This chapter is concerned with the nature of the relationship between language awareness and language change – in terms of interdependence and mutual influence. The basic division of the chapter builds on the possibility of raising the issue of influence from two perspectives: does change influence awareness, does awareness influence change?
This chapter is concerned with the nature of the relationship between language awareness and language change – in terms of interdependence and mutual influence. The basic division of the chapter builds on the possibility of raising the issue of influence from two perspectives: does change influence awareness, does awareness influence change?
Variation (and selection of variants) being the fundamental ‘mechanism’ of change (evolution) in all appearances of life on this planet – language being no exception – we may begin by noticing that ‘language change’ often will be more appropriately spelled out as ‘language variation and change’. Basically, language awareness is not only influenced by language variation and change – language awareness is the product of language variation and change, in the fundamental sense that we as human beings would not become aware of anything were it not for the fact that everything around us is subject to processes of variation and change. We may remain ignorant of the processes as such, but will register that they make a ‘difference’, and thus become aware of a ‘problem’ to be solved. However, as Dennis R. Preston – the leading scholar of ‘perceptual dialectology’ and ‘folklinguistics’ (Preston, 1989; Niedzielski and Preston, 2003) – always reminds us: in this area of research we need to distinguish between two kinds of human beings – linguists and ‘real people’. (See also Preston’s chapter in this volume.)
The role of change in arousing awareness of linguistic ‘problems’ to be solved is already evident in the earliest surviving text which deals with philosophy-of-language issues. Socrates’ discussion of the relative merits of viewing the names of objects as either natural/divine creations or man-made conventions (as told by Plato in Cratylus) emerges from an awareness of linguistic ‘difference’ resulting from variation and change. Notwithstanding their absurdity (from a modern point of view), Socrates’ many etymologies reveal an awareness that difference in meaning results from difference (substitution) in form: “Take, for instance, Διὶ φίλος; to change this from a phrase to a name, we took out the second iota and pronounced the middle syllable with the grave instead of the acute accent (Diphilus). In other instances, on the contrary, we insert letters and pronounce grave accents as acute” [399b]. 1 Socrates furthermore reveals awareness of language differences across space (i.e. dialectal variation) in discussions of what he calls ‘foreign names’ (see for example his discussion of the name Hestia [401b–c]), and across time: “on account of the lapse of time it may be impossible to find out about the earliest words; for since words get twisted in all sorts of ways, it would not be in the least wonderful if the ancient Greek word should be identical with the modern foreign one” [421d].
There seems to be no moral or aesthetic aspect to Socrates’ awareness that language varies and changes. He just observes ‘difference’ resulting from words having been ‘twisted’ in his own Greek in comparison with ancient Greek, and possibly more than in other dialects. However, as a companion to the advanced grammatical works of the Stoics in the centuries that followed, linguistic standards in terms of correctness and stylistic excellence emerged, with reference to the ancient writers. As the distance between this stabilized (written-language based) standard and the changing spoken language gradually increased, the awareness of variation and change as ‘difference’ turned into a conception of variation and change as ‘corruption and decay’. This conception saturated European linguistic thinking for the next two millennia and prevented any positive impact on language awareness from observations of facts pertaining to variation and change in language.
Nevertheless, it was indeed the introduction of new such facts to European linguistic thinking that eventually fuelled the showdown with the ‘corruption and decay’ doctrine and gave birth to the modern awareness of ‘genetic’ relationships behind language variation and change. In ancient India, too, the awareness of a language ‘problem’ had been aroused by the widening gap between written and spoken language, in consequence of which the language of the sacred texts of the Brahmin religion had been thoroughly described by Hindu grammarians and propagated for use in upper-caste speech under the name of Sanskrit. When the acquaintance with (the description of) Sanskrit spread among European scholars around 1800, its relationship with Greek and Latin and Persian (Iranian) languages was obvious, and hitherto confused notions of linguistic relationship quickly yielded to a common awareness that variation and change had to be studied in terms of systematic comparisons of how the form-and-content structuring of words corresponds across languages. Thus, even though Socrates (Plato) was already aware that many words appeared to be in a way both the same and different across geography and time, and was also aware that this ‘problem’ had to be solved by puzzling out how the expression of a content (the form of a meaning) had been constructed and ‘twisted’ differently, it was only with the works of Rasmus Rask, Franz Bopp and Jakob Grimm in the first decades of the 19th century that the assemblage of the puzzle bits was founded in a scientifically sound awareness of ‘genetic’ linguistic relationship. 2
At the time when the principles of systematic comparison were elaborated – in the historical linguistics of the 19th century, based on descriptions of Indo-European languages – the available technology for preserving the linguistic facts in which variation and change in speech could be studied had always been writing. The written texts exhibited the differences that had resulted from change, but did not disclose anything about the process of change. Quite naturally, change was in this situation held to be principally unobservable. Leonard Bloomfield stated in his book Language: “The process of linguistic change has never been directly observed; we shall see that such observation, with our present facilities, is inconceivable” (1933: 347). Awareness that observation of linguistic change is conceivable came with new facilities. Since the 1960s, the general availability of ever more advanced and powerful facilities for collecting, storing and analysing speech data (audio-recording and computer technology) has allowed sociolinguists, headed by William Labov (1972), to develop descriptions of how variation in speech is systematically involved with linguistic, social and ideological factors – in ways that arouse awareness that change in progress can be observed in empirical research, and induce trust that language variation and change can be scientifically explained in the long run.
When talk is of ordinary people (non-linguists) in linguistic literature, a more commonly used epithet than ‘real’ is ‘naïve’ – ‘naïve’ not in a negative sense, but as a neutral characterization of linguistically rather unaware people: i.e. people whose interest in language is to use it, not to study it and reflect on the ‘mechanisms’ that make it function, and change. In other words, we are talking of the great majority of people, and will consider whether their linguistic awareness is affected by language variation and change. As a first approach to the issue, it seems plausible to suggest that linguistically unaware people must be indirectly affected in the sense that the level of language awareness attained in linguistic thinking (as described in the section above) is likely to gradually become the common level. In that scenario, reference to ‘naïve’ people eventually becomes unjustifiable. Most people would be aware of the ‘mechanisms’ of language change, somewhat like most people today are aware that it is the earth that turns around the sun, rather than the sun turning around the earth. The general impression is, however, that most people are stuck in the two millennia long tradition of understanding change in moralistic terms of ‘corruption and decay’, while the basics of the scientific understanding developed during the last two centuries do not reach far beyond university courses in linguistics.
The explanation for the lack of indirect influence from scientific awareness on ‘naïve’ awareness no doubt has to do with the circumstance that the language awareness of all people (be they linguistically ‘naïve’ or not) is directly affected by their everyday experience with variation and change as users of language – a claim that simply follows from sociolinguistics’ understanding of language variation as a means of creating social difference, and thus a resource for negotiating social meaning. The human species has developed to what it is not least by developing language to a means of communication which is both sufficiently systematically homogeneous to allow for transfer of linguistic meaning and sufficiently systematically heterogeneous to allow for transfer of social meaning. Fundamental to the human condition, this dual communicative function of language (transfer of both linguistic and social meaning) secures the continuous dynamic evolution of social relationships, in terms of both cooperation and hostility. The conflicting needs for both sameness and difference bring about the ‘dynamism’ of variation and change in language, and there is much folk linguistic evidence for awareness of the problems and possibilities that emerge in this dynamism.
The Bible’s story about the Tower of Babel (in Genesis 11: 1–9) explains the existence of language heterogeneity beyond mutual intelligibility as God’s punishment of the human vanity which resulted when (would result if) all humans spoke the same language and were able to develop effective cooperation. More typically, such narratives focus on the role of linguistic difference in situations of hostility – naturally enough, because awareness of linguistic difference may be a matter of life or death: people who talk differently from ‘us’ may be enemies. The Bible (in Judges 12: 1–15) has a story about that, too: after the inhabitants of Gilead had defeated the invading tribe of Ephraim, the surviving Ephraimites, whose dialect did not contain the sound ‘sh’, were systematically killed after having been identified by their pronunciation of the word shibboleth as sibboleth. Shibboleth has become the international term for reference to words which by virtue of variation in pronunciation have been used to differentiate between ‘us’ and ‘them’, with many examples from contexts of war and persecution. 3
‘Holistic’ awareness of dialect difference may have this function too, as demonstrated in an anecdote about Ludvig Holberg, who in a period of war between Denmark and Sweden at the beginning of the 18th century was arrested in the vicinity of Copenhagen during a walk along the shores of Øresund (the strait between Denmark and Sweden) under suspicion of being a Swedish spy. Since the Viking Age, linguistic changes within the Scandinavian branch of North Germanic have created a dialect continuum stretching from the Danish-German border in the south to the Arctic coast of Norway in the far north. Holberg grew up in Bergen (Norway, which was then part of the Danish kingdom), but the soldiers who arrested him apparently thought he spoke Swedish. They were aware of a linguistic difference between themselves and Holberg, but not of its social meaning: he was not the Swedish enemy they thought he was. They did not kill him, luckily enough, as Holberg later became ‘the creator of the Danish language’ (written Danish, that is) in virtue of his theatre plays and various writings as a professor at Copenhagen University. It is the case still today that most Danes are unable to hear (i.e. they have no awareness of) the difference between Norwegian and Swedish – a difference which is clear enough to any Norwegian or Swede.
The main point so far is that the recognition by people of any instance of linguistic variation involves some kind of social evaluation, which in turn also involves recognition and evaluation of the social meaning of the variation. Thus, the scientific study of language awareness will deal with either the linguistic aspect or the social-meaning aspect (what is recognized? how is it evaluated?), or both. In this sense, lots of studies from a range of disciplines produce data which in various ways shed light on language awareness in ‘real people’: studies of language attitudes in social psychology (Giles and St. Clair, 1979), of dialect borders in ‘perceptual dialectology’ (Preston, 1989), of folk narratives in ‘folk linguistics’ (Niedzielski and Preston, 2003), of recognition and imitation in ‘sociophonetics’ (Hay, Warren and Drager, 2006), of stylization in sociolinguistics (Coupland, 2007). We might add numerous references for each of these disciplines, of course, but few studies would have been carried out with a language-change perspective – even fewer would have been concerned with asking whether folk linguistic awareness is differently affected by different facts of linguistic variation and change. The latter issue is central, however, to William Labov’s lifelong theoretical and empirical efforts to puzzle out ‘the principles of linguistic change’ (summarized in Labov, 1994, 2001, 2010).
Labov’s work belongs to the strong historical tradition which sees sound change as the major mechanism of linguistic change, and effect on awareness is a crucial parameter in Labov’s theorizing of sound change. Different types of sound change are characterized in terms of whether and how they affect awareness – in both its cognitive (‘internal’, linguistic-meaning related) and evaluative (‘external’, social-meaning related) aspect. On the one hand, empirical studies of ‘misunderstandings’ show that the cognitive consequence of sound change is a serious reduction in intelligibility both within and across dialects, and this hold true regardless of whether the change preserves, or subtracts from, the phonemic system’s capacity to make distinctions (2010: Ch.2–4). The consequence for the evaluative aspect of awareness, on the other hand, looks different, as types of sound change are found to differ a lot in their capacity for arousing social evaluation (2001: 25–28; 1994: 343).
Furthermore, sound changes are characterized in terms of whether they come ‘from below’ or ‘from above’ awareness, and linguistic variables (e.g. alternative pronunciations, i.e. variants, of the initial segment in shibboleth as either ‘s’ or ‘sh’) are said to be of three kinds with reference to this distinction: (i) indicators are variables that operate ‘from below’ and provoke no recognition/evaluation; (ii) markers are variables that rarely are talked about but available for recognition/evaluation as evidenced in ‘subjective reaction tests’; 4 stereotypes are variables that are commonly commented on in terms that are readily available in public discourse (1972: 178–179; 1994: 78; 2001: 196; 2010: 307). This availability-to-awareness framework plays a central role in Labov’s distinguishing between changes that originate within the speech community (they come ‘from below’) and changes that are introduced by borrowing from outside the speech community (they come ‘from above’). The framework is furthermore crucial to descriptions of how changes that originate within the speech community (may) go through a series of stages, largely defined by the degree of awareness involved (1972: 178ff.; 1994: 79ff.).
The conceptual work by Labov (outlined above) is by far the most elaborated and influential modelling of impact from language change on language awareness. One might argue that the nature of the impact appears unclear. The ‘from below’ versus ‘from above’ distinction seems to indicate an either/or conception of how various facts of language variation and change affect awareness: the variation is either available to awareness or not, awareness is either turned on or off. At the same time, the explications of how types of sound changes and variables differ (in terms of being variously ‘visible’ to evaluation) seem to indicate a ‘more–or–less’ conception of availability. Labov does not seem to share any feeling of opacity here. Throughout his entire work, the issue of how change affects awareness is treated at times in terms of dichotomy (‘below’ versus ‘above’), at other times in terms of degree.
Preston (1996, and see also his chapter in this volume) presents an elaborated model of folk linguistic awareness in a somewhat different perspective. Following up on an introductory credo – “I believe that a simple on–off characterisation of non-linguists’ awareness of language (or evidence of ‘knowledge’ at any level of awareness) cannot be made” (p. 40) – Preston develops a model that endows linguistic awareness with four dimensions (or ‘modes’): availability, detail, accuracy, and control. These are conceived of as relatively independent continua, which can be used to discuss and establish the relative ease or difficulty with which people (can) comment on linguistic topics, characterize speech or speakers, or perform a variety or an aspect of it. This approach to ‘measuring’ impact from various facets of language on awareness represents a contextualization of the issue, and Preston’s discussions lead him to conclude: “For every act of language production and language perception (including attitudinal as well as ‘processing’ perception), the mode and degree of awareness is an open question” (p. 45).
In Labov’s approach, the issue of how language change affects language awareness is dealt with at the level of variables (i.e. difference in the realization of a particular feature, as when shibboleth was pronounced with ‘s’ by some and ‘sh’ by others). In our studies in Denmark, we have dealt with the issue at the level of varieties (i.e. difference of a more holistic kind, as when Holberg’s speech made the Danish soldiers perceive him as a Swede).
Contemporary language change in Denmark is characterized by radical dedialectalization (Pedersen, 2003) in combination with nation-wide standardization: i.e. spread of Copenhagen speech (Brink and Lund, 1975; Kristensen, 2003; Maegaard et al., 2013). While changes within Copenhagen speech itself can be said to come ‘from below’ in labovian terminology, changes everywhere else will have to be described as wholesale replacement of the local dialects by the ‘standard language’ (Copenhagen speech), i.e. as change ‘from above’. As part of the LANCHART project, 5 we studied how the linguistic variation which results from the change (dedialectalization/Copenhagenization) is recognized and evaluated by young Danes in five communities across Denmark from east to west (Kristiansen, 2009). Our hypothesis was that the variation is available to awareness (recognition/evaluation) as three ‘accents’ of the standard, namely ‘advanced’, ‘less advanced’, and ‘locally coloured’ (the latter being different from ‘advanced’ and ‘less advanced’ Copenhagen speech only in terms of prosodic features). This was found to be the case. We call the three accents modern, conservative, and local. 6 In terms of social evaluation, local is strongly downgraded everywhere in comparison with modern and conservative. Importantly, however, this pattern appears only in a data elicitation context (a carefully designed and administered ‘speaker evaluation experiment’, see endnote 4) where the youngsters are unaware of reacting to (hypothesized) accentual differences. That is, if we apply labovian terminology, the pattern appears only when the evaluation comes ‘from below’. In data elicitation contexts where the local youngsters are aware of giving language attitudes away, and evaluations in that sense come ‘from above’, the local dialect is everywhere preferred to Copenhagen speech.
It seems, then, that the radical Danish language change (dedialectalization/ Copenhagenization) affects recognition/evaluation differently at (what I prefer to see as) two levels of ideological structuring or consciousness, corresponding to a distinction between ‘overt’ and ‘covert’ values in Labov’s framework. At the level of ‘overt’ values, organized and maintained in public discourse (and reproduced by aware youngsters in ‘consciously offered’ evaluations), the dying dialects receive attention and are talked about in terms of ‘respect’ and ‘love’ in ways never heard of when these dialects were strong and vital. At the level of ‘covert’ values (exhibited by unaware youngsters in ‘subconsciously offered’ evaluations), there is an ideological change in the opposite direction – a change of normative target – away from identification with ‘local’ values towards identification with ‘larger-society’. It is not obvious, though, that this change in language awareness (language-ideological structuring) – which took place primarily in the period 1960–1980 (Kristiansen, 1990, 2003) – should be seen as the result of influence from the facts of language change (dedialectalization/Copenhagenization). We are going to return to the question of whether it may have been the other way round: was it the change in language awareness that changed language use?
The question of how the concepts of consciousness/awareness should be understood in terms of degree or dichotomy may appear a rather ‘academic’ issue (of little consequence) as long as we focus on the influence from change on awareness/consciousness (as we have done so far). But the issue becomes otherwise ‘serious’ when the perspective is turned around and we focus on how awareness/consciousness influence change – more serious in the sense that it has substantial consequences for our theorizing of the forces that drive change ‘from below’. It is a matter of course that changes which are unavailable to (i.e. ‘below’) awareness/consciousness cannot be driven by awareness/consciousness, whereas a ‘degree’ conception of availability always leaves room for considering some driving force role for ‘subjectivity’ (awareness/consciousness).
Since in the case of Labov we are dealing with the major sociolinguistic account of linguistic change, it may seem surprising that the dominating tendency in Labov’s theorizing of driving forces is to emphasize the role of ‘mechanic’ forces and downplay the role of ‘subjective’ forces. In fact, Labov’s basic position has always been that “[i]n speaking of the role of social factors influencing linguistic evolution, it is important not to overestimate the amount of contact or overlap between social values and the structure of language” (1972: 251; 2001: 28). In dialogue with theories that emphasize the role of speaker intentions, he terminates the first Internal Factors volume of the Principles of Linguistic Change trilogy by arguing that “[t]here is a part of language behaviour that is subject to conscious control, to deliberate choice, to purposeful and reflective behaviour. But as far as I can see, it is not a major part of the language faculty, and it has relatively little influence on the long-range development of language structure” (1994: 598). About the second Social Factors volume he says that “[t]he main focus of this volume is on changes from below, that is, the primary form of linguistic change that operates within the system, below the level of social awareness. These include the systematic sound changes that make up the major mechanism of linguistic change” (2001: 279). The third and final volume on Cognitive and Cultural Factors states that “[o]n the whole, the most convincing and demonstrable determinants of language change are structural and mechanical” (2010: 244).
I think it can be argued (cf. Kristiansen, 2011) that Labov’s emphasizing of the ‘mechanical’ account is partly a consequence of his failure to establish ‘convincing and demonstrable’ evidence for his own initial belief in “the existence of an opposing set of covert norms, which attribute positive values to the vernacular” (1972: 249): “While such covert attitudes and beliefs may actually be involved in linguistic change, they are not usually supported by material evidence” (2001: 191). 7 This situation led Labov to foreground Bloomfield’s principle of density, which points to frequency of interaction as the explanation for how variation and change pattern in people’s speech. Subjectivities do not affect linguistic change directly, but possibly indirectly through changes in patterns of interaction:
There is no evidence that attitudes, ideologies, and opinions that people express in so many words will bear directly upon linguistic changes from below. These attitudes may influence who a person talks to and how often they talk, and so affect the flow of linguistic influence and the diffusion of sound changes within and across local social networksLabov, 2001: 409
It is not insignificant, of course, whether the derived proposal of dropping the search for covert norms in favour of relying on some frequency-of-interaction calculation is conceived of as ‘just’ a methodological choice, or as a theoretical statement regarding the nature of the forces which drive linguistic change. On the one hand, Labov’s foregrounding of interactional density is explicated in terms of methodological advantage. A social-network based calculation makes it unnecessary to continue the little successful search for covert attitudes: “The account based on covert attitudes is redundant to the extent that the network of daily interaction brings people into contact with the new form in proportion to their distance from the originating group” (2001: 192). In response to this suggested redundancy, it might be mentioned that Lesley and James Milroy, who based their Belfast study on a highly developed network model, have often pointed out that such a model “is not in itself sufficient to provide a full social explanation of linguistic change. What it proposes is a set of conditions that are necessary – but not sufficient – for linguistic change to take place. […] It is not about psycho-social attitudes to language” (J. Milroy, 1992: 204). Thus, the Milroys have suggested, in accordance with the findings of language attitudes research by social psychologists (e.g. Brown and Gilman, 1960; Ryan, Giles and Sebastian, 1982), that an integrated model of sociolinguistic structure must take into account the competing ideologies of solidarity and status (L. Milroy, 1987: 208–209; J. Milroy, 1992: 210, 213), and have furthermore stated that “models of social identity (Le Page and Tabouret-Keller, 1985), accommodation (Giles and Smith, 1979) and politeness (Brown and Levinson, 1987) will not be irrelevant [to the further development of our social model of language change]” (J. Milroy, 1992: 221).
On the other hand, Labov also points to theoretical implications regarding the nature of the driving forces in linguistic change: “The principle of density implicitly asserts that we do not have to search for a motivating force behind the diffusion of linguistic change. The effect is a mechanical and inevitable one; the implicit assumption is that social evaluation and attitudes play a minor role” (2001: 20). I shall not dispute whether it was Bloomfield’s understanding that the outcome of interaction is ‘mechanical and inevitable’, but this does not seem an obvious interpretation of the principle nowadays. What Bloomfield is talking about is accommodation: “Every speaker is constantly adapting his speech-habits to those of his interlocutors” (1933: 476). In accommodation theory, adapting to the interlocutor does not mechanically and inevitably mean convergence; it may also mean divergence – which strongly indicates the social/evaluative nature of the processes at work in interaction. In Labov (2010), the density principle is actually referred to as “Bloomfield’s principle of accommodation” (pp. 5, 155, 166), a change in terminology that might well be seen as a testimony to the foreword statement that “[t]he insights of Penelope Eckert on the social meaning of variation are fundamental to this volume” (2010: xxv).
Insights from both Eckert’s (2000) and his own work have made Labov accord a particular role to gender, as a crucial social factor in linguistic change (2001: Ch.8–9), and therefore also a factor of particular interest to the discussion of ‘mechanical’ effects of communicative patterns. Neither the highly social nature of gender, nor the highly frequent nature of inter-gender communication can be held in doubt, so “[i]f gender as a social factor is intimately involved with linguistic change, it is difficult to limit social factors to the mechanical effects of communicative patterns […], and one is inevitably led to the exploration of other social factors” (2001: 263). This applies also to incipient changes ‘from below’: “In the early stages of change – at the lowest levels of social awareness – gender operates as an independent and powerful factor. It follows that the forces active in qualitatively new changes include social factors, and that any effort to account for the initiation of change by purely internal arguments will fail to a significant degree” (2001: 322).
One might ask if what we see here is a theoretical recognition of some effect of ‘social awareness’ even in the early stages of change? Probably not, as Labov in another place addresses the common explanation of gender difference which refers to “women’s superior sensitivity to the social evaluations of language” – and explicitly rejects it on the grounds that “it assigns social sensitivity to early stages of change that are remote from levels of social awareness” (2001: 291). Labov’s theoretical claim is rather that effect from ‘social factors’ and from ‘social awareness’ are two different things. Incipient changes can therefore be affected by social factors, gender being a particularly powerful one, without any involvement of social awareness.
However, since the effect is not ‘mechanical’, some psychological process is needed to explain it. Attention is brought into play. Labov suggests that small children to start with are attentive to formality differences, not to social categories (such as gender). Rejecting the view that stylistic stratification is derived from social stratification (as proposed by Bell, 1984; Preston, 1989), Labov finds it “probable that [children] abstract a continuous stylistic dimension from a variety of speech contexts”:
Children 3 to 5 years old pay close attention to this dimension, since it indicates to them if they are being placed in the category of “good” or “bad” children and will be rewarded or punished for what they have done. I would therefore suggest that the formal/informal dimension is not a vague abstraction for children, but a useful scale of reference that is called upon many times during the day as the child responds to adults, deals with older kids, and tries to keep out of trouble
Children 3 to 5 years old pay close attention to this dimension, since it indicates to them if they are being placed in the category of “good” or “bad” children and will be rewarded or punished for what they have done. I would therefore suggest that the formal/informal dimension is not a vague abstraction for children, but a useful scale of reference that is called upon many times during the day as the child responds to adults, deals with older kids, and tries to keep out of troubleLabov, 2001: 420
The complex picture of how (degrees of) attention/awareness/consciousness relates to stylistic and social dimensions and affect linguistic change at successive stages is presented throughout Labov’s work in models of step-wise progression, with varying focus on the various parts of the complexity (e.g. 2001: 307ff., 437, 517–518). “The general proposal”, Labov says, “is that the use of a sociolinguistic variable is learned by association of variants with one or the other pole of an organizing principle of social life” (2001: 421).
Regardless of what kind and degree of ‘subjective’ processes (attention/awareness/consciousness) are involved in that ‘learning by association’, we are dealing with ‘perceptions’ – which, most importantly, are likely to be ‘pole-directed’. While a change from below develops as a gradual movement in continuous phonetic space, “the social perception of this process is more likely to be a polar opposition between ‘advanced’ forms and normal forms” (2001: 454). “It seems likely that what is perceived is not a specific target in terms of frequency or formant level, but a dimension or direction of shifting” (2001: 463–464).
The first and basic social polarity emerges as the formal/informal polarity is transformed into a conformity/nonconformity polarity (2001: 513–514). Subsequently, associations involving all kinds of more concrete social polarities are involved in aligning this first and basic social dimension with linguistic innovation (older/younger, higher/lower, local/outsider, jock/burnout, local/nonlocal, female/male, urban/rural, modern/old-fashioned (2001: 463, 513). “But the nonconformist/conformist polarity has a privileged position for congruence with the process of language change. It helps to explain its most general social characteristic” (2001: 513). The Nonconformity Principle states that “Ongoing linguistic changes are emblematic of nonconformity to established social norms of appropriate behavior […]” (2001: 516). The women who are identified as prototypical leaders of change from below are so in virtue of their nonconformity, not in virtue of their gender (2001: 516). In the case of most polarities, “the majority of speakers will be identified as in-betweens for any given polarity, and shift their behavior as their orientation toward the polar target changes” (2001: 463–464).
While the conception of ‘perceived polarities’ may seem easily reconcilable with any conception of subjectivities involved in language change, the insistence that the many concrete polarities share an abstract conformity/nonconformity polarity forms the basis for Labov’s special efforts to elucidate how ‘socially motivated projections’ can drive language change in the same direction across large territories and millions of people who have no connection with each other (2001: 511). Thus, the ‘driving forces’ discussed in Labov (2010: Ch.9) are of a subjective kind and include negotiations of local and social identity (pp. 185–186), efforts to “maximize […] status in a timely way” (p. 189), influence on peers from “opinion leaders” (p. 190), negative and positive evaluations on different levels of social awareness: overt stigmatization and rejection vs. covert positive forms of social motivation (pp. 191–192), unconscious recognition and preference/dispreference of things being ‘in/out of fashion’ (p. 195). Also, the role of individual ‘acts of identity’ as a mechanism that reifies social meaning and constructs group identity (pp. 193–194) is considered, but Labov mainly presents results that “raise serious obstacles to any proposal to explain sound change as a series of individual acts of identification with neighboring social groups” (p. 196), and to the view that “the ‘social meaning’ that drives sound change [is] transmitted by intimate face-to-face interaction” (p. 202).
It was the results of the Telsur project (the survey of language changes in progress in North America) that raised serious questions about the role of face-to-face interaction in change processes (whether ‘mechanic’ or ‘subjective’). Based on survey data from telephone interviewing in the early 1990s, the project for the first time provided a continental-wide Atlas of North American English (Labov, Ash and Boberg, 2006) and set the diffusion issue in a new ‘subjectivity’ perspective – that of ‘cultural factors’ – as it became clear that large regions “display an extraordinary homogeneity across great distances and across large populations” (2010: xxiv). In the Inland North, the changes known as the Northern Cities Shift (NCS) were found to be omnipresent, and spreading without traversing the old North/Midland boundary to the south (2010: Figure 8.3). It would still be possible to argue for a role for face-to-face interaction if the communication lines in the area were found to follow the east–west direction of the North/Midland boundary, and the spread was found to vary across city size in accordance with the ‘cascade’ model of spread (i.e. innovations are hypothesized to ‘cascade’ from bigger cities to smaller). However, relevant analyses showed communication lines in the area to go north–south (pp. 170–171), and city size to be of no significance (pp. 205–206). Thus, the general conclusion goes: “There is no doubt that language change may be local and reflect an immediate social motivation to reinforce local identity. But we have seen that language change in North America occurs on a much larger scale, where individual acts and motivations are irrelevant” (p. 244).
By contrast, facts about differences in present-day cultural patterns and political ideologies, with deep roots in different settlement histories, were found to show fairly much the same geographical distribution as the objective facts of the NCS. Thus, even though it is restated that “[o]n the whole, the most convincing and demonstrable determinants of language change are structural and mechanical”, it is at the same time added that “we must be alert to the possibility that ideology is a driving force behind change, as well as a barrier to its further expansion” (2010: 244).
After the municipal elections in Denmark in 2009, a map representation of the results showed that Jutland was politically divided along a north–south line, which coincided with a centuries-old dialect boundary. With only a couple of exceptions, all western mayors were (in Danish terminology) ‘bourgeois’ and all eastern mayors were ‘social-democrats’. As in the case of the US North/Midland boundary, the Jutland case is strongly indicative of how old differences to do with culture and local identity (can) significantly influence contemporary subjectivities. Unlike the US case, however, the Jutland cultural and ideological divide – which centuries ago created ‘Denmark’s most famous dialect boundary’ 8 and shows up in different electoral behaviour today – has no influence on linguistic behavior in present-day Jutland. All of Jutland, like all of Denmark, follows the lead of Copenhagen (Maegaard et al., 2013).
Yet, Labov’s point about ideology as a possible driving force is supported by the Danish case, as the LANCHART studies (see endnote 5) have documented that Denmark’s linguistic Copenhagenization has its correlate in a very strong subjective Copenhagenization. Everywhere outside Copenhagen, young people downgrade their own local accent (of Copenhagen/standard speech) in comparison with young ‘genuine’ Copenhagen speech, regardless of whether the latter is spoken in a conservative or a modern accent (see endnote 6), and regardless of whether the evaluative dimension is superiority or dynamism – a distinction which has great significance for the evaluation of the Copenhagen variation: modern (with traditional working-class features and a strong media position) is strongly upgraded on dynamism values, whereas conservative (reflecting the more traditional ‘standard language/public sector’ association) does as well or better on superiority values (see endnote 4). In our speaker evaluation experiments, we included eight ‘personality traits’ (seven-point scales) for each of the two dimensions (dynamism: self-assured, fascinating, cool, nice; superiority: intelligent, conscientious, goal-directed, trustworthy), and it was a most remarkable finding that the non-Copenhagen youngsters (from places across the whole country) evaluated conservative and modern in exactly the same way as Copenhagen youngsters, consistently on all eight ‘personality traits’. This can hardly be seen as a plausible result if the valorization of the involved variation was reconstructed locally as a concomitant to the linguistic Copenhagenization. We therefore argue that the subjective Copenhagenization precedes the linguistic one, and that ideology is a driving force of the process rather than a concomitant (Maegaard et al., 2013).
And we suggest that the nation-wide and copy-like acquisition of the Copenhagen-based language-related values can only be understood as a product of shared experience with how the language-norm-and-variation complex is treated in the modern media universe. We see TV and its role in developing a new national public sector from the 1960s on as particularly relevant to the discussion. The argument is not that the media influence speech directly – a view that is strongly rejected by most sociolinguists, Labov included (2001: 228) – but indirectly, by reshaping the language-ideological embedding of speakers and their speech, with consequences for language use (Kristiansen, 2014a, 2014b).
Neither in terms of how linguistic changes relate to awareness (coming ‘from below’ or ‘from above’ awareness) nor in terms of how they relate to values (driven by ‘covert’ or ‘overt’ values) is the Danish picture easily reconcilable with the Labovian framework, in which any import from outside the speech community (from other systems) is a change ‘from above’ – including imported features from the capital city’s traditional working-class variety, as in the case of modern Copenhagen. Labov explicitly mentions the comparable example of recent spread of traditional low-status features from London to other British cities as an example of ‘change from above’ (2010: 389). Thus, since change ‘from above’ must reflect ‘overt’ values, the linguistic Copenhagenization of Denmark appears an effect of ‘overt’ values within the Labovian framework. However, the evaluative pattern described above for the Danish accents emerges only when the evaluations are ‘from below’ (they are ‘subconsciously offered’). If such evaluations reflect ‘covert’ values, the spread of Copenhagen speech, spearheaded by modern features among young Danes, is driven by ‘covert’ values. It is a commonly held view that ideological favoritism strengthens the societal position of the standard language while ‘overtly’ low-prestige language (including dialects) may persist in virtue of ‘covert’ positive appraisal. 9 In contrast, our Danish picture shows the dialects to be ‘overtly’ protected in ideology – with no protective effect in production, the result being radical dedialectalization – whereas the corresponding radical Copenhagenization/standardization is driven by ‘covert’ values, favoring modern in particular as a dynamic way with language.
On the other hand, it makes good sense to talk about the Danish picture in terms of (assumed) ‘perceived polarities’. It is a reasonable assumption that the language-related local polarities of long ago (e.g. ‘western Jutland/eastern Jutland’) were gradually replaced by a global ‘Copenhagen/non-Copenhagen’ polarity from the time in the 17th and 18th centuries when Copenhagen developed to become Denmark’s indisputable centre of power and all kinds of standardizations, including language standardization. It is furthermore reasonable to assume that ‘Copenhagen/non-Copenhagen’ in subsequent centuries was constructed as the ‘status/solidarity’ polarity which contemporary social psychological research has found to be the common ideological counterpart of the linguistic ‘standard/non-standard’ polarity (Ryan, Giles and Sebastian, 1982). Although the evidence is sparse, we may with reasonable confidence continue to assume that the folk in their ‘covert’ attitudes favored their own dialect (or else it would not have persisted), even though they in ‘overt’ attitudes may have praised the language of the authorities (‘the best language’) and demeaned their own (‘our dialect is awful!’). What happened in the 1960s–1970s was a change in this language-ideological structuring.
Driven by general ‘covert’ ideological upgrading, the standard language (Copenhagen speech) got the better of the dialects, which in turn, as they died away, became an object of love and respect in ‘overt’ ideology, more highly appreciated than the standard. In ‘covert’ ideology, the ‘superiority/dynamism’ polarity developed in the construction of a new double-standard situation where the linguistic polarity conservative/ modern is associated with the public-sector polarity which results from the addition of the modern media institutions to the traditional institutions of education and business. The three polarities – superiority/dynamism, conservative/modern, education/media – may be seen as the ideological, linguistic, and societal aspects of a more abstract double-standard polarity. Only research in the future can tell whether this double-standard polarity is of a temporary or more permanent character in Denmark. Perhaps more interestingly, it is an open question for contemporary research whether the double-standard situation is a Danish specialty or an emergent polarity more generally in late-modern Europe. 10
Language attitudes; language ideology; causes of language change; language variation
The quotes in the paragraph are from Plato. Plato in Twelve Volumes (1921). Vol. 12, trans. Harold N. Fowler. Cambridge, MA: Harvard University Press; London: William Heinemann Ltd. The figures in sharp parentheses refer to the location of the quoted text in the Perseus Digital Library, www.perseus.tufts.edu/hopper.
‘Subjective reaction tests’ are experiments in which subjects listen to audio-recorded speakers and assess them on a number of scales which typically are supposed, or found, to represent underlying social-value distinctions, evaluative dimensions, such as social competence (job suitability, status) versus social likeability (friendship, solidarity), or social superiority versus social dynamism. Such tests are also known as ‘speaker evaluation experiments’, and the testing of evaluative reactions may be aimed at the use of the variants of a particular variable (as in Labov’s work), or at the use of different varieties, i.e. dialects, accents or languages (which is a more typical approach in social psychological research on language attitudes). We will return to the concepts of evaluative ‘dimensions’ or ‘polarities’ in later sections of the chapter.
The LANCHART project (LANguage CHAnge in Real Time; http://lanchart.hum.ku.dk) was funded for the ten-year period 2005–2015 by the Danish National Research Foundation, grant DNRF63 to Frans Gregersen. For collective presentations of LANCHART work, see Gregersen, 2009; Gregersen and Kristiansen, 2015.The attitudinal data referred to in this chapter was collected in 2005–2006. Detailed presentations of the LANCHART attitudes studies are found in Kristiansen, 2009.
The recognition/evaluation of modern and conservative is based on segmental variation, but our ‘holistic’ approach does not allow for any specification of possible differential contribution from segmental variants. Subsequent reaction tests based on technical manipulation of intonation contours do however allow us, in the case of local as spoken in Århus (Denmark’s second-largest city after Copenhagen), to specify intonation as the linguistic feature that triggers awareness of local as different from modern and conservative (Kristiansen, Pharao and Maegaard, 2013).
Further comments on the issue in 2001: 23–24, 105–106, 195–196, 217, 222, 329, 361, 512; and 2010:192.
Thorsen (1912) called it ‘Denmark’s most famous dialect boundary’ with reference to a most salient difference: post-positioned determiner on nouns in the east (as in the rest of Scandinavia, hus-et ‘the house’) versus pre-positioned in the west (as in German and English, a hus ‘the house’). Other salient differences (concerning the grammatical gender system and the prosodic stød-system) follow by and large the same dividing line.
Referring to the ‘or-else-they-would-not-have-persisted’ assumption (2001: 329, 223, 512) – as stated by other scholars (Ryan, 1979; L. Milroy, 1980) – Labov emphasizes the lack of supportive evidence for ‘covert’ values (see endnote 7), but originally shared the assumption (1972: 249).
The issue is being addressed by researchers in the pan-European SLICE network (see http://lanchart.hum.ku.dk and the network’s book series: Kristiansen and Coupland, 2011; Kristiansen and Grondelaers, 2013; Thøgersen, Coupland and Mortensen, 2016).