Skip to main content


Front. Psychol., 26 April 2022
Sec. Psychology of Language
This article is part of the Research Topic Sign Language Research Sixty Years Later: Current and Future Perspectives View all 33 articles

On the Reflexive KENDİ in Turkish Sign Language

  • Department of Linguistics, University of Michigan, Ann Arbor, MI, United States

Linguistic analysis is improved when it includes language beyond the spoken modality. This paper uses sign language data to explore and advance cross-linguistic typologies of reflexives, constructions expressing that co-arguments of a predicate are also co-referent. In doing so, we also demonstrate that the lexical item KENDİ in Turkish Sign Language (henceforth, TİD) can function as a traditional reflexive, in addition to its previously documented emphatic functions. We further show that KENDİ is a DP-type reflexive, which helps to explain the emphatic usages of KENDİ that have been the focus of previous research. We end by outlining a plan for future research that can further probe and unify the superficially distinct functions of KENDİ and the typology of anaphoricity across modalities. Data for the present research comes from recently conducted fieldwork interviews with two signers of the İstanbul dialect of TİD, both of whom have been exposed to TİD since birth.


Sensitivity to event participant structure is evident in the earliest stages of language acquisition (Pinker, 2013; Pace et al., 2016) and distinguishing the participants in an event is so fundamental to human language that it is present even in homesign (Goldin-Meadow et al., 2009). Interestingly, languages also universally have mechanisms for indicating that event participants are not distinct: reflexive constructions. Conceptually, reflexivity is a specific type of dependency relation between two arguments of a predicate where the two arguments are co-referent. Languages of the world have different strategies to mark this relation (Faltz, 2016/1977), as we discuss in more detail in “Background on Reflexivity.” In English, for example, this relation can be marked via reflexive pronouns (1).

Compared to the research on spoken languages regarding reflexivity (Frajzyngier and Walker, 2000; Büring, 2005b; Geniusiene, 2011; Faltz, 2016/1977, i.a.), there is limited literature on sign languages. This is partly because the linguistic study of sign languages is a fairly new endeavor. Though scattered earlier documentation exists (e.g., Desloges, 1779), linguistic analysis of sign languages began in earnest with Stokoe’s (1960) work on American Sign Language (henceforth, ASL). Though Stokoe made some observations regarding ASL syntax, his focus on the phonetics and phonology of signs left “much more” to do “in establishing exactly what are the structural principles of the sign language sentence” (Stokoe, 1960:32). Though much progress has been made since then (as one sees from all the work on sign languages cited in this paper as well as the other contributions to this volume), Stokoe’s (1960) statement still holds: there is still a tremendous amount of research to be done.

The research that has been done has shown that sign languages employ several strategies to form reflexive structures. For example, Kimmelman (2009) reports that ASL1, Russian Sign Language, Dutch Sign Language, Israeli Sign Language, and Croatian Sign Language all have reflexive pronoun strategies. Because reflexivity is present in some form across various languages and language families (Frajzyngier and Walker, 2000; Büring, 2005b; Geniusiene, 2011; König et al., 2013; Faltz, 2016/1977, to name a few) and across modalities, it may be a universal phenomenon of language. In that case, describing and analyzing reflexivity in understudied languages like sign languages provides an opportunity to expand existing accounts and holds a great deal of importance in testing our existing generalizations about argument structure and reflexivity. This is the aim of the present paper. We hope to contribute to the ongoing endeavors to document, describe, and analyze sign languages in the pursuit of a better understanding of human language.

Our focus will be the lexical item KENDİ in Turkish Sign Language (henceforth, TİD), which is produced by tapping in the middle of the chest with an open hand, fingers bent inward, as depicted in Figure 1.


Figure 1. KENDİ.

There has been little work on KENDİ in the existing literature, much of it limited to in-passing observation that the sign KENDİ exists. Zeshan (2002) and Sevinç (2006:16) briefly note that the form is attested but observe only an emphatic function (2a). In their recent grammar of TİD, Kelepir (2020a), citing data from Dikyuva et al. (2017), nevertheless label KENDİ as a reflexive pronoun (2b):

Our study looks to build upon these limited observations to examine if KENDİ can be used both as a traditional reflexive [as Dikyuva et al. (2017) and Kelepir (2020a) suggest] and/or as an emphatic marker [as Zeshan (2002) and Sevinç (2006) observed]. We focus on the İstanbul dialect of TİD. The layout of the paper is as follows: “Background on Reflexivity” summarizes the notion of reflexivity and provides an overview of the previous literature on reflexives in signed and spoken languages. “Methodology” explains our methodology of data collection. “KENDİ Marks Co-referential Relations” lays out the co-referential properties of KENDİ as a traditional reflexive. Having established that KENDİ can function as a traditional reflexive, we then turn to where KENDİ stands within reflexive typologies. Building on the data from earlier sections, “Co-referential Relations With KENDİ” further probes the syntactic and semantic properties of KENDİ in its traditional reflexive function. The functions of KENDİ beyond its traditional reflexive role is the focus of the final section before we close by summarizing our findings and laying out directions for future research.

Background on Reflexivity

In this section, we will overview basic properties of reflexivity that are relevant for this study and summarize the reflexivity patterns that have been documented in spoken and signed modalities.

Reflexivity in Spoken Languages

The literature on reflexivity suggests that it is a universal part of language, observed in many languages across different language families, albeit encoded with different grammatical mechanisms. With respect to the strategies that encode reflexivity, we can talk about two main kinds of reflexivity: lexical reflexivity and grammatical reflexivity. Lexical reflexivity, which is also sometimes called inherent reflexivity, is a phenomenon we observe on predicates that express events that are prototypically done to oneself, such as “bathe”. The default interpretation of an intransitive sentence like I bathed in English is reflexive, I bathed [myself]. Predicates like “bathe” can, however, express non-reflexive events, as with the transitive variant I bathed the dog in English. Lexical reflexives in English are also often compatible with (redundant) grammatical reflexivity, as in I bathed myself (emphatic interpretations, which we discuss below, may be more natural here). Here, the reflexivity relation is not only expressed via the semantics of the verb but also the argument structure of the utterance. Grammatical reflexivity manifests itself either via marking on the verb (a) or marking on the arguments (b):

In (3a), the reflexive marker {-un}2 is a verbal suffix. As for (3b), Faltz (2016/1977) observes that the reflexive pronoun “zich” in (3b) links the object of the verb to the subject, marking that the do-er and the patient are the same referent.

There is also variation within the categories of verbal and argumental reflexivity marking. Verbal reflexivity includes affixes like -un- in Turkish (3a) and clitics, such as the French se in s’habiller (“to dress oneself”). As for argument marking, we observe both free standing reflexive pronouns such as zich in (3b) as well as bound morphemes that shift a stem to a reflexive meaning (e.g., -self in English). It is not uncommon for languages to exhibit multiple reflexive strategies, or to combine them as part of a complex reflexive construction.

What unites these different strategies is that they are all subject to certain structural constraints. First of all, a reflexive requires an antecedent for co-reference, and the relationship between them is often called binding. Argument reflexives require being bound by a potential co-referent (4a-b) and are usually restricted as to what can bind them within what structural domain and/or configuration (4c-d), e.g., intra- or inter-clausal (see, among others, Chomsky, 1981; Reinhart and Reuland, 1993; Faltz, 2016/1977).

The literature on reflexives has long attempted to unify the structural constraints on the reflexivity. However, pinpointing these constraints isn’t always an easy task and is further complicated by the fact that reflexives are commonly observed to serve other functions (Reinhart and Reuland, 1993; Faltz, 2016/1977; Déchaine and Wiltschko, 2017 i.a.) that may not be subject to the same structural constraints as a traditional reflexive. Emphatic anaphors (5a) refer to one of the participants within the event, but their function is to put focus on or set apart a particular participant without affecting the argument structure of the verb. Logophors (5b), on the other hand, are anaphoric items that can get their co-reference from outside of the structural domain of a traditional reflexive. Thus, emphatic markers (5a) and logophors (5b) still express co-reference, which makes them anaphoric. However, they do not necessarily express co-reference between arguments of the predicate, so they are not reflexive.

Such multi-functionality has already been observed for KENDİ (Zeshan, 2002; Sevinç, 2006; Dikyuva et al., 2017), which we discuss in more detail below. However, what we first aim to show in this paper is that KENDİ can be used as a traditional reflexive. Before we move on to discussing that, however, we first provide a review of the existing reflexivity literature on sign languages.

Reflexivity in Sign Languages

As in other domains of linguistic structure, research on reflexivity in sign language is limited. However, the research that has been done observes key similarities across modalities. Sign languages, too, employ two main strategies to mark reflexivity: on the verb (6a), or on the argument of the verb (6b):

The verbal reflexivity of (6a) is expressed via the spatial agreement markers, indicated by the subscripted ‘‘a’’3. Here, the direction of the object agreeing verb’s path movement agrees with the subject, linking the grammatical object and the subject of the sentence in co-reference. Kimmelman (2009) also reports argumental reflexive pronouns in NGT, glossed as ZELF in (6b), similar to English -self. This parallelism across modalities is not unexpected considering that reflexivity reflects event participant structure, which may be conceptually and linguistically fundamental. This parallelism aside, what sign languages bring to the table is the way they use space for reference, which may lead to modality-specific effects on reflexivity. The role of space in modulating reflexivity is evident in (6a). These reflexivization strategies are not specific to the exemplified languages. Table 1 below shows various reflexivization strategies attested in sign languages.


Table 1. Reflexivization Strategies Attested in Sign Languages.

The multi-functionality of reflexive markers is also relatively well-documented in sign languages. As an example, SELF in ASL can also functions as copula (7a) and as an emphatic marker (7b):

All in all, the existing work, though limited, shows that reflexivity in sign languages is compatible with certain aspects of existing typologies. However, it is also the case that sign languages can provide unique and novel data to further explore the phenomenon of reflexivity, including how usage of space affects referential relations (Schlenker, 2018). Before turning to what TİD shows us about reflexivity, we first briefly explain how we collected and analyzed data in this study.


The data for this work comes from fieldwork sessions with two Deaf adult female signers of the İstanbul variety of TİD. Both have been exposed to TİD since birth. The sessions took place 2020–2021 and were conducted online via Facetime and Zoom due to the limitations on travel and in-person meetings during the COVID-19 pandemic. The language of interaction was TİD, though both consultants also have proficiency in written Turkish. We have utilized acceptability judgments and having the consultant describe contexts and situations using KENDİ as methods of data elicitation. Only one consultant at a time was present for each fieldwork session. Because of quality issues that can arise in videoconferencing, the data reported here were also recorded separately by one of the consultants and can be accessed in an online repository (file names correspond to example numbers):

KENDİ Marks Co-Referential Relations

In the previous section, we described the basic patterns of where and how co-referentiality and, specifically, reflexivity is marked in language. In this section, we will look further into co-referential dependencies in TİD, and the core properties of co-referentiality marked by KENDİ. Our aim here is to lay the groundwork for the more detailed discussion of the distribution of KENDİ in later sections.

As noted above, co-reference is structurally constrained and certain classes of DPs are restricted in the co-reference relations they can enter into. Büring (2005b) categorizes the possible co-referential relations between DPs as obligatory co-reference (8a), obligatory non-co-reference (8b), and optional co-reference, i.e., ambiguity (8c):

In (8a), the reflexive herself has to refer to Zelda, as indicated by the subscripted referential indices. In (8b), the pronoun she has to refer to an entity other than Zelda. Lastly, in (8c), the possessive pronoun her allows reference to either Zelda or an individual outside of the clause, a discourse referent. The Binding Theory (Chomsky, 1981) aims to account for these patterns of obligatory co-reference (Principle A), obligatory non-co-reference (Principle C), and optional co-reference (Principle B):

In (9), we document these patterns in TİD:

KENDİ in (9a) must refer to the subject, ELVAN, as indicated by the subscripted referential indices. This is obligatory co-reference, as one would expect of a traditional reflexive governed by Principle A. Skipping ahead to (9c), we again observe a familiar pattern: the sign name ELVAN and the 3rd singular subject cannot be co-referent. This is obligatory-non-co-reference, governed by Principle C. The structure that gives rise to optional co-reference ambiguity in (8c), however, patterns differently in TİD (9b) due to the spatialization of the possessive marker (see Cormier et al., 2013 for a discussion of pronominal spatialization in sign languages). POSS in TİD, as in many other sign languages, spatially indicates its referent (here, ELVAN). Thus, we have obligatory co-reference with a non-reflexive pronominal not so much because binding works differently in TİD (9b), but because spatialization can prevent certain ambiguities from arising in the sign modality (Quer and Steinbach, 2015).

Interestingly, optional co-reference is possible if the overt possessive is removed entirely, as in (10).

The ambiguity here is highly dependent on context and includes interpretation as definite nominal (“the son”). Ambiguity on a par with (8c), however, can arise due to the use of a null possessive. Because it is null, the possessive is not spatialized as in (9c). The null possessive can be co-referent with the overt (subject) argument (ELVAN) by default/without special context (on a par with the preferentially bound interpretations observed elsewhere, cf. Reinhart, 1983; Kehler and Büring, 2007). However, it can also refer to another contextually salient referent, as in (11):

Turning next to the structural distribution of these three co-referential patterns, Büring (2005b) shows that lack of an antecedent in a mono-clausal setting affects each type of DP differently (12):

There are also structural constraints on where an antecedent can be when it is present, and this is where notions of reflexivity are key. The sentences in (13) express a reflexive event. Unsurprisingly, that reflexive event can be described using an obligatorily co-referent reflexive pronoun (13a). The reflexive event cannot be described using an obligatorily non-co-referent R-expression (13c), nor can it be described using a non-reflexive pronoun (13b).

Parallel structural constraints hold in TİD. Like (12a), (14a) is ungrammatical because KENDİ requires a morphosyntactic antecedent (14a), unlike a non-reflexive pronoun such as IX3sg in (14b) or an R-expression such as ELVAN in (14c). Moreover, as in (13), we see that only KENDİ can be used in a reflexive environment like (15).

Thus, KENDİ appears to behave like a traditional reflexive that (i) requires a co-referential antecedent that is (ii) in the right structural configuration.

These morphosyntactic constraints pertain to what is meant by “binding domain” in Principles A-C. Binding domains appear to be sensitive to structural proximity, often called locality, which we illustrate here with clause boundedness. The sentences in (16) present the three types of DPs in the object position of a subordinated clause in English, with their potential antecedent in the subject of the matrix clause.

The boundary between the matrix clause and the subordinate clause seems to also function to demarcate binding domains. Thus, the antecedent for the reflexive in (16a), as compared to (8a) or (13a), is “too far away” to satisfy its binding requirements. In the case of a non-reflexive pronoun, the added distance of (16b) relative to (13b) allows the optional co-reference to emerge. Lastly, (16c) affirms that co-referential dependency, even when not local, between an R-expression and a potential antecedent is ungrammatical. Recall that this dependency was ungrammatical when it was local in (13c), too. Focusing on the comparison of (13a) and (16a), these data show that a traditional reflexive requires binding by an antecedent within its own local domain. This then would predict that if KENDİ is indeed a traditional reflexive, we should see evidence of structural sensitivity and locality constraints (though they may not match, exactly, those of English). The TİD equivalents of (16) are presented in (17), and the data in (17a) show that this prediction is borne out. (17a) is only grammatical when KENDİ is bound by an antecedent in its local domain (IX1sg_j) like in (16a), as opposed to an antecedent outside of it (ELVANi). As for the pronoun in (17b), we again observe the sentence is rendered grammatical as long as the pronoun is bound by a co-referent outside of its local domain, as opposed to being locally bound. Moreover, we see that an R-expression in TİD (17c) is degraded if it has an antecedent at all, even if that antecedent is in a different domain (though in TİD the judgment is that the sentence is highly marked, not fully ungrammatical):

Thus, in many respects, TİD aligns with the principles of binding theory. However, these principles are under debate, and a common critique is that there is not strict complementary in the distribution of the three types of DPs. Our key observation here, though, is that KENDİ is only grammatical when bound by an antecedent that is syntactically proximal to it, such as (IX1sg_j) in (17a), as opposed to an antecedent outside of it (ELVANi). The basics of co-referential relations among nominals in TİD tells us that KENDİ behaves like a traditional reflexive. However, reflexives aren’t a homogenous class, and reflexives with different morphosyntactic and semantic encodings may be subject to different restrictions (Thráinsson, 1976; Hellan, 1988; Reuland, 2011; Déchaine and Wiltschko, 2017). In the next section, we explore what type of traditional reflexive KENDİis, using the typology proposed by Déchaine and Wiltschko (2017). As will become clear in below, this typologically informed analysis of KENDİ can help us better understand previous claims in the literature.

Patterns With KENDİ Relative to Typologies and Analyses of Reflexivity

In the previous section, we summarized the basic co-referential relations in TİD and showed that KENDİ can function like a traditional reflexive. As we noted briefly above, however, reflexives are not a uniform class in many respects, including their syntactic category. Thus, we next ask what type of traditional reflexive KENDİ is, using the syntactic typology of reflexives proposed by Déchaine and Wiltschko (2017). Déchaine and Wiltschko (2017) divide reflexives into five types, and provide examples of languages with each type of reflexive (Table 2).


Table 2. Déchaine and Wiltschko’s (2017) typology of reflexives.

The main types of reflexives are: clitics, bound nouns, agreement markers, intransitivizers, and DP constituents. In addition to differing in syntactic category, the different types also exhibit slightly different structural patterns and contribute somewhat different semantics (despite all being a reflexive).

One such difference is that the multi-functionality of the reflexive marker differs depending on its type. The typology laid out by Déchaine and Wiltschko (2017) can be used to determine where KENDİ stands among reflexive markers, laying more solid groundwork for further analysis of KENDİ and co-reference relations in TİD. We begin by using Déchaine and Wiltschko’s (2017) diagnostics to identify the morphosyntactic category of the reflexive KENDİ. The results of these diagnostics show that KENDİ behaves like a DP-type reflexive. The classification of KENDİ as a DP-type reflexive leads to predictions about its behavior and functions beyond marking traditional reflexivity, which we discuss further in “The Function of KENDİ Beyond Traditional Reflexivity.”

KENDİ Is Not a Clitic

Much like traditional reflexives are semantically dependent on an antecedent, clitics are morphologically and phonologically dependent on a host. They cannot bear stress, and often come in a reduced phonological form, as exemplified by the reflexive clitic m’ in (18), which is a reduced form of the pronominal me. As the gloss suggests, the reflexive clitic m’ forms a phonological unit with its verbal host (auto-suggère). Moreover, though French is typically a postverbal object language and m’ is expressing the reflexive object of the predicate, the clitic appears in a preverbal position. Thus, the clitic has characteristic properties in terms of its morphophonology and its morphosyntactic distribution.

Turning to TİD, a default SOV language, note first that the typical position of object KENDİ is in the standard pre-verbal position, as illustrated in (9a) and similar examples. This distributional pattern is not inherently at odds with a clitic analysis, but if KENDİ were a clitic, we would expect it to form a phonological unit with a preceding or following element. However, as illustrated in (19), KENDİ can be morphophonologically separated from the verb (its following element) by an intervening adjunct (as illustrated in the repository video, the adjunct intervening in (19a) continues throughout the production of the verb, while the adjunct in (19b) is clearly sequential):

KENDİ is also morphophonologically independent from the elements that precede it. For example, KENDİ is adjacent to its antecedent, ELVAN in (19a), but separated from ELVAN by the intervening adverb NOW in (20).

Another piece of evidence that KENDİ is not a clitic is that it can dislocate to the left periphery, giving rise to a contrastive topic meaning (21).

In TİD, contrastive items are often observed with an eyebrow raise that is articulated simultaneously with the contrast-associated item (Gökgöz and Keleş, 2020 section 4.2). It is also the case that non-manual spread (annotated with the line above the manual sign glosses) (21) in sign languages has been associated with marking phrasal boundaries (Pfau, 2005, 2006). The non-manual marker associated with focus in (21) is not spreading over the subject ELVAN. Therefore, (21) not only shows that KENDİ can be linearly dislocated from what would have been its most plausible host, but also shows that it forms its own prosodic unit. This, then, backs up the narrative that KENDİ is a morphophonologically independent form.

Comparing KENDİ to other clitics that have been documented in TİD -namely, the clitic form of negationalso reveals differences. Zeshan (2003, 2004) and Kelepir (2020d: observe that manual negation in TİD can occur as a free form (Figure 2) or a clitic (Figure 3)5. In its cliticized form, negation loses its syllabicity (reduced movement, shorter duration) and assimilates to the location of its host, instead of the neutral signing position used in Figure 2.


Figure 2. Adapted from Makaroğlu and Dikyuva (2017, entry: “değil”). Open-source image available from the online TİD dictionary:


Figure 3. Adapted with permission from Kelepir (2020d: P4,

Unlike clitic negation, the signing position of KENDİ does not get assimilated to that of its host. This could be due to the body-anchored nature of KENDİ. However, there are other ways KENDİ differs from the negative clitic. Building on Zeshan and Kelepir’s analyses, Gökgöz (2009) found that the non-manual marker for negation, a head tilt, patterns differently with the clitic and non-clitic form. With cliticized negation, the head tilt associated with negation spreads onto its morphophonological host, as indicated by the line above KNOW and cliticized ^NOT in Figure 4. With non-cliticized, free negation, however, the non-manual marker only spreads over the negation itself (Figure 5).


Figure 4. Adapted with permission from Gökgöz (2009:20).


Figure 5. Adapted with permission from Gökgöz (2009:21).

Here, too, (21) provides the relevant evidence: we might expect the eyebrow raise in (21) to spread onto ELVAN too, had KENDİ cliticized to it. Thus, based on evidence from intervening items and the properties of clitics in TİD, the relationship between KENDİ and preceding or following elements is linear adjacency, not morphophonological dependency, as would be expected of a clitic.

KENDİ Is Not a Bound Noun

Bound noun reflexives are attested in languages such as in Halkomelem, where body part nouns attached to the predicate can be interpreted as co-referential with an argument, either the subject in an intransitive verb form (22a) or the object in a transitive verb form (22b). In the case of the predicate with an intransitive marker (-em) in (22a), only the reflexive interpretation is possible.

Data like that presented in the previous section also argue against a bound noun analysis of KENDİ’s. KENDİ does not display the morphophonological characteristics of a bound element.

KENDİ Is Not an Agreement Marker

Our next step is to check if KENDİ is a verbal agreement marker. As discussed earlier, marking reflexivity on the verb is a commonly employed phenomenon in signed (and spoken) languages. Moreover, sign languages often make use of space for modulating agreement, and we know that TİD is a sign language that marks agreement spatially (Gökgöz et al., 2020 section Spatial markers on predicates parallel agreement markers in tracking the event participants [see contributions to Lillo-Martin and Meier (2011) for discussion]. In fact, even intransitive predicates with a single argument can spatialize this way (Costello, 2016). However, we argue that this is not what KENDİ is doing for two reasons: (i) it marks reflexivity without being assigned a locus and (ii) still allows the verbal reflexive agreement marker to appear (if compatible with the predicate in general; Kelepir, 2020b section 3.1).

Before elaborating on those arguments, we first provide an example of a reflexive that is of the agreement marker type in Shona. The reflexive -zví in (23) has the distribution of an object agreement marker in the morphological template of Shona6 (note that Shona also has a non-reflexive -zvì that differs in tone from the high-toned reflexive -zví):

As discussed above, KENDİ, as compared to -zví, does not form a morphophonological unit with the verb. Indeed, KENDİ is not even obligatorily adjacent to the verb, as one might expect of an agreement marker in general. Moreover, KENDİ is body-anchored and does not make use of an assigned locus in the signing space. Thus, it’s quite unlike how agreement is marked in TİD (Gökgöz et al., 2020: and in other sign languages (Cormier, 2012:124–125; Sandler, 2012: 44–45). Finally, if KENDİ were a kind of less common agreement marker in language that does not form a unit with the predicate it’s marking “on”, and an almost unattested kind of agreement marker in sign language that does not make use of space, we would expect it to show up post-verbally, because that’s where functional items typically occur in TİD (Gökgöz, 2020).7 The default position of KENDİ, however, is preverbal. Therefore, as above, both cross-linguistic and language-internal patterns argue against this analysis of KENDİ.

KENDİ Is Not an Intransitivizer

As for the intransitivizer category, Déchaine and Wiltschko (2017) point to the -iso suffix in Cree as an example. They observe that patient/object marking is absent on the verb when the reflexive marker is present (24a), but present then the verb is used non-reflexively (24b):

As described above, TİD and other sign languages can incorporate spatial locations to mark agreement. TİD also displays object agreement on some verbs through palm orientation (Kelepir, 2020c section Figures 6, 7 illustrate how palm orientation marks object agreement in TİD. In Figure 6, the object is 1st person and the palm orientation is toward the signer, while in Figure 7 the palm orientation is toward an established spatial locus away from the signer’s body, the orientation of a 3rd person marker.


Figure 6. Adapted with permission from Kelepir (2020c: P4,


Figure 7. Adapted with permission from Kelepir (2020c: P4,

This palm orientation agreement marker is present in non-reflexive predicates like PROTECT or SUPPORT. Crucially, however, it is also present in reflexive usages of the same predicates. This is illustrated for reflexive and non-reflexive usages of the predicates PROTECT and SUPPORT in (25). In a sentence like (25a), the 3rd person palm orientation is toward a spatial locus away from the signers’ body, whereas in (25b) there is the reflexive marker KENDİ as well as palm orientation toward the signer’s body, just like Figure 6.8 Note that the signer has omitted the subject agreement marker in (25b), as has been observed elsewhere in sign languages (Padden, 1988, 1983).

Thus, KENDİ does not manipulate the argument structure like the intransitivizer -iso in Cree does, and we can conclude that KENDİ is not an intransitivizer because it co-occurs with a marker of transitivity, the palm orientation marker of object agreement.9

KENDİ Is a DP Reflexive

So far, we have seen that KENDİ does not align with the patterns of a clitic, bound noun, intransitivizer, or agreement type of reflexive. However, there is another kind of reflexive in the typology: a DP reflexive.

DP reflexives are basically reflexives that act like any DP, except for the specific dependency relation that they require a co-referring antecedent. Previous examples have shown that KENDİ is a reflexive that behaves like any DP (object) argument of the verb: it occupies an A position as the canonical object in the sentence and, moreover, can be dislocated to an A′ position the left periphery.

Recall that Déchaine and Wiltschko’s (2017) typology includes syntactically parallel items for each of the reflexive types. For DP type reflexives, the parallel they observe is possessors. Note that in some DP type reflexives, the connection with possessors is transparent; the reflexive form in English contains a possessive: myself. For KENDİ, the connection with possessives is twofold. One, KENDİ can be combined with an overt possessor like the English “xself”, as in (27).

Two, KENDİ itself can also be used as an independent possessive (28), though this usage isn’t very common.10

There are also cases of complementary distribution, where possessive usages of KENDİ block another possessive:

With respect to reflexive multi-functionality, additional functions of the DP type reflexives that Déchaine and Wiltschko (2017) observe are serving as emphatic anaphors (30a) and a logophors (30b):

As a reminder, emphatic anaphors are used to focus some aspect of the event or the referent’s role in the event, often from a non-argument position, and logophors are anaphoric elements that seemingly skirt binding principles by getting their co-referent from a discourse antecedent outside of their local domain. Thus, emphatic anaphors and logophors are reflexive forms that are in grammatical positions that are not associated with reflexivity (emphatic anaphor), or have non-prototypical antecedents (logophor). We are currently investigating whether a logophoric usage of KENDİ possible, but we already know from previous observations by Zeshan (2002) and Sevinç (2006) that KENDİ does indeed function as an emphatic anaphor. We elaborate on the emphatic anaphor usage of KENDİ in “The Function of KENDİ Beyond Traditional Reflexivity.”

To summarize, our fieldwork reveals that KENDİ in TİD can function as a traditional reflexive, and its previously observed usage as an emphatic anaphor is connected to its status as a DP-type reflexive, similar to English xself, and unlike other syntactic categories of reflexives discussed above. Moreover, we have seen that KENDİ shares other features that characterize DP-type reflexives, such as a structural parallelism with possessors. In the next section, we further explore the traditional reflexive usage of KENDİ, providing a more detailed description of its binding domain and its antecedents.

Co-Referential Relations With KENDİ

We have thus far provided some basic observations regarding co-referential relations in TİD and shown that KENDİexhibits behaviors consistent with a traditional reflexive. We have also argued that KENDİ is a DP-type reflexive. In this section, we will explore KENDİ as a traditional reflexive in more detail and discuss its relation to potential antecedents in local and long-distance binding domains. We first examine whether KENDİ can be bound by null antecedents as well as overt ones, and then the clausal location of these antecedents.


A defining characteristic of a traditional reflexive is that it requires an antecedent. Whether this antecedent must be overt or not depends on whether the language in question allows null arguments. This is illustrated for KENDİ by the contrast in (31), where (31a) contains an overt antecedent (ELVAN) for KENDİ and is grammatical, but (31b) lacks an overt antecedent and is ungrammatical.

Judgments like those for (31a) and (31b) are in “out of the blue” contexts. However, language is rarely used in truly out of the blue contexts. Given that TİD permits null arguments, we would predict that KENDİ can be licensed by covert antecedents. The data in (32)-(33) illustrate that this prediction is borne out (“e” glosses the position of the null argument):

These data show us that the required antecedent for KENDİ can be null arguments that are licensed by earlier portions of the discourse (32) or by contextual salience (33). Like KENDİ, the null argument that binds the anaphor, ei, is subject to its own licensing conditions (see Kayabaşı et al., 2020 for a detailed discussion of null arguments in TİD).

These data affirm that KENDİ patterns like a traditional reflexive in requiring an antecedent, though independent patterns of null argument licensing in the language mean that this antecedent need not be overt. These findings are in line with existing research on null arguments and reflexive pronouns in other sign languages (Lillo-Martin, 1986; Bahan et al., 2000; Koulidobrova, 2012; Kimmelman, 2018; Kayabaşı et al., 2020 i.a.).

The Structural Relationship Between KENDİ and Its Antecedent

We have already briefly described KENDİ’s relationship to overt and covert antecedents. In this section, we will talk about the structural logistics of KENDİ and discuss suitable structural positions for an antecedent to bind KENDİ, both in terms of hierarchy and proximity.

For a traditional reflexive like KENDİ, Principle A (Chomsky, 1981) is usually interpreted as requiring c-command of the reflexive by its antecedent. C-command is a structural relationship between two nodes, X and Y, neither of which dominates the other, but where every branching node that dominates X, also dominates Y (Reinhart, 1976; Büring, 2005b, i.a.). Figure 8 illustrates how A can bind B but not the other way around since A c-commands B but not vice versa.


Figure 8. Representation of C-command between A and B.

Crucially, there are patterns in language, like reflexivity, that show sensitivity to structural relationships like c-command. This is illustrated by the relationship between Charlotta and her/herself in (34). Note that the examples are structurally identical and Charlotta precedes the intended co-referential DP in both. Because Charlotta is “buried inside” the subject DP, however, it does not c-command the her/herself in either example. The resulting grammaticality differs depending on the use of a non-reflexive vs. reflexive pronoun. The non-reflexive pronoun in (34a) is grammatical because such pronouns are only optionally co-referent and do not need to be bound. However, a traditional reflexive is obligatorily co-referent, so the herself in (34b) does need to be bound. Consequently, expressing the co-reference via a reflexive, as in (34b), is ungrammatical.

A similar example for TİD is presented in (35). Here, there are two possible antecedents for KENDİ, the possessor ELVANi and the full possessive DP [ELVANi POSS3sg_i SISTER]j. Only [ELVANi POSS3sg SISTER]j, however, is in a c-commanding relationship with KENDİ. As predicted, this is the only DP that can bind KENDİ:

The fact that ELVANi cannot bind KENDİ tells us that TİD is no exception to the rule that the relation between the antecedent and a traditional reflexive is structurally determined, and one that cannot be characterized by linear precedence. Though critiques have raised questions about whether c-command accurately characterizes the structural constraints imposed on binding, what is relevant here is that (35) shows that there is a structural constraint at play. So far, we described what type of antecedents KENDİ can take and showed that it can be bound by both overt and null antecedents. Moreover, we have shown that the relation between KENDİ and its antecedent is subject to some type of structural constraint. To assess the proximity aspect of the antecedent-reflexive relation, we turn next to the pattern of KENDİ in subordinate sentences. In (36), KENDİ is in the object position of the subordinated predicate SEE, and co-reference with the subject (ELVAN) of the matrix predicate WANT is possible (either via direct binding or mediated by an intervening null argument, itself co-referent with ELVAN):

Note, however, there are no other viable antecedents in (36). In (37), we see how KENDİ behaves when multiple possible antecedents are present in the sentence.

Here, too, KENDİ is in the subordinate object position, but now two overt and distinct referents are available to serve as antecedents: the matrix predicate subject, IX1sg_j, and the subordinate subject, ELVANi. The only permitted antecedent for KENDİ in (37) is the closer subordinate subject. Note that the WANT type verbs in (36–37) are usually associated with non-finite sentential complements. In (38), we see the same binding patterns with THINK, a verb type that is often associated with finite sentential complements. There is currently very little known about finiteness in sign languages (and almost nothing known about this phenomenon in TİD) within the existing literature, but the patterns are the same across these predicate types.

Together, (35)-(38) show us that KENDİ is sensitive to familiar, structural binding constraints of hierarchy and proximity. Having elaborated a bit on the co-referential relations of KENDİ as a traditional reflexive, we will now turn to KENDİ’s function beyond traditional reflexivity, emphatic anaphoricity.

The Function of KENDİ Beyond Traditional Reflexivity

KENDİ has already been observed to be used as an emphatic anaphor (see Zeshan, 2002; Sevinç, 2006), as is commonly true of other DP-type traditional reflexives (Déchaine and Wiltschko, 2017). In this section, we will further explore the emphatic anaphor function of KENDİ.

Two Types of Emphatics

Emphatic markers are anaphors that co-refer to a participant of the given event to cast focus on it and to contrast it from other participants in a possible set of participants (Kemmer, 1995; Stern, 2004, i.a.). They usually occupy non-argument positions, and they do not express reflexivity. Focusing on English, Ahn (2010) notes two distinct usages of the emphatic anaphor xself (here, himself), exemplified in (39). In (39a), himself creates argument focus on its event participant antecedent, a ‘‘specifically John and not someone else’’ meaning. In (39b), however, himself modifies the event (not an argument), and emphasizes that John performed the given event without help or the contribution of another causer/agent.11

He labels the interpretation that arises from modifying the DP (39a) as “adnominal”, and the one that arises from modifying the DP as “adverbial” (39b).

In (40) we see the same two-way distinction with the emphatic function of KENDİ:

As the translations show, (39a-40a) and (39b-40b) resemble one another with respect to the emphatic contribution of himself and KENDİ.

Notice, though, that the different interpretations of himself in (39a-b) also correspond to different syntactic positions. Interestingly, the two distinct interpretations in (40) are possible with KENDİ in the same linear position. However, KENDİ in its emphatic function can also occupy different positions in the sentence (41), including the rightmost position (41c), which is ungrammatical for a traditional reflexive KENDİ as (42) shows (note also that emphatic KENDİ in (41) is optional):

Moreover, as the translations for (41a) and (41c) indicate, emphatic KENDİ has both adnominal and adverbial interpretive possibilities in each of these possible positions.

However, Ahn identified other diagnostics that distinguish adnominal and adverbial emphatic anaphora, such as: denying the event, specificity, thematic roles, context-free acceptability, stative verbs, prosody and stress. Here, we will use three of Ahn’s diagnostics to probe emphatic usages of KENDİ: (i) denying the event, (ii) specificity, and (iii) thematic roles of the co-referent (research is ongoing, and we do not have enough data to conduct all the diagnostics suggested by Ahn (2010) at this stage).

Denying the adnominal emphatic requires denying that the focused referent was, in fact, the relevant event participant at all; in (43a), this is accomplished by asserting that someone else did the activity. Denying the adverbial emphatic, however, doesn’t mean denying that the individual did the thing, just that they did the thing alone or without help. The felicitous adverbial denial pattern is given in (43b).

In (44), we see that both denials are felicitous with emphatic usages of KENDİ in each of the positions identified in (41)—that is, all three positions are apparently compatible with both adnominal and adverbial interpretations of KENDİ.

In addition to differences in deniability patterns, Ahn (2010) also observes that adnominal emphatic anaphors require a specific (but not necessarily definite) referent (45a,c), whereas no such restriction holds for adverbial emphatics (45b,d). “Specificity” here refers to the event participant being a unique entity as opposed to a generic one.

As for TİD, we again see a slightly different pattern. Non-specific referents—namely, WHICH CHILD (46) and SOMEONE (47)—are semantically compatible with the adverbial interpretation of emphatic KENDİ, as in English, but they are also compatible with adnominal interpretations. Here, too, these observations hold for KENDİ in multiple positions.

The third and final diagnostic that we will cover from Ahn (2010) concerns the thematic roles that are compatible with the emphatic anaphor. He observes that adverbial emphatic anaphors are only compatible with volitional and agentive subjects, whereas adnominal emphatic anaphors have no observed thematic role restriction. However, (48) shows not only that KENDİ is perfectly acceptable and grammatical with a non-volitional inanimate subject, but that it is compatible with an adverbial interpretation (EXTCL glosses an extension classifier sign):

So far, our assessment of the emphatic usages of KENDİshows us that they have more flexibility than (i) their traditional reflexive counterparts in TİD and (ii) their emphatic anaphor counterparts in English. However, it is important to note that the semantics and pragmatics of TİD is very understudied, as is the cross-linguistic typology of emphatic anaphors. Future investigation can investigate the source of these cross-linguistic differences and further assess if there are differences between adnominal and adverbial interpretations of KENDİ.

Ambiguity Between the Emphatic and the Traditional Reflexive: Optional Argumenthood

In this section, we will explore cases where KENDİ can be ambiguous between a traditional reflexive and an emphatic anaphor. Such cases are possible when the predicate of a sentence allows for object drop or null objects and a potentially reflexive event. The predicate VOTE GIVE can take a DP (49a) or PP (49b) as an object, and it also can be used intransitively (49c):

Thus, in (50), it is not clear whether KENDİ occupies an argument position as a traditional reflexive, or whether it’s an emphatic anaphor, and within the latter both adnominal and adverbial interpretations are possible.

We assume here that this is a case of structural ambiguity: despite KENDİ surfacing in the same linear position, it occupies different positions in the sentential structure. The intended interpretations can be contextually disambiguated but they can also be structurally disambiguated12. Examples of structural disambiguation are illustrated in (51), where the presence of a preposition (51a) or a separate DP (51b) unambiguously express the intended transitive interpretation (recall from (49a-b) that VOTE GIVE can take its object as a DP or PP, which is why the FOR is “optional”).

The above examples are cases where we see a given token of KENDİ that is compatible with different types of interpretations. However, there are also cases where we see multiple tokens of KENDİ within the same utterance, as in (52).

For the first interpretation of (52), it might be the case that the doubling of KENDİ functions in a similar way to focus doubling (Makaroğlu, 2012). The second interpretation, however, provides clear evidence that KENDİ can serve distinct functions, which can be combined within the same sentence.

Conclusion and Future Directions

In this paper, we discussed the co-referential properties and the syntactic category of the sign KENDİ in TİD, which we argue can serve both a traditional reflexive and emphatic function. We have shown that KENDİ in TİD can function as a reflexive in the traditional sense and we have used syntactic typologies to classify it as a reflexive of the DP type. As a reflexive, KENDİis subject to structural antecedence requirements. As a DP-type reflexive, KENDİ is able to serve functions outside of traditional reflexivity.

Importantly, this study lays the groundwork for further analyses of KENDİ as well as reflexivity in TİD and other sign languages in general. With respect to TİD, future research can expand our understanding of (i) the shared and different properties of the traditional reflexive and emphatic function of KENDİ, (ii) whether these properties are associated with distinct merge positions in the sentential structure, (iii) if logophoric usages of KENDİ are possible, (iv) non-manual characteristics of these distinct functions functions, and (v) potential language contact and bilingualism effects, among others. Moreover, future research can explore these issues in other signed languages, and further contribute to a cross-modal understanding of how co-reference is encoded.

Data Availability Statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

Ethics Statement

The studies involving human participants were reviewed and approved by Institutional Review Board of University of Michigan. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements. Written informed consent was obtained.

Author Contributions

DK conducted the fieldwork and drafted the manuscript, with critical revisions provided by NA. Both authors discussed the literature reviewed, designed the content and structure of the article, and approved the final version of the manuscript.


This research was funded by the University of Michigan Linguistics Department and Rackham Graduate School.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.


We would like to thank Gül Yiğit and Elvan Tamyürek Özparlak for participating in fieldwork interviews. Earlier versions of this work were presented at the Syntax-Semantics Discussion Group of the University of Michigan Department of Linguistics and at the 2021 Formal and Experimental Advances in Sign Language Theory conference. We would like to thank audience members at these events for their valuable feedback and discussion. Lastly, we would also like to thank Kim-Van Pham for her help with the formatting.


  1. ^ Koulidobrova (2009, 2011), Fischer and Johnson (2012), and Mathur (1996) observe that the reflexive pronoun “SELF” in American Sign Language also serves other functions. We discuss this phenomenon more in section “The Function of KENDİ Beyond Traditional Reflexivity”.
  2. ^ Note that the suffix noted here as -un is in fact {-(I)n} and becomes -un in this particular example due to vowel harmony (Göksel and Kerslake, 2004).
  3. ^ The subscripts a, b, c are used to express locus, and the subscripts i, j, k are used to express co-reference throughout the glosses. For the examples in which both locus and co-reference is marked, they are used as “a_i”.
  4. ^ Note that the string in (15b) could be grammatical as an intransitive structure in which the IX3sg that follows the subject is a post-nominal determiner that sets the locus for the subject (vs. a pronominal object). These strings can be disambiguated with pausing and other non-manual cues.
  5. ^ See Pfau and Quer (2008) for a detailed discussion and analysis of manual and non-manual negation cliticization using data from Catalan Sign Language (LSC) and German Sign Language (DGS).
  6. ^ Fv: final vowel. The final vowel on verbs alternate based on the features of its object in Shona, and therefore marked in glosses the same way it was marked in the cited source. See Storoshenko (2009) for a more detailed explanation.
  7. ^ Handshape may serve as a kind of gender agreement marker in Japanese Sign Language, Taiwan Sign Language, and Korean Sign Language, but these too spatialize when marking verbs (Fischer, 1996; Zeshan, 2006).
  8. ^ İşsever and Makaroğlu (2017) observes sentences like (25b), where KENDİ co-occurs with a verb that is marked for reflexiv(ized) agreement, are ungrammatical, and account for it by the Anaphor-Agreement Effect. However, one should keep in mind that their data comes from signers from Ankara, and ours from İstanbul, indicating there might be dialectal variation in this. Note that Kimmelman (2009) also observes an anaphoric reflexive being accompanied by a verb marked by reflexive agreement in RSL.
  9. ^ Moreover, comparing (a-b) below, we see in (b) that the movement for the verb starts from a locus in neutral signing space and moves toward the signer’s body, which is being used as a stand in for non-first person (in addition to its use as shifted first person under role shift, Meir et al., 2007). a. IX3sg IX3sg 3SUPPORT3   b. IX3sg KENDİ SUPPORT1. “He supports him.”             “He supports himself.”
  10. ^ Dikyuva et al. (2017:204) observe that KENDİ is accompanied by a non-manual “op” mouth gesture when used as a possessor, and that it is a restricted to valuable and precious entities. Though we leave the “op” mouth gesture as a matter for future research, we do note here that our consultants did not share the intuition that possessive KENDİ is incompatible with ordinary noun possessees, as illustrated in (a-b). The reader is referred to Barker (1991) for a discussion of semantic issues like these in possessive structures, and to Abner (2013) for a discussion of these issues in ASL. a. NOYANi_a IXa KENDİi BOOK aGIVE1. “Noyan gave his own book to me.” b. NOYANi_a IXa KENDİi BOOK KENDİi STUDENTb IXb aGIVEb “Noyan gave his own book to his own student.”
  11. ^ See Ahn (2010:10) for a discussion of how the plain adverbial emphatic himself in (39b) compares to the by himself phrasal modifier.
  12. ^ As noted, that context can disambiguate KENDİ with no added structural cue necessary: Context: There was an election in the university to be the department chair. Aslıi and Meltemj are candidates. Aslıi is the former chair. ASLIi ONCE-MORE DEPARTMENT CHAIR BE WANT. IX3sg KENDİi VOTE. IX3sg_i MELTEMj VOTE NOT. “Aslı wants to be the department chair again. She voted for herself, didn’t vote for Meltem”.


Abner, N. (2013). Gettin’together a posse: the primacy of predication in ASL possessives. Sign Lang. Linguist. 16, 125–156. doi: 10.1075/sll.16.2.02abn

PubMed Abstract | CrossRef Full Text | Google Scholar

Ahn, B. T. (2010). Not Just Emphatic Reflexives Themselves: Their Syntax, Semantics and Prosody. [Ph.D. Dissertation]. Los Angeles, CA: UCLA.

Google Scholar

Bahan, B., Kegl, J., Lee, R. G., MacLaughlin, D., and Neidle, C. (2000). The licensing of null arguments in American Sign Language. Linguist. Inq. 31, 1–27.

Google Scholar

Barker, C. (1991). Possessive Descriptions. Santa Cruz: University of California.

Google Scholar

Branchini, C. (2020). “Syntax: 2.1 The syntactic realization of argument structure,” in A Grammar of Italian Sign Language (LIS), 1st Edn, eds C. Branchini and M. Lara (Amsterdam: University of Amsterdam). doi: 10.1515/9783110757255-001

CrossRef Full Text | Google Scholar

Büring, D. (2005a). Bound to bind. Linguist. Inq. 36, 259–274. doi: 10.1162/0024389053710684

CrossRef Full Text | Google Scholar

Büring, D. (2005b). Binding Theory. Cambridge, MA: Cambridge University Press.

Google Scholar

Chomsky, N. (1981). “Principles and parameters in syntactic theory,” in Explanation in Linguistics, eds N. Hornstein and D. Lightfoot (London: Longman), 32–75.

Google Scholar

Cicilianić, T. A., and Wilbur, R. B. (2006). Pronominal system in Croatian sign language. Sign Lang. Linguist. 9, 95–132. doi: 10.1075/sll.9.1.07ali

PubMed Abstract | CrossRef Full Text | Google Scholar

Cormier, K. (2012). “11. Pronouns,” in Sign Language, eds R. Pfau, M. Steinbach, and B. Woll (Berlin: De Gruyter Mouton), 227–244.

Google Scholar

Cormier, K., Schembri, A., and Woll, B. (2013). Pronouns and pointing in sign languages. Lingua 137, 230–247. doi: 10.1016/j.lingua.2013.09.010

CrossRef Full Text | Google Scholar

Costello, B. D. N. (2016). Language and Modality: Effects of the Use of Space in the Agreement System of Lengua De Signos Española (Spanish Sign Language). [Doctoral dissertation]. Amsterdam: University of Amsterdam. doi: 10.1075/sll.19.2.06cos

PubMed Abstract | CrossRef Full Text | Google Scholar

Déchaine, R. M., and Wiltschko, M. (2017). A formal typology of reflexives. Studia Linguist. 71, 60–106. doi: 10.1111/stul.12072

CrossRef Full Text | Google Scholar

Desloges, P. (1779). Observations D’un Sourd Et Muet, Sur Un Cours Élémentaire D’éducation Des Sourds Et Muets, Published in 1779 by M. l’Abbé Descham (Chapelain de l’Église d’Orléans). Paris: Good Press.

Google Scholar

Dikyuva, H., Makaroğlu, H., and Arık, E. (2017). Turkish Sign Language Grammar. Ankara: Ministry of Family and Social Policies Press.

Google Scholar

Faltz, L. M. (2016/1977). Reflexivization: A Study in Universal Syntax. London: Routledge.

Google Scholar

Fischer, S., and Johnson, R. (2012). Nominal markers in ASL. Sign Lang. Linguist. 15, 243–250. doi: 10.1016/j.cognition.2014.11.029

PubMed Abstract | CrossRef Full Text | Google Scholar

Fischer, S. D. (1996). The role of agreement and auxiliaries in sign language. Lingua 98, 103–119. doi: 10.1016/0024-3841(95)00034-8

CrossRef Full Text | Google Scholar

Frajzyngier, Z., and Walker, T. (eds) (2000). Reflexives: Forms and Functions, Vol. 1. Amsterdam: John Benjamins Publishing.

Google Scholar

Geniusiene, E. (2011). The Typology of Reflexives. Berlin: De Gruyter Mouton.

Google Scholar

Gökgöz, K. (2009). Topics in Turkish Sign Language (Türk İaret Dili-TİD) Syntax: Verb Movement, Negation and Clausal Architecture. M.A. thesis. İstanbul: Boğaziçi University.

Google Scholar

Gökgöz, K. (2020). “Syntax: 4.2.2 The position of the possessive pronoun,” in A Grammar of Turkish Sign Language (TİD), 1st Edn, ed. M. Kelepir (Berlin: De Gruyter Mouton).

Google Scholar

Gökgöz, K., and Keleş, O. (2020). “Pragmatics: 4.2. Topic,” in A Grammar of Turkish Sign Language (TİD), 1st Edn, ed. M. Kelepir (Berlin: De Gruyter Mouton).

Google Scholar

Gökgöz, K., Sevgi, H., and Kayabaşı, D. (2020). “Syntax: Manual verb agreement,” in A Grammar of Turkish Sign Language (TİD), 1st Edn, ed. M. Kelepir (Berlin: De Gruyter Mouton).

Google Scholar

Göksel, A., and Kerslake, C. (2004). Turkish: A Comprehensive Grammar. London: Routledge.

Google Scholar

Goldin-Meadow, S., Cook, S. W., and Mitchell, Z. A. (2009). Gesturing gives children new ideas about math. Psychol. Sci. 20, 267–272. doi: 10.1111/j.1467-9280.2009.02297.x

PubMed Abstract | CrossRef Full Text | Google Scholar

Haegeman, L. (1994). Introduction to Government and Binding Theory. Wiley-Blackwell.

Google Scholar

Hellan, L. (1988). Anaphora in Norwegian and the Theory of Grammar. Dordrecht: Foris.

Google Scholar

İşsever, S., and Makaroğlu, B. (2017). “TİDde göndergesel nesneler ve uyum,” in Proceedings of the 31st National Linguistics Conference, (Ankara: Ankara University).

Google Scholar

Kayabaşı, D., and Gökgöz, K. (in press). The causative-inchoative alternation and age-of-acquisition effects on multi-predicate constructions in turkish sign language. Lang. Learn. Dev.

Google Scholar

Kayabaşı, D., Sevgi, H., and Özsoy, A. S. (2020). “Null arguments in Turkish sign language,” in Morphological Complexity within and across Boundaries: In honour of Asli Göksel, eds A. Gürer, D. Uygun-Gökmen, and B. Öztürk (Amsterdam: John Benjamins Publishing Company). doi: 10.1075/slcs.215.14kay

PubMed Abstract | CrossRef Full Text | Google Scholar

Kehler, A., and Büring, D. (2007). “Be bound or be disjoint,” in Proceedings of the 38th Meeting of the North East Linguistics Society, NELS 38, (Ottawa: University of Ottawa).

Google Scholar

Kelepir, M. (ed.). (2020a). A Grammar of Turkish Sign Language (TİD), 1st Edn. Brussels: European Commission.

Google Scholar

Kelepir, M. (2020d). “Morphology: manual markers,” in A Grammar of Turkish Sign Language (TİD), 1st Edn, ed. M. Kelepir (Brussels: European Commission).

Google Scholar

Kelepir, M. (2020b). “Morphology: 3.1. agreement,” in A Grammar of Turkish Sign Language (TİD), 1st Edn, ed. M. Kelepir (Brussels: European Commission).

Google Scholar

Kelepir, M. (2020c). “Morphology: object markers,” in A Grammar of Turkish Sign Language (TİD), 1st Edn, ed. M. Kelepir (Brussels: European Commission). doi: 10.1016/j.fct.2020.111490

PubMed Abstract | CrossRef Full Text | Google Scholar

Kemmer, S. (1995). “Emphatic and refexive-self: expectations, viewpoint, and subjectivity,” in Subjectivity and Subjectivisation: Linguistic Perspectives, eds D. Stein and S. Wright (Cambridge: Cambridge University Press), 55–82.

Google Scholar

Kimmelman, V. (2008). Sketch of Anaphora in RSL in comparison with ASL. Brussels: European Commission.

Google Scholar

Kimmelman, V. (2009). Reflexive Pronouns in Russian Sign Language and Sign Language of the Netherlands. [M.A. Thesis]. Amsterdam: University of Amsterdam.

Google Scholar

Kimmelman, V. (2018). Null arguments in Russian sign language. Form. Exp. Adv. Sign Lang. Theory 1, 27–38. doi: 10.1007/s11049-019-09448-9

CrossRef Full Text | Google Scholar

König, E., Siemund, P., and Töpper, S. (2013). “Intensifiers and reflexive pronouns,” in The World Atlas of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology, eds M. S. Dryer and M. Haspelmath (Munich: Max Planck Society).

Google Scholar

Koulidobrova, E. (2009). “SELF: intensifier and ‘long distance effects in ASL,” in Proceedings of ESSLLI, (Rīga: University of Latvia).

Google Scholar

Koulidobrova, E. (2011). SELF: Intensifier and ‘Long Distance Effects in American Sign Language (ASL). Ms Thesis. Storrs, CT: University of Connecticut.

Google Scholar

Koulidobrova, E. (2012). Parallelism revisited: The nature of the null argument in ASL as compared to the Romance-style pro. Sign Lang. Linguist. 15, 259–270. doi: 10.1075/sll.15.2.07kou

PubMed Abstract | CrossRef Full Text | Google Scholar

Liddell, S. K. (1980). American Sign Language Syntax. The Hague: De Gruyter Mouton.

Google Scholar

Lillo-Martin, D. (1995). “The point of view predicate in Americal sign language,” in Language, Gesture and Space, eds K. Emmorey and J. Reilly (Hillsdale, NJ: Lawrence Erlbaum Associates), 155–170.

Google Scholar

Lillo-Martin, D. (1986). Two kinds of null arguments in American sign language. Nat. Lang. Linguist. Theory 4, 415–444.

Google Scholar

Lillo-Martin, D., and Meier, R. P. (eds) (2011). On the linguistic status of ‘agreement’in sign languages. Theor. Linguist. 37, 95–142. doi: 10.1515/thli.2011.009

PubMed Abstract | CrossRef Full Text | Google Scholar

Loos, C. (2020). “Syntax: 2.1 The syntactic realization of argument structure,” in A Grammar of German Sign Language (DGS), 1st Edn, eds S. Prose, D. Nuhbalaoglu, A. Herrmann, J. Hosemann, and M. Steinbach (Brussels: European Commission).

Google Scholar

Makaroğlu, B. (2012). Türk İşaret Dili’nde soru: Kaş hareketlerinin dilsel çözümlemesi [Question in Turkish Sign Language: Linguistic Analysis of Eyebrows]. [M.A. Thesis]. Ankara: Ankara University.

Google Scholar

Makaroğlu, B., and Dikyuva, H. (eds) (2017). The Contemporary Turkish Sign Language Dictionary. Ankara: The Turkish Ministry of Family and Social Policy.

Google Scholar

Mantovan, L. (2020). “Lexicon: 3.7.4. Reflexive and reciprocal pronouns,” in A Grammar of Italian Sign Language (LIS), 1st Edn, eds C. Branchini and M. Lara (Brussels: European Commission).

Google Scholar

Mathur, G. (1996). A Pre-Suppositionality Marker in ASL. MS. Thesis. Cambridge, MA: MIT Press.

Google Scholar

Mehling, K. (2010). Heute Hier, Morgen Dort-Deixis in der Deutschen Gebärdensprache (DGS). [PhD Dissertation]. München: Ludwig Maximilian University of Munich.

Google Scholar

Meir, I. (1998). Syntactic-semantic interaction in Israeli Sign Language verbs: the case of backwards verbs. Sign Lang. Linguist. 1, 3–37. doi: 10.1075/sll.1.1.03mei

PubMed Abstract | CrossRef Full Text | Google Scholar

Meir, I., Padden, C. A., Aronoff, M., and Sandler, W. (2007). Body as subject. J. Linguist. 43, 531–563. doi: 10.1017/S0022226707004768

PubMed Abstract | CrossRef Full Text | Google Scholar

Navarrete-González, A., and Busto, R. V. Barberà, G., Zorzi, G.Cañas-Peña, S. (2020). “Lexicon: 3.7.4. reflexive and reciprocal pronouns,” in A Grammar of Catalan Sign Language (LSC), 1st Edn, eds J. Quer and B. Gemma (Brussels: European Commission), 230–237.

Google Scholar

Özsoy, A. S. (1983). Reflexivization in Turkish: A Syntactic, Semantic and Discourse Analysis. [PhD Dissertation]. Ann Arbor, MI: University of Michigan.

Google Scholar

Pace, A., Levine, D., Morini, G., Hirsh-Pasek, K., and Golinkoff, R. M. (2016). “The story of language acquisition: from words to world and back again,” in Child Psychology: A Handbook of Contemporary Issues, 3rd Edn, eds L. Balter and C. Tamis-LeMonda (Hove: Psychology Press), 43–79.

Google Scholar

Padden, C. (1983). Interaction of Morphology and Syntax in American Sign Language. [PhD dissertation]. San Diego, CA: University of California.

Google Scholar

Padden, C. (1988). Interaction of Morphology and Syntax in American Sign Language. New York, NY: Garland.

Google Scholar

Pfau, R. (2005). “Phrasal layers and prosodic spreading in sign languages,” in Paper Presented at the SIGNA VOLANT Workshop, (Toulouse: Milan).

Google Scholar

Pfau, R. (2006). “Visible prosody: spreading and stacking of non-manual markers in sign languages,” in Proceedings of the 25th West Coast Conference on Formal Linguistics, Cascadilla Proceedings Project, Seattle.

Google Scholar

Pfau, R., and Quer, J. (2008). “On the syntax of negation and modals in Catalan Sign Language and German Sign Language,” in Visible Variation, eds P. Perniss, R. Pfau, and M. Steinbach (Berlin: De Gruyter Mouton), 129–162.

Google Scholar

Pinker, S. (2013). Language, Cognition, and Human Nature: Selected Articles. Oxford: Oxford University Press.

Google Scholar

Quer, J., and Steinbach, M. (2015). Ambiguities in sign languages. Linguist. Rev. 32, 143–165. doi: 10.1515/tlr-2015-0001

CrossRef Full Text | Google Scholar

Reinhart, T. (1983). Coreference and bound anaphora: a restatement of the anaphora questions. Linguist. Philos. 6, 47–88. doi: 10.1007/bf00868090

CrossRef Full Text | Google Scholar

Reinhart, T., and Reuland, E. (1993). Reflexivity. Linguist. Inq. 24, 657–720.

Google Scholar

Reinhart, T. M. (1976). The Syntactic Domain of Anaphora. Doctoral dissertation. Cambridge, MA: Massachusetts Institute of Technology.

Google Scholar

Reuland, E. (2011). Anaphora and Language Design. Cambridge, MA: MIT Press.

Google Scholar

Sandler, W. (2012). Dedicated gestures and the emergence of sign language. Gesture 12, 265–307.

Google Scholar

Sandler, W., and Lillo-Martin, D. (2006). Sign Language and Linguistic Universals. Cambridge, MA: Cambridge University Press. doi: 10.1017/CBO9781139163910

CrossRef Full Text | Google Scholar

Schlenker, P. (2018). Visible meaning: sign language and the foundations of semantics. Theor. Linguist. 44, 123–208. doi: 10.1515/tl-2018-0012

PubMed Abstract | CrossRef Full Text | Google Scholar

Sevinç, A. M. (2006). Grammatical Relations and Word Order in Turkish Sign Language (TİD). [M.A. Thesis]. Dumlupınar: Middle East Technical University.

Google Scholar

Stern, N. (2004). The semantic unity of reflexive, emphatic, and other-self pronouns. Am. Speech 79, 270–280.

Google Scholar

Stokoe, W. C. (1960). Sign Language Structure. Studies in Linguistics: Occasional Paper 8. Berlin: De Gruyter.

Google Scholar

Storoshenko, D. R. (2009). “Investigating the Shona reflexive zvi,” in Proceedings of the 39th Annual Conference on African Linguistics (ACAL), eds A. Ojo and L. Moshi (Somerville, MA: Cascadilla Proceedings Project), 42–55.

Google Scholar

Thráinsson, H. (1976). “Reflexives and subjunctives in Icelandic,” in Proceedings of the Sixth Annual Meeting of the Northeastern Linguistics Society, (Brussels: European Commission), 225–239.

Google Scholar

Wilbur, R. (1996). “Focus and specificity in ASL structures containing SELF,” in Proceedings of the Winter Meeting of the LSA, San Diego, CA.

Google Scholar

Wilkinson, E. (2013). A functional description of SELF in American Sign Language. Sign Lang. Stud. 13, 462–490. doi: 10.1353/sls.2013.0015

PubMed Abstract | CrossRef Full Text | Google Scholar

Zeshan, U. (2002). Sign language in Turkey: the story of a hidden language. Turk. Lang. 6, 229–274.

Google Scholar

Zeshan, U. (2003). Aspects of Türk Isaret Dili (Turkish Sign Language). Sign Lang. Linguist. 6, 43–75. doi: 10.1075/sll.6.1.04zes

PubMed Abstract | CrossRef Full Text | Google Scholar

Zeshan, U. (2004). Hand, head and face-negative constructions in sign languages. Linguist. Typol. 8, 1–58. doi: 10.1515/lity.2004.003

CrossRef Full Text | Google Scholar

Zeshan, U. (2006). “Sign language of the world,” in Encyclopedia of Language and Linguistics, Vol. 11, eds K. Brown and R. E. Asher (Amsterdam: Elsevier), 358–365. doi: 10.1016/B0-08-044854-2/00243-1

CrossRef Full Text | Google Scholar

Zribi-Hertz, A. (1989). Anaphor binding and narrative point of view: english reflexive pronouns in sentence and discourse. Language 65, 695–727. doi: 10.2307/414931

CrossRef Full Text | Google Scholar

Zribi-Hertz, A. (1995). Emphatic or reflexive? On the endophoric character of French lui-m^eme and similar complex pronouns. J. Linguist. 31, 333–337. doi: 10.1017/s0022226700015632

CrossRef Full Text | Google Scholar

Keywords: Turkish Sign Language (TİD), reflexives, emphatic reflexives, co-referential relations, anaphora, typology, sign languages, syntax

Citation: Kayabaşı D and Abner N (2022) On the Reflexive KENDİ in Turkish Sign Language. Front. Psychol. 13:753455. doi: 10.3389/fpsyg.2022.753455

Received: 04 August 2021; Accepted: 12 January 2022;
Published: 26 April 2022.

Edited by:

Sabina Fontana, University of Catania, Italy

Reviewed by:

Seçkin Arslan, University of Groningen, Netherlands
Brendan Costello, Basque Center on Cognition, Brain and Language, Spain
Dave Kush, University of Toronto Scarborough, Canada
Vadim Kimmelman, University of Bergen, Norway

Copyright © 2022 Kayabaşı and Abner. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Demet Kayabaşı,

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.