Abstract
We examine spontaneous production data from the dialect of Modern West Thracian Greek (mwtg) (the local dialect of Evros) with regard to a hypothesis of syntactic borrowing of verbal transitivity. We argue that mwtg allows omission of the direct object with specific reference, in contrast to Standard Modern Greek (smg) and other Modern Greek (mg) dialects (spoken in Greece), but similar to Turkish. Object omission in mwtg is possible only in contexts where smg and other mg dialects show obligatory use of the 3rd-person clitic. We argue that syntactic borrowing in the case of language contact follows the transfer with second language learners: the relevant elements that host uninterpretable features are used optionally. Moreover, the definite article, in contrast to the indefinite article, is also affected by language contact. The 3rd-person clitic and the definite article are affected by contact as uninterpretable clusters of features. We claim that interpretability plays a significant role in transitivity in cases of language contact.
1 Introduction
In this study, we examine spontaneous production data from the dialect of Modern West Thracian Greek (mwtg), the local dialect of Evros, 1 in order to evaluate the formal factors involved in the integration of transitivity patterns into this dialect. We will show that the mwtg dialect, as represented in data from the local dialect of Evros, allows the omission of the direct object [object drop] with specific reference, in contrast to Modern Standard Greek (msg), and other Modern Greek (mg) dialects (spoken in Greece), but similar to Turkish. However, object omission is possible only in contexts where msg (and other mg dialects) shows obligatory use of the 3rd-person clitic. In this respect, we will argue that syntactic borrowing follows the characteristics of crosslinguistic transfer in cases of bilingualism and second language (L2) acquisition (according to the Interpretability Hypothesis; Tsimpli 2003a, b; Tsimpli and Dimitrakopoulou 2007; Tsimpli and Mastropavlou 2008). Accordingly, this study is the first attempt to include data from syntactic borrowing due to language contact in the comparison of production data from different populations, such as L2 learners, first language (L1) learners, specific language impairment (sli) speakers, and agrammatics. The aim of this comparison is an analysis that would account for the patterns of transfer / borrowing observed in the speech of different populations (for instance, cases of language contact as well as cases of second language acquisition or of specific language impairment) in a manner that can maximize similarities between the different populations (Tsimpli and Mastropavlou 2008; Genesee et al. 2004; Rice and Wexler 1996).
Western Thrace is considered a complex linguistic area of century-long presence of many linguistic communities (Adamou 2010, 2012). It is the area of a bilingual Greek-Turkish linguistic community and trilingual linguistic communities (Pomak
2
-Turkish-Greek, Romani
3
-Turkish-Greek and Armenian-(Turkish)-Greek), but also of a Greek dialect in constant strong contact with msg and Turkish (cf. Table 1; see also Section 2).
We follow Heine and Kuteva (2005) with regard to the concept of “borrowing”: with borrowing, we describe cases of transfer of forms, meanings, or constructions into a recipient language. Our study concerns cases of pat-borrowing [pattern borrowing] in terms of Matras and Sakel (2007): that is, cases where patterns (the distribution and mapping of meaning, but not the form) of one language are replicated in another language. 5 Borrowings can be found in the speech of monolingual speakers –in contrast to codeswitching phenomena. 6 But their origins can be in the speech of bilinguals and they need long-term conditions of contact for their introduction in the recipient language (Matras 2009). Matras and Sakel (2007), for instance, refer to the possibility of loss of features or categories (besides the various cases of grammaticalization) due to language contact. They do not discuss cases of object omission but do mention the reduction or loss of articles in dialects of Romani in contact with Polish and Russian. 7 Our case study also reflects a historical contact-induced innovation, 8 one that appeared at a time when Greek-Turkish bilingualism was more widespread in the area, either among Greek speakers of the dialect under examination or among speakers of Turkish, or even both, and later spread to monolingual mwtg speakers or speakers that were dominant in mwtg. See also fn. 6.
Lexical borrowing as a cause of language change has been well examined for several stages of the history of the Greek language (cf., among many others, Apostolou-Panara 1991, Christidis 2001). Syntactic borrowing is less examined: There are several studies on the Differential Object Marking (dom); however, the borrowing of verbal properties is understudied. Moreover, most of the past studies concern mainly nominal properties as well as borrowing into Greek dialects that are spoken outside Greece, e.g., Cappadocian Greek (for instance, for clitics and case morphology, see Janse (2001, 2006); see Spyropoulos and Tiliopoulou (2006) on case and definiteness in Cappadocian Greek, Spyropoulos and Kakarikos (2011) on Cappadocian Greek nominal inflection, among many others). 9
In Section 2, we describe language contact in the area of Greek Thrace, focusing on the use of Turkish and the Greek dialects spoken in this area. Section 3 presents the background on object clitics and their omission. We also analyze how object omission in cases of L2 acquisition can be explained with the Interpretability Hypothesis, and we state the hypothesis of the study with regard to language contact and object omission in the mwtg of Evros. Section 4 presents the data of the study, which are taken from a corpus of spontaneous production data from dialect speech. We show that object omission in mwtg is allowed only in the case of 3rd-person object clitics. Similarly, only omission of the definite article is permitted (the indefinite articles are not omitted). Individual variation and variation in the data from different villages are discussed. The data from the dialect are compared to results from relevant L2 studies. Section 5 summarizes the main conclusions of the study.
2 Language Contact in Greek Thrace
The history of language contact between Greek and Turkish in the area of Greek Thrace is long, including two significant historical periods: (a) the period when the Ottoman Empire used Turkish as the lingua franca of the Balkans and (b) the period after the recognition by the Greek state (in 1923) of Turkish as a minority language for Greek Thrace. The communities of Muslims of Greek Thrace, i.e., speakers of Turkish, Pomak and Romani, were exempted from the population exchange after World War I, according to the terms of the Lausanne Treaty of 1923, thereby gaining the right to bilingual education. As a result, Turkish became the current language of education, together with Greek, for the Muslim populations of the area. The acknowledgment of the right to bilingual Greek-Turkish schools also affected the non-Turkish linguistic communities of the area: Pomak and Romani speakers (of the Muslim religion) became bilingual speakers of Turkish, to an extent. The result is a homogenization process that made Turkish the first language of many Muslim communities of the area. This homogenization process resulted in a shift of Pomak and Romani speakers to Turkish (Sella-Mazi 1997, Adamou 2010).
Similarly to the Greek dialect of the area, which is the focus of this study, the local Romani and Pomak dialects were also influenced by Turkish. Komotini and Xanthi Romani, for instance, include many Turkish loan verbs that were borrowed with the Turkish person markers. 10 On the other hand, Pomak shows only a few cases of loan verbs (Adamou 2010, 2012). Sella-Mazi (1997) describes the linguistic situation in Greek Thrace as a combination of bilingualism and diglossia. 11 According to her, Turkish is mainly used by older speakers of Romani and Pomak, even when they speak to Greek-speaking people.
(1)kupardo ‘break sth.’ [koparmak]; divirdo ‘overturn, turn upside down’ [devirmek]; pis’manevu ‘repent, feel sorry’ [pişman olmak]; siasirdizu ‘be astonished’ [şaşırmak]; suklado ‘become cold, get cold’ [soğumak]
As with all cases of language variation, the movements of populations cause the emergence of a complex linguistic situation. The mg dialects of Western Thrace have been analyzed as the result of movements of Greek speakers, mainly from Eastern Thrace but also from Epirus, together with characteristics of local Greek dialects (Batzakas 2012). The local dialects are also influenced strongly by msg and, to a degree, by Turkish as spoken by the Muslim communities of speakers of Turkish and trilingual speakers of Pomak or Romani, who also speak Greek and Turkish. The complex status of mwtg is reflected in its vocabulary, which contains words from Koine Greek, from the mg dialects of Eastern Thrace, and loanwords from Turkish and Pomak.
We should notice the absence of studies on the mg dialects of Thrace – except for Iasmos, a Thracian dialect of Rodopi. The dialect studies that mention characteristics of the mg Thracian dialects refer to the possibility of absence of the article (among others, Kontosopoulos [1994] 2006; see below, Section 4.3, about the relationship between the system of articles with object clitics and clitic omission) – but language contact or other factors that may result in the absence of articles are not discussed in these studies.
In Section 3, we discuss the theoretical background on object omission and (un)interpretable features. We also present the Interpretability Hypothesis and we discuss what this hypothesis predicts for object omission in the case of language contact effects on mwtg.
3 Theoretical Background: Object Omission and the Interpretability Hypothesis
3.1 Object Omission
(2)- Do you like this article?
- *I love Ø.
(3)They have the ability to impress and delight Ø.
In msg (and the mg dialects that are spoken in Greece), only non-referential/indefinite objects can be dropped (Ex. 4), whereas the omission of an object with specific reference to a referential object is not permitted (Ex. 5) (Giannakidou and Merchant 1997; Tsimpli and Papadopoulou 2009).
msg requires the presence of an object clitic for direct objects that have specific reference 14 and disallows object omission for a referent that has been previously introduced.
The personal pronouns in msg can be strong (structurally independent personal) pronouns in the subject and object position and weak/clitic (structurally dependent personal) pronouns in the object position. Object clitics in msg precede the finite verb and are not stressed or focused (see 6). msg does not have subject clitics.
In contrast, Turkish
15
allows omission of objects with a specific reference, i.e., in a context where a referent has already been introduced into the discourse (Kornfilt 1987, 1997; Aygen 2001; Öztürk 2005, among others). See (7)–(8). Omission of an object with a specific reference can occur in Turkish, not only with 3rd-person pronouns but with 1st- and 2nd-person pronouns as well – even though there is a preference for omission of inanimate and nonhuman objects. An object pronoun, however, can appear in Turkish with a wide-focus question (Franks and King 2000). See (9).
Turkish lacks object clitics and weak pronouns (Franks and King 2000; Halpern and Fontana 1993). Similarly, Turkish lacks definite articles; case and number are marked as nominal suffixes. Moreover, bare nouns are allowed as arguments: bare direct objects are interpreted in most instances as nonspecific, whereas direct objects in the accusative are interpreted as specific (Kornfilt 1997; Enç 1991; Ketrez 2005; see (10)).
16
This means that specificity in Turkish is expressed by the use of the accusative case ending with the noun or by the introduction of a demonstrative. Nonspecificity is expressed by the absence of accusative case marking or by the presence of a numeral (but not when the modified phrase bears accusative marking: in this case, the numeral does not change the interpretation to non-specific).
17
In 3.2, we present the Interpretability Hypothesis and its relevance for language contact and, more specifically, object omission in contact settings.
3.2 (Un)interpretable Features and the Interpretability Hypothesis
Our approach is heavily based on the (un)interpretability status of formal features. Note that the following premises are essential for this approach: (a) The Minimalist Program distinction between interpretable and uninterpretable features, which is connected to how the features are read off by the lf (Logical Form) and pf (Phonetic Form) interfaces 18 ; (b) The observed developmental differences in accessibility of uninterpretable features between L1 and L2 learners and between typical and sli children; (c) The compensatory strategies that are used in the case of failure of appropriate analysis of the uninterpretable features (cf. Tsimpli and Mastropavlou 2008).
According to the Minimalist Program (Chomsky 1995, 2001), formal features are essential for the syntactic derivation as well as for the two interfaces, pf and lf. Cf. Figure 1. The Principle of Full Interpretation (fi), which determines whether a representation converges or not, is met when all features of the representation are interpretable at lf (they have semantic import, they contribute to the interpretation of the sentence). For instance, the [definiteness] or the phi-features (person, number, gender) on nouns are interpretable at lf. Phi-features on verbs and adjectives or Case are uninterpretable at lf. The uninterpretable features participate in the syntactic derivation only and are erased before lf. Accordingly, the phi-features on the verb are matched with the phi-features of the subject, and then they are erased.
The pf and lf Interfaces (Chomsky 1995, 2001)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
The pf and lf Interfaces (Chomsky 1995, 2001)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
The pf and lf Interfaces (Chomsky 1995, 2001)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
In this respect, Tsimpli (2003a, b and in subsequent work) has stated the Interpretability Hypothesis: uninterpretable features are inaccessible to adult L2 learners, whereas interpretable features are accessible. The misanalysis of the uninterpretable features, and the use of compensatory strategies by L2 learners in cases where the input is misanalysed, provide evidence that the uninterpretable features are inaccessible; the uninterpretable features cause learnability problems and make the relevant constructions more vulnerable to L2. In contrast, L1 learners have access to both uninterpretable and interpretable features from early on. According to the Interpretability Hypothesis, (i) the distinction between interpretable and uninterpretable features has consequences in L1 – L2 acquisition (as well as in the differences between typical and sli children), because uninterpretable features are not accessible after the critical period of acquisition; and (ii) the uninterpretable features, therefore, may not be analyzed correctly, and compensatory strategies are implemented.
For instance, Tsimpli and Mastropavlou (2008) have shown that the lack of accessibility to uninterpretable features results in a misanalysis of the 3rd-person clitic as well as the definite article in advanced L2 grammars. The nontarget performance with regard to the 3rd-person clitic and the definite article has been explained, according to Tsimpli and Mastropavlou’s analysis, in terms of the assignment of the interpretable feature of [referentiality/definiteness] on these elements by the advanced L2 grammar: 19 this misassignment occurs because the interpretable features are accessible (in all stages of language development) as well as necessary (because clitics and articles encode uninterpretable features that are not accessible). This is a type of a compensatory strategy: interpretable features are exploited in advanced L2 grammars, in other words, interpretable features are assigned on a misanalysed input. The particular features are drawn “from the repertoire of interpretable features that can be associated with the function of these elements” (Dimitrakopoulou et al. 2004). For instance, referentiality is a possible feature for clitics and specificity for articles. Similarly, the sli grammars show optionality with regard to the use of the overt morphology, in contrast to typical grammars, where the mapping between the formal features and their morphological realization is automatic.
According to Tsimpli, the role of the (un)interpretable features does not result in the absolute absence (and no transfer at all) of elements that bear uninterpretable features. “Compensatory strategies” lead the learnability process. For instance, the 3rd-person clitics and the definite articles would not be analyzed in a target way by the L2 grammar: the result would be the optional use of the relevant elements. In reality, this optionality is a superficial optionality: there is an incorrect specification of features to the L2 input. In other words, after the failure of correct input analysis, features are specified in a nontarget manner, which constrains the use of the L2 elements. The result is the assignment of interpretable features to the problematic elements, so that their distribution can be more regular.
To provide another example, according to Tsimpli’s (1997, 2003a, b) L2 studies, Greek L2 speakers of English misanalyse the L2 input and use subject and object English pronouns as resumptive elements (see 11a, b). Greek L2 learners of English accept as grammatical these particular sentences, especially if the pronoun is inanimate: the feature of [animacy] is not an L1 (Greek) feature for interrogative pronouns but affects L2, showing a restructuring of the L2 grammar. The L2 input leads Greek learners of English to impose a constraint on resumption, which is based on the interpretable feature of [animacy]: Animacy is realized on L2 pronouns (English wh- and personal pronouns distinguish between [+/–animate]: who vs. what and he, she vs. it) but not on L1 (Greek does not mark animacy but gender contrasts on wh- and personal pronouns). Moreover, English does not allow resumptive pronouns in the position of the gap in subject or object wh-questions, whereas Greek optionally permits resumptive clitics that are coindexed with the wh-phrase. Since the animacy features have a semantic contribution (interpretable features), they are used by L2 learners, as a compensatory strategy, to form a constraint on resumption, which is related only to uninterpretable features.
Compensatory strategies can be also found in sli data: for instance, the morphological deficit of absence of [+/-plural] in English sli grammars has been compensated with segmental and prosodic strategies in the production of plural (Paradis and Gopnik 1994, 1997).
The reason for our decision to examine clitics together with articles derives from the syntactic analysis of both clitics and articles as being clusters of case and agreement features (see Section 4.3): according to the theoretical framework of the Minimalist Program, these features are considered uninterpretable at the lf interface (cf., for instance, Tsimpli and Stavrakaki 1999). Moreover, it has been shown, based on other differences within the paradigm of clitics and articles, that there is a further distinction between (i) 1st-/2nd- and 3rd-person accusative object clitics, (ii) definite and indefinite articles and (iii) 3rd-person genitive clitics in the nominal domain and 3rd-person accusative clitics in the verbal domain [we will not include the latter difference in our study, because it is not relevant to the transitivity issues under examination]. The 3rd-person clitic and the definite article encode values of uninterpretable features only, whereas the 1st-/2nd-person clitics have the interpretable feature of [person], and the indefinite article has the interpretable feature of [definiteness]. We claim that the results of syntactic borrowing can be similar to the results of transfer in cases of L2 acquisition (or to the characteristics of the speech of sli speakers) in terms of (un)interpretable features. This is the reason we refer to the Interpretability Hypothesis (we argue that syntactic borrowing in cases of contact shares common properties to the results of transfer in the speech of L2 learners, if analyzed in terms of (un)interpretable features). In this respect, our study is a type of expansion of the Interpretability Hypothesis, to cover the case of syntactic borrowing as well.
With regard to the Evros variety of mwtg, according to the Interpretability Hypothesis, we predict differentiation in the language performance of speakers of the Evros variety of mwtg in 3rd-person versus 1st-/2nd-person clitics, in a way similar to the definite and the indefinite articles. 20 Recall that interpretable features are always accessible; uninterpretable features, after the critical period, are not accessible to L2 learners. Therefore, 3rd-person clitics and definite articles follow different patterns than 1st-/2nd-person clitics and indefinite articles in L2 acquisition. 21 Based on the Interpretability Hypothesis, our hypothesis is that the same effects should appear for the dialect under contact with Turkish-Greek bilingual speakers. Language contact with Turkish-Greek bilingual communities can lead to syntactic borrowing of clitic and article omission.
It is evident that, even though the results of language contact and of L2 or sli performance with regard to the uninterpretable features may show similar patterns, the similarities have a different source (syntactic borrowing due to language contact in the case of contact between languages/dialects, versus L1 transfer in the case of L2 acquisition, versus genetically related causes in the case of sli grammars). If the syntactic borrowing in the case of language contact is similar to the transfer with L2 learners, the uninterpretable features would be used optionally (similar to the case of L2 learners). The reason for this optional use would be the misanalysis of the relevant input.
Following the model of L2 studies, we have conducted an analysis of contexts where the 3rd-person object clitic and the definite article are obligatorily used in msg and other mg dialects. Our aim is to identify the distribution of these elements in the dialect speech. Our hypothesis is that object clitics can be omitted in the Evros dialect of mwtg, in contrast to other mg dialects that are not in contact with Turkish, as a result of a borrowing from Turkish, via Turkish-Greek bilinguals. Note that object omission is not possible for all person and number nps in msg and other mg dialects and is possible for all person and number nps in Turkish. The interpretability distribution predicts that the speakers of the dialect, in the case of contact with Turkish (Turkish-Greek bilinguals), will omit the 3rd-person object clitics more frequently than the 1st- and 2nd-person object clitics. In a similar way, the prediction would be that the definite article would be omitted more frequently than the indefinite article in obligatory contexts. Furthermore, differences in the frequencies of omission are expected depending on the syntactic context, if the 3rd-person object clitic (and the definite article) are (misanalysed and) associated with interpretable features in the dialect. 22
In Section 4, we discuss the data of this study that derive from a corpus of spontaneous speech. According to the data, object omission in the Evros variety of mwtg is permitted only in the case of 3rd-person object clitics. We also analyze individual variation as well as variation in the data from different villages.
4 Data and Discussion
4.1 The Corpus of this Study
The data for this study include spontaneous speech data, elicited in collective interviews, from 18 adult speakers of the Evros dialect of mwtg and msg. These speakers also have a limited knowledge of Turkish (See Table 2). The data were collected in five villages in Evros, which is in Greek Thrace, with the support of Aristotle University of Thessaloniki, Research Committee Grant 88012 (“Syntactic Borrowing as a Source of Language Change in Modern West Thracian Greek”), by an ma student, Mrs. Theodora Papadopoulou. Cf. also Papadopoulou (2013).
The data were collected in the period between November 2012 and February 2013 and derive from spoken production in the context of informal collective interviews/ dialogues with the participation of 3–6 speakers of the dialect. The researcher who collected the data is also a speaker of the dialect, but her utterances are not taken into consideration and are not included in the annotated corpus. We have followed Dimitrakopoulou et al. (2004) and Tsimpli and Mastropavlou (2008) in the methodology of collection of the data. The collective interviews/ dialogues were digitally recorded and transcribed. 23 They consist of three parts: (i) questions about the subjects’ biographical information (education and occupation), as well as their exposure to the mwtg dialect and to Turkish; (ii) instruction-giving tasks; and (iii) general discussion/dialogue. See also Table 3 with the details of the corpus.
Evros: The five villages of the corpus [Source [modified]: <http://gym-metax.evr.sch.gr/location.php> and <commons.wikimedia.org>]
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Evros: The five villages of the corpus [Source [modified]: <http://gym-metax.evr.sch.gr/location.php> and <commons.wikimedia.org>]
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Evros: The five villages of the corpus [Source [modified]: <http://gym-metax.evr.sch.gr/location.php> and <commons.wikimedia.org>]
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Greek Thrace [Source:http://www.maps-of-greece.com/thrace-map.htm]
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Greek Thrace [Source:http://www.maps-of-greece.com/thrace-map.htm]
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Greek Thrace [Source:http://www.maps-of-greece.com/thrace-map.htm]
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
4.2 Object Omission in the Corpus
The results confirm that Evros mwtg allows omission of objects with specific reference, in contrast to msg and other mg dialects spoken in Greece, but similarly to Turkish. According to the data, 24 omission of an object with specific reference is possible only in contexts where msg and other mg dialects show obligatory use of 3rd-person clitics. In this respect, mwtg shows variation with regard to the interpretable features in contrast to uninterpretable features: it appears that only items hosting uninterpretable features (3rd-person clitics [or definite articles – see below] but not 1st-/2nd-person clitics [or indefinite articles – see below) are candidates to be affected by language contact.
Figure 2 shows that the omission of the object is possible only in contexts where msg would show obligatory use of 3rd-person clitics (omission in 29/619 [4.68%] of the cases), 25 and not in contexts of use of 1st- and 2nd-person (no cases of object omission). In (12)–(14), we present examples of omission of the 3rd-person object clitic as well as of presence of 3rd-person and 2nd-person object clitics from the corpus of the dialect.
Use and omission of clitics in our dialect corpus
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Use and omission of clitics in our dialect corpus
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Use and omission of clitics in our dialect corpus
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Omission of 3rd person clitic:
Presence of 3rd person clitic:
Presence of 2nd person clitic:
We have also distinguished three groups on the basis of the age, to test the role of age in the production of the attested dialect characteristics. What we expected was that the younger speakers would probably be more influenced by msg. However, younger speakers’ more intense contact with the Turkish-Greek bilingual community is also a plausible hypothesis. The results in Table 4 confirm the hypothesis of a strong influence by msg with regard to object omission, which is more frequent in the case of older speakers and less frequent in the case of younger speakers. 26 msg would disallow object omission in these particular contexts.
Furthermore, in most of the cases, the omitted clitic is the only (referential) object, whereas only a few cases show topicalization of an np with clitic omission (that is, a clitic doubling construction with clitic omission). The majority of the cases with clitic omission as the only referential object refer to an inanimate entity. In only 15% of the examples with a clitic as the only referential object, does the omitted object clitic refer to an animate entity (Table 5). In (15)–(17), we present examples of these syntactic contexts of object omission: clitic doubling (15–16); the clitic as the only (referential) object (17).
With regard to the differences between the data from the different villages (Table 6), one observes significant variation: for instance, Lefkimi behaves similarly to msg and other mg dialects in not allowing object omission. On the other hand, Metaxades shows a tendency for a more frequent object omission; the 3rd-person clitic is omitted in 9.76% of the cases where it is obligatorily present in msg. 28
Variation is observed within individuals as well. This variation is not only related to age differences (Tables 7a-e). For instance, A.B. (83 years old) does not omit the object in any obligatory (for msg) context, whereas A.C. (87 years old) omits the 3rd-person object clitic in 9.52% of the examples. The high frequency of object omission in the data from Metaxades has its main source in the data of one speaker: R.B. (77 years old). 29 On the other hand, C.M. (80 years old) is a male speaker who shows no object omission, following the msg rule. The data from Lykofi contain the production of many young speakers of the dialect. 4/6 of the speakers do not omit the object in any case. Most of them are young (31 or 37 years old) but N.A.’s and E.B.’s data (62 and 70 years old, respectively) also do not include any case of object omission. Furthermore, T.B.’s data (36 years old) show one example of object omission. The second male of the corpus (P.A., 75, Lagyna village) generally follows the rule of msg and permits the omission of the object in just 1.01% of the cases. The two female speakers from Lefkimi (70 and 41 years old) follow the msg pattern and do not allow omission of the object, even though one of them belongs to the older generation.
In Section 4.3, we discuss the (un)interpretable formal features of clitics as well as articles, which can reveal properties of the cases of their omission.
4.3 Uninterpretable Versus Interpretable Formal Features: The Case of Object Clitics and Articles
We adopt the analysis proposed by Tsimpli and Stavrakaki (1999) and we assume that msg clitics are members of the D system: that is, they are categorially determiners, but clitics appear in the verbal domain, in contrast to articles that appear in the nominal domain. Furthermore, msg clitics are “severely deficient” (cf. Cardinaletti and Starke 1999); they are deficient in their morphosyntax as well as in their semantics, if contrasted with the strong pronouns. The strong pronouns have a functional layer related to referentiality: in other words, they have a richer structure.
We cannot claim that the differences between the 1st-/2nd-person clitics and the 3rd-person clitic are due to the fact that the selection of the appropriate referent for 3rd-person clitics is not restricted, but is restricted to speaker/hearer in the case of 1st-/2nd-person clitics. If this were the case, the dissociation between 1st-/2nd-person and 3rd person should also concern strong pronouns: such evidence is not available in the relevant L2 and sli literature. On the contrary, we assume that the referent for 1st-/2nd-person pronouns is unambiguously defined due to the person feature that these pronouns include. We further claim that 3rd-person pronouns are reference-assigned after lf, in the domain of pragmatic processing (Sperber and Wilson 1995). The lack of [person] feature permits them to be nonreferential. Evidence for the underspecification of the [person] feature in 3rd-person pronouns derives from their use as expletives, where they are permitted to not refer at all. On the contrary, in the case of 1st- and 2nd-person pronouns (clitics or strong forms), the referent is defined sententially in an unambiguous manner (Tsimpli and Mastropavlou 2008).
As mentioned, clitics are members of the D system, but appear in the verbal domain, whereas articles appear in the nominal domain. For articles, it has been argued that definite articles include a cluster of uninterpretable agreement and case features (see 18) but not the [definiteness] interpretable feature. Indefinite articles also bear an interpretable feature – in addition to the uninterpretable features of case and phi-features, which concern the definite articles as well. Indefinite features make a semantic contribution and are specified as [-definite]: evidence for this contrast between definite articles (without semantic contribution of [+/- definite]) and indefinite articles derives from the fact that indefinite articles cannot be used as expletives, that is in proper names, nominalized clauses and other constructions where no definiteness is conveyed (Cf. Karanassios 1992; Stavrou 1996; Giusti 1997; Giannakidou and Stavrou 1999; Tsimpli and Stavrakaki 1999; Panagiotidis 2002; Tsimpli and Mastropavlou 2008; Chondrogianni et al. 2015).
The following characteristics have been argued to arise from the differences in the formal features of the definite and indefinite articles (Tsimpli and Stavrakaki 1999):
(i) Nominalizations of clauses or adverbials with definite (but not with indefinite) articles (see Ex. 19): The definite article can appear as a nominalizer of full clauses, adverbials and other categories. In these constructions, there is no definite reading at all but the definite article has an expletive function, carries case features, and allows the phrase that follows to have the role of a subject or object. The indefinite article can precede only noun phrases and never has an expletive role (Vergnaud and Zubizaretta 1992; Longobardi 1994).
(ii) Use of the definite (but not the indefinite) article with proper names or generic arguments, without any semantic contribution (see Ex. 20): Definite articles must precede proper names in Greek, when in an argument position. There is no semantic contribution to the definiteness of the proper name, but the article has an expletive role again. In generic and habitual clauses, the definite article precedes the generic subject and functions as an expletive determiner again (Roussou and Tsimpli 1994).
(iii) The definite article can function as an expletive in object position as well – but then the constraints are lexical and its presence depends on the interpretation of the predicate, see Ex. (21).
The definite articles of msg are obligatorily present with proper names and count nouns (both singular and plural) in argument positions. Bare nouns are disallowed in the subject position and are only allowed in the object position if they have an indefinite interpretation; see Ex. (22).
(iv) Co-occurrence of the definite (but not the indefinite) article with a demonstrative (see Ex. 23): The definite article must precede a noun in cases when the noun is modified by a demonstrative. This happens even though demonstratives can be used on their own. The definite article in such cases is redundant, given that the demonstrative shows deictic properties.
(v) Doubling of the definite (but not the indefinite) article (“determiner spreading”) is possible – Ex. (24). See, among others, Alexiadou and Wilder (1998); Campos and Stavrou (2004); Kolliakou (2004); Lekakou and Szendrői (2007, 2014); Marinis and Panagiotidis (2011); Alexiadou (2014). On the other hand, the indefinite article can be used independently (and function as an indefinite pronoun), whereas the definite article cannot (see Ex. 25).
The morpho-phonological similarity between definite articles and 3rd-person object clitics [Table 8] led to an extension of the analysis of definite and indefinite articles to object clitics. Definite articles, similar to clitics, form a phonological unit with the word that follows and are unstressed, whereas the vowel of the neuter definite article can be elided in certain phonological environments (Ex. 26a). Note that they can also appear in a cluster of a preposition and an article (Ex. 26b). This similarity leads to the conclusion of a possible dissociation of the semantic (definiteness) from the morphosyntactic (Case and agreement) features and is related to the function of both elements (clitics and definite articles) as expletives (Sportiche 1996; Tsimpli and Stavrakaki 1999; Marinis 2000; Tsimpli 2003b).
3rd-person object clitics, similarly to definite articles, do not make a semantic contribution but only grammaticalize uninterpretable agreement and case features. On the contrary, 1st- and 2nd-person object clitics have a semantic content and are specified for the interpretable [person feature] (Tsimpli and Stavrakaki 1999; Torrego 1998; Manzini and Savoia 2001).
Moreover, the 3rd-person object clitics differ from genitive clitics in that the object clitics, but not the genitive clitics, can be found in idioms and can be used nonreferentially. Ex. (27).
In Table 9 (from Tsimpli and Mastropavlou 2008), the main characteristics of clitics and articles in msg are summarized.
In this section, we have discussed the relation of object clitics to the system of articles. We follow the proposal that 3rd-person object clitics, as well as definite articles, have only uninterpretable formal features, in contrast to 1st-/2nd-person object clitics and indefinite articles. The contrast between the uninterpretable features of the 3rd-person object clitics and the interpretable features of the 1st-/2nd-person object clitics can be in the basis of their different learnability patterns, as have been observed, for instance, with bilingual speakers or L2 learners. In the following section, we present the data of article omission (as compared with clitic omission) and we show that article omission is allowed only with definite articles.
4.4 Article Omission (Versus Clitic Omission) in the Corpus
We claim that the syntactic borrowing appears to follow the characteristics of syntactic transfer in cases of bilingualism and L2 acquisition (cf. Interpretability Hypothesis): the borrowing of transitivity characteristics in case of language contact affects uninterpretable features rather than interpretable features. 31 If this holds true, we expect similar patterns for the distribution of articles (definite versus indefinite articles) in this dialect. The results in Figure 3 confirm the hypothesis: there are no examples of omission of the indefinite article in the corpus. Moreover, the frequency of omission of the definite article is higher (10.89%) than the frequency of omission of the 3rd-person object clitic (4.68%). 32 Examples of definite article omission, as well as presence of definite/ indefinite articles in the corpus, are presented in (28)–(30). 33
Use and omission of definite/indefinite articles in our dialect corpus
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Use and omission of definite/indefinite articles in our dialect corpus
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Use and omission of definite/indefinite articles in our dialect corpus
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Omission of the definite article:
Presence of the definite article:
Presence of the indefinite article:
The 3rd-person clitic and the definite article appear to be affected by language contact as uninterpretable clusters of features. For instance, we have observed that clitics are dropped primarily when the reference is to an inanimate, rather than animate. Turning to the articles, the results show that the article is dropped primarily with proper names. On the contrary, there are only 20 examples (20/158; 12.66%) with nps with an adjective or a demonstrative but without a definite article. See Table 10 and (31)–(36) for examples of syntactic contexts of definite article omission.
Contexts of definite article omission: Proper names (in examples 31, 32)/place names/common nouns
Contexts of definite article omission: Nouns with an adjective/demonstrative
Turning to the three groups of ages, the picture is unexpectedly different than with the object omission (Table 11). There is a positive correlation between ages and omission of the definite article, but it is of the opposite direction than in the case of omission of the object clitic: the younger speakers omit the definite article in more cases than the older speakers (Figure 4). 34 This may be related to the intense contact of young speakers of the dialect with young bilingual speakers of the area. But this is the reverse case from object omission. By contrast, data from Group 2 are similar to the data from the group of young speakers, and this indicates that the definite article omission is significantly more frequent than the 3rd-person clitic omission for all groups other than Group 1 (the older speakers), which shows a similar tendency for both object and article omission. 35 We should notice that the significant fact showing the structural relation between object clitic omission and definite article omission is the absolute absence of omission of indefinite articles and 1st-/2nd-person object clitics from the data of all age groups and individuals. The different percentages of omission of definite articles and 3rd-person object clitics are related to the compensatory strategies that the different age groups and speakers follow. These strategies are affected by usage preferences, contexts, and other sociolinguistic factors (which may include awareness of the phenomena or social and local identity indexing). In this respect, it is of significance that there is no variation in omission of indefinite articles and 1st-/2nd-person object clitics with regard to either age groups or individual results.
Age and omission of 3rd-person object clitics versus omission of definite articles
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Age and omission of 3rd-person object clitics versus omission of definite articles
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Age and omission of 3rd-person object clitics versus omission of definite articles
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Similar to the L2 results (for instance, from a study on Turkish learners of Greek: Tsimpli and Mastropavlou 2008), the use of definite articles fluctuates more than the use of the indefinite article. To put it in a more precise way, the dialect data are absolute: it appears that there is a rule in the dialect for obligatory presence of the indefinite article in the same contexts in which it is obligatory in msg (see Table 12). Note that even the data from Lefkimi, which do not include any instances of object omission, show many examples of definite article omission (12.35%). 36
Again, there is individual variation in the data of definite article omission. This variation is not only related to age differences (Tables 13a–e). For instance, A.C.’s data (87 years old) show significantly more frequent omission of the definite article than the data from A.B. (82 years old; Lavara village). C.M. (80 years old), a male speaker from Metaxades, shows a frequent rate of omission of the definite article (22.22%), even though he has no examples of object omission. The data from Lykofi also include many instances of definite article omission, even in the utterances of young speakers. G.P. (37 years old) omits the definite article in 21.05% of the contexts where the definite article is obligatory for msg or the other mg dialects. On the other hand, 6.58% of the obligatory (for msg) definite objects are omitted in the data from T.B. (36 years old). The other male speaker of the corpus (P.A., 75 years old, Lagyna village) also allows more frequent omission of the definite article than omission of the 3rd-person clitic (8.41% versus 1.01%). With regard to Lefkimi, both speakers present omission of the definite article (even though they have a “msg behavior” with objects). The data from the younger speaker (C.H., 41 years old) include more cases of omission of the definite article (17.39% versus 11.51%).
In the data of L2 learners (Dimitrakopoulou et al. 2004), the omission of clitics is more frequent than the omission of definite articles in all groups of subjects (based on length of stay). This is in clear contrast to the data from Evros mwtg. See Figure 5 versus Figure 6. 37 Let us remember at this point that, when discussing L2 data, we do not assume any similarity in the speaker profiles. On the contrary, the groups (L2 learners vs. monolingual speakers of a dialect in a historical contact situation with bilingual speakers) and their characteristics are different; the hypothesis is that the results of the transfer as well as of the borrowing are similar in terms of (un)interpretable features – even though the sources and reasons for these results are not similar.
L2 adult learners of Greek (Russian and Serbian L1s): Omission rate of clitic pronouns and the definite article. From Dimitrakopoulou et al. (2004)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
L2 adult learners of Greek (Russian and Serbian L1s): Omission rate of clitic pronouns and the definite article. From Dimitrakopoulou et al. (2004)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
L2 adult learners of Greek (Russian and Serbian L1s): Omission rate of clitic pronouns and the definite article. From Dimitrakopoulou et al. (2004)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
wgt (Evros) dialect data: Omission rate of 3rd-person clitic pronouns and the definite article
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
wgt (Evros) dialect data: Omission rate of 3rd-person clitic pronouns and the definite article
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
wgt (Evros) dialect data: Omission rate of 3rd-person clitic pronouns and the definite article
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
We turn to a comparison between the Evros mwtg data with relevant data from L2 studies (see Figures 7a-b). In L2 data, the 3rd-person object clitic and the definite article are problematic, in contrast to the 1st- and 2nd-person object clitic and the indefinite article. Tsimpli and Mastropavlou (2008) have shown that adult L2 learners, as well as child L2 learners and sli children, follow a similar pattern: the 3rd-person object clitic contrasts with the 1st- and 2nd-person object clitic, and the definite article contrasts with the indefinite article. In these groups of speakers, the 1st-/2nd-person object clitic and the indefinite article are used almost targetlike, whereas the 3rd-person object clitic and the definite article demonstrate a developmental pattern and problems even for advanced L2 learners in the course of L2 acquisition. The similar asymmetry in L2 data and the dialect shows that the status of clitics and articles, as well as their acquisition and the borrowing through contact, is not the same for all members of this category. The uninterpretable characteristics of the 3rd-person clitics and the definite articles can account for the asymmetry with regard to the L2 data and the dialect under examination. The main difference between the data from L2 and the dialect data concerns the higher tendency for definite article omission than for 3rd-person clitic omission in the dialect; the L2 data show the reverse pattern (higher tendency for 3rd-person clitic omission than definite article omission). This can be a result of the variation between the different areas from which the dialect corpus is derived. For instance, the data from Lavara and Metaxades are quite similar to the L2 results: the object omission is affected by borrowing from Turkish in a higher or similar degree as the article omission. On the other hand, the data from Lefkimi and Lagyna show a clear preference for definite article omission in contrast to object omission (cf. Figure 6). This might be related to the particular speakers, the degree of knowledge of the dialect or the degree of contact with bilingual speakers of Turkish-Greek. This issue has to remain open.
L2 adult learners of Greek (Turkish and Russian L1s): Use of 3rd-person object clitic and definite article. From Tsimpli and Mastropavlou (2008)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
L2 adult learners of Greek (Turkish and Russian L1s): Use of 3rd-person object clitic and definite article. From Tsimpli and Mastropavlou (2008)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
L2 adult learners of Greek (Turkish and Russian L1s): Use of 3rd-person object clitic and definite article. From Tsimpli and Mastropavlou (2008)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Child L2 learners of Greek (Turkish L1): Use of 3rd-person object clitic and definite article. From Tsimpli and Mastropavlou (2008)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Child L2 learners of Greek (Turkish L1): Use of 3rd-person object clitic and definite article. From Tsimpli and Mastropavlou (2008)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
Child L2 learners of Greek (Turkish L1): Use of 3rd-person object clitic and definite article. From Tsimpli and Mastropavlou (2008)
Citation: Journal of Language Contact 12, 1 (2019) ; 10.1163/19552629-01201006
To summarize, the data have shown that mwtg allows omission of the object with specific reference, in contrast to msg and other mg dialects but similar to Turkish. Object omission is possible only in contexts where msg and other mg dialects show obligatory use of the 3rd-person clitic. See Table 14. The syntactic borrowing in the case of language contact appears to follow the transfer with L2 learners or bilingual speakers: the relevant elements that bear uninterpretable features are used optionally. For this reason, the definite article, in contrast to the indefinite article, is also affected by language contact. The 3rd-person clitic and the definite article are affected by contact as uninterpretable clusters of features.
5 Conclusions
This study consists of a first attempt to include data from syntactic borrowing due to language (/ dialect) contact in the comparison of production data from different populations (mainly L2 learners and bilingual speakers, but also sli speakers). We have examined data from a spontaneous speech corpus, elicited in collective interviews, from 18 adult speakers of the Evros dialect of mwtg and msg. We have argued that the mwtg dialect of Evros allows omission of the direct object with specific reference, in contrast to msg and other mg dialects (spoken in Greece), but similar to Turkish. Omission of an object with specific reference in clauses where msg and other mg dialects show obligatory use of a clitic is possible only in cases of use of a 3rd-person clitic in msg and the other mg dialects. It appears that only items bearing uninterpretable features (that is, 3rd-person clitics or definite articles, but not 1st-/2nd-person clitics or indefinite articles) are candidates to be affected by language contact. In this manner, syntactic borrowing follows the characteristics of transfer in cases of L2 acquisition, according to the Interpretability Hypothesis.
In this respect, we have compared the Evros mwtg data with relevant data from L2. It has been shown in L2 studies that adult L2 learners use the 1st-/2nd-person object clitic and the indefinite article almost targetlike, whereas the 3rd-person object clitic and the definite article demonstrate a developmental pattern. The data from the dialect under examination present a similar asymmetry as the L2 data. The main difference, however, between the data from L2 and data from the dialect under examination concerns the lower tendency for 3rd-person clitic omission than for definite article omission in the dialect. This can be a result of the variation between the different areas from which the dialect corpus is derived. For instance, the data from the Lavara and Metaxades villages are quite similar to the L2 results, and the data from the Lefkimi village are similar to the data from msg and other mg dialects in not allowing object omission.
Abbreviations
acc accusative; art article; cl clitic; def definite; F feminine; fi Full Interpretation; gen genitive; imp imperative; indef indefinite; lg Logical Form; M masculine; mat-borrowing matter borrowing; mg Modern Greek; msg Modern Standard Greek; mwtg Modern West Thracian Greek; N neuter; nom nominative; pat-borrowing pattern borrowing; pf Phonetic Form; pl plural; sg singular.
We would like to thank Eitan Grossman, Ilja Seržant, Alena Witzlack-Makarevich, Susanne Michaelis, and three anonymous reviewers for very useful suggestions and comments. As usual, all errors remain ours.
References
Adamou Evangelia . 2010. Bilingual speech and language ecology in Greek Thrace: Romani and Pomak in contact with Turkish. Language in Society 39(2): 147–171.
Adamou Evangelia . 2012. Social networks in Greek Thrace: Language shift and language maintenance. In Lindstedt Jouko and Wahlström Max (eds.), Balkan Encounters: Old and New Identities in South-Eastern Europe, 7–32. Helsinki: Slavica Helsingiensia.
Alexiadou Artemis . 2014. Multiple Determiners and the Structure of dps. Amsterdam / Philadelphia: John Benjamins.
Alexiadou Artemis and Wilder Chris . 1998. Adjectival modification and multiple determiners. In Artemis Alexiadou and Chris Wilder (eds.), Possessors, Predicates and Movement in the Determiner Phrase, 303–334. Amsterdam / Philadelphia: John Benjamins.
Anagnostopoulou Elena . 1999. Conditions on clitic doubling in Greek. In Henk van Riemsdijk (ed.), Clitics in the Languages of Europe, 761–798. Berlin: Mouton De Gruyter.
Anagnostopoulou Elena . 2003. Participles and Voice. In Alexiadou Artemis , Rathert Monika and Arnim von Stechow (eds.), Perfect Explorations, 1–36. Berlin: Mouton de Gruyter.
Apostolou-Panara Athena-Maria . 1991. English Loanwords in Modern Greek: An Overview. Terminologie et Traduction 1: 45–48.
Authier J.-Marc . 1992. Iterated cps and embedded topicalization. Linguistic Inquiry 23(2): 326–336.
Aygen Gülşat . 2001. Finiteness, Clause Structure, and Case. PhD dissertation, Harvard University.
Batzakas Georgios . 2012. Mia proti prosegisi tou idiomatos tu Iasmu Komotinis [The dialect of Komotini Iasmos: A first approach]. In Gavriilidou Zoe , Efthymiou Angeliki , Thomadaki Evangelia and Kambakis-Vougiouklis Penelope (eds), Selected papers of the 10th icgl , 973–982. Komotini: Democritus University of Thrace.
Brendemoen Bernt . 1998. Turkish Dialects. In Lars Johanson and Éva Ágnes Csató (eds.), The Turkic Languages, 236–241. London and New York: Routledge.
Campos Hector and Stavrou Melita . 2004. Polydefinite constructions in Modern Greek and Aromanian. In Mišeska Tomić Olga (ed.), Balkan Syntax and Semantics, 137–173. Amsterdam / Philadelphia: John Benjamins.
Vasiliki Chondrogianni , Marinis Theodoros , Edwards Susan and Blom Elma . 2014. Production and on-line comprehension of definite articles and clitic pronouns by Greek sequential bilingual children and monolingual children with specific language impairment. Applied Psycholinguistics 36(5): 1155–1191.
Christidis Anastassios-Fivos . 2001. Greek in contact with other languages. Introduction. In Christidis Anastassios-Fivos (ed.), A History of Ancient Greek: From the Beginnings to Late Antiquity, 721–732. Cambridge: Cambridge University Press.
Chomsky Noam . 1995. The Minimalist Program. Cambridge, ma: mit Press.
Chomsky Noam . 2001. Beyond Explanatory Adequacy. mit Occasional Papers in Linguistics 20. Cambridge, ma: mitwpl.
Cummins Sarah and Roberge Yves . 2005. A modular account of null objects in French. Syntax 8(1): 44–64.
Dawkins Richard M. 1916. Modern Greek in Asia Minor: A Study of the Dialects of Sílli, Cappadocia and Phárasa with Grammar, Texts, Translations and Glossary. Cambridge: Cambridge University Press.
Dimitrakopoulou Maria , Kalaitzidou Sophia , Roussou Anna and Maria Ianthi . 2004. Clitics and determiners in the L2 Greek grammar. Proceedings of the 6th International Conference on Greek Linguistics. http://www.philology.uoc.gr/conferences/6thICGL.
Elšik Viktor and Matras Yaron . 2006. Markedness and Language Change. Berlin: Mouton de Gruyter.
Enç Mürvet . 1991. The semantics of specificity. Linguistic Inquiry 2(1): 1–25.
Fosteris Dimitrios and Kesisoglou Ioannis I. . 1960. Lexilogio tu Aravani [Glossary of Aravan]. Athens: Institut Français d’Athènes.
Franks Steven Franks and King Tracy Holloway . 2000. A Handbook of Slavic Clitics. Oxford: Oxford University Press.
Gelderen Elly van. 2000. A History of English Reflexive Pronouns. Amsterdam: John Benjamins.
Gelderen Elly van. 2011. The Linguistic Cycle. Oxford: Oxford University Press.
Gelderen Elly van. 2013. Clause Structure. Cambridge: Cambridge University Press.
Genesee Fred , Paradis Johanne and Crago Martha B. . 2004. Dual Language Development and Disorders: A Handbook on Bilingualism and Second Language Learning. Baltimore, md: Brookes.
Giannakidou Anastasia and Merchant Jason . 1997. On the interpretation of null indefinite objects in Greek. Studies in Greek Linguistics 17: 141–155.
Giannakidou Anastasia and Stavrou Melita . 1999. Nominalization and ellipsis in the Greek dp. The Linguistic Review 16: 295–331.
Giusti Giuliana . 1997. The categorial status of determiners. In Haegeman Liliane (ed.), The New Comparative Syntax, 95–123. London: Longman.
Halpern Aaron and Fontana Josep M. . 1994. X0 Clitics and Xmax Clitics. Proceedings of the West Coast Conference on Formal Linguistics (wccfl xii): 251–266.
Heine Bernd and Kuteva Tania . 2003. On contact-induced grammaticalization. Studies in Language 27(3): 529–572.
von Heusinger Klaus and Kornfilt Jaklin . 2005. The Case of the Direct Object in Turkish: Semantics, Syntax and Morphology. Turkic Languages 9: 3–44.
Janse Mark . 2001. Morphological borrowing in Asia Minor. In Aggouraki Yoryia , Arvaniti Amalia , Davy J. I. M. , Goutsos Dionysis , Karyolaimou Marilena , Panagiotou Anna , Papapavlou Andreas , Pavlou Pavlos and Roussou Anna (eds.), Proceedings of the 4th International Conference on Greek Linguistics: 473–479. Thessaloniki: University Studio Press.
Janse Mark . 2002. Aspects of bilingualism in the history of the Greek language. In J. N. Adams, Mark Janse and Simon Swain (eds.), Bilingualism in Ancient Society. Language Contact and the Written Word, 332–390. Oxford: Oxford University Press.
Janse Mark . 2006. Object position in Cappadocian and other Asia Minor Greek dialects. In Janse Mark , Joseph Brian and Ralli Angela (eds.), Proceedings of the Second International Conference of Modern Greek Dialects and Linguistic Theory, 115–129. Patras: University of Patras Press.
Joseph Brian . 1999. Is Balkan comparative syntax possible? In Maria Luisa Rivero and Angela Ralli (eds.), Comparative Syntax of Balkan Languages, 17–43. Oxford: Oxford University Press.
Joseph Brian . 2000. Processes of spread for syntactic constructions in the Balkans. In Tzitzilis Christos and Symeonidis Charalambos (eds.), Balkan Linguistik: Synchronie und Diachronie, 139–150. Thessaloniki: University of Thessaloniki.
Karanassios Georgios . 1992. Syntaxe comparé du groupe nominal en grec moderne et dans d’autres langues. PhD dissertation, Université Paris viii.
Karatsareas Petros . 2013. Understanding diachronic change in Cappadocian Greek: The dialectological perspective. Journal of Historical Linguistics 3(2): 192–229.
Ketrez Fatma Nihan . 2005. Children’s Scope of Indefinite Objects. PhD dissertation, University of Southern California.
Kolliakou Dimitra . 2004. Monadic definites and polydefinites: Their form, meaning and use. Journal of Linguistics 40(2): 263–323.
Kontosopoulos Nikolaos G. [1994] 2006. Dialekti ke idiomata tis Neas Elinikis [Dialects and idioms of Modern Greek]. 4th edition. Athens: Grigoris.
Kornfilt Jaklin . 1984. Case Marking, Agreement, and Empty Categories in Turkish. PhD dissertation, Harvard University.
Kornfilt Jaklin . 1987. Turkish and the Turkic Languages. In Comrie Bernard (ed.), The World’s Major Languages, 619–644. Oxford: Oxford University Press.
Kornfilt, Jaklin. 1997. Turkish. London: Routledge.Kornfilt, Jaklin. 2008. dom and two types of dsm in Turkish. In Helen de Hoop and Peter de Swart (eds.), Differential Subject Marking, 79–111. Dordrecht: Springer.
Kostakis Thanasis P. 1968. To glosiko idioma tis Silis [The dialect of Silli]. Athens: Centre for Asia Minor Studies.
Kowaluk Agnieszka . 2001. The acquisition of Determiners and Pronouns in English L2. PhD dissertation, Cambridge, uk.
Kyranoudis Panagiotis . 1995. Fonitiki ton turkikon danion tu idiomatos Sufliu [Phonetics of Turkish loanwords in Soufli]. Thessaloniki: Aristotle University of Thessaloniki.
Lavidas Nikolaos . 2013. Null vs. cognate objects and language contact: Evidence from Hellenistic Greek. Acta Linguistica Hafniensia 44(2): 142–168.
Léglise Isabelle . 2013. The interplay of inherent tendencies and language contact on French object clitics. An example of variation in a French Guianese contact setting. In Léglise Isabelle and Chamoreau Claudine (eds.), The Interplay of Variation and Change in Contact Settings, 137–163. Amsterdam: John Benjamins.
Lekakou Marika and Szendrői Kriszta . 2007. Eliding the noun in close apposition, or Greek polydefinites revisited. ucl Working Papers in Linguistics 19: 129–154.
Lekakou Marika and Szendrői Kriszta . 2014. When determiners abound: Implications for the encoding of definiteness. In Patricia Cabredo Hoffher and Anne Zribi-Hertz (eds.), Crosslinguistic Studies in Noun Phrase Structure and Reference, 212–238. Leiden: Brill.
Longobardi Giuseppe . 1994. Reference and proper names. Linguistic Inquiry 25: 609–665.
Luraghi Silvia . 2003. Definite referential null objects in Ancient Greek. Indogermanische Forschungen 108: 169–196.
Manzini M. Rita and Savoia Leonardo M. . 1998. Clitics and auxiliary choice in Italian dialects: Their relevance for the person ergativity split. Recherches linguistiques de Vincennes 27: 115–138.
Manzini M. Rita and Savoia Leonardo M. . 2001. The syntax of object clitics: si in Italian dialects. In Cinque Guglielmo and Salvi Giampaolo (eds.), Current Studies in Italian Syntax: Essays Offered to Lorenzo Renzi, 233–264. Amsterdam: North Holland.
Marinis Theodore . 2000. The Acquisition of Clitic Objects in Modern Greek: Single Clitics, Clitic Doubling, Clitic Left Dislocation. zas Papers in Linguistics 15.
Marinis Theodore . 2002. Minimal inquiries and the acquisition of the definite article in Modern Greek. Proceedings of the 34th Colloquium of Linguistics, 73–82.
Marinis Theodore and Panagiotidis Phoevos . 2011. Determiner spreading as referential predication. Studia Linguistica 65: 268–298.
Matras Yaron . 2009. Language Contact. Cambridge: Cambridge University Press.
Matras Yaron and Sakel Jeanette . 2007. Investigating the mechanisms of pattern replication in language convergence. Studies in Language 31(4): 829–865.
Mavrogiorgos Marios . 2010. Clitics in Greek: A Minimalist Account of Proclisis and Enclisis. Amsterdam: John Benjamins.
McShane Marjorie . 2002. Unexpressed objects in Russian. Journal of Slavic Linguistics 10: 291–328.
Oikonomidis Dimosthenis I. 1958. Gramatiki tis Elinikis Dialektu tu Pontu [Grammar of the Greek Dialect of Pontus]. Athens: Akadimia Athinon.
Öztürk Balkiz . 2005. Case, Referentiality and Phrase Structure. Amsterdam: John Benjamins.
Panagiotidis Phoevos . 2002. Pronouns, Clitics and Empty Nouns. Amsterdam: John Benjamins.
Papadopoulos Anthimos A. 1955. Istoriki Gramatiki tis Pontikis Dialektu [Historical Grammar of the Pontic Dialect]. Athens: Epitropi Pontiakon Meleton.
Papadopoulou Theodora . 2013. Modern Western Thracian Greek. Unpublished ma Research Paper (supervised by Dr. Nikolaos Lavidas).
Paradis Michel and Gopnik Myrna . 1994. Compensatory strategies in familial language impairment. McGill Working Papers in Linguistics 10(1/2): 142–149.
Paradis Michel and Gopnik Myrna . 1997. Compensatory strategies in genetic dysphasia: Declarative memory. Journal of Neurolinguistics 10: 173–185.
Parodi Teresa , Schwartz Bonnie D. and Clahsen Harald . 2004. On the L2 acquisition of the morphosyntax of German nominals. Linguistics 42: 669–705.
Pérez-Leroux Ana T. , Pirvulescu Mihaela and Roberge Yves . 2006. Early object omission in child French and English. In Montreuil Jean-Pierre Y. and Nishida Chiyo (eds.), New Perspectives on Romance Linguistics. vol. i, 213–228. Amsterdam: John Benjamins.
Pérez-Leroux Ana T. , Pirvulescu Mihaela and Roberge Yves . 2008. A syntactic transitivity approach to null objects in child language. Lingua 118: 370–398.
Rice Mabel L. and Wexler Kenneth . 1996. Toward tense as a clinical marker for specific language impairment in English-speaking children. Journal of Speech and Hearing Research 39: 1239–1257.
Roussou Anna and Tsimpli Ianthi Maria . 1994. On the interaction of case and definiteness in Modern Greek. In Philippaki-Warburton Irene , Nicolaidis Katerina and Sifianou Maria (eds.), Themes in Greek Linguistics: Papers from the First International Conference on Greek Linguistics, 69–76. Amsterdam: J. Benjamins.
Roussou Anna and Tsimpli Ianthi Maria . 2007. Clitics and transitivity. In Alexiadou Artemis (ed.) Studies in the Morpho-Syntax of Greek, 138–174. Newcastle: Cambridge Scholars Publishing.
Sella-Mazi Eleni . 1997. Language contact today: the case of the Muslim minority in Northeastern Greece. International Journal of the Sociology of Language 126: 83–103.
Şener Serkan and Takahashi Daiko . 2010. Ellipsis of arguments in Japanese and Turkish. Nanzan Linguistics 6: 79–99.
Sioupi Athina . 2002. On the syntax and semantics of verb-complement constructions that involve ‘creation’: a comparative study in Greek and German. In Werner Abraham and C. Jan-Wouter Zwart (eds.), Issues in Formal German(ic) Typology, 263–284. Amsterdam: John Benjamins.
Sportiche Dominique . 1996. Clitic constructions. In Rooryck Johan and Zaring Laurie (eds.) Phrase Structure and the Lexicon, 213–277. Dordrecht: Kluwer.
Spyropoulos Vassilios and Tiliopoulou Marianna . 2006. Definiteness and case in Cappadocian Greek. In Janse Mark , Joseph Brian and Ralli Angela (eds.), Proceedings of the 2nd International Conference on Modern Greek Dialects and Linguistic Theory, 366–378. Patras: University of Patras.
Spyropoulos Vassilios and Kakarikos Konstantinos . 2011. A feature-based analysis of Cappadocian Greek nominal inflection. In Janse Mark , Joseph Brian , Pavlou Pavlos , Ralli Angela and Armosti Spyros (eds.), Studies in Modern Greek Dialects and Linguistic Theory, 203–213. Nicosia: Research Centre of Kykkos Monastery.
Stavrou Melita . 1996. Adjectives in Modern Greek: An instance of predication, or an old issue revisited. Journal of Linguistics 32: 79–112.
Theophanopoulou-Kontou Dimitra . 1986–1987. Kenes katigories ke klitika stin nea eliniki: i periptosi tu amesou andikimenu [Empty categories and clitics in Modern Greek: The case of the direct object]. Glossologia 5–6: 41–68.
Torrego Esther . 1998. The Dependencies of Objects. Cambridge, ma: mit Press.
Tsimpli, Ianthi Maria. 1997. Resumptive features and minimalism: Evidence from second language acquisition. In Elizabeth Hughes, Mary Hughes and Annabel Greenhill (eds.), Proceedings of the 21st Annual Boston University Conference on Language Development, 639–655. Somerville, MA: Cascadilla Press.
Tsimpli Ianthi Maria . 1999. Determiners and clitics in Greek. Working Papers in Theoretical and Applied Linguistics 6: 189–208.
Tsimpli Ianthi Maria . 2003a. Interrogatives in the Greek/English interlanguage: A minimalist account. In Mela-Athanasopoulou Elsa (ed.), Selected Papers on Theoretical and Applied Linguistics, 214–225. Thessaloniki: Aristotle University of Thessaloniki.
Tsimpli Ianthi Maria . 2003b. Clitics and determiners in L2 Greek. In Liceras Juana M. , Zobl Helmut and Goodluck Helen (eds.), Proceedings of the 6th Generative Approaches to Second Language Acquisition Conference (gasla 2002), 331–339. Somerville, ma: Cascadilla Press.
Tsimpli Ianthi Maria . 2004. Features in L1 and L2 acquisition: Evidence from Greek clitics and determiners. In Hendricks Henriëtte (ed.), Analyse comparative des processus d’acquisition en L1 et L2, Special Issue of aile 20: 87–128.
Tsimpli Ianthi Maria and Dimitrakopoulou Maria . 2007. The interpretability hypothesis: Evidence from wh‐interrogatives in second language acquisition. Second Language Research 23(2): 215–242.
Tsimpli Ianthi Maria and Mastropavlou Maria . 2008. Feature interpretability in L2 acquisition and sli: Greek clitics and determiners. In Liceras Juana M. , Zobl Helmut and Goodluck Helen (eds.), The Role of Formal Features in Second Language Acquisition, 143–183. Mahwah, nj: Lawrence Erlbaum.
Tsimpli Ianthi Maria and Papadopoulou Despina . 2009. Aspect and argument realization: A study on antecedentless null objects in Greek. Lingua 116(10): 1595–1615.
Tsimpli Ianthi Maria and Stavrakaki Stavroula . 1999. The effects of a morpho-syntactic deficit in the determiner system: The case of a Greek sli child. Lingua 108(1): 31–85.
Tsiplakou Stavroula , Papapavlou Andreas , Pavlou Pavlos and Katsoyannou Marianna . 2006. Levelling, koineization and their implications for bidialectism. In Hinskens Frans (ed.), Language Variation – European Perspectives: Selected Papers from the Third International Conference on Language Variation in Europe, 265–276. Amsterdam: John Benjamins.
Vergnaud Jean-Roger and Zubizarreta Maria Luisa . 1992. The definite determiner and the inalienable constructions in French and English. Linguistic Inquiry 23(4): 595–652.
Evros is the eastern part of the geographical region of Western Thrace. See Section 2 and Map 1, below.
Pomak is a Balkan Slavic dialect, akin to Bulgarian, spoken by Muslim inhabitants of the Rhodope Mountains in Greece (Adamou 2009).
Romani is an Indo-Aryan language.
We can use the term quatrilingual instead of trilingual (or trilingual instead of bilingual) in all of the cases, if we count separately the varieties of Greek. In this respect, we follow Adamou, who does not count the varieties of Greek separately. With regard to the question of the language in which these speakers are dominant, we should note that bidialectal speakers of community (a) are considered to be dominant in msg (but see the corpus data and the analysis below), speakers of community (b) in Turkish, speakers of community (c) in Romani, speakers of community (d) in Pomak, and speakers of community (e) in Armenian. On bidialectalism in msg and other varieties of Greek, see, among others, Tsiplakou et al. (2006).
By contrast, mat-borrowing [matter borrowing] describes the case where the forms and the phonological shapes are replicated from one language in another language.
The discussion that follows, according to our hypothesis, includes a comparison of our data (which derive from the speech of monolingual speakers of the dialect) with data from L2 learners, bilinguals or specific language impairment (sli) speakers. This comparison does not mean that we consider the speakers of the dialect bilinguals; we will emphasize below that our hypothesis is that the results of syntactic borrowing can be similar, in terms of syntactic (un)interpretable features, to the results of transfer in case of bilinguals or L2 learners or to some of the characteristics in the speech of sli speakers. We will also demonstrate that the characteristics of syntactic borrowing and the results of L2 or sli performance with regard to the uninterpretable features show similar patterns, even though these similarities have a different source: syntactic borrowing due to language contact in the case of contact between languages/dialects, versus L1 transfer in the case of L2 acquisition, versus genetically related causes in the case of sli grammars. For this reason, we do not discuss the profile of bilingual speakers, but the data of speakers of a dialect in contact settings. Our hypothesis is that there is a case of syntactic borrowing in the dialect (in a similar way that there is evidence of lexical borrowing; see below, Section 3.2).
For Matras and Sakel (2007), the results of the contact may or may not lead to language change; according to them, language change in this case depends on conditions relevant to exposure to structural variation or normative control, among other factors.
In this respect, and in relation to Matras and Sakel (2007), we agree with the reviewer who mentions the aspect of historical contact-induced innovations.
Cf. also Janse (2002: 387) for Greek dialects outside Greece in contact with Turkish, and the reverse case of borrowing of the characteristics of clitic objects into the Turkish dialects:
Remarkably, the Turkish Black Sea dialects have in their turn been heavily influenced by Pontic. Whereas Turkish normally omits the pronoun if it can be inferred from the context or situation, the Black Sea dialects not only use non-emphatic pronouns on the analogy of the Greek eps [enclitic pronouns], but they are usually postpositive vis-à-vis the verb on the analogy of the Pontic pattern as well […].
See also Brendemoen (1998). With regard to the aims of our study, the remark (in Brendemoen’s and Janse’s studies) that language contact in these cases affects the presence/ drop of the direct object is of significance. We will show that mg dialects in Thrace permit object drop; the opposite phenomenon (presence of a pronoun instead of the typical object drop of Turkish) has been attested for Turkish dialects that are in contact with Greek dialects of Pontus.
With regard to Komotini Romani, Adamou (2012) claims that the Turkish loanwords result from the “generalized bilingualism” during the Ottoman Empire. Adamou observes similar tendencies in other areas of Greece, where the Romani dialect of Christian Roma may include Turkish loan verbs without any contact between these speakers and Turkish during the current period. In Greek Thrace, current bilingualism helps the maintenance of the loanwords – but the influence may also have affected the relevant grammatical aspects of Romani (and mwtg; see Section 4).
Diglossia here refers to the functional distribution of two or more languages in the same community.
There are no previous studies on syntactic (besides the lexical) borrowing from Turkish into mwtg. It would be of significance for the analysis of the dialect under examination to have remarks on other aspects of the dialectal grammar that have been influenced by Turkish. We leave this issue (for instance, influence in the formation of wh-questions) open for further research.
See Cummins and Roberge (2005) for a detailed presentation of the literature on object omission and relevant examples. For a historical perspective on object omission in the diachrony of Greek, see Luraghi (2003) and Lavidas (2013).
In msg, clitics are associated with argument licensing (Roussou and Tsimpli 2007). The omission of the object clitic is not permitted, because the empty object is unrecoverable (Theophanopoulou-Kontou 1986–1987). The same rule holds for most Romance languages: this fact leads to the hypothesis of a relationship between the status of pronouns in these languages and object omission (Kowaluk 2001).
As well as Russian, Serbian, and Polish (among other languages). Cf. Kowaluk (2001), McShane (2002), among many others.
Cf. Tsimpli (2003b: 333): “Turkish marks specificity partly through case (Enç 1991; Kornfilt 1997). Slavic languages too seem to mark specificity through case in negative clauses, although this does not seem to be obligatory in all of them (Kowaluk 2001).”
We thank the reviewer who mentions a similar phenomenon with genitive subjects: cf. Kornfilt (1984); von Heusinger and Kornfilt (2005); Kornfilt (2008).
In terms of the Minimalist Program, interpretable features do not regulate syntactic derivations, are not erased and are present at lf; uninterpretable features are related to the syntactic computations exclusively. Interpretable features are features with semantic contribution to the interpretation of the sentence: for instance, the feature of [definiteness], phi-features (person, number, gender) on nouns, the [Q] feature in interrogatives. On the other hand, case, phi-features on verbs and adjectives are uninterpretable at lf, contribute only to the syntactic derivation, and are erased after having been matched and checked, before the lf (Chomsky 1995, 2001).
Cf. also, among others, van Gelderen (2013: 16):
The interpretable features are considered relevant for interpretation at Logical Form (lf), and include categorial and nominal [person, number, gender] phi-features. They are not deleted or erased after they are checked because they are relevant to the interpretative component. […] Uninterpretable features receive a value when they search and find an interpretable feature. These valued features are not interpreted at lf; they just go to the pf and, in English, involve the case features of nps and verbs and the phi-features of verbs.
According to this approach, the predictions for the L2 learners and article omission can be as follows: one expects target-like use of agreement and case morphology, but learnability problems in the mapping between morphology and uninterpretable formal features specification. Cf. Tsimpli and Mastropavlou (2008: 154):
Determiner use in L2 Greek is predicted to develop, although the set of obligatory contexts will differ from that of the native speaker, due to the nontarget feature-specification of determiners in the L2 lexicon. In order to constrain optional use of determiners in the L2, the learner imposes interpretable features such as [referentiality] or [specificity] on the Greek article. Thus, L2 use improves considerably but does not become nativelike.
With regard to language change, note that interpretable features tend to change into uninterpretable, except for the case of replacement of an element, where we can have the emergence of a new element with interpretable features (see van Gelderen 2000, 2011). The question that arises is how the dialect data are reconciled with the above claim. We leave this question open for future research.
We should note that the L2 studies have shown omission of the definite article in the data of L2 learners with an L1 that lacks a system of determiners that distinguishes between a definite and indefinite article (L1s are Slavic languages or Turkish). Moreover, these L1s allow for null objects with specific reference, and the object omission is possible for all person and number nps.
Notice that Elšik and Matras (2006:134–136) have shown, for Romani dialects, that 3rd-person markers are more likely to be borrowed than 1st- or 2nd-person markers. Moreover, Léglise (2013) has provided examples of language contact effects with regard to object omission in varieties of French, e.g., in French Guiana, in Burkina Faso, and in the Ivory Coast. Léglise’s corpora from French Guiana “only include examples of variation in the system of clitics”: The accusative is used instead of the dative for the 3rd-person plural, and no case marker is used instead of the accusative or the dative for the 3rd-person singular and plural.
The data were morphologically and syntactically annotated by Nikolaos Lavidas for the purposes of this study.
The data derive from our annotated corpus of spontaneous speech data. We leave open for future research the question whether other types of data (for instance, data from grammaticality judgments tasks) will show similar results.
This may seem as a rather small percentage of omission. On the other hand, there is a clear contrast with the fact that there are no cases of omission for 1st-/2nd-person. Moreover, we should notice that 3rd-person clitics can be used in such cases, but with the assignment of interpretable features – similar to their usage from L2 learners, for instance. See Sections 3.2 and 4.3.
Notice that the results (from a Fisher’s exact test) are not statistically significant for the comparison between group 1 and group 2 (χ 2=1.247, df=1, p=.276), as well as for the comparison between group 1 and group 3 (χ 2=1.419, df=1, p=.466).
This is an example with a topicalized object again. The object (prozimi ‘yeast’) has been mentioned in the previous clause – it is replaced by an object clitic in such cases (there is no repetition of the np), if not topicalized: […they have the yeast in their hands] and how do they prepare it? (pos tha na to pianune…).
The results (from a Pearson chi-square test) are statistically significant: for instance, for the comparison between Metaxades and Lagyna (χ 2=5.130, df=1, p=.024), with an effect size of φ=.124, which is a small size effect. See below, Section 4.3, for a further discussion of this comparison.
We should note that this speaker does not have either a better knowledge of Turkish than other speakers, or a family with this type of background. The aspect of discussion of the data of the individual speakers is connected with the broad discussion of the sources of syntactic borrowing. Our hypothesis concerns the results of syntactic borrowing, which we consider in terms of (un)interpretable features. The data probably reflect the results of a historical contact-induced innovation (see also fn. 8) – and, as the speakers of the dialect under examination show a restricted knowledge of Turkish, a plausible hypothesis for the sources could also include influence from Greek as spoken from the Turkish-Greek bilinguals.
With regard to crosslinguistic differences, note that the Greek clitics and articles are different from Romance clitics and articles, for instance, in that (i) Greek clitics and articles share both agreement and case features, whereas Romance clitics and articles host only common agreement features, (ii) Greek clitics and articles are marked for case, whereas Romance clitics, but not articles, are marked for case. There are further crosslinguistic differences with regard to interpretable features. For instance, in Greek, the features of topic or focus are syntactically active, but in English they are not. This means that whether a feature is or is not syntactically active (in other words, the grammaticalization of a semantic feature) is parameterized: in some languages, the particular feature affects derivations but in other languages, it does not (Tsimpli and Mastropavlou 2008).
Of course, we do not claim that either the syntactic transfer or the syntactic borrowing behave similarly in cases of bilingualism and L2 acquisition or language contact. We should repeat here that, according to our hypothesis, the results of the borrowing in cases of contact should be similar to the results of L2 transfer in terms of (un)interpretable features. This hypothesis is similar to the hypothesis of common results in terms of (un)interpretable features in L2 and sli. It is obvious that the sources of the characteristics of syntactic transfer in L2 and sli are different – but the results of the transfer can be described in a similar way, in terms of the role of (un)interpretable features.
The results (from a Pearson chi-square test) are statistically significant for the comparison between the omission of the 3rd-person clitic and the omission of the definite article (χ 2=20.323, df=1, p<.001), with an effect size of φ=.099, which is a small size effect.
We thank the reviewer for the discussion of the omission of the definite article in other mg dialects. We should note that the aim of the present study is to examine characteristics of transitivity in the wtmg dialect, which can be the result of borrowing from Turkish. The discussion of the omission of articles is significant for our proposed explanation, which is related to the (un)interpretable features. We leave a comparison between the two groups of dialects, Asia Minor Greek and mwtg, which cannot be discussed here due to space limitations, open for future research. A relevant comparison would require similar data on the hypothesis of a contrast between 3rd-person clitics vs. 1st- and 2nd-person clitics, as well as between definite and indefinite articles, from a similar corpus from the Asia Minor Greek dialects. The omission of forms of the definite article is attested in three of the Asia Minor Greek dialects, which have been influenced by Turkish: cf., among others, for Pontic: Papadopoulos (1933), Oikonomidis (1958); for Cappadocian: Dawkins (1916), Fosteris and Kesisoglou (1960); for Silliot: Dawkins (1916); Kostakis (1968). We should notice that Karatsareas (2013) has proposed an alternative analysis for the Asia Minor Greek dialects, according to which the origins of the phenomenon are phonological and are related to hiatus resolution.
Notice that the results (from a Pearson exact test) are not statistically significant for the comparison between group 1 and group 2 (χ 2=2.933, df=1, p=.087), or for the comparison between group 1 and group 3 (χ 2=2.001, df=1, p=.157).
The results (from a Fisher’s chi-square test) are statistically significant for the comparison between the omission of the 3rd-person clitic and the omission of the definite article: (i) for group 1 (χ 2=17.975, df=1, p<.001), with an effect size of φ=.274, which is a medium size effect, (ii) group 2 (χ 2=18.184, df=1, p<.001), with an effect size of φ=.113, which is a small size effect, and (iii) group 3 (χ 2=4.206, df=1, p=.048), with an effect size of φ=.146, which is a small size effect.
The results (from a Pearson chi-square test) are statistically significant, for instance, for the comparison between Lykofi and Lavara (χ 2=4.428, df=1, p=.035), with an effect size of φ=.113, which is a small size effect.
The results (from a Pearson chi-square test) are not statistically significant for the comparison between the omission of the 3rd-person clitic and the omission of the definite article, for instance, in Metaxades (χ 2=.084, df=1, p=.771), but they are statistically significant, for instance, for Lykofi (χ 2=6.141, df=1, p=.013), with an effect size of φ=.135, which is a small size effect.
We thank the reviewer who discusses the incorporation contexts where the indefinite article can also be omitted in Turkish. The dialect under examination does not show incorporation with indefinite articles, and this can also be a type of evidence of the role of the (un)interpretable features in case of syntactic borrowing. We have to leave a further examination of incorporation open for future research.