International Journal of Bilingualism

In this paper we provide a quantitative analysis of the behavior of nouns in two generations of speakers from a bilingual community in northern Cyprus. Diagnostics from three areas of grammar—phonology, morphology, and syntax—provide corroborating evidence that when a lone noun, of either English or Turkish origin, appears in contexts in which it is surrounded by the other language, it patterns systematically in accordance with its counterparts in that other language. On the other hand, when a lone noun, either English or Turkish, appears within a multiword fragment of English or Turkish, it patterns overwhelmingly with the language of its etymology. The strikingly different grammatical patterns highlight the fact that they represent two different types of behavior: borrowing in the former; codeswitching in the latter. These findings demonstrate that empirical investigation can disambiguate the community-specific status of language contact phenomena. Further, contrastive analysis across generations reveals that although the generations differ in the frequency with which they use Turkish and English, they employ the same strategies when they do. These strategies have far more to do with their preferred language of discourse, than with any purely linguistic factor.

In this paper I develop diagnostics for distinguishing word-internal codeswitching from borrowing, based on Ukrainian-English bilingual discourse; a typologically different language pair. Assuming that a codeswitched noun will retain its original characteristics, and a borrowed noun will acquire language-specific properties of the recipient language, I focus on conflict sites in the morphosyntactic structure of Ukrainian (a fusional language) and English (an analytical one). Using quantitative variationist methodologies, I show that English-origin nouns with overt Ukrainian morphology occurring in otherwise Ukrainian discourse do not behave as if they were English. Instead, they replicate the patterns of behavior of monolingual Ukrainian nouns with respect to gender assignment, modifier-noun agreement, inflectional variability and flagging. I found no evidence for lone item codeswitching within a word boundary, and conclude that all English-origin utterances occurring with overt morphology of the recipient language are borrowed.

This paper offers an empirical analysis of data from natural Igbo-English bilingual discourse which demonstrates how the two most important manifestations of language contact, namely, codeswitching and borrowing, can be unambiguously and consistently distinguished. The paper focuses on showing how inherently ambiguous lone English-origin nouns and verbs, incorporated into otherwise Igbo discourse, can be assigned language membership, not solely on the basis of their surface appearance as has been the case in the literature, but only by situating the ambiguous forms in the context of the entire system. Using diagnostics such as vowel harmony, inflection, and word order of verbs as well as modification structures of nouns, the paper demonstrates beyond a shadow of doubt that lone English-origin items are distributed among syntactic slots in the same way as native items of Igbo and differ from codeswitches to English which show distinct patterns. The inescapable conclusion is that the lone English-origin nouns and verbs in otherwise Igbo discourse cannot be classified as codeswitches but rather, must be considered to be borrowings.

In this paper we determine the status of ambiguous lone English-origin nouns in otherwise Persian discourse. Utilizing the variationist comparative method, we analyze their distribution and conditioning and compare them to those of their counterparts in unmixed English. The results of our diagnostic measures all show remarkable similarities between the treatment of native Persian nouns, attested loanwords, and unattested lone English-origin nouns. The Null Theory of codeswitching is assessed against these findings. The Null Theory's prediction that both English and Persian heads should structurally project was not confirmed while the nonce borrowing hypothesis fully accounts for the findings of this study. Thus, the Null Theory provides no better explanation for the parallel patterns exhibited by English nonce forms and attested loanwords in Persian discourse to unmixed Persian than the more straightforward Nonce Borrowing Hypothesis. Our results constitute overwhelming evidence that there is no distinction between established loanwords and nonce borrowing and that therefore, they should all be classed as borrowed elements. As such they cannot be considered as evidence for any theory of codeswitching.

The categorization of lone lexical items from one language embedded in another is often difficult due to their ambiguous status as either loanwords or codeswitches. Following variationist principles, we use a comparative method to disambiguate lone English-origin nouns in otherwise Acadian French discourse. Their exact status is important because it has been argued that the constraints on different manifestations of language contact differ. Incorporating the notion of structural conflict sites as diagnostics, we examine these items at the morphological (e.g., number marking), syntactic (e.g., adjective placement) and discourse (e.g., flagging) levels for evidence of system membership. Our method reveals that although most of these items can be unambiguously classified as borrowings, a small subset must be considered to be codeswitches.