The negative existential cycle in Turkish Sign Language (TİD)

1 Introduction

This study examines the relevance of the types and stages of Croft’s (1991) negative existential cycle (NEC) to Turkish Sign Language (TİD), a sign language that displays cross-modal variations in linguistic use and structure, namely language modality, in comparison to spoken languages. As suggested by the term “NEC”, numerous scholars, including Croft (1991), have put forth the proposition that the NEC is the consequence of language-internal mechanisms of negation renewal and that the input verbs in the cycle are existential verbs. Croft’s assertion is that the operation of the cycle is evident in three attested language types. Type A involves the negation of the existential predicate, which is carried out by the verbal negator. Type B features a special negative existential predicate that is unique and different from the verbal negator. In type C, there is a special negative existential predicate that is identical to the verbal negator (Croft 1991: 6).

TİD employs the verb var, meaning ‘to have’, as an existential predicate.^[1] In terms of its configuration, var ‘exist’ is produced with a 5-handshape , with the thumb pointing towards the chest of the signer and repeated movement. Its negative counterpart yok ‘not.exist’, produced with an open-B-handshape , is entirely distinct from the positive and is not derived from the positive form (see Figure 1).^[2]

Figure 1:

The existential marker var and its negative counterpart yok in TİD.

As shown in (2), the standard negator deĞİl ‘not’ cannot be used to negate an existential construction. In terms of its configuration, deĞİl consists of a single upward movement of both hands (or only the dominant hand from the wrist), mostly accompanied by a backward head tilt that occurs simultaneously (see Figure 2).

Figure 2:

The manual negator deĞİl ‘not’ in TİD (Dikyuva et al. 2017: 220).

In such instances, yok is the sole negative existential marker, as demonstrated in (3), where the existential predicate var is obligatorily omitted.

The ability of yok to express negative existence is evidence of its status as a unique form of the negative existential. In her crosslinguistic study, Veselinova (2013) defines prototypical negative existentials as constructions that express the absence or nonexistence of an entity in a given domain – typically in a specific location or within a particular time frame. These constructions are formally distinct from standard sentential negation and are not used to negate propositions; rather, they are employed to assert the non-presence of a referent. Prototypical negative existentials frequently manifest in verbless clauses or involve specific lexical items dedicated to existential negation. Adopting the definition put forth by Veselinova (2013: 118–139) allows us to conclude that the negative existential marker yok in TİD behaves in a manner analogous to a prototypical negative existential.

An examination of the characteristics of Croft’s cycle has been predominantly conducted in spoken languages. Consequently, the inquiries undertaken here in relation to the NEC will facilitate a more profound comprehension of the sources and pathways of verbal negation in the domain of signed modality. A hypothesis can be advanced that TİD may be considered as a language situated at Stage B of the NEC. This is due to the fact that it has distinct negative markers for verbal negation and for the negation of existential clauses. However, as evidenced by the corpus data, TİD exhibits uses of the negative existential marker yok which have developed into the encoding of standard/verbal negation, as illustrated in (4) and (5).

The current paper sets forth the argument that the negative existential function of the predicate yok and its functions as a standard negator are worthy of further examination. The aim of testing the NEC model is to delineate the progression of a negative existential into a standard negator, along with the mechanisms that govern the transition between stages. More precisely, the current paper addresses the following two research questions:

1. How did the negative existential negator yok develop to encode standard negation?

2. Within the framework of Croft’s type B ∼ C, in which specific contexts does the negative existential construction function as a means of negating verbal predications?

2 Croft’s negative existential cycle

The NEC was first proposed by Croft (1991) and subsequently expanded by Veselinova (2013, 2014, 2015, 2016). The NEC provides a means of modeling the evolution of standard negators from existential negators. Furthermore, this kind of development is cyclical in nature. Once an existential negator has undergone a transformation into a standard negator, the potential arises for a new expression dedicated to existential negation to emerge. This may then develop into a standard negator in its own right (for a detailed overview, see Veselinova and Hamari 2022).

Croft puts forth the proposition that the languages belonging to the abovementioned three stable types (types A, B, and C) evince minimal or no variation within their own linguistic systems. However, other language types demonstrate synchronic variation between their verbal negation and their negative existential predications. He categorizes these as A ∼ B, B ∼ C, and C ∼ A, proposing that collectively they offer an empirically testable methodology for ascertaining the existence of direct historical connections between synchronic language states within a typological classification of languages (1991: 3). In summary, Croft (1991: 6) presents the following description of the NEC:

1. Stage A: the negation of the existential predicate is carried out by the verbal negator

2. Stage A > B: the contraction or fusion of the existential negator is used alongside the verbal negator

3. Stage B: the negative existential is distinguished from the verbal negator

4. Stage B > C: the negative existential is employed for verbal negation

5. Stage C: the negative existential is identical to the verbal negator

6. Stage C > A: The negative existential fulfills all negation functions

A schematic illustration of the NEC is provided in Figure 3. The boxes with a full outline represent stable types, while those with a dashed outline represent transitional types. In the original visual representation of the Croft’s NEC, only the stable types were illustrated. In a family-based sample, type B is the most prevalent, being followed by the transitional types B ∼ C and A ∼ B (Veselinova 2016: 150).

Figure 3:

Croft’s negative existential cycle (modified from Croft 1991: 6).

It could be inferred from the graphic representation of the cycle that the stages outlined are necessarily sequential. However, Croft (1991: 22) states that there is no necessity for this to be the case; in fact, it is to be expected that there will be overlaps of different stages. He also puts forth a set of discrete characterizations for the transition from one linguistic type to another, contingent on the transitional phase under consideration. Accordingly, each kind of linguistic transition appears to be shaped by a distinct phenomenon within the domains of pragmatics, syntax, and morphology. This is schematically summarized as follows:

1. From type A to type B: fusion

2. From type B to type C: emphasis

3. From type C to type A: syntactic analogy

The notion of phonological fusion, as outlined by Croft (1991), seems to present fewer challenges than the precise role of emphasis and analogy in other phases of transition posited in the ideal model of the cycle. However, it merely prompts the question of the potential chronological sequence between the morphological process.

3 Negation strategies in TİD

TİD employs a sophisticated negation system comprising a multitude of manual and nonmanual negative markers. A substantial body of research has been conducted on the descriptive and theoretical aspects of standard negation in TİD (e.g., Dikyuva et al. 2017, 2024; Gökgöz 2009, 2011; Kubuş 2008; Makaroğlu 2021, 2023; Pfau 2016; Zeshan 2003, 2006). The language also exhibits a notable degree of diversity in the expression of standard negation, evident in both its phonological and morphological forms and the number of markers employed in negative constructions. In terms of word order, it has been suggested that TİD is a SOV language and that the basic manual negator deĞİl ‘not’ occupies a clause-final position (see Kubuş 2008; Pfau 2016; Zeshan 2003, 2006). Additionally, the manual gesture palm.up is frequently observed in negative contexts, often functioning as a marker of epistemic negation or absence.

Although numerous studies on standard negation have been conducted from diverse theoretical perspectives within the TİD literature, there is a paucity of research on existential negation. The existing studies on this topic are scarce and lack analytical depth, focusing instead on descriptive accounts. TİD has both a positive existential particle var ‘exist’ and a negative existential particle yok ‘not.exist’ (Makaroğlu 2024).

With respect to the syntactic function of negative existentials as markers of negation, a considerable body of evidence from a wide range of languages exists which demonstrates that negative existentials function syntactically in the same slot as their affirmative counterparts, as illustrated by examples (6) and (7) in TİD (for detailed overview on existentiality in sign languages, see Zeshan and Perniss 2008). This finding offers compelling support for the view that negative existentials can be considered to be lexically and syntactically equal to affirmative suppletive paradigmatic relation to their counterparts, rather than negators, negative particles, or negative polarity items.

Zeshan (2006) posits that the yok particle is typically located at the end of a clause, except when a pronominal index pointing or palm.up marker is present, as in (8). In such cases, nonmanual marking may encompass both signs or else the scope will be limited to the negative particle alone.

The etymology of yok in TİD remains unclear, as the origin of this negative existential marker has not been discussed in earlier literature. Additionally, it is evident that, despite the extensive research conducted on standard negation in TİD, studies examining negative existentiality remain limited.

4 Methodology

5 The negative existential marker yok in TİD

5.1 The distribution of the standard negation function of yok

Based on a corpus analysis of the 453 tokens of yok, the distributional results of this study provide insights into the frequency of yok functional features. As illustrated in Figure 4, the statistical analysis indicates that the existential function of yok is the most prevalent in the database, constituting 70.6 % (n = 320) of the examples in the database. The possessive function of yok, on the other hand, accounts for 13.7 % (n = 62) of the examples. The standard negation function of yok accounts for 11.0 % (n = 50) of the examples. The remaining functions of yok, including interjection, emphatic, and prohibition, collectively account for 4.6 % of the examples in the database (n = 21). This statistical distribution reveals that the functions of yok are typically seen in three basic types and that its use in the standard negative function is not a rare phenomenon.

Figure 4:

The distribution of the functions of yok in the dataset.

In the domain of corpus linguistics, log-likelihood scores have been employed to assess the statistical significance of variations in the size of language corpora for the purpose of conducting detailed corpus profiling (Rayson and Garside 2000). In the present study, the presence of unexpected values in three distinct age groups is investigated. The log-likelihood ratio of the standard negation function of yok was calculated as illustrated in Table 3.

Table 3:

Log-likelihood ratio of the standard negation function of yok across age groups.

Comparison	Group 1	Total tokens (n)	No. of tokens of yok	Group 2	Total tokens (n)	No. of tokens of yok	Log-likelihood	Difference (%)
Young versus Middle-aged	Young	44,584	30	Middle-aged	29,467	15	0.80	32.19
Young versus Older	Young	44,584	30	Older	19,100	5	4.73	157.04
Middle-aged versus Older	Middle-aged	29,467	15	Older	19,100	5	1.83	94.45

In accordance with the established p < 0.05 level of significance for the log-likelihood ratio of 3.84, it was determined that there was a statistically significant difference between the young adult and older adult groups. However, no significant difference was observed between the young adult and middle-aged adult groups, nor was there one identified between the middle-aged adult and older adult groups. These observations indicate that the standard negation function of yok is more prevalent in young TİD signers, suggesting that the grammaticalization of this function is a more recent development in the language’s evolution.

It is worth noting that grammatical variation remains a subject of considerable uncertainty due to the dearth of research focusing on the linguistic factors that shape the spectrum of variants observed in any signed language (for detailed discussion, see Johnston and Schembri 2010).

As previously mentioned, the present study’s sample is limited to 14 cities in Türkiye. The analysis reveals that yok was used in the standard negation function in 12 cities (indicated in blue in Figure 5) but was not used in this way in İstanbul or İzmir (marked in red). This observation lends robust evidence to the supposition that there is an absence of regional variation. However, it is also evident that the inclusion of data from Türkiye’s eastern cities is necessary to test the hypothesis.

Figure 5:

The distribution of the standard negation function of yok across cities in Türkiye.

6 Conclusions

The present study is grounded in synchronic evidence derived from Croft’s (1991) negative existential cycle, a theoretical model postulating a connection between negation and the existential predicate as a driving force behind the evolution of standard negators. It demonstrates that TİD presents characteristics of stage B ∼ C, wherein yok is progressively substituting for the verbal negator in the grammatical system. However, it has not yet been extended to the whole grammatical system and is currently limited to specific contexts.

In the context of the theory and modeling of language change, it is crucial to emphasize that the NEC is a variationist model. Its stages do not progress sequentially; rather, they tend to coexist. The synchronic data from the TİD provide substantial support for this assertion, demonstrating that standard negativity can be encoded in alternative ways. Croft’s (1991) theoretical framework, particularly the identification of the intermediate B ∼ C stage, provides a valuable insight into the processes of negation and the connection between negation and negative existential predication, along with their relationship to other domains within the grammar. In the context of TİD, two domains of inquiry are the perfective aspect and the emphatic form. Applying this model suggests that the negative existential sign yok has progressed along the following pathway:

negative existential marker > negative perfective marker > standard negator

To conclude, in the following examples, (29) and (30), the standard negative marker deĞİl and palm.up cooccur with the negative existential mouthing for a particular instance of standard negation. Within the framework of grammaticalization, units characterized by distinct linguistic content, originating from two discrete channels, have the capacity to undergo integration within the domain of visual-gestural modality. The preliminary observations concerning the role of negative existential mouthings in standard negation thus give rise to a number of theoretical inquiries regarding the characteristics of multimodal grammar in signed modality.

Furthermore, the interaction of existential mouthings and negative manual signs in NEC offers a novel perspective on the concept of “multimodal grammar”, thereby supporting the notion that every construction has a multimodal potential in sign languages. Therefore, it is argued that these issues should be regarded as a motivating factor for reevaluating and refining the conception of “construction”, rather than for excluding multimodality in sign languages. Future research on negative existentials and their multimodal properties in TİD and other sign languages will provide further insights into the nature of the Croft cycle.