- Shopping Bag ( 0 items )
1 The study of dialect convergence and divergence: conceptual and methodological considerations
Frans Hinskens, Peter Auer, and Paul Kerswill
Dialect change can have several different manifestations. Among these, dialect convergence (dc) and dialect divergence (dd) noticeably affect the relationships between related dialects. Dc and dd have probably been present for as long as dialects have existed. Various historical developments, including the 'modernisation' of society, have left their mark on the very nature of dialects and have partly changed the dynamics of dc and dd; moreover, they have broadened them to dialect - standard language convergence.
This chapter sets the stage for the various aspects of the study of dc and dd presented in this book, in that it both provides a general introduction and constitutes a springboard for the discussion of the themes and approaches which play a role in the individual chapters. As an introduction, the chapter presents the central terminology (section 2), provides the background information necessary for the interested non-specialist (section 3), sketches what we see as the main research methods (section 4), and binds together the issues featured in the various chapters (section 5).
2 Definitions of the Key Concepts
We will use the notion of 'dialect' to refer to a language variety which is used in a geographically limited part of a language area in which it is 'roofed' by a structurally related standard variety; a dialect typically displays structural peculiarities in several language components (cf. Chambers and Trudgill 1998: 5), though some of the authors in this book deal mainly with phonetic (or 'accent') features. Usually dialects have relatively little overt prestige and are mainly used orally. Lacking in this definition is the fact that the dialects of a certain language area (including the standard variety) maintain very specific historical relationships (cf. Agard 1971: 21-24).
The notions of dc and dd can be defined, respectively, as the increase and decrease in similarity between dialects. Whereas dc involves the linguistic unification, focusing (sensu Le Page and Tabouret-Keller 1985), and homogenisation of the linguistic repertoire, including the traditional dialects,1 dd amounts to linguistic diversification, growing diffuseness and heterogenisation - although divergence may lead to focusing in a repertoire, making the varieties which survive the process more distinct from each other. Weinreich (1954: 395) defines convergence as 'partial similarities increasing at the expense of differences' (though, in his view, divergence is the main subject matter of diachronic dialectology). As should be clear from these definitions, convergence and divergence are relational notions, referring to either processes or the results of processes.
Dialect convergence may lead to simplification (Trudgill 1986) and to the reduction of intrasystemic, especially 'quantitative', variation. However, in most studies of dialect convergence or divergence, attention is only paid to the question of how processes of linguistic change affect intersystemic variation, i.e. differences between dialects. These differences can pertain to either categorical or quantitatively variable features.
Sometimes dc and dd are two sides of the same coin. Gilles (1998b), for instance, shows that the dc of Letzebuergesch towards the central Luxemburg variety implies its giving up of east Luxemburg features and, hence, a divergence from Mosel Franconian dialects of German. Ó Curnáin (1998) demonstrated how, in the same West Galway vernacular of Irish, in the segmental phonology dc and dd can coexist. Pedersen (1998) showed how, in the course of the nineteenth century in Copenhagen and Stockholm, the convergence of the stylistically marked differences between urban dialect and the spoken standard and divergence of the socially marked differences between both systems occurred simultaneously.
Dc and dd can change the relationships between the dialects involved and may, hence, necessitate the reclassification of the dialects involved (cf. Samuels 1972: 92).
3 Background and Conceptual Frameworks
To bring the concepts of dc and dd more clearly into focus, we will now present a rough overview of the historiography of the study of dc and dd (section 3.1) as well as a short discussion of related concepts in contemporary approaches to dialectology, in two branches of sociolinguistics, and in the study of language contact (section 3.2).
3.1 Historiographical sketch
Here, we briefly discuss some of the main models, theories, proposals, findings, and individual observations which are relevant to, and can sometimes retrospectively be seen as precursors of, much of the present-day study of dc and dd. We will largely concentrate on the areas of historical linguistics and traditional dialectology.
The manifestations of dialect divergence are most visible in language history. The long-lasting process of the diversification of Proto-Indo-European into (what are retrospectively referred to as) language families, and of language families into languages, etc., largely took place in linguistic prehistory. The results of divergence are represented visually in the branching lines in the family tree diagrams of historical linguistics.
Undoubtedly, the most influential school of historical linguistics is that of the Neogrammarians, one of the main spokesmen being Hermann Paul. Applying a partial analogy from nature, Paul recognises only the language of the individual, the idiolect, which is the product of ontogenesis and phylogenesis. 'Dialect split means, simply, the increase of individual differences beyond a certain measure' (Paul 1920: section 22; our translation). He asks the question of why it is that 'a greater or lesser amount of agreement is maintained in this group of individuals which is constituted in this or that way'. The reason is that language habits ('Sprachusus') are determined by human interaction ('Verkehr'), which has either a levelling or a differentiating effect (Paul 1920: section 23). 'Each change in language use is the product of the spontaneous behaviour of single individuals on the one hand, and the nature of interaction on the other. If instances of spontaneous behaviour are very differently distributed in the various districts, then the levelling (to the extent that it is necessary) taking place in districts which are remote from each other and have no mutual interaction must necessarily lead to different results' (Paul 1920: sections 22-25, our translation).
The Neogrammarians distinguished between language change in the strict sense and borrowing. Language change has language-internal origins. Formal (rather than semantic) change can take the form of either sound change, which is achieved spontaneously, or analogical change. When a change is not achieved autonomously, that is, when it does not have an internal origin, it can either stem from another language or 'from within the same speech area', as stated by Bloomfield (1933: 444), who referred to the latter type as 'dialect borrowing'. Sound change was claimed by the Neogrammarians to be lexically exceptionless, hence the designation 'sound laws'.
Only a few historically attested instances of sound change appear to be completely exceptionless, however. The fact that, in the grammar and lexicon of individual dialects, regular and exceptional ('residual') forms can often be found to exist side by side has led many scholars to subscribe to the idea that 'sound laws', especially, tend to operate sporadically, which leads some, including Schuchardt, to conclude that sound laws do not exist.
A process of language change that has not come to completion in some respect leaves behind language variation, either intrasystemically (as in, for example, lexically diffuse sound change) or between closely related language varieties (e.g. dialects or style levels). In traditional dialectology much attention is paid to intersystemic variation. Natural and man-made borders were typically looked upon as explanations of the location of dialect boundaries as the outcomes of dialect divergence (or, rather, non-convergence).
Apart from divergence, lexical dialect mixing (Mischung) and the levelling of variation (differences) between dialects (Ausgleich)2 were thought to be the key mechanisms that destroy regularity and the alleged exceptionlessness of the 'sound laws' and thus made it impossible to reconstruct historical developments from the geographical distribution of particular forms, the original aim of nineteenth-century historical linguistics and dialectology (Dauzat 1922: 22). The insight developed that the forces constituting individual dialects and dialect landscapes are not only the human linguistic 'hardware' (to use a modern expression), such as the articulatory organs (the possibilities and limitations of which were held responsible for 'sound laws') and the 'software' located in the brain (cognition being held responsible for analogy), but also social interaction, social networks, contact between places, etc., leading to all types of what the Neogrammarians labelled dialect borrowing. 'The maps showed . . . that local dialects do not exist in a state of isolation from one another' (Bynon 1983: 185).
As early as 1870 Schuchardt, who, after the publication in 1885 of his Über die Lautgesetze, gegen die Junggrammatiker, was the first leader of the opposition against the Neogrammarian views, distinguished between two opposite forces working on language. What he labelled 'centrifugal force' (Zentrifugalkraft) leads to the differentiation of language, whereas 'centripetal force' (Zentripetalkraft) aims at unity. Centripetal force exerts its influence through such institutions as the school, the church, and the state. In Schuchardt's later writings, these notions occur under the headings of Spaltung (split) or Divergenz (divergence) and Ausgleich (levelling) or Konvergenz (convergence), respectively.3 Reflecting on the mutual influence between the standard variety and a dialect ('langue littéraire et idiome local'), one of the founding fathers of modern linguistics, de Saussure, writes that language history is a continuous struggle between 'la force d'intercourse et l'esprit de clocher', i.e. between the tendencies towards unification and those towards particularism and cultural fragmentation.4 German dialectologists coined the notion of Abbau to refer to levelling in the dialect-standard language dimension.
Applying these insights to historical dialectology, Frings (1936) accounted for the emergence of a German Gemeinsprache (common language) as a consequence of the convergence of the Middle German settlers' dialects in what was to become the Upper-Saxonian area. Migration as a force for levelling will be a recurring theme in this volume. Historical dialectology provides evidence for divergence, too. Goossens (1970) points to the following trends, which led to the divergence of the dialects of Dutch and German:
1. specific linguistic elements or structures in the German 'dialect cluster' underwent changes that did not occur in the Dutch cluster;
2. the Dutch cluster underwent changes that the German one did not undergo;
3. both clusters underwent different changes.
Little by little the dialect-geographical investigation of dialect boundaries was given up in favour of the study of the history of individual words, leading to extreme positions such as the one expressed in the famous dictum 'chaque mot a son histoire'. In the eyes of many linguists, traditional dialect geography is characterised by atomism and, in the worst cases, complete abstinence from theoretical reflection. That this extreme position was perhaps rare is indicated by the fact that most dialect atlases contain maps based on phonological parameters, implying that there is a general rule behind the change in the phonological shape of the words.
3.2 Fencing off dialect convergence and divergence from related concepts
In this section, dc and dd will be compared to closely related notions from sociolinguistics, especially the social psychology of language; from both traditional and more modern approaches to dialectology, including levelling and koineisation; from pidgin and creole studies; and, finally, from the study of 'mixed languages'.
3.2.1 Accommodation and variation Convergence and divergence both have short-term and long-term manifestations. Their short-term manifestations are often discussed under the heading of accommodation, and, in Giles' et al.'s (1987) model, are the opposite of non-accommodation, though more usually (e.g. Trudgill 1986) accommodation is associated just with convergence. Short-term convergence is exemplified by the observation that in babies' utterances F0 often has lower values when the infant is interacting with the father than during interactions with the mother (Giles and Powesland after Daan et al. 1985: 72; see Kerswill 2002a for further examples and discussion). In adults, short-term convergence can be either 'upwards', i.e. from a geographical or a social dialect towards the standard language (although there is usually an upper boundary that cannot be transgressed without sanctions) or 'downwards' (as when members of the local elite speak dialect (Daan et al. 1985: 76; Voortman 1994), a phenomenon for which there are such telling labels as 'magistratenplat' ('magistrate's dialect').5 Of course, Giles' description presupposes a neatly hierarchically structured society in which dialect and standard can be related in a straightforward way to social position, but this may not be the general pattern. A demonstration of short-term accommodation is Coupland's (1984) account of a Cardiff travel agent's response to her clients. Coupland argues that the accommodation on various phonetics variables is not a mechanistic matching of frequencies, but rather an attempt at 'identity projection'. (See Kerswill 2002a; and AUER AND HINSKENS.)
Motivations for short-term divergence may range between strictly situational (the desire to distance oneself from one's conversational partner) or more general (the need to develop, maintain, or stress social or personal identity, or to demarcate the ingroup from the relevant outgroup). In the longer term, language can thus become the symbol of an entire minority group (as in the case of Welsh, Basque, Catalan, Frisian, and maybe also in the case of Letzebuergesch).6 The divergence of African American Vernacular English (Labov and Harris 1986) from white dialects, which resulted from the fact that the AAVE speakers have not participated in any of the sound shifts characteristic of the white vernaculars, may originally also have had this motivation.
Both short-term convergence and short-term divergence can take place psychologically and/or linguistically. Psychological accommodation (convergence or divergence) has to do with the communicative intentions and attitudes of a speaker towards his interlocutor or audience, and may not result in actual linguistic accommodation. While linguistic convergence can be described as the linguistic manifestation of speakers adapting 'to the speech of others to reduce differences' (Siegel 1985: 367), divergence is the exploitation of differences, for example by using different features more often and thus making them more salient. In sum, 'according to this theory, people may adjust their speech with others either to reduce or to accentuate linguistic (and hence social) differences between them' (Siegel 1987: 240).
While Labovian sociolinguistics associates language use primarily with social structures and social behaviour, research on linguistic accommodation is grounded in theories of 'social action' (how social meaning is produced from interaction), more specifically 'rational action' (Turner 1996). Linguistic accommodation is analysed as the outcome of more or less conscious choices on the part of rational social actors, the choices being tailored to expectations about their extralinguistic consequences. If the boundaries between linguistically distinct groups are permeable (Mummendey 1999), the speaker may benefit by moving closer to the other group by converging linguistically, either by the avoidance of salient features of the speaker's own dialect or by the adoption of features of the interlocutor's dialect. This can affect the interlocutor's attitudes and behaviour in positive ways. As in Giles' theory, accommodation by Trudgill's (1986) more restricted definition of linguistic convergence may take the form of the reduction of differences or the adoption of features from the dialect spoken by the interlocutor.
Some of Gilles' (1998a: 73) findings show that psychological convergence (or divergence, for that matter) is not necessarily expressed in linguistic convergence (or divergence, respectively). Gilles' findings do not provide any evidence for short-term, interactional convergence between speakers of different dialects of Lëtzebuergesch: 'We are dealing with a process of convergence which can be located solely in the speakers' mind, but has no effect on their actual verbal behaviour' (73).7 Blom and Gumperz's (1972) finding that students from Hemnesberget (a village in Norway) who had been living in the city claimed to speak the local dialect yet had adapted their speech to one of the standard varieties, shows how psychological non-accommodation can go hand in hand with long-term objective linguistic divergence.
In the case of Serbian/Croatian, psychological divergence between several ethnic and religious groups seems to be leading to growing structural divergence between the respective dialect groups (cf. Janich and Greule 2002; Grčević 2002; Gvozdanović in press). Psychological convergence may be the reason why some linguists (Angelov 2000) have come to regard Macedonian as a dialect of Bulgarian. However, national ideologies probably also play an important part in this judgement.
3.2.2 Dialectology Whereas in connection with the analysis of dialect borrowing the focus is on the overall effects on the 'recipient' dialect, in connection with geographical diffusion (or expansion or areal diffusion), the focus is on specific dialect features. The German dialectologist Theodor Frings (e.g. Aubin et al. 1926) can be called the main protagonist of the 'expansionist' approach in dialect geography, which was essentially an elaboration of Johannes Schmidt's 1872) 'Wellentheorie' (wave theory). As Bynon (1983: 192-193) points out, Schmidt's wave model can also account for dc through 'the elimination of specific isoglosses which previously served to differentiate . . . two dialects through the spread of features from one dialect area over the territory of the other . . . The degree of such convergence will clearly depend both upon the length of time during which they previously underwent separate developments as well as the length of time during which they were subsequently subjected to the influence of a common centre' such as the growing influence of the standard language. (See Britain 2002a.)
A textbook example of the wave-like areal diffusion of an innovation is Kloeke's (1927) account of the spread of the diphthongisation of West Germanic /uː/ to /œyː/ (via an intermediate /yː/) from the cities in the northwestern Netherlands to the more peripheral parts of the language area in the sixteenth and seventeenth centuries. With respect to this phoneme, three rather than two 'clusters' of dialects resulted from the incomplete spread of the change, leading to dd.
Trudgill (1983) developed his 'gravity' formula (which we will sketch in section 5.4 below) in order to model the areal diffusion of innovations. Innovations are supposed to jump from large, influential cities to smaller, less influential ones, in order of decreasing size (the 'urban hierarchy'). To judge from Trautmann's (1880) and Trudgill's (1983) account of the spread of the uvular 'r' in northwestern Europe, this can occasionally even have Sprachbund-type effects. Trudgill (1992) compares the (1) dialectological, (2) macrosociolinguistic/geolinguistic, and (3) microlinguistic approaches to diffusion. In connection with (1), he discusses isogloss bundles and transition zones; in connection with (2) corridors of variability; and with respect to (3) linguistic accommodation. He points out that accommodation is usually not perfect. 'At the micro level, the best-known form of imperfect accommodation is hyperadaptation, and the best-known form of this is hypercorrection' (78). He illustrates this with data for the so-called Bristol 'l', which refers to the addition of /l/ word-finally in words such as idea and Norma, giving forms homophonous with ideal and normal. (See also Britain 2002b, 622-627; 2003.)
Bailey et al. (1993) introduce quantitative techniques to analyse the areal diffusion of grammatical, phonological, and lexical innovative dialect features of English in Oklahoma. Their findings led them to conclude that 'different patterns of diffusion are tied to the different social meanings that linguistic features carry' (386), and that 'innovations that diffuse hierarchically represent the encroachment of external norms into an area, whereas features that diffuse in contrahierarchical fashion represent the revitalisation of traditional norms'. We will return to this issue in section 5.4 below. Kerswill (2003) discusses differences in the rate and extent of diffusion of vowels and consonants, finding that, in Britain, consonantal features spread rapidly across the whole country, while innovations in vowels seem restricted to smaller regions.
In his work on the structural consequences of language contact, Van Coetsem (1988) draws a general distinction between source versus recipient linguistic systems, assuming that either can take the 'agentivity' role. Which linguistic system is proactive as the 'agent' depends on dominance, that is, on the bilingual's relative proficiency in the two languages. In this model, borrowing is a matter of recipient-language agentivity, while imposition8 stems from source language agentivity (2000: 5, 32). Strictly speaking, the notion of dialect borrowing refers to the process of one dialect copying an element or structure from another dialect; a long-term result can be the convergence of the recipient dialect with the source dialect. Kruijsen (1995) discusses examples of phonological (stress patterns) and morphological traits which were imported with French loan words in the Limburg dialects of Dutch spoken in the Belgian region of Haspengouw/Hesbaye near the Dutch-French language border.
A mechanism countering dialect borrowing is sociolinguistic polarisation. 'This force can act defensively, by retarding structural borrowing, but also offensively, by engendering developments diametrically opposed to what is found in other dialects or by bringing about something like hypercorrections in reverse' (Hock 1991: 428). The first type of effect comes down to resistence to convergence; the second one results in divergence towards the other dialects through hyperdialectisms. It would seem that a precondition for sociolinguistic polarisation, be it defensive or offensive in nature, is a certain level of awareness of the spreading feature in the consciousness of the speakers of the 'threatened' dialect. This may have played a role in the history of Hiberno-English (cf. Hinskens, Kallen, and Taeldeman 2000: 4). The defensive or offensive reaction may well have sociopsychological motivations, particularly non-integrative attitudes towards the speakers of the 'threatening' dialect. Some effects of polarisation in the creation of hyperdialectisms in the Flemish context are discussed in Taeldeman 2000.
Initially, because of extensive borrowing, dialect contact often leads to abundant variation as a result of dialect mixing, the partial merging of the lexicons and grammars of different but related dialects. Logically, the effects are visible only in areas where the original dialects used to be different. An example from historical dialectology is the enormous pool of variation resulting from interdialectal contact in sixteenth-century Judeo-Spanish, after the expulsions of Jewry from the Iberian Peninsula and their migration to the Balkans, Asia Minor, and North Africa (Minervini 1998).
Whenever dialect mixing leads to the stabilisation of the variants that are typical of the respective 'pure' lects along with additional 'compromise' variants, one usually speaks of fudging (cf. Chambers and Trudgill 1998: 110-118; Britain 2002, 2004). A well-known lexically intermediate or compromise form from the Cologne area in the German Rhineland and the dialects in the neighbouring transition area between Ripuarian and East Limburg dialects of Dutch is 'öllich', /ʌlıç/, or 'öllik', /ʌlık/, 'onion', which has been analysed as a fusion (a 'mixed compound', as Singh 1981 has baptised this type of formation) of the lexical variants 'ön' /ʌn/ (< Lat. unio), used in the dialects west and southwest of Cologne, and the more northern 'look' /lok/, related to standard German Lauch (Aubin et al. 1926: 32-33). As is evident from the latter example, convergence (and divergence for that matter) has consequences for the dialect landscape. Dialectal transition zones can result from partial cross-dialectal convergence (cf. Mazur 1996). In his discussion of what he labels interdialect, Trudgill (1992: 77-78) presents examples of intermediate forms in the phonological component (vowel quality in certain lexical sets of dialects of East Anglia) and the lexicon. Trudgill goes on to discuss the variation between German dialects in the word for 'potato', viz. Grundbirne, lit. 'ground pear', which is used in an area in between the areas with Erdbirne, 'earth pear', and Erdapfel, 'earth apple', respectively;9 a similar, more recent, example from British English, discussed by Trudgill, concerns central and southern take away, the northern variant carry out, and the intermediate take out, which is used in the southern part of northern England.
Phonologically intermediate forms are exemplified by the spread of '/oi/ instead of standard German /ai/, replacing the base dialectal form /a/ in the Rheno-Franconian area' (Ziegler, after Auer 1998a: 5). A similar example comes from the dialects of Dutch spoken in the extreme southeast of Limburg. The easternmost dialects have undergone dorsal fricative deletion with compensatory lengthening, yielding forms such as /naːt/, 'night', and /liːǝt/, 'light', which do not occur in the dialects spoken west of this area, which have preserved /naxt/ and /lıçt/ (which are identical with the standard variants). In a subset of the relevant items, the dialects in an intermediate area show vowel lengthening but no dorsal fricative deletion, hence /naːxt/ and /leːçt/ (cf. Hinskens 1992: section 12.2.1; 1998a: 47-48).
Independently of whether fudging occurs, in situations marked by heavy dialect mixing, after a certain period of time a process of selection usually takes place. After all, 'many mundane events suggest that people have a deeply ingrained attraction to linguistic conformity. The stigmatization of certain dialect features appears to be an overt attempt by communities to stamp out certain variants' (Chambers 1995: 209). This sifting through dialect levelling sometimes, though not necessarily, leads to koineisation, i.e. structural convergence between closely related linguistic systems, eventually leading to the stabilisation of some compromise variety (Hinskens 2001: 200; cf. Kerswill 2002a).
|1||The study of dialect convergence and divergence : conceptual and methodological considerations||1|
|2||Internal and external factors in phonological convergence : the case of English /t/ lenition||51|
|3||Dialect/standard convergence, mixing, and models of language contact : the case of Italy||81|
|4||Convergence and divergence in grammar||96|
|5||Phonology, grammar, and discourse in dialect convergence||135|
|6||Processes of standardisation in Scandinavia||171|
|7||The birth of new dialects||196|
|8||Dialect convergence in the German language islands (Sprachinseln)||221|
|9||Political borders and dialect divergence/convergence in Europe||236|
|10||The influence of urban centres on the spatial diffusion of dialect phenomena||263|
|11||Subjective factors in dialect convergence and divergence||287|
|12||How similar are people who speak alike? : an interpretive way of using social networks in social dialectology research||303|
|13||The role of interpersonal accommodation in a theory of language change||335|