Investigating the Effect of Socio-Cognitive Salience and Speaker-Based Factors in Morpho-Syntactic Life-Span Change

Abstract

This article explores changes across the life-span of the individual speaker. Three morpho-syntactic variables that operate at different levels of socio-cognitive salience—quotation, stative possession, and future temporal reference—are traced across longitudinal trend and panel data. The analysis of a panel corpus that spans forty-two years reveals that linguistic malleability is contingent on the variables’ embedding in speakers’ cognitive-evaluative structures but also on a range of speaker-based factors, such as the speakers’ socio-economic trajectory, personality profile, as well as their type and intensity of contact with children and younger speakers. The study thus supports research that has shown that the adaptive behavior across the life-span results in complex outcomes that cannot be characterized by wholesale convergence or divergence.

Keywords

panel research life-span change morpho-syntax socio-cognitive salience quotation stative possessive future temporal reference

1. Introduction

Many of our hypotheses regarding the rate and trajectory of language change rely on the assumption of a critical period for language acquisition followed by relative stability (Lenneberg 1967). However, research across the life-span of the individual has accrued an increasing body of evidence that adults’ linguistic tendencies are less fixed than previously assumed, even past critical age (e.g., Arnaud 1998; Sankoff 2006; Sankoff & Blondeau 2007; Harrington, Palethrope & Watson 2000; Sankoff & Wagner 2011). Surveying the evidence, Labov (2006:239) concludes that “in every [follow-up study of changes in progress] [. . .] there was some advance in the adult population.” Modifications in individual grammars can occur in the direction of ongoing community-wide change (“life-span change”; Sankoff 2005) as well as against the direction of ongoing change (“retrograde change”; Sankoff & Wagner 2006, 2011).¹

To date, the majority of research on linguistic lability across the life-span focuses on speakers’ adjustment to ongoing phonological changes (e.g., Yaeger-Dror 1993; Sankoff 2004; De Decker 2006; Sankoff & Blondeau 2007; Wagner 2008; Bigham 2010; Gregersen & Barner-Rasmussen 2011; Prichard & Tamminga 2012). We know much less about the range and limits of intraspeaker lability in variables above the level of phonology (see, however, Yaeger-Dror 1989, 1993; Sankoff & Wagner 2006, 2011; Rickford & Price 2013). The level of linguistic structure at which a changing feature is situated is important since some panel research suggests that speakers adapt to ongoing morpho-syntactic changes differently than to changes in phonology. Blondeau, Sankoff, and Charity (2002:16) hypothesize that “the adoption of phonetic innovations might be more difficult than the adoption of innovations at other levels.”² Comparing results gleaned from investigations of variables at different levels of modularity has thus the potential to lead to a more encompassing picture of linguistic malleability across the life-span of the individual (Sankoff 2005; Bowie & Yaeger-Dror 2015). This is especially pertinent since research on the socio-psychological determinants of online processing has suggested that a variable’s situatedness in linguistic structure is related to its salience, a socio-cognitive mechanism that governs perceptual awareness (Labov 1993; Levon & Buchstaller 2015).

Whereas variability in phonetic structure is assumed to be available for acquiring social associations, research on the socio-cognitive salience of features at the level of morphology or (morpho-)syntax is relatively sparse (Chapman 1995; Hollmann & Siewierska 2006; Jensen 2014). Moreover, while research on phonological structure has revealed important social and stylistic correlates between the level of awareness of individual phonetic variables and speakers’ linguistic choices across their life-span (De Decker 2006; Bigham 2010; Prichard & Tamminga 2012), we currently know very little about the importance of socio-cognitive salience on linguistic malleability in the area of morpho-syntax and syntax.

The present paper seeks to add to the investigation of changes across the life-span of the individual by examining six panel speakers’ behavior against the backdrop of a trend data-set capturing the population benchmark in the North East of England. The research reported here examines speakers’ (non)adjustment to three ongoing morpho-syntactic changes which are situated at different levels of socio-cognitive salience. One change involves quotative be like, which has acquired considerable notoriety in the past two decades (D’Arcy 2007). In contrast, the other changes, which occur in the systems of stative possession (e.g., have vs. got) and future temporal reference (e.g., will vs. going to), have—to my knowledge—not attracted prescriptive commentary in the popular press or in other public fora. But whereas the forms involved in the variable expression of future time reference seem to be entirely devoid of ideological and social associations, the high salience of stative possessive got, which is widely perceived as an American feature, seems to have affected speakers’ socio-cognitive awareness of have got, the other incoming form in the system. Capturing the effect of salience in panel research thus requires a model that considers both the indexical load of the incoming form as well as of its competitor forms.

The present article explores the impact of socio-cognitive salience on the usage of variable forms across a time-span of forty-two years. The relatively small, ethnographically informed nature of the panel sample gives us the opportunity to explore the role of speaker-based factors, such the individual speakers’ personalities, their individual social trajectories as well as their type and amount of contact with children (Rogers 2003; Rickford & Price 2013). The analysis is guided by the following questions: do post-adolescent speakers adopt morpho-syntactic changes that are currently making headways in the community in which they are embedded? If yes, which speakers and which variables exhibit post-adolescent linguistic change? How does salience interact with life-span change? To what extent can a speaker’s personality, their social trajectory, and contact with children act as explanatory parameters for ongoing life-span change?

2. Salience

Salience is a notoriously under-defined concept. While most models conceptualize salience as the result of multiple factors (including cognitive, language-internal and language-external determinants; Kerswill & Williams 2002), recent approaches have taken a processual approach, focusing on the cognitive implications of social perception (Jensen 2014:185; Levon & Fox 2014). Such models consider both the type and token frequency of the variable and its variants as well as their ideological weightiness, including the amount of “socially marked information” associated with a particular linguistic feature (Labov et al. 2006:107).

The field of sociolinguistics has long recognized socio-cognitive salience as an important factor in language variation and change. Salient features have been argued to be prone to language change (Labov 1972:320, 1994:300). By the same token, research on linguistic diffusion tends to consider social attention and evaluation as the motor for change in progress, with increasing public awareness sparking the acceleration of the S-curve (see Nevalainen & Raumolin-Brunberg 2003:53-54). Also the concept of age grading, where speakers modulate away from linguistic features that are considered inappropriate at certain life stages, fundamentally relies on the notion of socio-cognitive salience: the avoidance of negatively evaluated variants is contingent on (i) the indexical meaning of the linguistic feature in question as well as (ii) speakers’ awareness of these social associations (Campbell-Kibler 2010, 2011).

Regarding speakers’ malleability across the life-span, Prichard and Tamminga (2012:94) have found that social awareness plays an important role in “determining which [phonetic] variables are subject to correction”: individuals who are bound for nationally prestigious universities modulate away from incoming sound changes that are above the level of social awareness and thus available for moderation (see Labov 2001:108, 272, 438, for stylistically induced correction). Sound changes below the level of awareness, on the other hand, proceed regardless of the speakers’ educational orientation (see also Bigham 2010). But whereas sociophonetic research is starting to establish a causal link between speakers’ level of awareness of certain articulatory contrasts and their phonetic choices across the life-span (De Decker 2006; Bigham 2010; Wagner 2012b), we do not yet have enough evidence from panel studies on morpho-syntactic changes that might allow such generalizations.

Sankoff and Wagner (2006, 2011) suggest that speakers modify their speech toward certain morpho-syntactic variants provided those variants carry the social meanings they want to associate themselves with. Their research on the Montreal French future reveals that speakers in the higher socio-economic bands move toward forms that have taken on formal and archaic associations. Thus, similarly to phonetic heterogeneity, the adoption of and/or modulation away from incoming morpho-syntactic variants during the life-span of the individual seems to be shaped by the speaker’s awareness of the variability within the system as well the social indexicalities of the available linguistic forms. But we do not know to what extent Sankoff and Wagner’s (2006, 2011) findings can be generalized across other variables in the area of morpho-syntax. To this aim, this study examines the use of three ongoing morpho-syntactic changes over the life-course of six speakers in the North East of England.

Obviously, the issue of salience forces us to consider the role of speaker agency. While the creative construction and negotiation of speaker stances, styles, and personae has only been recently incorporated into panel research, Rickford and Price (2013) draw attention to the shifts in positionality and identity management individuals undergo during their life-times (see also Rickford & McNair-Knox 1994). They argue that panel research needs to pay more attention to stylistic variation, weighing the contributions of circumstances that are typical for certain life stages versus individual stance-takings (see also Wagner & Tagliamonte, forthcoming). Thus, rather than trying to disentangle life-span change and stylistic variation, panel research forces us to acknowledge the fact that language variation across the life-span is embedded in “social [and age-appropriate] meaning-making” (Rickford & Price 2013:172). The present article adds to this line of research by drawing on the role of personal factors to explain fluctuating usage of variables that are situated at different levels of perceptual awareness. By doing so, this paper aims to contribute to the recent turn toward ethnographic research and the increasing interest in socio-cognitive salience in sociolinguistic panel studies.

3. Data and Sampling

As Wagner (2012a) points out, the literature on life-span change can be broadly characterized as either large samples structured by broad socio-demographic categories (e.g., Sankoff & Blondeau 2007; van Hofwegen & Wolfram 2010) or small samples with detailed interpretation of local social meaning (e.g., Sankoff 2004; Rickford & Price 2013). The present article contributes to ongoing research on changes across the speakers’ life-course by presenting a study which relies on a panel sample of six speakers in combination with a longitudinal trend sample. Similarly to De Decker (2006), these two data-sets will be enriched by questionnaire data and ethnographic observations from within the community.

The trend study draws on two pre-existing real-time corpora. The older dataset relies on the Tyneside Linguistic Survey (TLS) recordings conducted in the late 1960s and early 1970s. The newer data come from the ongoing Diachronic Electronic Corpus of Tyneside English project (DECTE; see Corrigan et al., forthcoming), which has yielded an average of sixty-five recordings per year since the mid-2000s. For the present analysis, I chose a socio-demographically balanced sample (by age, gender, and socio-economic class) of sixteen speakers from the TLS and the DECTE corpora. In Buchstaller (2014), I have argued that comparative analysis of these recordings is viable insofar as both corpora consist of spontaneous talk-in-interaction collected under the auspices of sociolinguistic projects and following the format of unstructured or semi-structured sociolinguistic interviews. With very few exceptions, the interviews were conducted at the informants’ homes, and both sets of recordings contain a range of genres, including narratives, argumentation, projections on the future, and grammaticality judgments. The main factor that distinguishes the TLS data from the later interviews is that the earlier recordings were conducted as one-on-one interviews with a local fieldworker whereas the DECTE recordings rely on dyads of speakers with a more or less involved interviewer. This tended to result in a slightly less spontaneous interaction overall in the TLS data.

The panel sample relies on six different speakers from the TLS corpus, all of whom were originally interviewed in 1971. As part of an ongoing project on life-span changes, we created a second time slice by re-recording them in 2013. The re-interviews carefully replicated the original interview situation of the TLS, including the setting, location, interviewer, and the types of questions asked (see Gregersen & Barner-Rasmussen 2011). Also, the fieldworker of the 2013 panel study matched the social characteristics of the initial interviewer. Both were highly educated males from the North East, whose accents many informants found difficult to localize.³ The second set of interviews aimed to recreate the structure of the original recordings as closely as possible, asking informants whether they remember and indeed agree with their answers to the questions initially asked in 1971. Additional questions added in 2013 enquired about the informants’ lives since the original interview and their opinions about changes in the Tyneside area. The combination of trend and panel data thus gives us the opportunity to investigate longitudinal language change in the individual and compare it with ongoing changes in the community in which they are embedded. The sampling frame for the longitudinal trend and panel data is given in Table 1.

Table 1.

Sampling Frame for the Diachronic Investigation of Language Change Longitudinally as well as Across the Lifespan of the Individual Speaker

The North East has experienced far-reaching socio-demographic changes since the 1960s. While the Tyneside community was relying on the heavy industry characteristic of many northern cities when the TLS project collected the first data-set, the conurbation has since been transformed to an economy based on retail, education, and customer services (Beal, Burbano-Elizondo & Llamas 2012). As we will see, the trajectory of some of the panel speakers mirrors these changes from industrial to postindustrial economy.

4. The Variables under Investigation

The three morpho-syntactic variables explored in this article are currently in the process of changing in the North East of England and indeed across the English-speaking world. All three can be conceptualized as sets of paradigmatic alternates within one syntactically defined slot, a criterion that is often adduced for the definition for a morpho-syntactic variable (Rickford et al. 2007).

Quotation has been defined as “all strategies used to introduce reported speech, sounds, gesture and thought by self or other” (Buchstaller 2006:5). Whereas older speakers tend to rely on say and unframed quotes (1a) and (1b) for reported speech and think for reported thought (1c), many younger (and increasingly also middle-aged) speakers favor innovative forms. This article focuses on the most notorious of these incoming variants, be like (1d), a vigorous, global newcomer which has altered the quotative landscape in the past twenty years (e.g., Rickford et al. 2007; Tagliamonte & D’Arcy 2007).

(1) a. He says “if there was an explosion in Low Fell we wouldn’t hear it”

(DECTE_GM)

b. [. . .] and then listen to the Queen’s speech [informant puts on a high

pitched voice] “this is The Queen” (DECTE_7_025)

c. I used to think “Oh, it must be marvellous” (DECTE_GM)

d. I asked to have a look and they were like “no you’ll just have to take it”

(DECTE_005)

In contrast to the rapid influx in the system of speech and thought reporting, the competition amongst the forms expressing stative possession is the result of a gradual layering process (Tagliamonte, D’Arcy & Jankowski 2010). Whereas the oldest variant have has expressed static ownership since Old English (2a), “the idiom have got” (Huddleston & Pullum 2002:112) is attested since the Middle English period (2b). Over time, its phonetic substance eroded, resulting in the contracted forms ’ve/’s got typical of spoken language and/or spontaneous registers (2c). This erosion process has come to completion in some varieties, leading to the complete elision of have and the concomitant expression of stative ownership via got (2d).

(2) a. Well I have friends at college (TLSG_20)

b. Bede College have got a very good rugby Team (TLSG_20)

c. I’ve still got the old LPs you know (DECTE_A1)

d. I think they got a lot more opportunities (DECTE_A2)

Have got has increased throughout the last century (Kroch 1989), especially in the British Isles, where the competition in the system of stative possession plays out principally between have and have got (including its reduced forms).⁴ The newest variant, got, is generally considered US English and, as of yet, rare in British English and its derivatives. There is evidence, however, that contemporary British speakers have recently started to use got as a minority form (Tagliamonte, D’Arcy & Jankowski 2010; Tagliamonte 2013).⁵

Two variants in the system of future temporal reference, will (3a) and shall (3b) have been part of the variable since late Old English (Visser 1963-1978:1582, 1698), expressing modal meanings of obligation and volition from which their future meanings have developed. In contemporary British Englishes, the use of shall with futurate meaning, which prescriptive rules limit to first person subjects, is extremely rare, stylistically restricted to formal registers, almost entirely limited to the South (Tagliamonte 2013:125), and practically non-existent in the data.

The development of future time reference out of go-lexemes (3c) is one of the paradigm cases of grammaticalization and it has been reported for many related and unrelated language families (Heine & Kuteva 2002). In the English language, the form was not used “with any real frequency as a [clear] future marker” until the seventeenth century, however, and it is thought to be increasing in frequency ever since (Poplack & Tagliamonte 1999:318). Tagliamonte (2013:124) suggests that the “acceleration of going to is relatively recent.”⁶

(3) a. They’ll put you on a section (DECTE_A1)

b. I shall give him all the help he needs (Visser 1978:1603)

c. What sort of a life is that boy going to have (DECTE_04A)

Descriptive and prescriptive grammars tend to differentiate will and going to in terms of epistemic stance, modal meaning, or proximity. However, Quirk et al. (1985:213-218) concede that while will is “the closest approximation to a colourless, neutral future,” the differences between the two “should not be exaggerated. The choice of one construction [. . .] rather than another [. . .] [tends to have] a scarcely perceptible effect on meaning”. In fact, empirically orientated research has shown that it is often difficult, if not impossible, to establish objective means to identify or even measure such subjective and often highly contextualized notions as speaker intention and listener inference (Fleischman 1982; Poplack & Tagliamonte 1999). This supports Heine’s (1993:92) contention that “in spite of their different verbal sources [. . .] ‘there is no demonstrable difference’ (Palmer 1974:163)” between these two forms in contemporary English (see also Danchev & Kytö 1994).⁷

5. Testing for Salience

An analysis that considers language use across the life-span needs to establish the level of awareness at which a variable and its variants operate as well as the socio-indexical associations they carry. In this project, socio-cognitive salience was defined, following Levon and Fox (2014) and Preston (2011), as speakers’ awareness of the social indexicalities of a linguistic form. It was established via three separate measures: firstly, I investigated the variants’ “social embedding in the community” (Jensen 2014:30), which was operationalized as the existence of prescriptive commentary or overt correction in the public sphere. Forms that have achieved a high level of social awareness tend to be remarked upon (see Milroy & Milroy’s 2003:26 “complaint tradition”). To this aim, blogs, opinion pieces in the press, or indeed any published material decrying the slipping standards of the English language were scouted for “public recognition in the form of stigma or prestige” (Labov 1993). I also conducted an attitudinal questionnaire with seventy-six British native speaker informants (about 70 percent live in the North of England and Scotland) ranging in age from the late teens until the mid-70s with a roughly equal gender split. The questionnaire contained two parts (see the appendix). A recognition task tested whether informants can identify certain linguistic forms as non-standard when they occur in a text of naturally occurring speech. Informants read a short text which contained the morpho-syntactic features investigated in this project. They were then asked to circle every word they “think is not standard (i.e., not used by national newsreaders), colloquial, or typical of the language used by young people.” The second part of the survey consisted of a social attitudes questionnaire, which tested for the socio-evaluative load of individual variants. Informants were asked to indicate whether they thought that fourteen sentences (such as I got two younger brothers) would be typically “used by national newsreaders,”⁸ by younger or older, more or less educated, and by British speakers. The following paragraphs consider public commentary on the use of these features, the results of the recognition task and the social attitudes questionnaire.

The innovative quotative be like has reached the level of a stereotype, a socially marked form that is above the level of public consciousness, commented on, prominently labeled and “subject [to] public discussion” (Labov 2001:272). Indeed, almost “any tirade about the state of the language is sure to say something about like as a plague on the language” (Bierma 2005; see also D’Arcy 2007). My informants’ reactions support this contention: in the recognition task, quotative be like was the third most circled feature, with 74 percent of all respondents marking the variant as non-standard. The social attitudes questionnaire further corroborates the results: as Table 2 demonstrates, the variant was overwhelmingly associated with non-standard (99 percent), younger (92 percent), and less educated speakers (86 percent), the characteristic profile of a stigmatized feature that has reached the level of a stereotype (Buchstaller 2014: ch. 5). Hence, if we assume that “salience is linked with [social] indexicality [. . .], [i.e., the] social values [that] emerge in a community” (Jensen 2014:96), these findings suggest that be like has achieved a very high level of socio-cognitive salience.⁹ The variant is also associated with non-British speakers (see also below). Note in this respect that the “conventional” quotative say (Buchstaller 2014:119) seems to be entirely inconspicuous, receiving not a single marking in the awareness test and, as Table 2 reveals, being overwhelmingly considered standard and typical of older, educated, and British speakers.

Table 2.

Responses to the Social Attitudes Questionnaire (Percentages Calculated out of 72-76 Ratings)

Variant	Non-Standard (%)	Younger speakers (%)	Less educated (%)	British English (%)
Say	12	13	6	85
Be like	99	92	86	45
Got	99	90	86	50
Have got	30	26	28	11
Have	17	11	8	1
Going to	11	3	6	98
Will	17	12	9	93

The socio-cognitive awareness of the ongoing changes in the system of stative possession is fundamentally different in nature. Tagliamonte, D’Arcy, and Jankowski (2010:150) report that the variable is used “without conscious awareness,” and that “informants do not know whether they use have or have got a cat.” This claim is supported by reports in the literature that gradual community-wide changes tend to be imperceptible to the individual speaker (Sankoff, Thibault & Wagner 2004). As far as I can tell, the popular press does not contain any prescriptivist comments on the system of stative possessives, nor indeed any commentary on the aesthetic and pragmatic differences between the two variants (contrary to the system of future time reference; see below). However, Huddleston and Pullum (2002:112) claim that the stative possessive variants pattern according to stylistic parameters, pointing out that “have got is restricted to informal style, but is otherwise very common, especially in Br.E.” The results of my attitudes questionnaire suggest that respondents from northern England and Scotland are not entirely unaware of the variability in the system of stative possessives. In the recognition task, where informants were asked to circle the words they considered “not standard [. . .] colloquial, or typical of the language used by young people,” have was completely inconspicuous, receiving some of the lowest scores in terms of social recognition (13 percent).¹⁰ However, 22 percent of all informants circled I have got problems with nuts too and a full 58 percent circled got in the construction My family got a walnut tree.

The social attributes test supports this finding: as illustrated in Table 2, the majority of respondents marked have as standard (only 17 percent non-standard ratings) and the variant was not particularly associated with younger (11 percent) or non-educated (8 percent) populations. In contrast, the vast majority of my respondents indicated that got was non-standard English (99 percent), and it was generally connected with younger (90 percent) and less educated speakers (86 percent). Have got patterns in-between the highly salient got and the non-salient have with ratings of 30 percent non-standard, 26 percent younger speakers, and 28 percent uneducated.

Stative possessive got is not only clearly above the level of consciousness, it is also widely perceived as an American feature (Tagliamonte, D’Arcy & Jankowski 2010): 50 percent of my respondents associated got with non-British Englishes (the second highest ratio after be like). Jensen (2014) has argued that clear spatial associations promote the socio-cognitive salience of a linguistic form (see also Honeybone & Watson 2013; Beal & Cooper 2015). It might thus be the case that the socio-spatial indexicality of got affects British speakers’ perceptual awareness of its competitor variant, have got, which is isomorphic in form and includes the lexeme got. Indeed, while it is difficult to measure the exact extent of informants’ hand-drawn marking of the words they considered non-standard, many of the circles in the construction I have got problems with nuts too seem focus around the lexeme got.

Regarding the competition in the system of temporal future reference, the prescriptive and descriptive literature boasts a long tradition of deliberating the stylistic, textual, and pragmatic differences between will and shall, and, more recently, between will and going to (see above; also Joos 1964; Leech 1971; Comrie 1985). But while the fact that the system is variable seems to be widely known and recognized, there is little evidence that speakers attach any social meaning in the form of stigma or prestige to this heterogeneity. I am not aware of any public discussions circling around the social meaning of the variants in the system of future time reference, neither in the media nor in any other non-linguistic fora. And while pedagogical web pages dedicated to the English tense-aspect system tend to refer to the standardly adduced differences in immediacy, attitude, or point of view, to my knowledge, none attach any social or ideological meaning to the occurrence of the future variants.¹¹ Hence, notions of correctness or social appropriateness do not seem to be part of the discourses surrounding the variability in the system.

The results of my survey further suggest that the future variants are entirely below the socio-evaluative radar of my informants. In the recognition task, will was only circled once (thus left unmarked 99 percent of the time), and going to was never marked at all. As Table 2 reveals, both variants were overwhelmingly judged to be standard (89 percent and 83 percent) and British (over 90 percent each). Very few respondents associated the forms with younger (3 and 12 percent) and less educated speakers (6 and 9 percent respectively).

Based on the findings reported here, we might thus argue that quotative be like is maximally salient, being frequently commented on in the popular press, picked out by my informants in the social awareness task, and associated with a cluster of social indexicalities (Jensen 2014:36, see also Johnstone & Kiesling 2008). Within the system of stative possessives, got operates above the level of socio-cognitive awareness amongst my informants but seems to have largely escaped prescriptive public commentary. Inclusion in a variable system that includes a relatively salient and structurally isomorphic variant appears to have influenced the perceptual awareness of have got, which might have inherited some of the cognitive-evaluative structures attached to got. In contrast, the variability in the system of future time reference seems to be almost entirely devoid of socio-ideological associations.

6. Longitudinal and Life-span Change

This section explores the three ongoing changes in trend and panel data. For each of the variables, I will first investigate the changes in progress across the trend data. The behavior of the individual panel speakers will then be compared and contrasted to the community that surrounds them. The panel data analysis is beleaguered by a caveat that plagues the lion’s share of longitudinal research, especially in the area of morpho-syntax: the reliance on low token counts (Sankoff 2004; Rickford & Price 2013). The findings reported here thus need to be confirmed on the basis of larger, independent trend or panel samples (which are extremely difficult to construct; see Cheshire 1982, 2005; Bailey, Wikle & Tillery 1997). Throughout the article, the significance of results will be reported via chi-square tests (χ²).

6.1. Longitudinal and Life-span Change in the System of Quotation

Trend data reveal that be like is making swift progress in the quotative system. The sixteen TLS speakers, who were interviewed in 1969-1971, at least a decade before Butters’s (1982) first attestation of be like, produce only one instance of like preceding a quotation (0.4 percent, see Table 3). The form occurs in the construction Personal Pronoun + Be + like + that + GESTURE (4), which has been interpreted as related to the more familiar be like cases (Buchstaller & van Alphen 2012; Macaulay 2001).

(4) if he hadn’t a cigarette for an hour he was like that “[Informant makes hand-gestures].” (TLSG_17)

Table 3.

Quotative Distribution in the 1960s/1970s

	%	n
say	63	164
think	6	15
unframed	21	55
other^a	10	25
Total	100	259

Including n = 1 (0.4 percent) be like

As we would expect with a vibrant change in progress, be like frequencies increase to 21 percent in the 2000s data shown in Table 4 (χ² = 85.234, p < .001). Apparent time analysis supports this finding. As has been pointed out in Buchstaller (2014:166), younger speakers in the 2000s have much higher be like ratios (39 percent) than older speakers (3 percent).

Table 4.

Quotative Distribution in the 2000s

	%	n
say	35	190
think	9	48
unframed	21	116
like	21	113
other	14	76
Total	100	543

Given the fact that be like is only incipient around the time the TLS data were collected, it is not surprising that none of the six panel speakers recorded in 1971 produce any tokens of the variant. But as Table 5 reveals, by 2013, they seem to have taken on board the quotative innovation, albeit at very low frequencies. While small token numbers call for great caution in interpreting these results, older speakers picking up on trends that sweep the community around them have been interpreted as “life-span change” (Sankoff 2005:1011; Sankoff & Blondeau 2007).

Table 5.

The System of Quotative Introducers Amongst the Tyneside Panel Speakers Recorded in 1971 and Again in 2013

	1971		2013
	%	n	%	n
say	51	49	62	221
unframed	32	30	21	75
other	11	11	9	30
think	7	7	7	26
like	0	0	2	7
Total	100	97	100	359

When we split up these numbers by the individual speaker, it becomes immediately obvious that not all panel members participate in the ongoing trend in the quotative system. Two speakers, Fred and Aidan, do not go along with the change in progress. Three other speakers, Rob, Edith, and Anne, started using be like in the forty-two years that lie between the two interviews, albeit very infrequently (n = 1 and n = 2). Nelly, finally, uses the form four times in 2013.

There is a body of evidence that community-wide changes away from the standard tend to occur mainly amongst speakers on the lower end of the socio-economic spectrum (Yaeger-Dror 1994; Sankoff & Blondeau 2007:581). Changes toward the prescriptive standard, on the other hand, tend to be adopted by the middle class speakers (Sankoff & Blondeau 2007; Sankoff & Wagner 2011). The Tyneside panel speakers corroborate this finding: Rob, Edith, and Anne, the speakers who have adopted be like by 2013 are firmly working class, both by their own assessment and by independent measures.¹²

Fred and Aidan, on the other hand, the speakers who do not adopt the incoming form have both experienced considerable upward social mobility across the forty-two years that divide the two data-sets. Both men originally hail from working-class families: Aidan trained as a welder, and Fred, a self-declared “kid from the rough end of Gateshead,” started out as a low-level clerk. When they were first interviewed in 1971, Fred was undergoing teacher training at Durham University, and Aidan had just started out as a welding lecturer. By 2013, both have established themselves in respectable middle-class careers. Fred just retired as a religious education teacher, and Aidan retired as a lecturer at a technical college. The fact that neither of the two men has taken on board the incoming stigmatized variant might thus be seen in the light of recent results on the impact of higher education on speaker’s vernacularity (De Decker 2006; Wagner 2008; Prichard & Tamminga 2012). More importantly, perhaps, both Fred and Aidan have ended up in language sensitive occupations (i.e., professions in which the use of “the legitimized language [. . .] is relatively important,” Sankoff and Laberge 1978:241). Hence, given the professionally precarious ideological load of quotative be like (D’Arcy 2007), we should not be surprised that individuals in the educational sector avoid a linguistic form that is widely stigmatized as “inarticulate” (Schneider 1999) and “verbal trash” (Zernicke 1999). Both Fred and Aidan mention the linguistic scrutiny under which they found themselves, and Fred in particular comments on the degree to which he has changed his accent as a consequence of normative pressures. Fred and Aidan’s trajectory might thus remind us of Sankoff’s (2004) finding from the Seven Up series that Nicholas, another working-class boy who grew to be a lecturer, became less vernacular in the course of his life.

Nelly, on the other hand, increased her use of the variant alongside the community change, producing the highest use of quotative be like (n = 4) in 2013. Nelly is middle class when considering a number of social indicators such as housing, attitudes, and aspirations. Her social background thus might not fit with her adoption of a highly stigmatized variant.¹³ But as I have pointed out in Buchstaller (2015), it is another aspect of Nelly’s personal circumstances that might hold a clue to her adoption of be like: Nelly spent her entire life surrounded by children and young people of all ages. She trained as a nursery nurse, and she worked in kindergartens and preparatory schools until the age of twenty-eight, when she quit her job to bring up her own children and later to help raise her grandchildren aged sixteen, fifteen, and nine in 2013. Nelly continues to have an unusually close and intimate relationship with two of her grandchildren who live nearby (the third lives in London): they tend to spend the afternoons with her, and she reports having long conversations with them, especially with her grandson. Indeed, maybe due to her constant and intimate interaction with young people, Nelly displays very positive attitudes toward innovation and change in general. At seventy-two years of age, she is the only woman in the panel who reports using a mobile phone. She even programs her husband’s phone because she is more “technical,” she communicates via email, and she uses the internet. Hence, Nelly’s linguistic behavior, adopting a variant overtly stigmatized as [+young], might be explained by her close contact to younger speakers in combination with her remarkably positive attitude (given her age) toward innovations and change more generally (see Rogers’s 2003:247 adopter categories). In Chambers’s (2003:95) categorization, Nelly might be classified as an “insider [. . .] who [orients to or] is more similar to the people in the next generation [. . .] linguistically” (consider also the case of Vera in Kochetov 1999 cited in Chambers 2003).

In conclusion, the change in progress in the quotative system involves both generational advancement of the incoming form as well as “the conversion of [some] older speakers to an innovative form, even if they learned an older form as children” (Boberg 2004:266). And while the socially upward mobile speakers Aidan and Fred fail to participate in the highly salient and socially stigmatized change in progress, working-class speakers and speakers with very close contact to children go along with the flow by adopting the innovative quotative be like.

6.2. Longitudinal and Life-span Change in the System of Stative Possessives

Table 6 shows the community-wide changes in the system of stative possession. In the 1960s/1970s recordings, have got occurs in 49 percent of the variable contexts. This result finds a parallel in Tagliamonte’s (2013) apparent time findings from Wheatley Hill, a locality approximately 20 kilometers from Gateshead, where older speakers use have got at a rate of under 60 percent. By the 2000s, have got figures have increased in the Tyneside community corpus to 63 percent (χ² = 15.715, p < .001), which is fully in line with results reported from other communities in the North, such as Buckie and York (Tagliamonte 2013). Apparent time evidence, which is not detailed in Table 6, further supports the hypothesis that the data presented here form part of a larger trend toward increased rates of have got across Northern English localities: in the 1960s/1970s data, younger speakers use have got 71 percent of the time compared to only 31 percent for the older speakers. In the 2000s, while frequencies have increased, the age-gap remains: 73 percent have got amongst the younger speakers compared to 48 percent amongst the older speakers (but see below). Also, got is on the rise, albeit at low levels of frequency (from 2 percent in the 1960s/1970s to 6 percent in the 2000s).

Table 6.

The System of Stative Possession Amongst the Tyneside Trend Sample in the 1960/1970s and in the 2000s

	1960/1970s		2000s
	%	N	%	n
have got	49	64	63	159
have	50	65	31	77
got	2	2	6	16
Total	100	131	100	252

Table 7 reveals that the panel speakers increase their use of got across their life-span, from only one token (1 percent) in 1971 to n = 16 (8 percent) in 2013 (χ² = 4.418, p < .05). But they do not follow the ongoing trend toward increasing ratios of have got, producing 66 percent in 1971 and 61 percent in 2013. Why would this be the case?

Table 7.

The System of Stative Possession Amongst the Tyneside Panel Speakers Recorded in 1971 and Again in 2013

	1971		2013
	%	n	%	n
have got	66	49	61	115
have	32	24	31	59
got	1	1	8	16
Total	100	74	100	190

The panel results in Table 7 are reminiscent of Tagliamonte’s (2013) findings from Wheatley Hill, where apparent time results suggest that speakers in the youngest age bracket (ages thirty-six and below) have retreated from have got. Variationist sociolinguistics has brought to light a number of cases where ongoing linguistic changes are arrested, often due to the high socio-cognitive salience of the form (such as quotative be all, Rickford et al. 2007; Philadelphia tense (æh), Labov, Rosenfelder & Fruehwald 2013). However, as the robust rise of have got ratios across apparent time and real time trend data suggests, the larger Tyneside community continues to propagate the ongoing change in progress.

Indeed, there are good reasons to believe that individual level variation—rather than a general move away from the incoming form—might be able to explain the seemingly inconsistent findings in Tables 6 and 7. Splitting up the results for have got by the individual speakers, as Table 8 does, reveals that behind the aggregate panel data lie very different and highly speaker-specific behaviors: Anne and Rob, the two working-class speakers who had picked up on the stigmatized form be like by 2013, also participate in the ongoing trend in the system of stative possession, increasing their use of have got from 63 percent to 71 percent and from 75 percent to 90 percent. While these individual changes are not statistically significant, Anne and Rob do significantly surpass the community benchmark in the 2013 recordings (χ² = 6.542, p = .0105).

Table 8.

Panel Data on Stative Possessive Use: Comparison Between 1971 and 2013 by Speaker

	Have got		Got		1971	Have got		Got		2013
	%	n	%	n	Total statives	%	n	%	n	Total statives
Aidan	80	12	0	0	15	76	38	10	5	50
Fred	67	8	0	0	12	39	7	0	0	18
Anne	63	5	0	0	8	71	22	10	3	31
Rob	75	9	8	1	12	90	27	7	2	30
Nelly	17	2	0	0	12	14	6	10	4	42
Edith	87	13	0	0	15	79	15	11	2	19

Above I have argued that Fred and Aidan’s social ascent is marked linguistically via conservative behavior toward innovative be like. Table 8 reveals that while these two social risers start off relatively high in terms of have got rates, they do not participate in the community-wide trend toward this variant, which seems to operate at least to a certain extent above the level of consciousness. Below we will discuss this finding in more detail.

Interestingly and somewhat surprisingly, both Edith and Nelly, the most prolific adopter of be like, fail to participate in the ongoing change toward stative possessive have got. Nelly in particular remains at a very low level of uptake across the two recordings (17 percent in 1971 and 14 percent in 2013), slightly reducing her frequencies in opposition to the community trend. Crucially, Nelly’s use of have got is significantly different from all other speakers combined at the .001 level (in 2013, χ² = 12.066).¹⁴ Her conservatism vis-à-vis the changes in the system of stative possession is unexpected given her readiness to embrace the potentially much more costly form be like (see Buchstaller 2014: ch. 4). But I would like to argue that her uptake of the newest form, got, might be used to explain her non-participation in the change in progress toward increasing rates of have got: Nelly embraces the highly salient variant with n = 4 (10 percent), alongside her progressive peers. While this pattern does not reach significance, I interpret it to mean that Nelly, the speaker who has the closest contact with children and adolescents (Rogers 2003; Sankoff 2006), picks up on the most recent, highly salient variants (be like, got) but not so much on slowly incoming forms. Does this mean that the slow, gradual incrementation (i.e., progressive increase) of have got across the larger community is below Nelly’s level of consciousness? We can only speculate. It might be the case that Nelly has not noticed (Preston 2011) the change in progress toward increasing rates of have got. Indeed, her lack of style shift toward increasing rates of have got might suggest that she has not,¹⁵ which would hamper her ability to participate in the community-wide change (Trudgill 1986). It might equally be the case that have got operates above Nelly’s threshold of salience but that she eschews slowly incoming variants while simultaneously embracing highly salient forms indexed as [+young], such as be like and got. Her early and eager uptake of non-linguistic innovations suggests that this might be the case.

All panel speakers, apart from Fred, have taken on board the stative possessive newcomer got. Rob, Edith, and Anne’s adoption of the variant might not be particularly surprising—recall that they have started using be like, which bears a much greater social risk (see Buchstaller 2014)—but Aidan’s use of got might seem surprising. Aidan and Fred, the social risers who ended up in educational professions, have been shown to make consistently conservative quotative choices, and they also do not participate in the community-wide increase in have got frequencies. But whereas Fred remains reluctant to embrace the newest variant in the system of stative possessives, Aidan is amongst the most prolific users of innovative got (n = 5, 10 percent). Although he does not increase his proportional use of have got, he remains at high frequencies of usage in 2013.

While numbers are small and need to be supported by further evidence, Aidan’s linguistic choices might be explained by the different socio-ideological indexicalities of the individual variants in conjunction with his self-positioning as a lecturer in a technical college. Aidan seems to rely on the social meaning of stative possessive got, along with a set of other non-standard forms (see below), in order to create a liberal, socially progressive, and anti-authoritative stance. Similarly to be like, stative possessive got has the indexical load [+young] and [+US]. In contrast to be like, however, got has not attracted the wrath of prescriptive grammarians, making it a much safer choice for professionals (Sankoff & Laberge 1978). What this effectively means is that stative possessive got is available at a relatively low social cost to speakers who experience substantial marketplace pressures but want to portray themselves as au fait with innovative linguistic trends (and thus avoid being perceived as stuffy or old-fashioned).

Aidan’s comments in the 2013 interview might be used to support this point. In (5), he points out that changes in the teacher-student relationship have resulted in a noticeable shift in interactional style, which used to be highly formal and hierarchical at the beginning of his career but which has since become much less authoritative, a development he seems to support. Aidan, crucially, reveals himself as rather astute in picking up on the symbolic significance of stylistic choices for the creation and reflection of interactional stances within the context of his technical college. Thus, similarly to the avoidance of ties and formal terms of address, his adoption of innovative got and continuously high rates of incoming have got might be part of the more egalitarian and anti-authoritarian persona he aims to project. This hypothesis might be supported by Aidan’s strategic uses of “hyper-vernacular” forms, such as ain’t and multiple negation (as in 6; see Cutillas-Espinosa, Hernández-Campoy & Schilling-Estes 2010).

(5) In the seventies you went to work with a collar and tie [. . .] but the type of kids that we got through our hands resented ties eh their teacher wore a tie [. . .].

I thought the tie had this sort of symbol [. . .] but gradually ehm [. . .] the majority of teaching staff stopped wearing ties and wore a pullover open-necked shirt that that sort of thing. And things did change and the attitude. You were “Mister Fulham” in the early days “Mister Fulham this Mister Fulham that” and then that didn’t lie particularly well with me. I mean you know I’m Aidan that’ll do me [. . .]

And then it—towards the end it changed, it became “Aidan, Aidan this Aidan that”

There was a change—a social change in the eh—in the attitudes of staff and pupils and students. (A_2013)

(6) a. You ain’t going to get nowhere in the civil service. (A_2013)

b. She’s a little bit less academically able than he is. But ehm she ain’t going to let it faze her. (A_2013)

Fred, on the other hand, the retired religious education teacher, continues to avoid the innovative form got and he has reduced have got frequencies since his first interview, which was conducted when he was still in college (from 67 percent to 39 % not statistically significant due to low token numbers). In contrast to Aidan, Fred adopts a rather critical stance toward the changes in the educational system, bemoaning the increasingly lackadaisical attitude and lack of discipline in public schools. As (7) illustrates, Fred would like to see a return to more structured and performance-orientated classes. Also, while Aidan displays a fairly positive attitude toward societal changes and their repercussions on the school system, Fred projects a very different persona throughout his interview, describing himself as a “a nerd” who “doesn’t like modern stuff” (see 8 and 9). His cautious attitudes toward changes in educational policy as well as his rather retro attitude toward cultural and technological innovations might explain why he—as the only speaker in the panel sample—has not picked up on stative possessive got. In Rogers’s (2003:284) adopter categorization, Fred could thus be characterized as a laggard, that is, someone whose “point of reference [. . .] is the past and [who is] suspicious of innovation and change.”

(7) I’ve got a fairly negative view of the way education has gone.

I think the experiment that we had in the forties, fifties and sixties after the Second World War seems to me to have gone off beam really [. . .]

From what I remember from my education it was a very formal education, a boys’ grammar school, Gateshead Grammar School. We were formally taught [. . .]

I thoroughly enjoyed it. You know it was what I needed. I don’t think kids get that these days in secondary school. But I think there isn’t enough academic pressure in ordinary state comprehensive schools [. . .] (F_2013)

(8) I had a paper round and apart from buying a bike I also bought a second-hand ehm record [. . .] and I the first track that I bought was Beethoven’s fourth symphony which I can’t believe. Sorry I was always a nerd.

Right ehm so that’s been a lifelong thing [. . .] I learned to play the cornet or the flugelhorn [. . .] I enjoy playing the classical guitar which again is a typical nerd’s instrument really. (F_2013)

(9) Now I’ve got a little CD player which is not the same really so no I’m not that happy. I’ve still got the old LPs you know the old vinyl [. . .]

I like a good bass you know and surround sound and I don’t like modern stuff well the stuff you buy in the normal shops [. . .]

I don’t know yes it may be nostalgia yeah it may be. When I started it wasn’t vinyl it was Bakelite. I should have kept those as well. I used to be interested in Gilbert and Sullivan at one time, again a nerdy thing. I’m just obviously a nerd yeah. (F_2013)

6.3. Longitudinal and Life-Span Change in the System of Future Time Reference

While speakers might be aware that they have more than one choice to express future time reference, they do not seem to attach any socio-ideological associations with the competition between will and going to (i.e., they do not classify the variability; see Preston 2011). This section seeks to investigate whether the dearth of social indexicality in the system of future time reference manifests itself in a lack of post-adolescent modulation across the life-span.

Considering first the trend data, Table 9 reveals a relatively rapid shift toward increasing use of going to in the Tyneside speech community, from 20 percent in the 1960s/1970s to 46 percent in the 2000s (χ² = 18.514, p < .0001). Tagliamonte’s (2013:124) contention that the incursion of going to into the system has recently accelerated seems to be borne out by the data. Apparent time analysis (not shown in the table) supports this hypothesis: younger speakers produce consistently higher going to ratios, in the 1960s/1970s (29 percent versus 9 percent) as well as in the 2000s data (53 percent versus 25 percent). Do the panel speakers go along with the community-wide change in progress? Table 10 suggests that the aggregate going to rates increase from 20 percent to 30 percent between the 1971 and the 2013 recordings (though the results are not statistically significant). Unfortunately, splitting up the results by the individual speaker reveals that, apart from Anne, the panel members produce only very few instances of future tokens in the TLS. The following sections will thus focus on Anne, a speaker who has demonstrated substantial life-span change.

Table 9.

The System of Future Time Reference Amongst the Tyneside Trend Sample in the 1960/1970s and in the 2000s

	1960/1970s		2000s
	%	n	%	n
Going to (gonna)	20	20	46	97
Will/’ll	79	79	55	116
Total	100	99	100	213

Table 10.

The System of Future Time Reference Amongst the Tyneside Panel Speakers Recorded in 1971 and Again in 2013

	1971		2013
	%	n	%	n
Going to (gonna)	20	6	30	23
Will/’ll	80	24	70	54
Total	100	30	100	77

Anne’s proportional use of the going to future stays entirely stable across the forty-two years that divide the two interviews (13 percent, n = 2). On the one hand, this might seem unexpected since Anne has followed other community-wide trends, increasing her use of have got and got and picking up on potentially socially costly quotative be like (Buchstaller 2014). What all of these variants have in common is that they are above the level of socio-cognitive awareness and thus readily available for incrementation across the life-span. Her non-increase of going to might thus be contingent on the lack of social indexicality of the change in the system.

While the dearth of future tokens in the 1971 data makes the other speakers’ developmental trajectory difficult to interpret, note that Nelly, who previous sections have shown to shift toward highly salient incoming forms, is equally stable across the two data-sets (40 percent, n = 2). The lack of incrementation in the system of future time reference across the individuals’ life-span might thus be explained by the fact that the community-wide trend toward increased ratios of going to is inaccessible for speakers’ socio-cognitive processing (Labov et al. 2011) and consequently not available for incrementation. This finding suggests that postadolescent speakers only participate in ongoing changes in progress they have access to, i.e., provided that they are aware of the social meaning of their changing nature (Trudgill 1986; Preston 2011).

7. Conclusion

This article explored three morpho-syntactic changes in progress across longitudinal trend and panel data. Sociophonetic research has revealed important and regular correlations between the level of awareness of the social meaning of phonological changes and speakers’ modulations during their life-span. But we know very little about the impact of socio-cognitive salience on changes at other levels of linguistic modularity. The study reported here examined three changes in progress at different levels of social perception and indexicality: ultra-salient (Jensen 2014) and socially stigmatized quotative be like; highly salient but devoid of prescriptive commentary stative possessive got, moderately salient and not socially indexical have got; and the change in the system of future temporal reference toward going to, which seems to be largely devoid of social associations. The results, while exploratory in nature, support recent findings emerging out of morpho-syntactic panel research which have demonstrated the key role of socio-cognitive salience on linguistic changes (Sankoff & Wagner 2011).

Token numbers were sometimes precariously low, but an analysis of the individual panel members allowed us to examine the importance of speaker-based factors such as contact with children, the individuals’ social trajectories, and their presentation of self. Individual-level analysis reveals that the panel speakers fall into three broad linguistic profiles: two upwardly mobile speakers, whose choice of profession means they have found themselves under considerable linguistic scrutiny throughout their careers, tend to resist ongoing trends. Neither Fred nor Aidan pick up on quotative be like and they do not participate in the community-wide increase in have got rates. Fred in particular comments on the marketplace pressures he has faced throughout his life, which—in combination with his reticent attitude toward the adoption of innovations in general—might account for the fact that he is the only speaker in the panel sample who has not picked up on stative possessive got. Aidan on the other hand adopts the got variant, using it in conjunction with other non-standard forms to portray himself as a modern, liberal educator who is in touch with incoming trends.

Rob, Anne, and Edith, the three working-class speakers, shift toward new or incoming forms, provided that they are available for incrementation. With respect to be like, their adaptive behavior entails a change away from conventional quotatives and toward a highly stigmatized form. With respect to moderately salient have got, Rob and Anne follow the community in increasing their ratios of the variant, which comes at a relatively low social cost. All three also pick up on new and incoming got to a certain degree. Their behavior can thus be described as life-span change (Sankoff 2005; Sankoff & Wagner 2011), a specific form of age-grading, where older speakers go along with ongoing changes in community norms.

Finally there is Nelly, the middle class speaker whose professional choices and personal trajectory have meant that she has had the regular and intimate contact with children and young people which is typical for involved grandparents but fundamentally different from the teachers Fred and Aidan. Nelly is very open toward innovations in general—recall that she programs her husband’s cell phone. I suggested that Nelly could be characterized as a hip “insider” (Chambers 2003:95) who readily picks up on highly salient youth trends (be like, got), but who does not necessarily go along with slowly progressing community-wide changes (such as stative possessive have got). This might be because Nelly is not aware of more gradual changes, a hypothesis which is supported by the fact that her have got ratios do not increase in the course of her interview (recall also Tagliamonte, D’Arcy & Jankowski’s 2010:150 claim that speakers are not aware whether they say they have or have got a cat). Another possible explanation might be that the adoption of linguistic forms that index [+young] is a conscious, agentive choice by a speaker with ample, longstanding, and intimate contact with adolescents and young adults.

While low token numbers precluded a detailed analysis of the full set of panel speakers, even individuals who have been shown to pick up on other ongoing community-wide trends failed to participate in the change in the system of future temporal. Above I have argued that this lack of incrementation might be explained by the low level of socio-cognitive salience of the variability in the system, which would make it unavailable for modification across the speakers’ life-spans.

Let us now consider three main research questions posed in the introduction:

Do post-adolescent speakers go along with the morpho-syntactic changes that are currently making headway in the community in which they are embedded (i.e., do they manifest life-span change)? The present paper contributes to the increasing amount of research which suggests that some speakers modify morpho-syntactic variables in the course of their life-span. Similarly to findings on phonological variables, speakers’ adaptive behavior across their life-histories results in complex outcomes that cannot be characterized by wholesale convergence or divergence (Bigham 2010; Prichard & Tamminga 2012). Morpho-syntactic malleability during a speaker’s life-span is contingent on a multitude of factors, including the socio-cognitive salience of the changes involved (both of the individual variant as well as of the variable in which it is embedded), the socio-economic dynamics of the community and of the speaker, as well as on a range of speaker-based factors, such as the amount and intensity of contact with younger speaker and the individual’s interactional stance-taking (Rogers 2003; Denis 2011).

Which speakers exhibit post-adolescent linguistic change across the forty-two years covered by the data? At a very general level, we might characterize the findings reported here as follows: social risers in linguistically sensitive jobs do not follow changes that are socially costly. Working-class speakers tend to follow ongoing changes. The middle class individual who maintains unusually close and intimate relationships with younger speakers, Nelly, is especially astute in picking up youth trends. These findings allow for the possibility that the changes across a speaker’s life-span might not be the low-level automatic convergence effects suggested by Trudgill (2011) but rather conscious adaptive behavior towards changing social norms (as argued in Prichard & Tamminga 2012). If the effects were merely a matter of mechanistic convergence, we would not expect to see Aidan avoiding be like while simultaneously adopting the socially less costly got or indeed Nelly’s non-adjustment to the community-wide increase in have got rates. In the same vein, Prichard and Tamminga (2012:94) suggest that “complex accommodation outcomes may reflect speaker’s strategic use of both [old and new] [. . .] forms to maintain ties to their home social networks while simultaneously connecting to” different networks and audiences. The speakers in the research reported here seem to juggle conflicting linguistic identities, namely their own, age-appropriate language use as well as ongoing community-wide changes resulting in new and incremental targets, which are represented not only by the generation of their grandchildren, to whom many have a very close relationship, but also by the fieldworker, who falls into the generation below them. In order to account for this complex picture, the research reported here suggests that we need to draw on a combined approach, relying on both the socio-cognitive salience of the linguistic features as well as on speaker-based factors. Future research will need to explore the micro-stylistic choices panel speakers make throughout the moment-by-moment interaction, examining factors such as footing, identity management, and topic development (Sharma & Rampton 2015; Gregersen & Barner-Rasmussen 2011).

How does socio-cognitive salience interact with (non)adoption? Labov’s (1994:111-112) hypothesis that “variables operating at high levels of social awareness [can be] modified throughout a speaker’s lifetime” is fully corroborated by the research reported here. Indeed, the present article aimed to unpack the concept of socio-cognitive salience by investigating four variants at different levels of community- and speaker-based awareness. Highly salient (got) and openly proscribed forms (be like), while entirely avoided by the careful upwardly mobile speaker, are moderately embraced by speakers in contact with children, with less severe marketplace pressures or by those who choose to portray themselves as progressive and in touch with incoming trends. Even moderately salient forms such as have got are amenable to change across the life-span (but recall Nelly and Edith’s non-participation in the trend). The incoming going to-future, on the other hand, seems to be situated below the level of awareness. Even the most enthusiastic adopter of other changes, Anne, does not increase her going to ratios beyond the frequencies she set at a young age (or indeed when she was first recorded at age twenty-three). These findings seem to suggest that only those variants that are fully embedded in listeners’ cognitive-evaluative structures show malleability across the life-span of the individual speaker (Bigham 2010; Prichard & Tamminga 2012). Variants that are not socio-cognitively accessible might thus not be available for modification across a speaker’s life-span.

Research on the transmission of linguistic forms has confirmed that while “change from below [the level of consciousness can be] [. . .] driven by the continuous process of incrementation” (Labov 2007:346, 379), the diffusion of such changes to other populations, such as geographically and socially isolated individuals, tends to be sporadic, less regular, and at a much slower rate (see also Sankoff & Wagner 2011). There is thus a fundamental difference between acquisition through transmission, where “successive cohorts and generations of children advance the change beyond the levels of their caretakers” (Labov 2007:346, 379) and post-adolescent appropriation, which relies, amongst other things, on speakers’ awareness of the variability and its social indexicalities. We can thus appeal to socio-cognitive salience as an explanatory factor for the occurrence of linguistic malleability across the life-span of the individual. Put more generally, we might hypothesize that socio-cognitive salience is a crucial parameter in establishing the “scope and limitations of speakers’ abilities to change their speech” (Bowie & Yaeger-Dror 2015). Awareness of the social indexicality of linguistic resources is thus a maximally parsimonious explanation which captures both the linguistic malleability we observe across the speaker’s life-span as well as stylistic shifts across parameters such as topic, interlocutor or attention to speech. It therefore satisfies Occam’s Razor.

But the findings reported here also support the importance of speaker-based factors, including the individual’s social networks and communities of practice but also more interactional considerations such as speaker stance and identity formation: only variants that can be fruitfully integrated into an individual’s presentation of self will be taken up and incorporated into their linguistic habitus.

Footnotes

Appendix

Acknowledgements

I would like to thank Adam Mearns, Steff Otte and Anja Auer for their help in extracting and coding the data. I would also like to thank the audience at the Northern Englishes Workshop at Lancaster University as well as two anonymous reviewers. All remaining errors are obviously my own.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This paper could not have been written without the generous support of a European Commission CIG grant (2013-2017: Diagnostics of linguistic change: Mapping language change in real and apparent time. FP7 Marie Curie European Research Grant).

Notes

Author Biography

Isabelle Buchstaller is Professor of Varieties of English at Leipzig University. She works on sociolinguistics, language variation and change, and language contact. Her recent research examines the extent to which individual speakers’ grammars change in relation to ongoing community-wide changes. Her books include Quotatives: New Trends and Sociolinguistic Implications.

References

Aarts

Bas

Joanne

Leech

Geoffrey

Wallis

Sean

(eds.). 2013. The verb phrase in English: Investigating recent language change with corpora. Cambridge: Cambridge University Press.

Arnaud

René

. 1998. The development of the progressive in 19th century English: A quantitative survey. Language Variation and Change 10(2). 123-152.

Bailey

Guy

Wikle

Tom

Tillery

Jan

. 1997. The effects of methods of results in dialectology. English World-Wide 18(1). 35-63.

Beal

Joan

. 2009. Three hundred years of prescriptivism (and counting). In van Ostade

Ingrid Tieken-Boon

Van der Wurff

Wim

(eds.), Current issues in Late Modern English, 35-56. Bern: Peter Lang.

Beal

Joan

Burbano-Elizondo

Lourdes

Llamas

Carmen

. 2012. Urban North-Eastern English: Tyneside to Teesside. Edinburgh: Edinburgh University Press.

Beal

Joan

Cooper

Paul

. 2015. The enregisterment of Northern English. In Hickey

Raymond

(ed.), Researching Northern Englishes, 27-50. Amsterdam: Benjamins.

Biber

Douglas

Johansson

Stig

Leech

Geoffrey

Conrad

Susan

Finegan

Edward

. 1999. The Longman grammar of spoken and written English. Harlow, Essex: Longman.

Bierma

Nathan

. 2005. It’s been, like, 10 whole years since “Clueless” helped spread Valley slang. Chicago Tribune. http://articles.chicagotribune.com/2005-07-20/features/0507190293_1_clueless-slang-amy-heckerling (18 August 2014).

Bigham

Douglas

. 2010. Mechanisms of accommodation among emerging adults in a university setting. Journal of English Linguistics 38(3). 193-210.

10.

Blondeau

Hélène

Sankoff

Gillian

Charity

Anne

. 2002. Parcours individuels et changements linguistiques en cours dans la communauté francophone Montréalaise. Revue Québécoise de Linguistique 31(1). 13-38.

11.

Boberg

Charles

. 2004. Real and apparent time in language change: Late adoption of changes in Montreal English. American Speech 79(4). 250-269.

12.

Bowie

David

Yaeger-Dror

Malcah

. 2015. Language change in real time. In Honeybone

Patrick

Salmons

Joseph

(eds.), The Oxford handbook of historical phonology, 603-618. Oxford: Oxford University Press.

13.

Buchstaller

Isabelle

. 2006. Diagnostics of age-graded linguistic behaviour. Journal of Sociolinguistics 10(1). 3-30.

14.

Buchstaller

Isabelle

. 2014. Quotatives: New trends and sociolinguistic implications. Boston: Wiley.

15.

Buchstaller

Isabelle

. 2015. Exploring linguistic malleability across the life-span: Age-specific patterns in quotative use. Language in Society 44(4). 457-496.

16.

Buchstaller

Isabelle

D’Arcy

Alexandra

. 2009. Localized globalization: A multi-local, multivariate investigation of quotative like. Journal of Sociolinguistics 13(3). 291-332.

17.

Buchstaller

Isabelle

van Alphen

Ingrid

. 2012. Introductory remarks on new and old quotatives. In Buchstaller

Isabelle

van Alpen

Ingrid

(eds.), Quotatives: Cross-linguistic and cross-disciplinary perspectives, xii-xxx. Amsterdam: Benjamins.

18.

Butters

Ronald

. 1982. Editor’s note [on be like ‘think’]. American Speech 57(2). 149.

19.

Campbell-Kibler

Kathryn

. 2010. Perception in sociolinguistics. Language and Linguistics Compass 4(6). 377-389.

20.

Campbell-Kibler

Kathryn

. 2011. The sociolinguistic variant as a carrier of social meaning. Language Variation and Change 22(3). 423-441.

21.

Chambers

J. K.

2003. Sociolinguistic theory. Oxford: Blackwell.

22.

Chapman

Carol

. 1995. Perceptual salience and analogical change: Evidence from vowel lengthening in Modern Swiss German dialects. Journal of Linguistics 31(1). 1-13.

23.

Cheshire

Jenny

. 1982. Variation in an English dialect: A sociolinguistic study. Cambridge: Cambridge University Press.

24.

Cheshire

Jenny

. 2005. Syntactic variation and spoken language. In Cornips

Leonie

Corrigan

Karen

(eds.), Syntax and variation: Reconciling the biological and the social, 81-106. Amsterdam: John Benjamins.

25.

Comrie

Bernard

. 1985. Tense. Cambridge: Cambridge University Press.

26.

Corrigan

Karen

Buchstaller

Isabelle

Mearns

Adam

Moisl

Hermann

. Forthcoming. The talk of the toon: A linguistic time capsule for the Google generation. In Corrigan

Karen

Mearns

Adam

Moisl

Hermann

(eds.), Digitizing corpora, vol. 3. London: Palgrave.

27.

Cutillas-Espinosa

Juan Antonio

Hernández-Campoy

Juan Manuel

Schilling-Estes

Natalie

. 2010. Hypervernacularisation and speaker design: A case study. Folia Linguistica 44(1). 31-52.

28.

Danchev

Andrej

Kytö

Merja

. 1994. The construction be going to + infinitive in Early Modern English. In Kastovsky

Dieter

(ed.), Studies in Early Modern English, 59-77. Berlin: Mouton de Gruyter.

29.

D’Arcy

Alexandra

. 2007. “Like” and language ideology: Disentangling fact from fiction. American Speech 82(4). 386-419.

30.

De Decker

Paul

. 2006. Real-time investigation of social and phonetic changes in post-adolescence. University of Pennsylvania Working Papers in Linguistics: Selected Papers from NWAV 34 12(2). 65-76.

31.

Denis

Derek

. 2011. Innovators and innovation: Tracking the innovators of and stuff in York English. University of Pennsylvania Working Papers in Linguistics 17(2). 61-70.

32.

Fleischman

Suzanne

. 1982. The future in thought and language: Diachronic evidence from Romance. Cambridge: Cambridge University Press.

33.

Gregersen

Frans

Barner-Rasmussen

Michael

. 2011. The Logic of comparability: On genres and phonetic variation in a project on language change in real time. Corpus Linguistics and Linguistic Theory 7(1). 7-36.

34.

Harrington

Jonathan

Palethrope

Sallyanne

Watson

Catherine

. 2000. Monophthongal vowel changes in received pronunciation: An acoustic analysis of the Queen’s Christmas broadcasts. Journal of the International Phonetic Association 30(1/2). 63-78.

35.

Heine

Bernd

. 1993. Auxiliaries. Oxford: Oxford University Press.

36.

Heine

Bernd

Kuteva

Tania

. 2002. The world lexicon of grammaticalization. Cambridge: Cambridge University Press.

37.

van Hofwegen

Janneke

Wolfram

Walt

. 2010. Coming of age in African American English: A longitudinal study. Journal of Sociolinguistics 14(4). 427-455.

38.

Hollmann

Willem B.

Siewierska

Anna

. 2006. Corpora and (the need for) other methods in a study of Lancashire dialect. Zeitschrift für Anglistik und Amerikanistik 54(2). 203-216.

39.

Honeybone

Patrick

Watson

Kevin

. 2013. Salience and the sociolinguistics of Scouse spelling: Exploring the phonology of the contemporary humorous localised dialect literature of Liverpool. English World-Wide 34(3). 305-340.

40.

Huddleston

Rodney

Pullum

Geoffrey

. 2002. The Cambridge grammar of the English language. Cambridge: Cambridge University Press.

41.

Jensen

Marie M.

2014. Salience in language change: A socio-cognitive study of Tyneside English. Newcastle: Northumbria University dissertation.

42.

Johnstone

Barbara

Kiesling

Scott

. 2008. Indexicality and experience: Exploring the meanings of /aw/-monophthongization in Pittsburgh. Journal of Sociolinguistics 12(1). 5-33.

43.

Joos

Martin

. 1964. The English verb: Form and meanings. Madison: University of Wisconsin Press.

44.

Kerswill

Paul

Williams

Ann

. 2002. ‘Salience’ as an explanatory factor in language change: Evidence from dialect levelling in urban England. In Jones

Mari C.

Esch

Edith

(eds.), Language change: The interplay of internal, external and extra-linguistic factors, 81-110. Berlin: Mouton de Gruyter.

45.

Kochetov

Alexei

. 1999. Language and social change in a rural Russian community. Toronto: Generals paper, Department of Linguistics, University of Toronto.

46.

Kroch

Anthony

. 1989. Reflexes of grammar in patterns of language change. Language Variation and Change 1(3). 199-244.

47.

Labov

William

. 1972. Sociolinguistic patterns. Oxford: Blackwell.

48.

Labov

William

. 1993. The unobservability of structure and its linguistic consequences. Ottawa: NWAV 22 paper.

49.

Labov

William

. 1994. Principles of linguistic change, vol. 1. Internal factors. Oxford: Wiley.

50.

Labov

William

. 2001. Principles of linguistic change, vol. 2. Social factors. Malden, MA: Blackwell.

51.

Labov

William

. 2006. The social stratification of English in New York City. 2nd edn. Cambridge: Cambridge University Press.

52.

Labov

William

. 2007. Transmission and diffusion. Language 83(2). 344−387.

53.

Labov

William

Ash

Sharon

Ravindranath

Maya

Weldon

Tracey

Baranowski

Maciej

Nagy

Naomi

. 2006. Listeners’ sensitivity to the frequency of sociolinguistic variables. University of Pennsylvania Working Papers in Linguistics 12(2). 105-129.

54.

Labov

William

Ash

Sharon

Ravindranath

Maya

Weldon

Tracey

Baranowski

Maciej

Nagy

Naomi

. 2011. Properties of the sociolinguistic monitor. Journal of Sociolinguistics 15(4). 431-463.

55.

Labov

William

Rosenfelder

Ingrid

Fruehwald

Josef

. 2013. One hundred years of sound change in Philadelphia: Linear incrementation, reversal, and reanalysis. Language 89(1). 30-65.

56.

Leech

Geoffrey

. 1971. Meaning and the English verb. New York: Longman.

57.

Lenneberg

Eric

. 1967. Biological foundations of language. New York: Wiley.

58.

Levon

Erez

Buchstaller

Isabelle

. 2015. Perception, cognition and linguistic structure: The effect of linguistic modularity and cognitive style on sociolinguistic processing. Language Variation and Change. 27(3). 319-348.

59.

Levon

Erez

Fox

Sue

. 2014. Salience and the sociolinguistic monitor: A case study of ING and TH-fronting in Britain. Journal of English Linguistics 42(3). 185-217.

60.

Macaulay

Ronald

. 2001. You’re like “why not?” The quotative expressions of Glasgow adolescents. Journal of Sociolinguistics 5(1). 3-21.

61.

Milroy

James

Milroy

Lesley

. 1985. Linguistic change, social network and speaker innovation. Journal of Linguistics 21(2). 339-384.

62.

Milroy

James

Lesley

Milroy

. 2003. Authority in Language. Investigating Standard English. London and New York: Routledge.

63.

Nevalainen

Terttu

Raumolin-Brunberg

Helena

. 2003. Historical sociolinguistics: Language change in Tudor and Stuart England. London: Longman.

64.

Palmer

Frank R.

1974. The English verb. London: Longman.

65.

Poplack

Shana

Tagliamonte

Sali

. 1999. The grammaticization of going to in (African American) English. Language Variation and Change 11(3). 315-342.

66.

Preston

Dennis R.

1985. The Li’l Abner syndrome: Written representations of speech. American Speech 60 (4). 328–336.

67.

Preston

Dennis

. 2011. The power of language regard: Discrimination, classification, comprehension and production. Dialectologia, Special Issue II. 9-33.

68.

Prichard

Hilary

Tamminga

Meredith

. 2012. The impact of higher education on Philadelphia vowels. University of Pennsylvania Working Papers in Linguistics 18(2). 87-95.

69.

Quirk

Randolph

Greenbaum

Sidney

Leech

Geoffrey

Svartvik

Jan

. 1985. A comprehensive grammar of the English language. New York: Longman.

70.

Rickford

John

Buchstaller

Isabelle

Wasow

Thomas

Zwicky

Arnold

Traugott

Elizabeth

. 2007. Intensive and quotative ALL. Something old, something new. American Speech 82(1). 3-31.

71.

Rickford

John

McNair-Knox

Faye

. 1994. Addressee- and topic-influenced style shift: A quantitative sociolinguistic study. In Biber

Douglas

Finegan

Edward

(eds.), Perspectives on register: Situating register variation within sociolinguistics, 235-276. Oxford: Oxford University Press.

72.

Rickford

John

Price

Mackenzie

. 2013. Girlz II women: Age-grading, language change, and stylistic variation. Journal of Sociolinguistics 17(2). 143-179.

73.

Rogers

Everett

. 2003. Diffusion of innovations. New York: Free Press.

74.

Sankoff

David

Laberge

Suzanne

. 1978. The Linguistic market and the statistical explanation of variability. In Sankoff

David

(ed.), Linguistic variation: Models and methods, 239-250. New York: Academic Press.

75.

Sankoff

Gillian

. 2004. Adolescents, young adults and the critical period: Two case studies from “seven up.” In Fought

Carmen

(ed.), Sociolinguistic variation: Critical reflections, 121-139. Oxford: Oxford University Press.

76.

Sankoff

Gillian

. 2005. Cross-sectional and longitudinal studies in sociolinguistics. In Ammon

Ulrich

Dittmar

Norbert

Mattheier

Klaus J.

Trudgill

Peter

(eds.), Sociolinguistics: An international handbook of the science of language and society, 1003-1013. Berlin: Mouton de Gruyter.

77.

Sankoff

Gillian

. 2006. Age: Apparent time and real time. In Brown

Keith

(ed.), Encyclopedia of language and linguistics, 2nd edn., 110-116. Amsterdam: Elsevier.

78.

Sankoff

Gillian

Blondeau

Hélène

. 2007. Language change across the lifespan: /r/ in Montreal French. Language 83(3). 560-588.

79.

Sankoff

Gillian

Thibault

Pierrette

Wagner

Suzanne E.

2004. An apparent time paradox: Change in Montréal French auxiliary selection, 1971-1995. Ann Arbor, MI: NWAV 33 paper.

80.

Sankoff

Gillian

Wagner

Suzanne E.

2006. Age grading in retrograde movement: The inflected future in Montréal French. University of Pennsylvania Working Papers in Linguistics 12(2). 203-216.

81.

Sankoff

Gillian

Wagner

Suzanne E.

2011. Age grading in the Montréal French future tense. Language Variation and Change 23(3). 275-313.

82.

Schneider

Alison

. 1999. Taking aim at student incoherence: Spread of speech programs across the curriculum irks some communication professors. Chronicle of Higher Education 45(29). A16-A18.

83.

Sharma

Devyani

Rampton

Ben

. 2015. Lectal focusing in interaction: A new methodology for the study of superdiverse speech. Journal of English Linguistics 43(1): 3-35.

84.

Tagliamonte

Sali

. 2013. The roots of English. Cambridge: Cambridge University Press.

85.

Tagliamonte

Sali

D’Arcy

Alexandra

. 2007. Frequency and variation in the community grammar: Tracking a new change through the generations. Language Variation and Change 19(2). 199-217.

86.

Tagliamonte

Sali

D’Arcy

Alexandra

Jankowski

Bridget

. 2010. Social work and linguistic systems: Marking possession in Canadian English. Language Variation and Change 22(1). 149-173.

87.

Torres-Cacoullos

Rena

Walker

James A.

2009. The present of the English future: Grammatical variation and collocations in discourse. Language 85(2). 321-354.

88.

Trudgill

Peter

. 1986. Dialects in contact. Oxford: Blackwell.

89.

Trudgill

Peter

. 2011. Sociolinguistic typology: Social determinants of linguistic complexity. Oxford: Oxford University Press.

90.

Visser

Fredericus

. 1963-1978. An historical syntax of the English language. Leiden: E.J. Brill.

91.

Wagner

Suzanne E.

2008. Language change and stabilization in the transition from adolescence to adulthood. Philadelphia: University of Pennsylvania dissertation.

92.

Wagner

Suzanne E.

2012a. Age grading in sociolinguistic theory. Language and Linguistics Compass 6(6). 371-382.

93.

Wagner

Suzanne E.

2012b. Real-time evidence for age grad(ing) in late adolescence. Language Variation and Change 24(2). 179-202.

94.

Wagner

Suzanne

Tagliamonte

Sali

. Forthcoming. What makes a panel study work? Researcher and participant in real time. In Wagner

Suzanne

Buchstaller

Isabelle

(eds.), Panel studies of variation and change. New York: Routledge.

95.

Yaeger-Dror

Malcah

. 1989. Real time vs. apparent time change in Montreal French. York Papers in Linguistics 13. 141-153.

96.

Yaeger-Dror

Malcah

. 1993. Linguistic analysis of dialect “correction” and its interaction with dialect salience. Language Variation and Change 5(2). 189-224.

97.

Yaeger-Dror

Malcah

. 1994. Sound change in Montreal French. In Keating

Patricia

(ed.), Phonological structure and phonetic form: Papers in laboratory phonology 3, 267-292. Cambridge: Cambridge University Press.

98.

Zernicke

Kate

. 1999. Talk is, like, you know, cheapened. Boston Globe. 31 January 1999.