Borrowing: the synchrony of integration*

The notion of loanword assimilation is operationalized in a number of d~flerent ways, focusing on both linguistic and social aspects. The indices of integration thus constructed are applied to a set of lexical data elicited from Puerto Ricatz children and adults from East Harlem, New York. The results of this survey are analyzed statistically using the method of principal components. We interpret the output in terms of the social and linguistic trajectory of words during the borrowing and integration process. Of particular importance are the relatively close relationship between increase in usage frequencies and the processes of phonological ititegration, the transient nature of inconsistencies in gender assignment, and the fates of competing lexical items for a single referent. The lexical stock of languages may contain a considerable proportion of words borrowed from one or more other languages. The historical record, together with methods of historical and comparative linguistics, can help us infer which words were borrowed, from what language, and approximately when. On the synchronic level, however, making such inferences can be more difficult, particularly because there is no unequivocal way of deciding when a lexical item from one language, used during discourse in another language, whether by a single speaker, or repeatedly in a community, should be considered a loanword. It may constitute all or part of a code-switch, which is a phenomenon quite distinct from borrowing. It may be a manifestation of incomplete acquisition of one of a bilingual's two languages. It might be a momentary lapse of the type often classified as 'interference'. Or we might want to characterize it as still another kind of result of language contact. It has been claimed that 'from synchronic examination [i.e. without comparative or etymological evidence] no loans are discoverable or describable' (Fries and Pike 1949; see also Haugen 1950a; Weinreich Linguistics 22 (1984), 99-135 0024-3949/84/0022-0099 $2.00 @ Mouton Publishers 100 S. Poplack and D. Sankoff 1953), presumably because they are so perfectly assimilated to recipientlanguage patterns. We dispute this all-or-nothing viewpoint of the borrowing process, focusing specifically on the mechanisms by which an item is gradually converted from a foreign element to a nativized one. It is during this transition that it is difficult to recognize and distinguish loanwords. Though there is a large literature on the topic of borrowing, there remain many unanswered questions about the linguistic, sociological, acquisitional, and other aspects of borrowing, as well as the subtle methodological dilemrnas involved in detecting, defining, identifying, and characterizing loanwords. In this paper we first review the pertinent literature, focusing on three aspects of the borrowing process: the linguistic mechanisms involved, the social dynamics, and the definitional and analytic problems in studying lexical transfer. We will abstract from this work a number of concepts and hypotheses which have been developed to describe and account for the importation of forms from one language to another. We will proceed to operationalize these concepts on the basis of a corpus of potential loanwords collected from a group of bilingual Puerto Rican children and their parents. We then analyze these data statistically to evaluate the dimensionality of the borrowing process the degree to which processes of phonological, syntactic, lexical, and sociological integration of a foreign element occur in unison or proceed independently. We pay special attention to the intergenerational transmission of borrowed material, and in particular to the roles of older and younger, monolingual and bilingual speakers in its propagation and eventual phonological shape. Traditional as well as more recent studies of borrowing stressed that adaptations of foreign items of any linguistic level to the pattern of the recipient language, such as incorporation of verbal and nominal suffixes, assignment of gender, etc., were indications that the forms had been integrated into that language (e.g. Haugen 1950a: 396, 440; also implicit in Bloomfield 1933: 450, 453; see also Hyman 1970; Lovins 1974). Fries and Pike incorporate phonological, grammatical, and social criteria in their assun~ption (1949: 39) that a loan sequence of phonemes can be considered completely assimilated when (a) it parallels the sequences occurring in native materials, or is analogous to them; when (b) its occurrence in relation to grammatical boundaries is the same as sequences in rlative words; and when (c) the words containing it are in common use by monolinguals; or a loan sequence may be considered completely assimilated when it serves as a pattern for the development of new sequences in the native language. This state of complete assimilation, however, obviously does not come about instantaneously. And during the period, months, years, or generaBorrowing: the synchrony of integration 101 tions, over which it is attained, little is known about which of these criteria is satisfied first, or last, or about the intermediate stages in this attainment. The linguistic integration of loanwords is but one aspect of their assimilation into a language. The sociological process of acceptation is another. Studies of languages in contact have focused on two basically sociological or sociolinguistic distinctions as being relevant to the incorporation of borrowed material into the linguistic repertoire of the community. One is the differential role, as carriers of innovations, of monolinguals versus bilinguals of varying degrees of competence; the other involves the changing shape of loanwords across successive generations of speakers. A common view is that bilinguals tend to use loanwords before monolinguals, who learn them from the former. Bilingual speakers are also thought to assimilate new sounds sooner than their monolingual counterparts (Fries and Pike 1949: 39). Haugen distinguishes between monolinguals and those who became bilingual as adults on the one hand and childhood bilinguals on the other. Observing that borrowed items tend to retain an uncertain linguistic status for some time after their first adoption (1956: 5 9 , he attributes part of this vacillation to the awareness on the part of the bilingual of the origin of the borrowed word, and presumably to his indecision as to whether to produce it according to recipient or donor language rules. In addition, both monolinguals and nonfluent ('adult') bilinguals make phonic adaptations (or 'distortions') of loanwords, while fluent ('childhood') bilinguals reproduce the patterns of the donor language (Haugen 1950a, 1956). These individual differences in ability and desire are presumably responsible for the alternative forms often shown by loans reported by Haugen. He posits three stages of loanword integration (e.g. 1969: 394) ranging from 'pre-bilingual', when forms are reproduced according to recipient language patterns with great irregularity in the results, to 'adult bilingualism', at which point the loans are produced more systematically, through 'childhood bilingualism', during which sound types (and presumably other patterns) from the donor language are introduced into the recipient language. Indeed, childhood bilinguals or younger speakers (who seem to be thought of as indistinguishable), according to this schema, arc the spcokcrs responsible for introducing NEW patterns into the recipient language. Whichever speakers first make use of loanwords, and whatever the mechanism of spread within the lexicon through time and throughout the conlmunity, an important diagnostic for the incorporation of a form into the native lexicon is the increased frcquency of its usilge. Even the degree to which the loanword is linguistically integrated has been attributed to 102 S. Poplack and D. Sankof the frequency of its use within the community (Kreidler 1979: 143). Holden (1976: 131) has suggested that 'most loanwords which show even a minimal degree of acceptance by the target language immediately assume a phonetic shape which is identical to that of the native vocabulary'. Independent of these hypotheses correlating frequency of use with morphophonemic adaptations, frequency of use in itself has been considered a criterion of integration. Hasselmo suggests, for Swedish-English bilinguals, that 'some of the English discourse units introduced into Swedish discourse are used with such regularity that it may be necessary to regard them as in some sense irztegrated with an American Swedish mode of speaking' (1970: 179). More specifically 'if certain instances of interference are repeated often enough in discourse in a certain language to be regarded as habitualized, the forms and/or patterns involved can be referred to as (socially) integrated with the language of the community' (1970: 179; see also Mackey 1970). But as Mackey has observed (1970: 204), there is no necessary indication in the (single) occurrence of a given element in discourse of whether it represents a case of interference or a case of integration. 1.0. Analytical distinctions At first glance it might seem to be an easy matter to detect material borrowed from one language in the discourse of the other, merely by comparison with the standard variety of the recipient language, or with communities not in contact with the donor language. This type of external comparison is, howcvcr, quite inadequate. The co-occurrence of forms from two languages may be due to a number of processes other than borrowing, the most important of which are code-switching and incomplete second-language acquisition by native speakers of another. Codeswitching is simply the alternate use of the two languages in discourse, and even in a sentence, without any necessary influence of one language on those str&ches of discourse realized in the other. Partial acquisition of a second language may lead to the use of first-language items in intended second-language discourse, but on an idiosyncratic basis. They may be considered borrowings on the individual-speaker level, but not on the level of the community speech variety. As Haugen (1969: 371) has pointed out, innovations made by language learners do not spread to native speakers of the language they learn; it is the innovations they make in their own language which spread. Linguists have also tried to define borrowing disjunctively by virtue Borrowing: the synchrony of integration 103 of being neither code-switching nor momentary or idiosyncratic uses of first-language terms to f i l l gaps in intended second-language competence. 'The latter behavior, along with others, has generally been subsumed under tlic label popularized by Weinreich (1953), 'interfcrence', and occupies some nebulous area in between the other two extremes. In Haugen's (1956) schema these phenomena are located along a continuum of code distinctiveness, with switching representing maximal distinctness, integration representing maximal levelling of distinctions, and interference rcferring to an overlapping of two codes, contrary to contemporary norms. In determining whether adaptation had occurred or not (i.e. whether integration or switching was involved), Haugen suggested that the phonological and morphological shape of the borrowed form were the determining factors. However, it is rather the bilingual ability of the speaker which determines the pronunciation of the second language, so that this criterion will misidentify code-switches as loanwords and vice versa. Shaffer's (1978) claim that integration is morc accurately measured by syntactic considerations also fails to unambiguously distinguish loanwords from (one-word) code-switches. Indeed, as Hasselmo observes, althougll the intention of the speaker may be a binary choice between switching and integration, the stretches of speech actually produced are often ambiguous. Since code-switching is not identifiable on the basis of linguistic phonological, morphological, or syntactic features alone, the occurrence of a borrowed item that shows a high degree of social integration (i.e. acceptance and use by community members) could be interpreted as an instance of a loanword, while one that shows a low degree of social integration would be an instance of code-switching (1970: 180). In a similar vein, Mackey (1970: 21 1) suggests that the more an item is integrated (used), the less likely that its appearance is a case of interference. 2.0. An operational framework We may abstract froni the previous considerations four basic types of criteria for the characterization of loanwords: 1. Frequency of use. By this measure, as used by Fries and Pike, Hasselmo, Mackey, and Murphy, the more frequently a specific donorlanguage item is used in recipient-language discourse and by more people, the more reasonable it is to consider it as having become a bona-fide term of the recipient language. 2. Native-language synonym displacement, as measured by the translatability test used by Hasselmo and Mackey's availability test, and implicit 104 S. Poplack and D. Sank08 in Weinreich's discussion of lexical integration. If a borrowed term can be shown to displace in usage an indigenous term for the same concept, it can be considered to have taken over the latter's role in the lexicon. 3. Morphophonemic and/or syntactic integration. If a borrowed term takes on a phonological shape typical to the recipient language, acquires the morphological affixes appropriate to that language, and functions in sentences as a native word of some syntactic category, then it can be considered a well-established borrowing. This approach is embodied in the work of Fries and Pike, Bloomfield, Weinreich, Haugen, and others. 4. Acceptability. If native speakers judge a donor-language word to be an appropriate designation whether or not they are aware of its etymological origins, this is indicative that it may occupy a place in the recipient lexicon. Not all of these criteria, however, will be satisfied in all cases which we may want to consider loanwords, and each of them may be satisfied by words which are not. For example, a word from one language [nay be used frequently in discourse which is predon~inantly of another language, but only because it occurs often in code-switches (e.g. the determiner the occurs frequently in switched NPs). A borrowed word may be phonologically, n~orphologically, and syntactically integrated into the recipient language but only because the speaker has little productive competence in the donor language or simply because of interlingual coincidence between donor and recipient codes. Acceptability is notoriously misleading, especially in contexts where the recipient language is socially inferior to the donor. Even in cases where neither language is stigmatized, Hasselmo documents for Swedish-English bilingualism cases where items were identified as being of English origin, yet showed low translatability, but high acceptability (1969: 71), results which are dillicult to interpret in terms of integration into the linguistic repertoire, short of arbitrarily assigning supremacy to one of the criteria involved. Similar difficulties were encountered in a later replication of this study among Chicano bilinguals by Murphy (1974), leading him to suggest that not only these tasks but a'lso the very attempt to get bilinguals to establish language boundaries are inappropriate (1974: 63-64). Synonym displacement may be a solid criterion, but only when a single borrowed word is displacing a single well-identified native word, a situation which is not necessarily the rule and which is very difficult to demonstrate, since the precise referent of the word may be impossible to reconstruct at the time of the analysis. Most of these criteria are based on anecdotal evidence, albeit on the part of observant and highly insightful scholars, and remain empirically unsubstantiated. For example, the proposed correlation between frequency of use and degree of linguistic integration, or between degree of Borrowing: the synchrony of integration 105 acceptability and linguistic integration, quantitative hypotheses both, have never been quantitatively tested. The role of bilingual versus monolingual, or older versus younger speakers, in introducing and propagating loanwords has never been empirically investigated or established.' Parenthetical comments on the vacillations which mark the incorporation of transferred words into the vocabulary cannot capture the orderly heterogeneity which has been documented for all aspects of speech behavior at the community level. Conversely, the empirical studies which do exist, largely on acceptability arid availability, do not take into account the more strictly linguistic concomitants of loanword assimilation. Nevertheless, the four types of criteria discussed above are useful in that they are abstracted from the key processes which make up the phenomenon of lexical borrowing. It is in fact reasonable to assume that as a borrowed word is more and more used, it tends to become phonologically and n~orphologically integrated, to displace competing recipient language forms,' and at least eventually, to be accepted by its native speakers. 3.0. Data and methods 3.1. The contact situation The data on which this paper is based were collected from 14 children and eight adult residents of a stable bilingual (Spanish-English) community in East Harlem, New York, one of the oldest continuous Puerto Rican communities in the United States. Though English has coexisted with Puerto Rican Spanish to some degree since the American occupation of the island in 1898, the most intense contact has occurred since the end of World War I1 with a massive population influx to the United States, concentrated almost wholly in the geographically circumscribed area in which this comn~unity is located. Despite widespread unfavorable attitudes of the non-Hispanic population at large toward Puerto Rican Spanish language and culture, and generally strong pressure to assimilate to the ways of the mainstream, there is as yet no conclusive evidence in this comnlunity of language shift, even anlong third-generation bilinguals, partially due to a pattern of circulatory movement between Puerto Rico and the United States and to a more recent influx of Spanish speakers from other parts of the Caribbean, both of which have the eHect of replenishing the stock of Spanish-dominant or monolirigual speakers and hence revitalizing the language. Puerto Ricans in New York may be 106 S. Poplack and D. Sankof said, then, to be undergoing intense, though short-term 'cultural pressure' from source-language speakers (Thomason 1981: 14), a situation which generally results in copious borrowing of (at least) donor-language lexical items if not also foreign s t r ~ c t u r e . ~ Though several aspects of the grammatical structure of local Spanish have in fact been shown to be free of influence from English (Poplack 1981), incorporations from that language, in the form of both code-switching and borrowing, are numerous enough to be remarked upon and stereotyped by Puerto Ricans and non-Hispanics alike.