‘Paolo did not sign the contract.’ [LIS | Geraci 2005]

The majority of sign languages form negation according to a different, a non-manual dominant pattern. In those languages, the use of a non-manual marker (headshake) represents the main negation strategy. Despite the fact that the manual negator is often co-produced, it remains optional. As demonstrated in (7) and (8), the headshake negates the sentence and the manual negator may be omitted (Zeshan 2004). The headshake usually accompanies manual signs, but the exact timing and the scope of the headshake are determined by language- specific constraints. In DGS, the headshake must accompany at least the predicate, regardless of whether the manual negator is used or not (7), and can optionally spread to all constituents except the subject (Pfau 2008). In American Sign Language (ASL), however, the headshake must spread onto the object and accompanies the predicate when the manual negator NOT is absent (Neidle et al. 2000). Moreover, in ASL it is felicitous for the headshake to have scope over only the manual negator, whilst this is ungrammatical in DGS. The licensed scope of the headshake is indicated with brackets in (7) and (8).

(7) ___hs (__hs) WOMAN FLOWER BUY (NOT) ‘The woman does not buy a flower.’ [DGS | Pfau 2008: 46] (8) a. ___hs (______hs) JOHN NOT BUY HOUSE b. ______hs JOHN BUY HOUSE ‘John is not buying a house.’ [ASL | Neidle et al. 2000: 44f]

To summarize, sign language negation can be grouped into a binary systems according to the obligatory presence of the manual or the non-manual negator and the scope of the non- manual element: a non-manual dominant and a manual dominant pattern. At the current state of research, non-manual dominant negation appears as the typologically unmarked form. Besides this dichotomy, a considerable amount of within-group variation exists in terms of the scope and the timing of the non-manual element (Pfau 2008, 2016; Zeshan 2004, 2006).

2.3 Negative Concord Regardless of language modality, multiple negative elements can occur within a single sentence. Whenever more than one negator does not change the polarity of the negative clause back to affirmative, we are dealing with negative concord (NC). Crucially, the negative elements do not cancel each other out (Corblin et al. 2004). In other words, both negative items do not interact with each other and are conceived as if they were a single one. Standard French and Afrikaans are examples of typical NC languages, i.e. NC is used to express standard negation (de Swart 2010). NC can materialize in two ways: standard negation in French consists of two different negative markers ne…pas (9), while the same negator nie is repeated in Afrikaans (10). In contrast, Russian and Italian are considered examples of non-typical NC languages. Standard negation in both languages employ a single negator, but NC is required whenever a n-word is used (11).

(9) Jean ne lit pas le livre J. NEG read.3SG NEG the.M book ‘John doesn’t read the book.’ [French] (10) Hy is nie moeg nie He is NEG tired NEG ‘He is not tired.’ [Afrikaans | Biberauer & Zeijlstra 2012: 359] (11) Gianni non ha telefonato a nessuno Gianni NEG has called to n-body ‘Gianni didn’t call anybody.’ [Italian | Penka & Zeijlstra 2010: 779]

The visual-spatial language modality allows for at least three different kinds of NC: two manual negators, two non-manual negators, or one manual and one non-manual negator. As explained in Section 2.2.2, the combination of a manual and a non-manual element represents the default in sign language negation. The case of two manual negators has been termed ‘negator-doubling’ (Zeshan 2004). This strategy gives rise to an additional pragmatic meaning in Brazilian Sign Language (Libras). Specifically, it has an emphatic function, as shown in example (12) (Nunes et al. 2008). In the case of negator-doubling, the two non- manual markers may fuse as a result of a harmony phenomenon, i.e. an assimilatory process over long distance.

(12) ______neg __neg I (NO) WILL BUY CAR NO ‘I will not buy a car.’ [Libras | adapted from Nunes et al. 2008: 178]

In KK, however, the repetition of the manual negator within the same sentence does not convey any additional pragmatic meaning. For this reason, I propose referring to this type as ‘negator-copying’ rather than negator-doubling. Although the reiteration of the same negative particle resembles example (10) from Afrikaans, NC in both Libras and KK is strikingly different: the reduplication of the negative particle is optional in both signed languages, while both negative particles need to be used at all times in Afrikaans. At this point, no signed language has been reported to exhibit obligatory NC involving multiple manual elements. Similarly, the combination of multiple non-manual elements as a result of NC remains unattested. Nevertheless, one may hypothesize that the general potential for simultaneity may facilitate the combination of more than two negative elements, both within and across channels. In contrast to NC languages, German exhibits double negation (DN), where every negator is sensitive to additional negative elements within the same sentence (Corblin et al. 2004). Consequently, an even number of negative elements yields a positive reading. As shown in (13), the combination of two negative words nicht (not) and niemand (nobody) results in an affirmative interpretation.

(13) Ich habe nicht niemanden gesehen I have.1SG NEG nobody seen ‘I have seen someone.’ (lit. ‘I have not seen nobody.’) [German]

(14) ?? SMOKE CANNOT NOBODY ‘Everybody must smoke.’ [LIS | Geraci 2005]

As for DN in signed languages, Geraci (2005) reports an example from LIS, where two negative manual signs, namely CANNOT and NOBODY, appear to give rise to a DN reading (14). Nevertheless, DN within the same channel is uncommon. In conclusion, similar to some spoken languages, sign languages may combine multiple negative markers. Notably, NC across two channels, namely manual and non-manual, is

widespread among sign languages. DN is clearly less common than NC, yet not completely unattested in signed languages.

3 Kata Kolok (KK)

3.1 Linguistic and Demographic Overview KK is a sign language isolate that arose in a rural enclave in Northern Bali, Indonesia, due to an exceptionally high incidence of hereditary deafness (Friedman et al. 2000, Winata et al. 1995). Deaf villagers bear deaf children and many hearing villagers also carry the recessive gene causing deafness. Genetic evidence indicates that the first cohort of deaf siblings was born into the village approximately 125 years ago, which marks the emergence of KK (Table 2; Friedman et al. 2000; Winata et al. 1995). From that point up to the present day, KK has been used and acquired by at least six generations of deaf children.

Table 2. Timeline for the emergence of KK. Table adapted from de Vos (2012: 47).

Event Generation Biological of signers generation

Spread of mutated At least 11 generations ago mutation of gene DFNB3 8 generations from 2011 I First deaf individuals affected by DFNB3

7 generations ago II

6 generations ago First cohort of deaf siblings 1st III Emergence of KK

Development of KK 4 generations ago 2nd IV 3 generations ago 3rd V 2 generation ago 4th VI 1 generation ago 5th VII

KK in its current form 6th VIII

In 2012, 2.2% of the village population was deaf, but no deaf baby had been born in the village between 2005 and 2015 (de Vos 2012, 2016). As a result of the high prevalence of deafness, the community has culturally as well as linguistically adapted to deafness in various ways. Hearing villagers speak Balinese and Indonesian and many also use KK on a daily basis (bimodal bilinguals). The degree of KK proficiency among hearing villagers varies according to various factors such as kinship relation, the amount of interaction with the deaf, or the geographic location of the home. Nevertheless, KK serves as means of communication

in social, political, and religious contexts, and deaf children are immersed in a rich language acquisition setting from birth in which not only immediate caretakers, but also relatives, teachers, neighbors, and peers can sign (de Vos 2012). Since 2007, KK has been implemented in deaf education at a local primary school and the international fingerspelling was introduced for schooling purposes (de Vos 2012). Nowadays, deaf youngsters may pursue further education in other parts of Bali where Indonesian signing varieties are used. Nonetheless, KK has developed under virtually no influence from any other signed languages, at least until 2007 (de Vos 2012; Marsaja 2008). This might have led to some typologically unusual characteristics such as an enlarged signing space, the use of an absolute frame of reference, and the absence of mouthings, i.e. silently mimicking spoken words (de Vos 2012; Marsaja 2008). According to Marsaja (2008), the basic word order in KK is SVO, yet de Vos (2012) shows that KK signers do not primarily rely on word order as subject and object are frequently omitted in spontaneous interaction.

3.2 Negation in KK Compared to other signed languages, the range of signs considered to be negative in KK is small: NEG, NOT-YET, and FINISH (Marsaja 2008; Perniss & Zeshan 2008). Negative signs, in particular NEG, occur predominantly in post-predicative or clause-final position (de Vos 2012; Marsaja 2008; Zeshan 2006). Although some examples in Marsaja (2008: 197) include negator-copying, it is not explicitly mentioned in the context of negation. NEG, a side-to-side handwave as shown in Figure 2, functions as default negator. Thus, besides marking standard negation, NEG serves to negate pronouns as well as imperatives (Marsaja 2008).

Figure 2: Manual negator in KK. Side-to-side handwave with 5-handshape. Image retrieved from de Vos (2012: 142).

Marsaja (2008) states that NEG is mandatory in all negative utterances, while non-manual markers alone never negate a clause. This would classify KK as a manual dominant system. Marsaja (2008) makes an even stronger claim by suggesting that standard negation actually requires NEG and only very rarely involves a subtle headshake. As all sign languages make use of both kinds of markers, this pattern would be typologically highly marked (Zeshan 2004, 2006). KK could even be a typological rarity: a manual-only system has not been attested to this point (see Table 1).

Furthermore, the negative markers NOT-YET and FINISH occur in more specialized contexts: NOT-YET functions as negative incompletive and FINISH serves as completive as well as negative existential. NOT-YET is formed with a padding movement, the palm of the hand facing the ground (de Vos 2012; Marsaja 2008). For FINISH, the palms of the hands are rapidly turned up and outwards. FINISH expresses a negative existential meaning and can be used interchangeably with NEG in most contexts of negative existence and possession (Perniss & Zeshan 2008). To the best of my knowledge, no in-depth study has tried to delineate the exact function of the different negators. NOT-YET and FINISH will not be included in the core of my study. FINISH is, however, be briefly mentioned in Section 5.5.

3.3 Present Study The current study offers a revised account of negation in KK. I challenge Marsaja’s (2008) classification of KK as manual-only sign language based on two pieces of evidence: (i) elicited examples in Marsaja (2008) include clearly articulated headshakes; (ii) fieldwork- and corpus-based observations indicate that headshakes frequently occur in negative contexts,

sometimes even without a manual negator. Moreover, a previous study suggests that the multifunctional marker tongue protrusion is frequently involved in negation as well (see also Figure 2; Lutzenberger in prep.). Hitherto, we lack a thorough study to confirm these observations. This study represents an in-depth analysis of intergenerational corpus data, aiming to disentangle the role, the relevance and the relation of manual and non-manual negation markers. I address the following research questions: (i) Is KK manual dominant in terms of negation? (ii) Are non-manual elements involved in KK negation? (iii) If yes, which non-manual elements and what are their roles?

The manual negator NEG is among the most frequent signs in the KK corpus, but headshakes also occur often. It is hypothesized that KK follows a manual dominant rather than manual- only system. I anticipate that the negative particle NEG functions as an obligatory negator that is (predominantly) accompanied by (a) non-manual element(s), specifically a headshake and/or tongue protrusion. Furthermore, KK is comparably young and emerged rapidly. A previous investigation of intergenerational data suggests that tongue protrusion shows characteristics of grammaticalization throughout three generations of KK signers (Lutzenberger in prep.). Besides uses as gestural, lexical, death marker, and negative evaluative, the use of tongue protrusion in the context of negation increases in the youngest generation of signers. On this ground, I hypothesize that non-manual negators may grammaticalize within three generations. This may lead to an increased involvement of non- manual markers and/or even the use of specific non-manual elements as independent negation marker. Specifically, younger signers are expected to use headshakes and/or tongue protrusion more frequently as a negator than older signers. With this study, I thus contribute to both the typology of sign language negation and the study of diachronic language change.

4 Methodology

4.1 Data The current study is based on the KK Corpus, a naturalistic data set of deaf KK signers of generation II through VI. The corpus is stored and archived in The Language Archive at the Max Planck Institute for Psycholinguistics, The Netherlands (König 2011). The KK Corpus comprises 50 hours of monologues and interactional data, both semi-spontaneous dialogues and multi-party conversations, and elicited data such as the Frog Story (based on Mayer 1969; de Vos 2016). Recording sessions were carried out by de Vos during fieldtrips and by a hearing research assistant, a fluent KK signer and active member of the community. Various communicative settings are featured such as festive ceremonies, social gatherings, story telling, and informal gossiping. The levels of transcription detail vary across the recordings, but 4.5 hours have been transcribed comprehensively by de Vos and (local) research assistants. For the purpose of this study, three semi-spontaneous dialogues with a relatively high level of transcription detail were selected. The most important selection criterion was to cover signers from different generations. Given that there are very few recordings of a single generation II-signer available and that the youngest signer (generation VI) is still an infant, the final data set consists of signers from generation III through V. Details of the sample are found in Table 3.

Table 3. Detailed information on the sample used in the present study.

Participant Signer 1 Signer 2 Signer 3 Signer 4 Signer 5 Signer 6

Gender male male female female male male Length of 61 min 18 min 25 min Recording Dyad Dyad I Dyad II Dyad III

Generation III IV V

Evidently, the sample is biased in gender and length of recording. Nevertheless, gender has generally not been reported to affect the grammatical realization of negation. Palfreyman (2015) is the first to discover a correlation of gender with the syntactic position of the negator and the predicate in two urban signing varieties of Indonesia. An effect of gender is, however, unexpected for KK, since occupational or social roles of villagers are rarely

restricted to a certain gender. The variance in the length of recordings will be accounted for through reporting values of negation per minute rather than absolute frequency.

4.2 Coding Although the selected data included detailed transcriptions, all files were enriched by manual coding, using the annotation software ELAN (Crasborn & Sloetjes 2008; Wittenburg et al. 2006).1 While the focus of this study is clearly on standard negation, other negative forms such as negative interjections, negative existentials, etc. have been included in the initial coding. In addition to coding the manual and non-manual activity, I dedicated separate tiers to selected functional and analytic information. An example of a coded ELAN file is provided in Figure 3. Details about the coding scheme can be found in Appendix A.

Figure 3. Example of a coded ELAN file.

During coding, I encountered several challenges. I will briefly discuss the two most frequent problems. Like Spanish and some other spoken languages, the means of negative interjection and the clause negator are formally identical in KK. Furthermore, KK relies heavily on shared knowledge and context, which makes the omission of sentential constituents a very

1 The software can be downloaded for free from http://tla.mpi.nl/tools/tla-tools/elan/.

common pragmatic strategy. For this reason, elliptic standard negation and negative interjections are not always unambiguously distinguishable. I coded conservatively by excluding instances with a subtle articulatory break or changes within the accompanying non- manuals, as these features indicate separate prosodic domains (Sandler 1999). Furthermore, word classes are not always easily assigned in signed languages. For example, the KK sign glossed as MONEY can be used as a noun or a verb, and sometimes both interpretations are possible in the same context. Lexical signs in KK are semantically broader than in most urban sign languages (Schwager & Zeshan 2008). Thus, at times it is not straightforward whether the negative element operates on a declarative, as in ‘I did not buy this/anything’ or a negative existential as in ‘I did not have any money’. The differentiation between verbal or nominal predicates is therefore not always clear-cut. Every instance that clearly did not involve a negative existential meaning was coded as standard negation. Instances of negator- copying are noted as such in the comment tier, but counted as a single negative sentence. Similarly, immediate repetitions of the same negative utterance were counted as a single instance of negation. 10% of the data (11 minutes) was re-coded by the same researcher to provide an intra- coder reliability measure. This ensures the validity of the coding and thus, the findings of the present study. I calculated Cohen’s Kappa, using the psych package (Gamer et al. 2012) in R (R Core Team 2016). Cohen’s Kappa determines the intra-coder agreement, i.e. the stability of the coding, as κ-values between 0 and 1, 1 indicating the perfect agreement (Cohen 1960; Fleiss et al. 2003). Calculating Cohen’s Kappa on each rating from the re-coded subset yielded a substantial intra-coder agreement between both rounds of coding (κ = 0.951; z = 15.6; p < 0.05; Fleiss et al. 2003).2

2 The overall κ comprises ratings for the identification of an instance of negation, identification of negation type, the identified manual and non-manual markers per instance, the clausal position, and the presence of headshake spreading. As for individual κ values per factor, the occurrence of negation (κ = 0.958; z = 7.04; p < 0.05), the combinatorial negation pattern (κ = .0825; z = 6.06; p < 0.05), the syntactic position of the manual negator (κ = 0.906; z = 6.66; p < 0.05) and headshake spreading (κ = 0.941; z = 6.93; p < 0.05) show very good intra-rater agreement. The type of negation yielded a lower value (κ = 0.713; z = 5.26; p < 0.05). This is, however, still considered as a good agreement (Altman 1991; Fleiss et al. 2003). The lower rating for negation type relates to the aforementioned difficulties in distinguishing between specific types (e.g. standard negation vs. negative existential).

4.3 Procedure I coded all the data in three rounds, following the coding scheme in Appendix A. An initial round targeted only the negation tier, marking all instances of negation and the signer who produced the negative utterance. Afterwards, I added coding of the remaining tiers, before reviewing the coding in a final round. In addition, the data were systematically checked for missed tokens by searching for NEG, headshakes and tongue protrusion. I used the ELAN function Multiple File Export with negation as a reference tier to extract all data automatically into Excel. Some instances had to be excluded due to reasons such as (i) absence of a felicitous translation of the utterance, (ii) fuzziness of the recording, or (iii) bad light conditions in the video. Furthermore, as negative interjections, negative imperatives and interrogatives, existentials, modals and completives all fall outside the definition of standard negation (see Section 2), they were excluded from the analyses. Unlike Dahl (2010), I decided to include not only verbal predicates but also nominal and pronominal predicates due to the challenges of determining the word classes in KK. As such, the definition used resembles the criteria for basic clause negators in Zeshan (2004), namely the most neutral form to reverse the polarity. I wish to acknowledge the fact that this study does not claim to be exhaustive, but explicitly focuses on standard negation.

5 Results

An overview of results of standard negation is provided before discussing the attested manual and non-manual markers in turn, starting with the most frequent one: the manual negator. Aspectual and other negatives that were coded but lie beyond the focus of this study will be touched upon briefly in Section 5.5. Throughout Section 5, the information presented concerns the pattern found across the entire data set, i.e. generalizing over different generations of signers. Generation-specific observations are indicated accordingly.

5.1 Overview of Results The final data set yielded a total of 420 negative propositions, including standard negation (39%), negative imperatives (19%), negative existentials (21%), negative interjections (16%) and other negatives (5%). The distribution of the different types of negation across all three generations is illustrated in Figure 4.

Negation type per dyad 100% N = 11 N = 6 N = 5 90% N = 15 N = 18 N = 36 80% N = 31 70% N = 18 other 60% N = 59 N = 7 negative interjection 50% N = 5 negative existential 40% N = 14 negative imperative N = 48 30% N = 69 N = 45 standard negation 20%

Relative frequency of negation types frequency Relative 10% 0% Generation III Generation IV Generation V Figure 4. Distribution of negation types per generation.

Standard negation (N=162) occurred on average 1.6 times per minute. Rates within the individual dyad varied between 1.1/minute in generation III, 2.7/minute in generation IV and

1.8/minute in generation V. In the majority of tokens (86%; 139/162), standard negation involves the manual clause negator NEG – by itself or in combination with non-manual markers (Table 4). It represents the most frequent marker and is found in the clause-final position in most cases (75%; 104/139). 14% of the time (22/162), NEG is unaccompanied by headshake or tongue protrusion. Nevertheless, non-manual markers were also seen frequently. A negative headshake occurs nearly as frequently (80%; 130/162) as NEG, while

tongue protrusion is used notably less often (19%; 30/162). The negative headshake predominantly combines with the manual negator NEG (57%; 92/162). In addition, tongue protrusion is combined with NEG or NEG and headshake in 15% (25/162). Lastly, 14% (23/162) of the data are negated by means of non-manual markers alone while the manual negator NEG remains unused.

Table 4. Absolute frequency of manual and non-manual markers used in standard negation.3

manual particle non-manual element Total number of standard negation tongue NEG headshake protrusion 56 61 5 Generation III 69 (81.2%) (88.4%) (7.2%) 41 35 10 Generation IV 48 (85.4%) (72.9%) (20.8%) 42 34 15 Generation V 45 (93.3%) (75.6%) (33.3%) 139 130 30 Total 162 (85.8%) (80.2%) (18.5%)

5.2 Manual Marking

In line with Marsaja’s (2008) observation, NEG is the only manual negator attested for diverse kinds of negative meanings such as ‘not’, ‘never’, ‘nobody’, etc. Furthermore, the combinatorial patterns are striking (Figure 5). The pattern in example (144) is the most common one: the manual negator is accompanied by a headshake in 57% (92/162) of all tokens. Considerably less frequently, the manual negator NEG combines with tongue protrusion (tp) (4%; 7/162), illustrated in (15), or both non-manual markers (11%; 18/162). Note that example (15) contains two negative particles as a result of full repetition. A manual-only pattern is attested as well (16). In 14% of cases (22/162), NEG is unaccompanied by any negative non-manual markers.

3 Values do not sum up to 100% since manual elements may combine with one or even both non-manual markers listed here.

4 BI1 represents the gloss for a first-person-pointing with a flat-B-handshape.

(14) ___hs

BI1 IX‘locative’ IX‘locative’ COFFEE NEG ‘I don’t take my coffee over here.’ [GD3jan7 00:27:55.880] (15) ___tp ___tp SPRAY RAIN NEG RAIN NEG ‘No rain came after the pesticides had been sprayed.’ [SuJu16jan7 00:16:21.209]

(16) DRINK TEA NEG ‘She does not drink tea.’ [PiKe4jan7 00:16:14.000]

Although examples (14) and (16) are semantically and syntactically similar, only the former contains a headshake. A wide range of subjects such as food- or work-related issues, the weather, and different activities are discussed in the data set, and the topic of conversation does not appear to influence the distribution of the headshake. Thus, there seems to be no relationship between the presence or the absence of the headshake and specific semantic context or types of predicates.

100% N = 13 N = 7 N = 3 90% N = 6 80% N = 8 N = 8 N = 9 70% N = 5 N = 4 non-manual-only 60% N = 43 N = 4 N = 3 manual-only 50% N = 25 N = 24 NEG +headshake + tongue protrusion 40% NEG + tongue protrusion 30% NEG + headshake 20% Frequcny of cominations of markers of cominations markers Frequcny 10%

0% Generation III Generation IV Generation V Figure 5. Combinatorial patterns of relevant manual and non-manual markers shown per generation.

Figure 5 shows the distribution of all patterns attested in negation across the three generations. Combining NEG with one or both non-manual markers is the primary strategy in all signers. NEG associated with a headshake is the most prevalent strategy, accounting for half of the data in both generation III and generation IV and for almost two thirds in generation V. Usage of the manual negator unaccompanied by any non-manual marking (manual-only negation) is similar across dyads: 12% in generation III, 17% in generation IV and 13% in generation V.

Now we turn to the clausal position of the manual negator. As explained in Section 3.1, constituent order in KK is flexible, and determining the word order is aggravated by the fact that all kinds of constituents, i.e. subjects, objects, pronouns, and verbs, are frequently dropped. Therefore, it is more insightful to determine the syntactic position of NEG in terms of clause-final or non-final position, than its relation to the subject or the object. Immediate repetitions (N=13) were counted as a single negative utterance and thus they can be classified as clause-final or non-final. In the case of negator-copying, as in example (17), the position of the original negator is unclear. These instances were treated as a separate group. A total of 33 instances included either a full repetition of the negative utterance as shown in (15) or negator copying (N=20). Table 5 provides an overview of the positions of NEG found in the data set.

Table 5. Syntactic position of the manual negator NEG.

Generation III Generation IV Generation V Total 42 25 38 105 clause-final (75.0%) (61.0%) (90.5%) (75.5%) 3 9 2 14 non-final (5.4%) (22.0%) (4.8%) (10.1%) 11 7 2 20 negator-copying (19.6%) (17.1%) (4.8%) (14.4%) 139

NEG absent 13 7 3 23 (NMM only) Total 69 48 45 162

Overall, sentences in KK are rather short, some comprising a single and others just a few more signs. In 76% (105/139) of all negated clauses that contain NEG, the manual negator occupies a clause-final position. Typically, NEG negates the preceding predicate, as in (14),

(16) and (18). In a few cases, NEG occurs in initial or medial position (10%; 14/139). When the manual negator is placed in medial position, the clause-final element may function as reinforcement or specification. In (19a), the celestial pointing towards the sky indicates the lack of rainfall, while the locative pointing in (19b) refers back to the mother or her house. As such, the clause-final element represents a pragmatic strategy to stress information that has been provided already, or to make implicit information explicit. It is possible that the clause- final element in these examples is actually optional. There are, however, not enough data to

confirm this. Importantly, the pointing in examples (19a) and (19b) is different from the pointing in (14), where the locative pointing precedes NEG and provides new information. Note that the manual negator cliticizes with the preceding constituent in (19a).

(17) ___hs NEG CUT-GRASS NEG ‘He does not work.’ [GD3jan7 00:28:328.500] (18) ___hs WORK NEG ‘He does not work.’ [SuJu16jan7 00:03:28.310]

(19) a. PAST-TIME#ONE IX2 RAIN^NEG IX‘celestial’ ‘Yesterday, it did not rain.’ [PiKe4jan7 00:10:53.111] b. MOTHER NEG IX‘locative’ ‘He doesn’t live at his mother’s house any longer.’ [PiKe4jan7 00:14:55.400]

(20) NEG COOK | GO-AWAY ‘I didn’t cook, I went there (to the house).’ [PiKe4jan7 00:00:57.500]

The manual negator occurs in sentence-initial position in (20) and is followed by two verbal predicates ‘cook’ and ‘go’. Superficially, sentence-initial NEG may resemble negative interjections (‘No, …’). However, there are no manual pauses or non-manual indications of clause boundaries. Instead, the manual negator is often followed by a verbal, nominal or pronominal predicate with positive polarity. As negative interjections are prosodically distinguished elsewhere in the data, there is little reason to treat examples like in (20) as such. Despite the fact that the number of non-final instances overall is negligible, the data from generation III included the highest degree of variation.

5.3 Non-Manual Marking Across all data, the majority of examples (86%) include non-manual markers, specifically, a negative headshake and tongue protrusion. The high frequency of headshakes identifies it as the main non-manual negation marker, whereas the status of tongue protrusion is somewhat less clear. I discuss some evidence that tongue protrusion might be associated with non- existence in Section 5.5. The previous section (5.2) stated that the combination of NEG and headshake represents the most frequent pattern found in the data. Thus, I will first present

results of how those two elements are coordinated and to what extent the headshake spreads (Section 5.3.1) before providing details about the use of tongue protrusion (Section 5.3.2). Finally, in Section 5.3.3, I turn to regularities in the choice of the particular non-manual element.

5.3.1 Scope of the Headshake As explained in Section 5.2, the headshake is engaged in the expression of negation in 80% of the data. Specifically, in 68% of cases (110/162) both the manual negator as well as a headshake are involved. The simultaneous articulation of the headshake with manual signs allows for it to spread over sequences of signs. Nonetheless, in KK, the headshake typically does not spread but accompanies only NEG (74.6%; 82/110), as shown in (21a). In 26% of the tokens (28/110), the headshake has scope over the manual negator and an adjacent sign such as in (21b). The combination of headshake spreading with the verbal predicate GIVE in (21b) represents a frequent construction, accounting for 6/17 tokens (35%). Nevertheless, only in 16% of cases (17/110) we are dealing with real spreading. In six examples, the spreading can be considered a harmony phenomenon as a result of negator-copying or repetition of the entire negative clause; in two cases, a manual sign cliticizes with NEG, which makes it impossible to distinguish spreading from a lexical headshake bound to NEG; three cases include co-articulatory, thus phonetic, headshake spreading, e.g. a locative point includes a head movement that fuses with the headshake.

(21) a. ___hs MONEY NEG ‘He does not give me money.’ [SuJu16jan7 00:04:06.600] b. ______hs MONEY GIVE NEG ‘They were not given money.’ [SuJu16jan7 00:12:13.400]

In KK, the scope of the headshake is minimal. It rarely covers multiple adjacent signs. Generally, only the immediately neighboring sign is under the scope of the headshake.

Moreover, the onset of the headshake usually precedes the (clause-final) negator NEG. The fact that KK utterances are prototypically short and often consist of a single sign may explain this tendency. Even if greater spreading were possible, it is overshadowed by the shortness of the utterances in the present sample. From an articulatory point of view, the non-manual

element is able to precede the manual negator because the hands can produce signs only sequentially. Generally, the minimal scope over the preceding sign may be co-articulatory (phonetic). However, this is unlikely, since the timing and duration of headshake spreading is syntactically defined and aligns with manual signs. My data may therefore document the emergence of the spreading domain in KK negation.

5.3.2 Tongue Protrusion Tongue protrusion features in a total of 30/162 instances of negation (18.5%). It is attested as a combined form with NEG as in example (15), repeated here as (22); with NEG and headshake (N=17) as shown in (23); as sole negator (N=4), illustrated in (24); and co-articulated with a headshake (N=2) unaccompanied by the manual negator (25). As will be explained in Section 5.5, there is some evidence that tongue protrusion is related to non-existence. As with the examples provided in (22) through (25), 83% of the data including tongue protrusion may be interpreted as (examples of) negative existentials.

(22) ___tp ___tp SPRAY RAIN NEG RAIN NEG ‘No rain came after the pesticides had been sprayed.’ [SuJu16jan7 00:16:21.209] (23) ___tp+hs

SIGN-NAMEA SIGN-NAMEB NEG IX‘locative’ ‘A and B are not coming to the event over there.’ [GD3jan7 00:08:54.100] (24) ______tp ______tp

FISHING SIGN-NAMEA GOOD FISHING ‘A did not catch anything.’ [SuJu16jan7 00:13:03.780] (25) ______hs+tp BI1 DRINK IX‘locative’ ‘I hadn’t been drinking over there, but they have (he said).’ [PiKe4jan7 14:31.500]

The non-existential meaning in (22) is obvious: the absence of rainfall. Example (23) deals with the fact that two villagers are not coming, which entails that they will not be present during a specific social event. The villager in (24) returns home with empty hands: FISHING is negated with a protruded tongue to express that villager A did not catch any fish when going on a fishing trip with other villagers. Similarly, tongue protrusion is co-produced with a

negative headshake in (25) in order to express that a particular signer did not gather with peers to consume alcohol. Besides non-existence, it is also possible that tongue protrusion relates to another, previously identified function: negative evaluation (Lutzenberger in prep.). This interpretation is compatible with all examples provided above. The lack of rain after spreading the pesticides in example (22) prevents the chemicals from diffusing in the soil, and thus minimizes its effects. In (23), villagers may be expected to attend the particular meeting due to communal rules. The negative evaluation of (24) is straightforward: the lack of prey means that the family will not have food to eat. Lastly, the negative judgment in example (25) may arise from the incidences that happened as a consequence of drinking, since drinking itself is not necessarily always regarded negatively. Similarly to non-existence, testing whether tongue protrusion expresses negative evaluation reveals the following pattern: although 60% of all instances of tongue protrusion are compatible with an interpretation of negative evaluation, this is not necessarily the primary one. Essentially, both negative evaluation and non-existence are inferential, therefore more implicit than explicit. Nevertheless, the use of tongue protrusion as a pragmatic marker for negative evaluation and as a specified negative element may have co-evolved, and now co-exists in KK. Further support comes from the examples without any non-manual elements (manual-only) for which (inferential) negative evaluation is usually not available. A single instance in which the signer talks about the madness of another villager may be interpreted in this fashion. Similarly, a maximum of five instances of manual-only negation may be compatible with (inferential) non-existence. This suggests that tongue protrusion may be treated as a pragmatic marker of negative evaluation and potentially as a negation marker that is used for a specific type of negation, namely non-existence.

5.3.3 Choice of Non-Manual Marking The non-manual elements at the focus of this study are headshake and tongue protrusion, whereby the former occurs much more frequently. Both of the markers are most commonly co-articulated with NEG. However, they also occur independently, i.e. without an accompanying manual sign – be it in combination or on their own. Given that diverse combinatorial patterns are attested, the choice of non-manual marking could be driven by factors such as the semantic context or specific predicates. The examples (19a) and (15), repeated below as (26) and (27), as well as example (28) are situated within the same semantic domain, namely weather conditions. Whilst all three examples involve the manual

negator NEG, the choice of non-manual markers varies. Example (26) does not involve any non-manual element; tongue protrusion is used in (27) without an apparent impact or change in the negative interpretation; finally, NEG cliticized to RAIN in (28), and as a result, both signs are co-articulated with a negative headshake. Evidently, there is no obvious cue for when combined forms are preferred over manual-only negation, nor for which conditions promote the use of a specific non-manual element.

(26) PAST-TIME#ONE IX RAIN^NEG IX‘celestial’ ‘Yesterday, it did not rain.’ [PiKe4jan7 00:10:53.111] (27) ___tp ___tp SPRAY RAIN NEG RAIN NEG ‘No rain came after the pesticides had been sprayed.’ [SuJu16jan7 00:16:21.209] (28) ______hs RAIN^NEG ‘It does not rain.’ [GD3jan7 00:50:23.200]

Besides the combinations with NEG, a substantial amount of the data (14%; 23/162) represents purely non-manual negation, i.e. NEG remains unused. The distribution across the three generations looks as follows: thirteen examples occur in generation III, seven are attested in generation IV, and three in generation V. Use of an independent headshake is most common (82.6%; 19/23) and is illustrated in (29) and (30). Generally, the headshake is co- articulated with the clause-final constituent (29). Nevertheless, in some cases, the headshake and the preceding sign are not fully synchronized (30). In other words, the manual sign precedes the onset of the headshake or vice versa. As a result, the headshake starts only in the middle of the manual sign. This example highlights the challenge of analyzing maximally reduced instances of sentential negation, as is common in KK.

(29) ______hs BI1 TWICE-MARRIED ‘I won’t get married a second time.’ [GD3jan7 00:30:31.119] (30) ___hs RICE ‘I don’t bring any rice.’ [PiKe4jan7 00:13:41.000]

(31) ______tp ______tp

FISHING SIGN-NAMEA GOOD FISHING ‘A did not catch anything, did he?’ [SuJu16jan7 00:13:03.825] (32) _____hs MOTORBIKE ‘He does not like driving the motorbike.’ [GD3jan7 00:49:02.470]

In very few cases such as in (24), repeated here as (31), tongue protrusion can function as a negation marker on its own. Nevertheless, the examples are too few in number to draw definite conclusions about specific conditions for this function. In addition, four cases are attested in generation III-signers in which the headshake is produced completely independently of manual signs, i.e. it is articulated by itself between two manual signs, be it a consecutively produced negative particle or another sign (32). This has been reported for a few other sign languages (e.g. Hendriks 2008 for Jordanian Sign Language).

5.4 Statistical Trends Following the presentation of the mainly qualitative results above, I now add some remarks on the statistical significance of the results. I designed two linear mixed effect models using R (R Core Team 2016) and the lme4 package (Bates et al. 2015) to perform an analysis of the predictive power of generation on (i) the negation patterns and (ii) the scope of the headshake. According to Pfau (2015), sign languages start out as manual systems, go through a stage with a combined pattern, and result in a system that highly favors non-manual marking. The scope of the headshake leads towards non-manual dominance. In the present context, the youngest signers are predicted to show the highest degree of combination of NEG with non-manuals, as well as headshake spreading. The first model examines the involvement of the manual negator across all generations, and the second determines inter-generational divergences in headshake spreading. In both models, generation was entered as a predictor (fixed effect) and random intercepts were set by participant. The significance value was determined at 0.05. Contrasts were defined to compare the youngest generation (V) against generation III and generation IV, as well as the two older generations against each other. The implications of the results below will be discussed in Section 6. The first model concerns the involvement of the hand (dependent variable). The model tested whether young signers make use of the manual negator more frequently than signers

from older generations. It did not yield a significant result. In other words, this model provides no evidence that younger generations rely significantly more on the hand than older signers or that generation III and generation IV differ in their involvement of NEG. However, there seems to be a trend towards using more NEG over time (see Figure 6): the difference between generations III and IV compared to generation V (z = 0.115) is smaller than the difference between generations III and IV (z = 0.548). To sum up, the model suggests that signers progressively combine manual and non-manual elements, and, at the same time, reduce the use of independent non-manuals. A high degree of inter-signer variation may have obscured any potential cross-generational differences. Possibly, significant differences in negation patterns could surface when enlarging the sample size.

100% N = 13 N = 7 N = 3 90% N = 6 80% N = 8 N = 8 N = 36 70% N = 48 N = 33 60% non-manual-only 50% manual-only 40% NEG + NMM 30%

20%

Relative frequency of negation patterns frequency Relative 10%

0% Generation III Generation IV Generation V Figure 6. Visualization of the involvement of the hand (model 1) across three generations.

In the second statistical analysis, I calculated whether signers from different generations vary significantly when it comes to spreading of the headshake A second linear mixed effect model tested whether generation predicts the use of headshake spreading (dependent variable). This model revealed a significant effect for generation V when compared to generations III and IV (z < 0.05). Thus, the headshake spreads significantly more often in the youngest signers than the older signers. This is illustrated in Figure 7.

*** 100% n.s. 90% 80% N = 41 N = 25 eading

r 70% 60% 50% no spreading N = 16 equncy of sp spreading

r 40%

30% N = 11 20% Relative f 10% N = 4 N = 2 0% Generation III Generation IV Generation V Figure 7. Distribution of headshake spreading across the three generations (model 2).

Given that each signer were defined as random intercepts, the inter-signer variability in the data seems to be negligible. This suggests that the observed differences between signers from different generations are unlikely to be caused by idiosyncratic variation. This is corroborated by the relative frequency of scope that is considerably higher for both generation V-signers than for older signers. In addition, I checked for potential effects of gender in both models by defining gender as fixed effect. In both cases, gender does not seem to influence the use of the manual negator and headshake spreading, i.e. gender did not yield a significant effect.

5.5 Additional Observations For the sake of completeness, I will briefly illustrate the most common negative forms that go beyond standard negation, namely negative existentials, negative imperatives, and negative interjections. Table 6 provides an overview of the distribution in the data set.

Table 6. Overview of the distribution of common negation types. standard negative negative negative other Total negation existential imperative interjection 162 89 78 69 22 420

38.6 % 21.2 % 18.6 % 16.4 % 5.2 %

Besides standard negation, negative existentials are common in the data (21%; 89/420). As explained in Section 4, it is not always unambiguous whether an utterance is an example of standard negation or negative existential, since NEG is used in both contexts. Many examples of negative existentials deal with the fact that the villagers do not have food or money, e.g.

(33). In line with Perniss & Zeshan (2008), NEG and FINISH were attested in such contexts, yet

NEG is clearly more frequent. Notably, these authors do not mention tongue protrusion, a marker that appears crucial in this data set due to its high frequency and its ability to function as a sole negation marker (34).

(33) ___tp FOOD NEG HUNGRY ‘There is no food. We will stay hungry.’ [GD3jan7 00:18:43.000] (34) ____tp FOUR ‘None of the four (villagers) had any chickens.’ [SuJu16jan7 00:06:10.400]

KK signers frequently express prohibition by using negative imperatives. Most commonly, the structure follows “X. Don’t do that.” This structure is exemplified in (35) with the verbal predicate HIT and including negator-copying. An interesting observation in this context is that examples may involve tongue protrusion or may be followed by the lexical sign BAD, both of which can express negative evaluation at an individual as well as a normative (societal) level.

Most commonly attested are NEG co-articulated with a headshake, and some additional non- manuals such as backwards body leans and a facial expressions of discontentment, similar to the ‘not face’ (Benitez-Quiros et al. 2016).

(35) ___open mouth PREGNANT NEG HIT NEG ‘Don’t hit the pregnant woman.’ (lit. ‘The pregnant woman, don’t you hit her!’) [GD3jan7 00:22:53.000]

Finally, negative interjections are marked by a prosodic break or by a sentence-initial NEG that does not have scope over the subsequent affirmative clause (36). Negative interjections are generally expressed by NEG combining with a headshake, or by an independent headshake. Moreover, signers often produce pairings of a (rhetorical) question followed by a negative answer (37), or a negative-affirmative construction within the same signer. Tongue protrusion features only rarely in negative interjections. In these cases, it functions as marker of negative evaluation or is used lexically, e.g. with BAD.

(36) NEG | TEN-THOUSAND-RUPIAH ‘No, I got just 10,0000IDR.’ [PiKe4jan7 00:01:38.000] (37) ______brow raise ___hs TIME NINE | NEG TIME-EIGHT ‘At nine o’clock? No, at eight.’ [PiKe4jan7 00:04:19.000]

To conclude, NEG is used in all three types of negative constructions (negative existentials, negative imperatives, negative interjections), stressing its status as the default negation marker. Despite some variation, tongue protrusion is highly prevalent in negative existentials and often functions as a marker of negative evaluation in negative imperatives and interjections. Tongue protrusion thus may have developed into a specified negation marker for non-existence. This is supported by cross-linguistic findings: Palfreyman (2015) observes idiolectal uses of tongue protrusion for negative existentials in an Indonesian signing variety used in Central Java (Solo), and tongue protrusion is reported to accompany the sign

NOTHING in New Zealand Sign Language and Flemish Sign Language (Zeshan 2004).

6 Discussion

The results presented in Section 5 are interpreted in the following section. First, I place the findings from KK in a typological context (Section 6.1). For this purpose, I generalize across different generations of signers. Then, I discuss the findings from the perspective of diachronic language change by drawing comparisons between generations (Section 6.2).

6.1 KK in Typological Perspective This study has shown that standard negation is most commonly expressed by combining the negative particle NEG with a negative headshake. The headshake does not usually spread onto adjacent signs. Generally, the negative particle occurs in clause-final position. Nevertheless, there is a substantial amount of data where the manual and the non-manual marker are not co- articulated. Specifically, 14% manual-only negation and a further 14% non-manual-only negation were attested. Before evaluating the research questions stated in Section 3.3, it is worth restating that manual dominant and non-manual dominant sign languages are differentiated by (i) the presence of an obligatory manual marker and (ii) the scope of the headshake (Zeshan 2004, 2006). Manual dominant systems are characterized by the fact that a negative particle must be used at all times and that the headshake does not spread beyond this particle. In contrast, the manual negator is optional and the headshake may have scope over adjacent signs in non- manual dominant systems. Interestingly, the pattern attested for KK negation is compatible with both systems to some extent.

In line with Marsaja (2008), NEG represents the most frequent marker, present in 86% of the data. Thus, a large part of KK negation conforms to a manual dominant system. Similarly, the potential for headshake spreading is typical for non-manual dominant sign languages. In KK negation, headshake spreading is uncommon and if present, severely restricted. This further supports the classification of KK negation as a manual dominant pattern. Despite the optionality of the manual negator in non-manual dominant sign languages, one must not forget that it is still commonly used (e.g. Oomen & Pfau 2017 for

NGT): in this data set in 86% of cases. Furthermore, NEG is absent in 14% of all negation tokens. Hence, the present results are not incompatible with a non-manual dominant pattern either. In both classificatory scenarios, 14% of non-manual-only negation would be dismissed as exceptions when classifying KK negation as manual dominant system, and 14% of manual-only negation when categorizing it as non-manual dominant system. It is by no

means valid to disregard a substantial amount of data. The fundamental question arises as to whether it is legitimate to omit an obligatory element comprising a sizeable amount of data. If so, how much divergence is justifiable to still classify as a manual dominant or indeed as a non-manual dominant system? The established classification in manual dominant and non- manual dominant negation seems to entail a continuum. Notably, KK integrates characteristics from both extremes. Although the results from this study indicate that KK patterns somewhat more alike manual dominant languages, KK negation is certainly neither a manual-only nor as strictly manual dominant as reported previously (Marsaja 2008). It is even questionable whether KK negation can be classified as a manual dominant system at all. Thus, the contribution to sign language typology made by this study is not just in adding a sign language of the lesser-studied type to the picture, but in scrutinizing established classifications. Specifically, it raises the question of whether there is a threshold for allocating a specific sign language into one of the established negation types. As explained in Section 2.2, there exists considerable cross-linguistic variation within the two major types of sign language negation. Table 7 provides an overview of negation patterns across seven different sign languages, taking into account the classificatory difficulties for KK negation explained above. The KK pattern is both in line and in disagreement with cross-linguistic tendencies.

Table 7. Typological comparison of negation patterns across seven sign languages (adapted from Oomen & Pfau 2017; KK added). DGS LSC ASL LIS TID NGT KK country of usage Germany Catalonia USA Italy Turkey Netherlands Bali constituent order SOV SOV SVO SOV SOV SOV flexible manual dominant? - - - + + - (+)

NEG clause-final? + + +/- + + +/- + hs only on NEG? - + + + + ? + hs only on predicate (if NEG is absent) + + - - - + + hs spread onto object? + + + - - + +/- hs spread onto subject? - - +/- - - - +/-

Negative Concord - + + - + + ?

In Table 7, KK is the only sign language that is not used in an urban setting. SOV is the most common constituent order for all the sign languages in Table 7, except KK and ASL. While for ASL, SVO had been suggested, no robust constituent order had been established for KK. Like other sign languages, KK shows a great deal of flexibility in terms of the syntactic position and the omission of constituents (Sandler & Lillo-Martin 2006). Possibly due to the high context-dependency of this language, eliding constituents is even more common than in other signed languages. The results from this study remain inconclusive as to whether the manual or the non-manual element is obligatory. For the reasons explained above, I wish to leave it open as to whether the distributional patterns are sufficient to categorize KK negation, as indicated by the brackets in Table 7. Importantly, KK integrates characteristics from both a manual dominant and a non-manual dominant system. The co-existence of both systems might indicate a transitional phase. Regardless of which marker is obligatory, negative particles occur predominantly in clause-final position across all sign languages in Table 7, including KK. Nonetheless, KK exhibits a language-specific pattern when it comes to the use of the headshake: although headshake spreading is uncommon and if present, has scope over maximally one adjacent sign, the kind of constituents is much more flexible than attested in other sign languages. Only in KK can the headshake accompany the negative particle, spread onto the subject or the object when NEG is present and over the (verbal and non-verbal) predicate when NEG is absent. This observation may be explained by the fact that many utterances comprise very few signs, irrespective of the type of constituent. Thus, it is possible that headshake spreading is associated with the clause-final position rather than a specific word class. To conclude, KK negation conforms to and diverges from cross-linguistic tendencies.

The most frequent pattern in KK (combination of NEG and headshake, no spreading) is characteristic of a manual dominant system. Still, this would also be compatible with a non- manual dominant pattern (see e.g. LSC in Table 7). It is possible that KK negation is currently in a transitional stage where different systems co-exist. In any case, KK negation points out the gradient nature of established classifications, which raises the question of a classificatory threshold.

6.2 Aspects of Grammaticalization The sampling method used in this study provides us with intergenerational data: the three dyads represent three different generations of signers. Thus, it is possible to discuss the

attested patterns from a perspective of diachronic language change across three generations of signers. Five key observations can be made:

(i) The combination of NEG and headshake is the most common strategy across all three generations. Frequency fluctuations are minimal. (ii) The engagement of the manual particle increases slightly over the three generations. (iii) Signers from the youngest generation make use of a greater range of combinatorial

variants of NEG, headshake or/and tongue protrusion. (iv) The use of independent non-manual markers slowly decreases over time. (v) The frequency of headshake spreading increases significantly in generation V. Although the effect of generation on different negation patterns did not reach statistical significance, the tendency points towards more combined forms and greater presence of the manual negator. Given that headshake spreading was significantly more frequent in generation V-signers than in older signers, we may interpret this as evidence of language change. Nonetheless, there are at least three alternative explanations: (i) headshake spreading is characteristic of a sociolinguistic youth-variant. Potentially, this feature is used only at a certain age while it is abandoned again when growing older (Labov 1965, 1994; Sankoff 2006); (ii) headshake spreading in generation V may be caused by a single lexical item: the verb GIVE precedes NEG in six out of eleven instances of headshake spreading in generation

V, while GIVE accompanied by headshake is not attested in any other generation. Another statistical model could detect possible interactions of signer and predicate; (iii) in spite of the significant result in the linear mixed effects model, the possibility of idiosyncratic inter- signer variation cannot be excluded. The number of examples including headshake spreading per signer is too small to ultimately eliminate the possibility of spreading as an idiolectal variant. Therefore, increasing the sample size could corroborate the effect. Lastly, one may hypothesize that the emerging headshake spreading in generation V represents a language contact phenomenon. This is, however, unlikely, since the headshake plays a minor role in negation in signing varieties across Indonesia (Palfreyman 2015), including the variety used in Singaraja, a nearby city (p.c. with Nick Palfreyman). I now offer some speculations on possible diachronic scenarios in the expression of negation. Pfau (2015) proposes that sign language negation conforms to the key principles of the Jespersen’s Cycle, found for example in French. According to this theory, negative elements are reinforced through the use of a second negative particle and then weakened again by losing one of the two (Jespersen 1917). Pfau (2015) hypothesizes that sign

languages emerge as manual dominant systems, and associate a grammaticalized manual gesture with a negative headshake in a second step. Afterwards, the headshake detaches from the manual element, thus becoming an independent marker of negation that is capable of spreading, and may subsequently take over the status of the obligatory element. Ultimately, this results in a non-manual dominant pattern. Due to the shared modality of gesture and sign, the grammaticalization of manual gestures is common in sign languages (van Loon et al. 2014; Pfau & Steinbach 2011), and especially prominent in negation (Zeshan 2004). For KK, it is likely that NEG originates in a manual gesture used by (hearing) members within the community (Marsaja 2008). Nevertheless, non-manual elements function as fundamental elements in signed languages (Sandler & Lillo-Martin 2006; Pfau & Quer 2010). Thus, the visuo-gestural modality also favors the integration of non-manual gestures into sign language grammar (Pfau & Steinbach 2011). However, what motivates the assumption that manual gestures precede non-manual ones? Pfau (2015) argues with the aid of linguistic typology: cross-linguistically, the existence of particles is universal. Given that manual negators are used as particles, it seems likely that they arise before non-manual markers do. Nevertheless, this proposal is not grounded in empirical evidence. In the present data set, both manual and non-manual elements are used independently and in combination across signers from all generations. Thus, the KK data do not support Pfau’s (2015) hypothesis. Withal, it seems plausible that manual (handwave) and non-manual (headshake) gestures entered the linguistic system of KK around the same time after the language’s emergence. In this scenario, KK initially made use of a range of diverse variants based on manual

(NEG) and non-manual (headshake, tongue protrusion) elements all of which originated in culture-specific gestures (Fridlund 1994; Kendon 2002; Kettner & Carpendale 2013; Marsaja 2008; Meltzoff & Moore 1977, 1989; Pfau 2015; Rozin & Fallon 1987; Spitz 1957). Later, signers start to converge on different, yet functionally redundant markers. Sign languages strive for simultaneity where possible in order to increase language efficiency (Pfau 2015). The KK data set endorses this: the use of independent non-manuals decreases alongside an increase in combinations. In favor of enhancing language efficiency, and to reduce redundancy, individual markers begin to specialize, as in the case of tongue protrusion, which develops a negative existential meaning. The fact that non-manual-only negation decreases whilst the proportion of manual-only negation remains stable might indicate that KK has reached the stage of a manual dominant system by generation V. Although the increase in headshake spreading in generation V may even delineate a first step to free the negative

particle, a reduction in the use of the manual-only pattern would be expected if KK were to move towards a non-manual dominant system. One can envisage at least three scenarios for future generations: (i) Headshake spreading was a side effect of losing the non-manual-only pattern. As a result, the scope reduces to a single sign, namely the manual negator, which functions as the obligatory marker. In other words, the manual dominant system stabilizes. (ii) Headshake spreading remains, and becomes more systematic and productive. At the same time, non-manual-only negation steadily decreases until it has disappeared completely and the manual negator stabilizes as an obligatory element. This leads to a hitherto unattested typological type: manual dominant, plus spreading. (iii) The systematicity of headshake spreading increases and manual negators are progressively dropped. As a result, the headshake becomes obligatory and the system moves towards a non-manual dominant one. The different scenarios are visualized in Figure 8.

Figure 8. Sketch of possible grammaticalization scenarios of KK negation.

Although the findings from this study do not necessarily suggest a manual dominant origin for KK negation, it is possible that this pattern precedes the analyzed data. Thus, a manual dominant system may have been characteristic of the language use of signers from generations I and II. In that case, however, it remains unclear what motivated the use of independent non-manual markers and why this is considerably more frequent in generation III than in younger signers.

7 Concluding Remarks

In an effort to elucidate standard negation in KK, I conducted a study of 1.73 hours of naturalistic data drawn from the KK Corpus. Adopting an intergenerational approach, I have shown that negation in KK is typologically unusual, but in a different way than previously suggested (Marsaja 2008). In line with the predictions stated in Section 3.3, the manual negator is used very often, but a negative headshake and tongue protrusion play an important role as well. Specifically, the headshake occurs as independent negator, whilst tongue protrusion seems to mark non-existence. The corpus data from this study clearly do not support the tentative claim that KK exhibits negation according to a manual-only type. To the contrary, KK negation includes characteristics of both, a manual dominant and a non-manual dominant pattern (Marsaja 2008). As such, the KK pattern does not entirely fit the existing classifications. This may be due to theoretical reasons, e.g. thresholds defining the established classifications, or a result of diachronic language change across three generations. I wish to emphasize that the data in this study represent a small sample that may not be representative of the entire population. Therefore, I encourage the reader to treat the discussion about diachronic language change in KK negation with caution. Hopefully, a future study can extend the data set and corroborate the speculative proposals about different pathways that were advanced here. Furthermore, comparing this data set with real-time recordings of the same pairings of signers would allow us to evaluate ongoing language change more accurately and minimize the possible effects of age-grading (Labov 1965, 1994; Sankoff 2006). Ideally, data of the oldest and youngest generation of signers would be included in the analyses as well. Nevertheless, the extreme age difference between signers from generation II and generation VI may blur the findings. Importantly, this study does not claim to be exhaustive in accounting for all possible variables. Time depth has been discussed as a main factor to affect language change. Nevertheless, social or geographic circumstances such as kinship relations certainly have a major impact on language change as well. Especially in communities like the one KK is used in, this should be taken into consideration in future studies. Lastly, the study of the relationship between standard negation and other negative forms, e.g. negative interjections, imperatives, existentials, incompletives and completives, entails a promising contribution to the study of diachronic language change. In particular, the use of specific non-manual elements is of interest, as the glimpse of the data provided in Section 5.5 may indicate a grammaticalization continuum for all negative non- manual markers used in negative contexts.

References

Altman, D. G. (1991). Practical statistics for medical research. London: Chapman & Hall. Antzakas, K. (2006). The use of negative head movements in Greek Sign Language. In U. Zeshan (ed.). Interrogative and negative constructions in sign languages (pp. 258- 270). Nijmegen: Ishara Press. Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67(1), 1-48. doi:10.18637/jss.v067.i01. Benitez-Quiroz, C. F., Wilbur, R. B., & Martinez, A. M. (2016). The not face: A grammaticalization of facial expressions of emotion. Cognition 150, 77-84. http://doi.org/10.1016/j.cognition.2016.02.004. Biberauer, T., & Zeijlstra, H. (2012). Negative Concord in Afrikaans: filling a typological gap. Journal of Semantics 29, 345-371. doi:10.1093/jos/ffr010. Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and psychological measures. 20, 37-46. Corblin, F., Déprez, V., de Swart, H., & Tovena, L. (2004). Negative concord. In F. Corblin & H. de Swart (eds.), Handbook of French semantics (pp. 417-455). Stanford: CSLI Publications. Crasborn, O., & Sloetjes, H., (2008). Enhanced ELAN functionality for sign language corpora. Proceedings of the 3rd Workshop on the Representation and Processing of Sign Languages: Construction and Exploitation of Sign Language Corpora, 39-43. Dahl, Ö. (2010). Typology of negation. In L. R. Horn (ed.), The expression of negation. (pp. 9-38). Berlin: Mouton de Gruyter. de Swart, H. (2010). Expression and interpretation of negation: An OT typology. Studies in natural language and linguistic theory, Volume 77. Dordrecht: Springer. de Vos, C. (2012). Sign-spatiality in Kata Kolok: How a village sign language of Bali inscribes its signing space. PhD dissertation. Nijmegen: Max Planck Institute for Psycholinguistics. de Vos, C. (2016). Sampling shared sign languages. Sign Language Studies 16(2), 204-226. doi:10.1353/sls.2016.0002. Fleiss, J., Levin, B., & Paik, M. (2003). Statistical methods for rates and proportions. 3rd Edition. New York: Wiley & Sons. Fridlund, A. (1994). Human facial expression. An evolutionary view. London: Academic Press. Friedman, T. B., Hinnant, J. T., Fridell, R. A., Wilcox, E. R., Raphael, Y., & Camper, S. A. (2000). DFNB3 Families and shaker-2 mice: Mutations in an unconventional myosin, myo 15. Advances in Oto-Rhino-Laryngology 56, 131-144.

Gamer, M., Lemon, J., & Singh, I. F. P. (2012). irr: Various coefficients of inter-rater reliability and agreement. R package version 0.84. https://CRAN.R- project.org/package=irr. Genetti, C. (2007). A grammar of Dolakha Newar. Berlin: Mouton de Gruyter. Geraci, C. (2005). Negation in LIS (Italian Sign Language). In L. Bateman & U. Cherlon (eds.), Proceedings of the North East Linguistic Society (NELS 35). Amherst, MA: GLSA. Hendriks, B. (2008). Jordanian Sign Language: Aspects of grammar from a cross-linguistic perspective. PhD dissertation. Utrecht: LOT. Jespersen, O. (1917). Negation in English and other languages. Copenhagen: Høst. Kendon, A. (2002). Some uses of the headshake. Gesture 2(2), 147-182. doi: 10.1075/gest.2.2.03ken. Kettner, V. A., & Carpendale, J. I. M. (2013). Developing gestures for No and Yes: Head shaking and nodding in infancy. Gesture 13(2), 193-209. doi:10.1075/gest.13.2.04ket. König, A. (2011). The Language Archive. Retrieved 23 June 2017, from https://tla.mpi.nl/. Labov, W. (1965). On the mechanism of linguistic change. In C. W. Kreidler (ed.), Report of the sixteenth Annual Round Table Meeting on linguistics and language studies. (pp. 91-114). Washington: Georgetown University Press. Labov, W. (1994). Principles of linguistic change. Volume 1: Internal factors. Oxford: Blackwell. van Loon, E., Pfau, R., & Steinbach, M. (2014). The grammaticalization of gestures in sign languages. In C. Müller, A. Cienki, E. Fricke, S. H. Ladewig, D. McNeill, & S. Tessendorf (eds.), Body – Language – Communication: An international handbook on multimodality in human interaction (pp. 2133-2149). Berlin: Mouton de Gruyter. Lutzenberger, H. (in prep.). The polyfunctionality of tongue protrusion in Kata Kolok (KK). Marsaja, G. I. (2008). Desa kolok: A deaf village and its sign language in Bali, Indonesia. Nijmegen: Ishara Press. Mayer, M. (1969). Frog, where are you? New York: Dial Books for Young Readers. Meltzoff, A. N., & Moore, M. K. (1977). Imitation of facial and manual gestures by human neonates. Science 198(4312), 74-78. Meltzoff, A. N., & Moore, M. K. (1989). Imitation in newborn infants: Exploring the range of gestures imitated and the underlying mechanisms. Developmental Psychology 25, 954-962. doi:10.1037/0012-1649.25.6.954 Miestamo, M. (2005). Standard negation. The negation of declarative verbal main clauses in a typological perspective. Berlin: Mouton de Gruyter. Miestamo, M. (2007). Negation – An overview of typological research. Language and Linguistics Compass 1(5), 552-570.

Miestamo, M. (2013). Subtypes of asymmetric standard negation. In M. S. Dryer & M. Haspelmath (eds.), The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chapter/114, Retrieved 12 April 2017). Neidle, C., Kegl, J., MacLaughlin, D., Bahan, B., & Lee, R. (2000). The syntax of American Sign Language. Functional categories and hierarchical structure. Cambridge: The MIT Press. Nunes, J., & de Quadros, R. M., (2008). Phonetically realized traces in American Sign Language and Brazilian Sign Language. In J. Quer (ed.), Signs of the time: Selected papers from TISLR 2004 (pp. 177-190). Hamburg: Signum Press.

Oomen, M., & Pfau, R. (2017). Signing NOT (or not): A typological perspective on standard negation in Sign Language of the Netherlands. Linguistic Typology 21(1), 1-51. Palfreyman, N. B. (2015) Sign language varieties of Indonesia: A linguistic and sociolinguistic investigation. Unpublished PhD dissertation. University of Central Lancashire. Payne, J. R. (1985). Negation. In T. Shopen (ed.), Language typology and syntactic description, Volume I – Clause structure (pp. 197-242). Cambridge: Cambridge University Press. Penka, D., & Zeijlstra, H. (2010). Negation and polarity: an introduction. Natural Language & Linguistic Theory 28(4), 771-786. Perniss, P., & Zeshan, U. (2008). Possessive and existential constructions in Kata Kolok. In P. Perniss & U. Zeshan (eds.), Possessive and existential constructions in sign languages (Sign Language Typology Series No. 2) (pp. 127-149). Nijmegen: Ishara Press. Pfau, R. (2008). The grammar of headshake: A typological perspective on German Sign Language negation. Linguistics in Amsterdam 1, 37-74. Pfau, R. (2015). The grammaticalization of headshakes: From head movement to negative head. In A. D. M. Smith, G. Trousdale & R. Waltereit (eds.), New directions in grammaticalization research (pp. 9-50). Amsterdam: John Benjamins. Pfau, R. (2016). A featural approach to sign language negation. In P. Larrivée & C. Lee (eds.), Negation and negative polarity. Cognitive and experimental perspectives (pp. 45-74). Dordrecht: Springer. Pfau, R., & Quer, J. (2007). On the syntax of negation and modals in German Sign Language (DGS) and Catalan Sign Language (LSC). In P. Perniss, R. Pfau & M. Steinbach (eds.), Visible variation: Comparative studies on sign language structure (pp. 129- 161). Berlin: Mouton de Gruyter. Pfau, R., & Quer, J. (2010). Nonmanuals: their prosodic and grammatical roles. In D. Brentari (ed.), Sign languages (pp. 381-402). Cambridge: Cambridge University Press.

Pfau, R., & Steinbach, M. (2011). Grammaticalization in sign languages. In H. Narrog & B. Heine (eds.), The Oxford handbook of grammaticalization (pp. 1-15). Oxford: Oxford University Press. R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. Rozin, P., & Fallon, A. E. (1987). A perspective on disgust. Psychological Review 94(1), 23- 41. doi:10.1037/0033-295X.94.1.23. Sandler, W. (1999). The medium and the message: Prosodic interpretation of linguistic content in Israeli Sign Language. Sign Language & Linguistics 2(2), 187-215. Sandler, W., & Lillo-Martin, D. (2006). Sign Language and Linguistic Universals. Cambridge: Cambridge University Press. Sankoff, G. (2006). Age: Apparent time and real time. In K. Brown (ed.), Encyclopedia of language and linguistics. Volume 1. (2nd ed.). (pp. 110-116). Cambridge: Elsevier. Schwager, W., & Zeshan, U. (2008). Word classes in sign languages: Criteria and classifications. Studies in Language 32, 509-545. doi:10.1075/sl.32.3.03sch. Spitz, R. L. A. (1957). No and Yes: On the genesis of human communication. New York: International University Press. Tang, G. (2006). Questions and negation in Hong Kong Sign Language. In U. Zeshan (ed.) Interrogative and negative constructions in sign languages (Sign Language Typology Series No. 1) (pp. 198-224). Nijmegen: Ishara Press. Winata, S., Arhya, I. N., Moeljopawiro, S., Hinnant, J. T., Liang, Y., Friedman, T. B., & Asher, J. H. (1995). Congenital non-syndromal autosomal recessive deafness in Bengkala, an isolated Balinese village. Journal of Medical Genetics 32, 336-343. Wittenburg, P., Brugman, H., Russel, A., Klassman, A., & Sloetjes, H. (2006). ELAN: a professional framework for multimodality research. Proceedings of the 5th International Conference on Language Resources and Evaluation (LREC 2006), 1556-1559. Yang, J. H., & Fischer, S. D. (2002): Expressing negation in Chinese Sign Language. Sign language and Linguistics 5(2), 167-202. Zeshan, U. (2004). Hand, head and face – Negative constructions in sign languages. Linguistic Typology 8 (1), 1-58. Zeshan, U. (2006). Negative and interrogative constructions in sign languages: A case study in sign language typology. In U. Zeshan (ed.) Interrogative and negative constructions in sign languages (Sign Language Typology Series No. 1) (pp. 28-68). Nijmegen: Ishara Press.

Appendix A

Coding Scheme tier name function Code manual glosses (dom./non-dom) non-manual tongue protrusion non-manual_head headshake translation translation into English negation marks instance of negation indicates the signer negation category specifies the negation category refusal, non-existence, denial constituent order position of the negator in relation to predicate, subject and object combined form manual negator co-articulated with 0 (absent) non-manual elements 1 (present) pt function function of tongue protrusion 0 (absent), gestural, lexical, death marker, negative evaluation, negation negation type specifies the type of negation standard negation negative imperative negative completive negative modal negative interjection negative existential negative contrast NEG presence presence of manual negator 0 (absent) 1 (present) hs presence presence of headshake 0 (absent) 1 (present) pt presence presence of tongue protrusion 0 (absent) 1 (present) hs spreading presence of headshake spreading 0 (absent) 1 (present) comment additional comment