EDITOR: Peter Siemund TITLE: Linguistic universals and language variation SERIES TITLE: Trends in Linguistics. Studies and Monographs [TiLSM] 231 PUBLISHER: De Gruyter Mouton YEAR: 2011
Anish Koshy, Department of Linguistics & Phonetics, The English & Foreign Languages University, Hyderabad, India.
INTRODUCTION The present volume consists of 13 contributions from 21 scholars plus an introduction, organized in 4 major sections, namely, ‘Varieties and cross-linguistic variation’ (4 papers), ‘Contact-induced variation’ (2 papers), ‘Methodological issues of variation research’ (3 papers), and, ‘Variation and linguistic theory’ (4 papers).
SUMMARY The editor’s introductory note provides a sketch of the volume’s structure and notes that most of the research projects reported here study both the limits and origins of variation. They also address methodological issues apart from taking up questions on reliability of data while making universalist claims from variation studies. Several of the papers originate from a workshop on Linguistic Universals and Language Variation (UniVar), held at the University of Hamburg’s Research Centre on Multilingualism, July 2007.
Diessel and Hetterle argue that causal adverbial clauses, unlike temporal and conditional clauses, are closer to coordinated structures as they are only loosely connected to their main clauses. Adverbial clauses are seen to (a) use the same verb forms and arguments like ordinary main clauses; (b) be placed after the semantically associated clause, and (c) be commonly expressed by a separated intonation unit in conversation. These three patterns are said to be motivated by “their communicative function in speaker-hearer interactions”, where “causal clauses are commonly used to support a problematic statement” (24), that is, to support interactional disagreement or misunderstanding. The three major patterns are determined statistically. Semantic and pragmatic features are analyzed through analysis of conversational discourse.
Loporcaro investigates two Euroversals, often considered independent of each other, namely, the alternation between ‘be’ and ‘have’ in perfective constructions, and the pattern of accusative alignment . It is argued that the crucial choice between ‘be’ and ‘have’ is often dictated by the system of alignment that a language makes use of, and hence, claims of an overall pattern of accusative alignment for European languages cannot be maintained. Auxiliaries become relevant to the system of alignment when seen as part of the cross-referencing system of subjects/objects on verbs. It is observed that perfect auxiliaries are not accusatively oriented and are rather part of an active/inactive alignment.
Exploring parametric variation in languages with regard to allowing or disallowing determiner-possessive combinations in possessive Noun Phrases (NPs), Kupisch and Rinke investigate the diachronic development of possessive NPs in modern Italian and Portuguese and argue that modern varieties are of the AG-type even though 13th century varieties also exhibited bare possessives. The parametric difference between AG (Adjective-Genitive) and DG (Determiner-Genitive) languages is whether the possessive/genitive is treated as an adjective or as a determiner, the former allowing articles in such phrases, and the latter ruling them out. Statistical analysis of a corpus of Italian and Portuguese possessive Determiner Phrases (DP) from texts from the 13-19th centuries shows a gradual process of the possessive NPs becoming strictly AG type in the modern variants. A structural change leading to the grammaticalization of the demonstrative and its subsequent reanalysis as a definite article is considered a more crucial determining factor in a language being AG or DG, than the categorical status of the possessive itself. That the final shape of variation could be guided by structural changes in other related domains is thus interestingly brought out.
Investigating variation in the default agreement pattern in the use of past tense ‘be’ in thirteen different varieties of English, with ‘was’ representing default agreement, Tagliamonte explores the universality of conditioning constraints for agreement. The possibility of a hierarchical order in the operation of these constraints, and the role of social and other factors favouring or disfavouring certain constraints thereby also promoting or inhibiting variants is also explored. In many varieties of English, in spite of the subject being plural, the agreement on the be-verb is singular (default agreement). A statistical analysis of the data finds that existential constructions, 2nd singular ‘you’, Negation, and 3rd plural NP are often seen to promote default agreement. Often explained in the literature as a regularization process, the paper shows that default agreement is not a modern phenomenon and has existed for centuries. That regularization is not a feature limited to vernacular forms of languages is also brought out in the paper.
Contesting the view that contact does not always lead to reduction in the sense of simplification, Enger explores the reduction in the gender system in Scandinavian usually attributed to contact. Problematizing the simplistic view of reduction due to contact, the paper argues for a system that works by compensation -- the possibility that simplification in one domain may lead to or may occur parallel to complexity in another. Looking at plural declensions, verb inflections, lexical and referential gender, Enger claims that diachronic change represents reduction of irregularities and that numerical increase does not necessarily mean increased complexity. It is strongly argued that language internal factors like sound laws need not work to the exclusion of language external factors like contact, and that a purely “phonological account of the loss of the feminine plural agreement” (194) does not suffice and that “contact has mattered” (195).
Exploring whether certain categories are easier to borrow, and if so, what semantic/pragmatic factors favour this ease, Matras understands contact as not just a prolonged use of a pair of languages, but also as some degree of bilingualism. While a review of literature on borrowing covers borrowing scales, structural autonomy, semantic transparency, low paradigmaticity, etc. as some possible factors, the paper postulates borrowing as the “removal of an invisible demarcation line that separates subsets within the linguistic repertoire” (204). Two large-scale samples exploring borrowing hierarchies suggest the independence of meaning and structural autonomy as aiding factors. This is expanded to include semantic gaps, overburdening due to parallel structures, factors of prestige, speakers’ control over form-selection and the blurring of demarcations.
Contending that all linguistic theories must take into account functional and social factors, Bisang argues that the requirements of reproducibility and falsification cannot be met in variation studies due to the nature of the contributing factors. With social factors like power, prestige and identity determining linguistic choices, language cannot be a self-contained system. Since speakers do not necessarily produce the same utterance every time and a corpus never provides the full social background of the data, the two most widely used methods of collecting data become irreproducible -- a combination of text based research and elicitation may thus be more useful. That social and functional factors are important even for typological generalizations is underlined by the potential inaccessibility of relevant facts due to the probability sampling requirements and due to non-inclusion of dialectal varieties which may be at considerable structural variance to the standard variety.
Drawing up various parameters that influence morphosyntactic variation in World Englishes, Kortmann and Szmrecsanyi attempt to explore through the notion of Vernacular Universals, morphosyntactic features and strategies employed by individual types of varieties in any language. Understood in the literature to refer to phonological and grammatical processes recurring in vernaculars wherever they are spoken, Vernacular Universals are not attributable to sociolinguistic factors, and are considered unlearned and innate features of vernacular dialects independent of their typological makeup or areal features. Varieties of English are explored in terms of high and low contact between the varieties; as different variety types like L1, indigenized non-native L2, pidgins and creoles; as spoken and written varieties and also in terms of their analyticity and syntheticity.
Taking language contact, dialect contact and multilingualism as causative factors, Davydova et al. explore varieties of English to analyze recurrent cross-dialectal grammatical patterns in domains like pronominal systems, tense/aspect systems, and embedded questions. Many grammatical features like subject-auxiliary inversion in embedded clauses, copula-drop etc., shared by non-standard varieties of English in spite of no obvious historical contact, are claimed to strengthen the arguments for vernacular universals and angloversals. Comparison across varieties is proposed along three dimensions: typological hierarchies revealing speakers’ access to ordered relations of cognitive domains; proximity of a variety to some reference variety exploring substrate influence or looking at contact-related origin scenario or imperfect L2 acquisition leading to simplification or overgeneralization, and; source/origin of a non-standard grammatical phenomenon to see effects of contact.
Investigating variation across speakers, dialects, modes and genres, in the placement of the finite verb in main clause declaratives in Norwegian, considered a strict verb second (V2) language, Eide and Sollid look at each individual speaker as a multilingual with multiple parallel grammars/sub-grammars at command. Intra-individual variation is attributed to a number of simultaneously internalized grammars. If the same language has multiple internalized grammars, then the validity of grammaticality judgments would be compromised, unless multiple systems that co-exist are not allowed to interact and contaminate each other. Corpus studies reveal that some V3 structures are used by speakers to express in-group identity and solidarity, while most V1 and V2 and other V3 structures are explained in terms of verb-raising.
Approaching variation and change in Romance languages through the minimalist approach, Mensching and Remberger propose that variation must be handled via a system of parameters defined by the feature composition of functional categories, along with lexical items stored in the mental lexicon. With a powerful lexicon, distinction of two or more varieties of the same language spoken by an individual is reduced to as many different lexicons, with the only issue left for analysis being which kind of differences demand postulating separate lexicons. Aspects of the Minimalist program with special relevance to syntactic variation studies as well as the state of the art in comparative Romance syntax form part of the discussions. Parameterization at the level of the lexicon for synchronic and diachronic variation is expressed in terms of EPP (Extended Projection Principle) features, ø - features (agreement features), HAF (Head-Attraction Feature), etc.
Stressing that universals of different degrees of particularity must always be understood against a starting point of potential individual language variation, Joseph advocates a localistic approach to the study of universals and variation with the speaker at the centre, who always favours/prefers local solutions or small-scale generalizations. Ascribing a larger role to functional and cognitive factors, variation is seen as a result of incomplete generalizations by the speakers, and universals as a result of speakers reacting to the same “sorts of stimuli” and coming at those stimuli “with similar cognitive preferences” (404) as seen in the preference for symmetry/patterns and filling of gaps. While variation has often been granted a functional basis, the paper makes an important claim by arguing that even universals are rooted in the human act of using language for interaction, making the role of innateness almost peripheral. The paper advocates a combination of the innateness and derivative approaches.
Investigating language variation with respect to sounds and its relation to syntax, morphology and the lexicon, Hinskens evaluates both rule-based and usage-based approaches to variation. While usage-based models give particular importance to large scale storing of information in the lexicon as speakers encounter the information and see language structure as emerging from language use, rule-based approaches postulate innate abilities to generalize and categorize, conceptualizing the lexicon as a list of exceptions (after Bloomfield). Usage-based models see the lexicon rather as a “network of prototype-wise organized words, phrases and constructions” (438). Variation and change are accounted for in terms of distributional and usage frequencies. Presenting a quantitative case study of reduction and deletion of unstressed vowels in spoken modern standard Dutch and variable deletion of word-final /t/ or /d/ in several Germanic languages, the paper claims that the latter cannot be rule-based or dependent on the structure of the language and hence cannot be derived from underlying full-forms, thus underlying the role of frequency of occurrence/use.
EVALUATION The volume highlights ongoing research and its importance in drawing/exploring the relationship between linguistic universals and language variation, a timely topic with wide implications. Both the functionalist-inductive (language typology) and the formalist-deductive (Generative grammar) approaches representing the two major strands of research in language universals get a fair representation in the volume; as also do all the stakeholders in variation research: typologists, sociolinguists and dialectologists, as well as those involved in issues of language contact, language acquisition and language change. The book comes out as a fairly successful one in what it had set out to achieve in terms of bridging the “rather artificial theoretical divide” (5) between the formal and functionalist approaches. And thus, the volume would be of interest to both practitioners of formal and functional linguistics.
Most studies reported here are empirical, and this explains the statistical nature of many of the conclusions reached. Universals as tendencies, established statistically as implicational and non-implicational universals (applicable also to Vernacular Universals), is a common thread in the volume. Statistical tools lead to postulation of implicational hierarchies, an indication of how best to understand claims about both universals of language and variation. A majority of papers in the volume stress that facts about language can at best only be tendencies, not absolute laws.
An important thread running through the papers is a focus on dialectal and vernacular variants of standard languages. These papers not only bring out their significant features but also question the rationale and validity of typological and other universal generalizations which ignore their distinctiveness and assume that valid statements on the universal properties of language can be made without taking them into consideration. Loporcaro’s disagreement with those who claim European languages exhibit routinely similar patterns, brings to focus the need for larger studies taking serious note of dialectal variations. Bisang problematizes sampling requirements and the non-inclusion of dialectal variants in investigations, leading to the potential ignoring of certain important typological generalizations. Kortmann and Szmrecsanyi also criticize typologists and sociolinguists for having completely ignored vernacular varieties of most standard European languages as well as pidgins and creoles. This has amounted to giving no role to language contact in the shaping of universals in theories. Davydova et al. also take up this issue forcefully.
There is also a concerted effort to advocate for an integrated approach to understanding variation and universals drawing on the best that functional and formal approaches have to offer. The integration of approaches leads to superior explanations over those based on compartmentalized approaches. While doing so, many have also brought out the limitations that either approach suffers from. Loporcaro castigates functional approaches for their narrow perspectives on grammatical matters and generative approaches for glossing over variation in smaller varieties. Eide and Sollid opt to explain variation using an integrated approach that leaves space for both generative (P&P approach) and sociolinguistic (accommodation theory) perspectives, whereby the speaker is said to have a choice and depending on the social context of use chooses the appropriate form. Joseph also places the speaker at the centre of variation and universals research by claiming that universals are products of what speakers do with their languages. In fact, Mensching and Remberger’s theory-driven paper stands alone as an exception to the integrated approach advocated by many.
Variation is approached from various perspectives in the papers. The diachronic perspective is brought out by Kupisch and Rinke, underlining how variation characterized languages in their older stages as much as they do in their current forms. Davydova et al. emphasize that varieties are not random in their structural forms, and hence require detailed in-depth studies including unearthing historically relevant details from even seemingly similar structures in different varieties.
The question of what leads to variation has been approached in most papers, not in an either … or manner with respect to language internal (structural) and external (socio-historical) factors, but rather as an end-result of both these factors playing out their respective roles leading to variation in patterns that finally shape universal patterns. Tagliamonte highlights this with multilayered constraints which sometimes work in cross-purposes. Enger highlights this using a highly recessive system like gender problematizing the common understanding of simplification in grammars in terms of numerical increase/reduction. Hinskens underlines the importance of usage-based approaches to variation especially when variation is not determined by the structure of the language. Bisang argues that exact reproducibility and hence falsifiability may not be possible in a methodology that combines the two factors, even though he still finds it a useful premise to prevent premature conclusions while evaluating linguistic generalizations. Bisang elevates social-functional factors as being crucial not only at the level of language but also for typological generalizations.
Borrowing, a prime factor in variation, is also approached from fresh perspectives. Borrowing hierarchies are posited more as trends than absolute laws. Matras links borrowing to the multilayered repertoire of linguistic structures of speakers, thereby making it contingent on pressures of interaction contexts, factors of accommodation and motivation of bilingual speakers.
If one has to labour to point out a drawback of the volume even though a minor one, it has to be regarding the representation of languages and language groups. Methodologically, only a few papers are seen to make use of large database, in spite of the almost axiomatic assumption in variation and universal studies today that conclusions about crosslinguistic variation must be built on careful samples. For a volume claiming on the back cover to be correlating linguistic variation of different kinds (cross-linguistic, regional, diachronic, contact-induced and socially-conditioned), in its attempt to highlight universal patterns underlining surface variations, it is indeed surprising to see that many papers make claims regarding variation with implications for universals, based solely on languages from a particular stock, reminiscent of the days when generative linguists worked on the assumption of unraveling universal tendencies among languages based on in-depth study of a single language as representing a token instantiation of Universal Grammar. Diessel and Hetterle’s and Matras’ papers are exceptions to the above. That the volume follows from a Workshop in Hamburg, might have dictated the shape the volume has finally taken and the languages that have been represented. The title of the volume could have had a sub-title making it more explicit than the grander one it has chosen to retain. This minor drawback aside, that most of the papers present original research findings and not mere speculations of theory make it a worth-while collection. That many of the beholden principles and received wisdom of issues related to variation get questioned is a positive direction for language universal and variation studies, which have often been crippled by an unquestioned adherence to theoretical baggage.
Most work on language universals or variation is driven by theoretical persuasion. So while most works on variation take a sociolinguistic perspective, universals are largely dealt with by either typologists or generative linguists who have often approached the subject as though there could be no meeting ground. Much emphasis here is on integrated approaches, bringing the best of the formalist and functional approaches together for better explanations, making it a timely publication in linguistics where much time has been wasted in being more loyal camp followers than in being critically appreciative of divergent views and explanations. The book holds much promise in the shaping of future research methodologies and research goals.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Anish Koshy is an Assistant Professor in Linguistics in the Department of
Linguistics & Phonetics, The English & Foreign Languages University,
Hyderabad, India. His research interests lie in the lesser-studied
languages of India and South Asian languages from a typological
perspective. He is currently working on the typological nature of clitics
in the Austroasiatic languages of India, namely the Munda and the Khasian