The individual in the semiotic landscape

I present the extreme proposal that change spreads by virtue of its role in a system of social meaning. And since individuals cannot construct meaning on their own, they can play no elemental role in sound change. Based on ethnographic-variationist studies of sound change among preadolescents and adolescents, I challenge two common assumptions in the study of variation and change: (1) that sound change is autonomous, and (2) that change spreads from individual to individual, by imitation and in isolation. Whatever its origins, whether from linguistic pressures (“change from below”) or social pressures (“change from above”), sound change spreads by virtue of its incorporation into a semiotic landscape, as non-referential material is recruited into signs articulating social distinctions. Participation in this landscape connects the individual to the immediate community and to the larger social order and it is through participation in this landscape that speakers produce and perceive – and accelerate – changes in progress.


Introduction
William Labov has long argued that the individual is not a suitable object of linguistic study because the speech patterns of the individual do not exist independently of the larger patterns of the community.I carry this claim one step further, with the proposal that change spreads by virtue of its role in a system of social meaning, and meaning is not constructed by individuals.This is an extreme hypothesis, and while I cannot say that sound change never progresses without taking on social meaning, I have never seen a contemporary example of one that did.The evidence of complex indexicality in variation is sufficiently strong at this point (e.g.Zhang 2005;Podesva 2011;Podesva & Van Hofwegen 2016) that I would not be satisfied with the claim that a sound change was meaningless unless every effort had been made to prove otherwise.
Before I turn to my argument, I should point out that I am engaging in a different discourse from the one that examines cognitive and articulatory abilities or styles that are individually, not socially, variable such as those discussed in several papers in this volume (MacKenzie To appear; Yu To appear).Some of these differences may well lead to social differences but in order for them to affect the spread of change, they need to rise to a social level, transcending the individual.As Alan Yu (To appear) points out, the spread of change ultimately depends on the innovative speaker's social influence, and this is unlikely to be a direct result of differences in modes of production and processing.
I will be arguing that however they originate, sound changes enter into a semiotic landscape, taking on meaning as components of styles.These styles are part of the construction of personae that inhabit and define the social landscape.Underlying this argument is the recognition that style is not trivial or an add-on to language, but fundamental to social -hence linguistic -life.In my focus on the semiotic landscape, I challenge two basic approaches to variation and change.The first is the treatment of sound change as autonomous.We examine a sound change as having a life of its own, whether alone or in the company of related sounds, as in a chain shift.But a sound change, or even an entire shift, does not occur on its own in speech but in the company of an endless variety of other variables, specifically as a component not just of a dialect but of styles.The second is the linear perspective in which a change is seen as spreading through social networks, from individual to individual by imitation or accommodation and as a function of contact.I do not mean to minimize the importance of social networks in the spread of change, but question any view of networks as mere conduits, and contact as the sole force in the adoption of change (e.g.Bloomfield 1933;Bermúdez-Otero To appear).This mechanistic approach, I note, is not the perspective of most sociolinguistic work on networks (e.g.Milroy 1980;Cheshire 1982;Bortoni-Ricardo 1985), which focuses on motivations for adoption of change in and across local clusters.While social networks clearly structure exposure to change across large populations, such as between cities or between city and country, network clusters are commonly sites of local practice in which speakers jointly orient to their proximal and distal surroundings (Eckert 2000;2004).It is not individualto-individual contact that spreads change at this local level, but the joint indexical process that establishes a change as locally meaningful.I argue that whatever its origins, sound change spreads by virtue of its incorporation into the social-semiotic landscape.

The semiotic landscape
The social stratification of variables found in urban studies shows the spread of sound change, such as the raising of (aeh) and (oh) in New York (Labov 1966), via class-based social networks.The resulting class stratification applies not only to sound changes in progress but to variables that are relatively stable over time, such as (-ing) and (dh-stopping).This pattern is the result of the centrality of class to the social order.But while all of these variables correlate on the macrosocial level with class, they do so differently.(oh) and (aeh) raising in New York City show a complex class pattern known as the "lower middle class crossover", while (-ing) and (dh-stopping) show a regular class and stylistic stratification.And even (-ing) and (dh-stopping) show different gender-by-class correlations (Labov 2001), with only (dh-stopping) showing gender crossing over with class.These differences lie in the fact that these variables are expressing slightly different things.Linguistic correlations with membership in macrosocial categories are not directly caused by these abstract structures, but are the result of behavior patterned by the constraints that this structure imposes on social life across the social order.The resulting patterns are a reflection of ideological differences both major and minor, ultimately local manifestations of global issues.The landscape perspective is not a replacement for the macro-social structure in the study of variation, but an elaboration of it -a perspective on that structure from the ground up.The macrosocial structure creates constraints on life at different places in the political economy, as does the physical environment, migration, sexuality, life stage.People in different places live differently, work differently, do different things, talk about different things.They have different perspectives on the world and to a great extent different needs and desires, all in response to material and other differences in the conditions of their joint everyday lives.And they express these differences in subtle and not-so-subtle stylistic patterns.
The acoustic frequency of /s/ is well known as a stereotype of sexuality -a high frequency (or fronted) variant is popularly known as the "gay lisp."In a study across California, Podesva and Van Hofwegen (2016) have shown that not only do women have fronter /s/ than men, and gay men have fronter /s/ than straight men, but in the rural north of California, town-oriented speakers have fronter /s/ than country-oriented speakers.A change is currently in progress, whereby younger country-oriented men are fronting their /s/, perhaps in a move away from hyper-masculinity.The sheer complexity of this pattern, and its relation to change is a clear indication that sound change moves through a social-semiotic landscape in which ideological connections -in this case among gender, sexuality, and rurality -motivate indexical construals and indexical moves.Hence, the mere finding that a change in progress does not correlate with the usual macrosocial categories is not an indication that it has no social meaning, for there are changes in progress that show a variety of social patterns related to local and global concerns.
Sound change emerges in multiple ways.We generally make a distinction between changes from below and from above the level of consciousness.In the former case, phonological pressures offer up changes, which are then recruited into ideological projects.In the latter, ideological projects go looking for linguistic material.This distinction between changes from above and from below is a useful one, but like the vernacular construct, it is too strict a dichotomy.Certainly the sound changes that we tend to be interested in are those that emerge from pressures in the phonological system.And these get called changes "from below."And we have changes that are commonly called changes "from above" -such as the insertion of rhoticity in New York or the reversal of (ay) centralization on Martha's Vineyard -which seem to come about by virtue of their potential for social meaning.But how about contact variables like (dh stopping), which originate in bilingual communities and become nativized?The belief in a strict conscious-unconscious divide is an unsubstantiated one, but it pervades the field, with change from below enjoying status by virtue of its purely language-structural origins, independent of social agency.This link between consciousness and agency makes many uncomfortable with the picture I'm presenting because stylistic practice involves agency, and there is a common belief that agency must be conscious.In fact, work in cognitive science (e.g.Smith & Kosslyn 2007) has made it clear that most of what we do, we do unconsciously.And if we never ask whether the use of a passive construction or dative alternation is the result of conscious agency, why would we ask that of the use of a sociolinguistic variable?Sociolinguistic variables are indexical signs (Peirce 1931-6), that is signs in which the relation between form and content is one of association in the world.An indexical sign points to, or indexes, something in the context, and a sociolinguistic variable indexes something about the speaker in the situation.While a sound change may begin as a perturbation in form, it becomes a sign as soon as it begins to take on meaning.And I would say that it takes on meaning as soon as it has a social distribution that can be construed in some way.This begins quite early in the change process: Tamminga (2016) has shown that hearers can be unconsciously sensitive to the social distributions of variables that show no stylistic variability -i.e.Labov's (1971) indicators.The process of construal begins with the initial social distribution of the sign, so while a sound change starts out indexing the population in which it originates, it may well be re-construed as indexing some associations with that population.This process of reconstrual gives rise to new orders of indexicality (Silverstein 2003), and competing re-construals may emerge to give rise to an ideologically structured array of potential meanings -an indexical field (Eckert 2008a).The role of sound changes in the stylistic system is no doubt not random, but structured at least by the locality of their origins, which in some sense predetermines the kinds of indexicality they are likely to take on (Eckert 2018).And sound changes that are part of a single larger process are likely to have similar meanings, as the meanings of earlier members of the process are likely to bleed into subsequent members.Thus we may find that adjacent members of a chain shift have similar meanings (Eckert & Labov 2017).
But as sound changes take on this indexicality, they combine with variables indexing other aspects of the speaker's self in the moment -stance, mood, desired qualities -the result of a process of bricolage.This process is not random (pace Guy & Hinskens 2016) but structured by the semiotic potential of different kinds of variables (Eckert 2018).
This should not be too much to accept when we recognize that language is not simply a referential system, but quite centrally an expressive one as well (e.g.Jakobson 1960).Human sociality requires that we be able to communicate some kinds of information nonsententially.And variability, particularly phonetic variability, provides non-referential material with which we can do this.The social meaningfulness of variation, in other words, is not an incidental or marginal fact, not a side-product of the spread of change, but a design feature of language.Sociolinguistic variables combine into speech styles, and these styles in turn combine with other semiotic systems (e.g.clothing, movement, demeanor) in the construction of personae.It is at this stylistic level, the level of the persona, that variation most clearly connects to the social world, playing out the details of identity and practice in the local spaces that populate the macrosocial categories that give structural sense to variation.
The social-semiotic landscape is an imagined array of social types, distinguished on the basis of social issues and grounding linguistic variability in ideology.It is through participation in this landscape that individual speakers produce and perceive -and accelerate -changes in progress.Speakers engage in stylistic practice to construct personae in the moment, based on their orientation to the range of possibilities offered in the landscape.Thus the adoption of a change is a performative act, not necessarily a conscious one, by which individuals resolve their immediate place in the social landscape.The landscape is structured by stylistic landmarks in the form of what Agha (2003) calls Characterological Figures.These figures are widely conventionalized stereotypes (e.g.Valley Girl, Hippy, Cholo) that articulate social distinctions specific to time and place.The speaker, a stylistic agent, parses figures in the landscape, relating linguistic differences to social differences.Thus change doesn't just "arrive" at a speaker; it is already there in the landscape, and the speaker adopts it if it is useful.
It is also at the stylistic level that variation plays a role in social change.For social change importantly involves changes in the array of personae that make up the social world, as hippies disappear and punks appear, as hobos give way to migrant workers, and as male lawyers, doctors, professors and politicians give way to female ones.Perhaps the most powerful demonstration of the relation between style and social change is Zhang's (2005;2017) study of the emergence of Yuppies in Beijing, as part of China's move into the global economy.Zhang documents the construction of a new cosmopolitan style of Beijing Mandarin among young managers in foreign-based financial companies.This style took on meaning in distinction from other Beijing styles, most centrally, the more conservative style of financial managers in state-owned businesses.But the stylistic moves that resulted in the cosmopolitan style involved linguistic choices that referred to a variety of other social types, such as "smooth operators", "alley saunterers", and Taiwan-Hong Kong capitalists.In other words, the Yuppies looked to the semiotic landscape around them to locate their choice of resources to construct a new and needed persona.And with their choices, they changed the landscape itself.For while one might say that this new speech style mirrored, or resulted from economic change, it also played a role in bringing this change about.Although stylistic practice is not the cause of social change, it is a necessary vehicle for change.The Yuppies' cosmopolitan speech style served as cultural capital for the foreign businesses, whose distinctiveness and prestige depended on a visibly and auditorily cosmopolitan workforce.And the Yuppies' elaborate lifestyles contributed to the general awareness of China's increasing economic divide, and drew consumerism into Chinese culture.
Landscape is a metaphor that pops up all over social science in the form of ethnoscapes, technoscapes, financescapes, mediascapes (Appadurai 1996), even selfscapes (Hollan 2014).It is a useful metaphor not only because spatializing social relations is a cognitive aid (Camp 2007), but because every landscape embodies a perspective.The maps of our dialect atlases provide a bird's eye view of the material facts, but meaning is enregistered in virtue of local perspectives.An individual's, or a community's, landscape is constructed in the interests of producing a local that embodies relations to the non-local (e.g.Johnstone et al 2006).Indeed, the places through which change spreads all have their own characters in the wider landscape, which contributes to the meanings of changes as they take off or pass through.Appadurai (1996: Chap. 9) speaks of localness as an active construction, which involves what the local is not, as well as what it is.One could say that a stylistic landscape is constructed in the service of the production of locality, combining resources that are available through direct everyday exposure with resources encountered at a greater social (and geographic) distance.Stylistic practice picks out and arranges elements to create a "here," in relation to diverse "theres," which will be constrained, but not determined, by physical and social locality.
It is not uncommon to think of sound changes as flowing along a network, arriving de novo in one community after another.But ethnographic evidence clearly indicates that people hear these changes coming.Studies of sound change have tended to focus on speech communities, with little attention to the wider and immediate surroundings.But the identity of every community is based in its relation to its surroundings, and this relation is key to the everyday lives and identities of community residents.A landscape is a perspective on one's surroundings, and one's social-semiotic landscape develops throughout life as one's time, place and social engagement expand.Every individual has a unique landscape, but what is individual is located within -and constructed in dialogue with -those who co-occupy the landscape.This process begins in infancy, and one's linguistic productive and receptive repertoire is a continual process of making one's way through the landscape.

Language development and the semiotic landscape
Language development from the very start is based in the individual's emerging socialsemiotic landscape, and the meaning of variation is both prior to, and continuous with, the development of the referential system.The association of language variability with distinctions in the social world is just a fundamental aspect of the constant search for patterns that underlies our linguistic competence.Children's earliest linguistic distinctions, no doubt, are in prosody and voice quality (Fernald 1989), as they come to associate patterns with caretakers' expressions of affect.This pattern recognition expands to individuals, and to age and role differences within the family, and beyond.Andersen's (1990) study of small children's role playing observes stylistic portrayals of character types emanating from family life (doctors, fathers, mothers) that attend to phonetic differences.No doubt kids develop early on a sensitivity to "baby talk," as not being a baby emerges as a major social concern for children.Inasmuch as age is a fundamental hierarchy, it is not surprising that features of this "baby talk" style emerge as important resources among adults for expressing intimacy as well as belittlement.So far, studies of small children's patterns of variation (e.g.Foulkes et al 2005;Smith et al 2009) have focused on parentchild interactions, and an understanding of how very small children's patterns change as their social worlds expand is yet to come.There is no question that quite early on, chil-dren become aware of status among peers, stemming not only from age but from what one might call coolness.
From the very earliest years, children are subject to what I have called (Eckert 2000) the developmental imperative -a desire to be older, to move on to the next stage.This is encouraged by adults up to a point, as they praise children for the development of new abilities, and even for physical maturation.It is also motivated by the desire to acquire new skills, freedoms, and prerogatives, and a desire for adventure.These specific desires may have their own places in the social landscape.I have often heard colleagues talk about sound change spreading by kids imitating or emulating older kids.There is no question that small children imitate older people, but imitation is considered childish, and among older kids it's definitely considered uncool.But even imitation is selective, done in virtue of what the imitator associates with the person -or the feature -they're imitating.Little girls put on high heels, lipstick and hats.Little boys swagger and carry big things around.They aren't trying to be adults but to be like adults in some way, to be older in virtue of what they value about being older, the rights or qualities they wish to get older for.And maturation involves developing a sense of what those things are and where they live out in the social landscape.In other words, maturity brings stylistic engagement not so much with individuals as with the landscape.As I approached each life stage, it was as important to me to differentiate myself from my parents and siblings as to emulate them.My differentiation of myself from elders was embedded in ongoing social changes, and all the personal practices that went with those changes made me strive to be a "more advanced" kind of person.One might see my having moved into a different social milieu than my older sister as just a result of different choices, but those particular choices had not been available to my sister eight years before.We are indeed quite different from each other, but some of those differences are a result of historical circumstance.We spent our adolescences in very different social-semiotic landscapes: Benny Goodman vs. Elvis.We do not just imitate our elders; we also move away from them, and we compete with them.In the process, I might emphasize, we participate in social change.
While small children are competent sociolinguistic actors, they take on a more dramatic role in change as they move into adolescence.There is a clear acceleration of sound change during preadolescence and adolescence, a pattern commonly called the "adolescent peak" (Tagliamonte & D'Arcy 2009) which, it is important to point out, is part of a more general acceleration of the use of other kinds of expressive, notably nonstandard, features (Holmes 1992;Eckert 2000).Sociolinguists have focused on the acceleration of sound change and nonstandard features in this period, but the role of adolescents in innovations in other kinds of expressive forms such as quotatives (Tagliamonte & D'Arcy 2004) and intensifiers (Tagliamonte 2008;Beltrama & Casasanto 2017) indicates the increased social intensity of this life stage.My ethnographic work with preadolescents and adolescents, among other things, follows the opening up of the social landscape and its concomitant structuring and elaboration of the semiotic landscape.

The preadolescent landscape
"We seem s -I-I feel myself as a different person I don't know why.Lots of different changes in my life.Wow.Now I'm beginning to call myself a life." The above was Selena's 1 insightful response to my asking how she felt being a sixth grader.Twelve years old, Selena and her peers had moved into the final year of their elementary school career.They were now the oldest in the school, and on their way 1 All names of research participants and schools are pseudonyms.
Eckert: The individual in the semiotic landscape Art.14, page 7 of 15 to being teenagers.My ethnographic study of preadolescents 2 was a study of the most intense developmental imperative, as the cohort found itself at the huge crossroad from childhood to adolescence.Some were reluctant to make this crossing, while others were eager, and these eager ones saw themselves as change agents, leading their cohort into adolescence, pioneering new practices and creating additional structure and perspective for the landscape.As suggested by Selena's reflection, the big move in preadolescence is the creation of an integrated social market, in which individuals come to see themselves as having social value as they come to commodify individuals, contacts, knowledge, language.(Agha 2011 provides a thorough treatment of the larger process of commodification that subsumes this activity.) Moving from childhood to adolescence is a move from small friendship groups to a peerbased social order.Constructing this social order involves drawing social control from adults into the cohort itself, a process that takes off seriously in the United States in late elementary school -in fifth and then big time in sixth grade.I followed this process in two elementary schools in San Jose, California, examining how this took place and how sociolinguistic variation figured in the process (see e.g.Eckert 2011).The two schools were in neighborhoods not five miles apart, but radically different in population.Fields Elementary served a predominantly white middle class area, while Steps served a predominantly poor and ethnically diverse one.
During fifth grade in both schools, small friendship groups merged to form a community of practice intent on leading their peers into adolescence.They constituted a popular crowd, based in the most salient teenage practice -creating a heterosexual market.The crowd engineered boy-girl pairs, and individual value on the market was based above all on whom one got paired with, and on one's role in negotiating the pairs.These pairings were statusful precisely because they were the work of the crowd, not of the individuals being paired up.Any boy and girl who decided on their own to pair up were not only illegitimate, but scorned.And indeed, the pairings were not so much about relationships between the members of a pair; they were often very brief, and purely transactional.As one girl at Fields said, talking about what couples did together, "Sometimes, you know, the boys won't even talk to you and stuff."And at Steps, Selena, finished up a narrative of the final episode of a long negotiation that ended in the boys sending her boyfriend over to do the deed: "He said 'do you want to be with me?' and I said 'Yes' and he said 'Okay' and then he just left." The crowd constituted a kind of going public -of moving from private lives onto a stage from which they and their activities could be seen by the larger elementary school public.These activities generated knowledge: knowing who was with whom, who were friends or on the outs, along with what to wear, how to move and how to talk became cultural capital in the emerging social order.A hierarchy emerged, at the top of which were people who got to be paired up, followed by those who were friends with them and who might also participate in deciding on the pairing, and were included in activities such as slam books, 3 and finally those who were not directly connected to the crowd but had access to the cultural capital that the crowd generated.All individuals found themselves placed in an integrated landscape, in which social status legitimated innovation.Within the school, the social landscape was also a physical landscape, inasmuch as the crowd dominated the 2 This work was funded by a Major Grant from the Spencer Foundation.central part of the playground, and the other kids arrayed themselves around the marginal spaces and in "child" spaces like the jungle gym.
The crowd at each school was somewhat ethnically heterogeneous, but the cultural capital that each traded in reflected the locally dominant ethnicity.The crowd at Fields was dominated by white middle class Anglos and dealt in mainstream privilege.The crowd at Steps was dominated by relatively poor Latinos and knowledge of gang culture and identification with the Norteños (see Mendoza-Denton 2008) were important cultural capital.The social landscape in each school was part of a larger landscape, and ethnic practice at each school was bound up with the other.The opposition between Anglo and Latino is a complex of embedded oppositions, what Irvine and Gal (2000) refer to as fractals, touching on every aspect of life from the neighborhood to the city, regional, state, and national level.The kids at Steps and Fields were embedded in a racialized landscape, they were aware of each other, and knew that they were going to find themselves in the same high school.A few Fields parents were planning to send their kids to private high schools for that reason.And while Steps and Fields went to different middle schools in seventh grade, they knew -anticipated -that their own middle school would bring them into a cohort with kids from schools very different from their own.While they are less than five miles apart, and few of the kids from either school had any direct contact with the other, they were critically aware of -and oriented to -each other as occupying places in a wider landscape in which ethnicity and relations of power are co-constructed.The occasional real -even distant -contact in malls, streets, events and other public places, strengthened and elaborated this landscape.Some of these contacts were more intimate.A prominent crowd member at Fields was a Chicana who lived in the Steps catchment area.But her parents, wanting to keep her away from the gang culture, had registered her at Fields under the address of a relative who lived in the Fields catchment area.And there were some from Fields and Steps who ended up for complicated reasons in the same middle school, making my fieldwork in middle school slightly more difficult, as kids from each of the two elementary schools were horrified that I would have anything to do with the kids who had gone to the other.
The activities in late elementary school are preparation for secondary school, where a cohort that has been together since kindergarten will emerge in a much wider landscape consisting of kids from about five elementary schools.At this point, one could say that the elementary cohorts find themselves in a more public environment, and the crowd members will have to compete with unknown people for space on the stage.Fields kids were preparing for this in their participation in institutions outside the school -sports teams and classes, cheerleading classes, and other activities that their parents supported.The kids at Fields were sharply aware of the gang culture in the larger society, at Steps, in high school.And tough kids at Fields, particularly boys, were said to be "gang."A Fields crowd member, Rachel, recounted a street encounter with a "gang" guy who knew her older brother.
"… he's all, "Hey, I know you."I'm all, "Oh, gosh, I wanna run, I wanna run."I almost like, I felt like I wanna cry so bad because he was near me.I thought like maybe, you know, he'd try to jump me or you know, cuz I was like really close to him.I was like this close to you.And, um, he's all, "I know you.You, your-your name's [Rachel]."I'm all, "Ye-, ye, yeah."Uh, cuz he knows my brother.He used to hang like, around my brother.Cuz my brother and his friends were like the cool kind of people." While Rachel claims fear in the face of this guy, she also recognizes the coolness of his place in the landscape.It certainly can't be ignored that coolness attaches to gang culture at Fields as well as at Steps.Among other things, it represents an extreme of autonomy, independence from adults and authority, that are valued elements of the developmental imperative.Meanwhile, kids at Steps were taking forays on their own into wider public spaces, such as malls, where gang affiliation could be acted out on a more public stage.Talking about encountering Sureñas at the movies, Bertha gleefully recounted her eagerness to engage: "They say "come on come on" "Ah -I don't want to kick your ugly ass okay so bye ho's" and I was screaming like this…I like like getting in fights I know.I'm all "you got a problem with me?" Then I'm all "Step up punks, bitches, ho's" and all that."These differences played out robustly in linguistic style.One day in class at Steps, a disliked white girl raised her hand and said something in a distinct California white girl speech style.Carlos turned to me and said "I hate her -why can't she talk normal?"This girl was disliked by Anglos and Chicanos alike, and Carlos took her use of aggressively white girl speech as a personal claim to white entitlement.The California Vowel Shift (CVS) has at least two versions.Anglos in California show a nasal split for trap, diphthongizing and raising the nucleus before nasals, and lowering and retracting it elsewhere.Latinos, on the other hand, lower and retract all occurrences of trap.Also, Anglo speakers front the nucleus of goose and goat, while Latino speakers tend to avoid this fronting.Crucially, though, these "ethnolectal" features constitute more general resources for the expression of more complex meanings than ethnicity (see Fought 1999;Eckert 2008b).Overall, kids at Fields front goose and goat more than kids at Steps and they have a greater nasal split than the kids at Steps.Within each school, the crowd leads their peers in the local pattern, so the crowd members at Fields show more goose and goat fronting as well as a greater nasal split than their peers.Meanwhile, at Steps, nobody fronts goose, but the crowd shows LESS goat fronting than their peers and less of a nasal split.There are non-crowd kids at Steps who identify as Latino every bit as intensely as members of the crowd, but who show some nasal split.And there are Anglo and Asian American crowd members at Steps who show none.Meanwhile, some white Anglo boys at Fields whom their peers characterize as "gang" show their toughness in the selective use of Chicano patterns (particularly light /l/).These sound changes, in other words, have been recruited locally in each school into a regime of coolness, in which ethnicity is salient but not the primary concern.

The adolescent landscape
By the time kids get to high school, the structure emerging from the preadolescent heterosexual market expands into a well-established peer-based social order.What follows is based on two years of ethnographic work in one high school in the Detroit suburbs, which I call Belten High, and shorter ethnographies in three schools in other parts of the suburban area.(See e.g.Eckert 1989;2000).While my preadolescent work took place in California in the nineties and my adolescent work in Michigan in the eighties, the two constitute a regular and enduring pattern of progression taking into account the significant local and regional differences.The kids at Fields and Steps functioned in a landscape that was located primarily in their own school, but with engagement with local social distinctions within a wider discourse, and with daring forays into the actual public scenes where this discourse played out.By the time they reach high school, kids become part of those public scenes.
In my ethnographic study of variation in the Detroit suburbs, 4 the major social distinction within each school was not ethnicity but class.The suburbs were a white socioeconomic continuum, as the predominantly African American city gave way precipitously to almost entirely white suburbs, where affluence rose as one moved outward from the edge of the city.The Detroit suburban population was to a great extent a result of white flight and social distinctions in the suburbs served among other things to erase race.The kids in the suburban schools had almost no contact with African Americans, viewing them as living in another universe.And indeed, they rarely went to Detroit proper, but to the almost entirely white towns that butted up against Detroit, commonly viewing them as part of Detroit.The socioeconomic continuum of the conurbation constituted the larger landscape of the suburban high schools, and the identity of each school was based on its place in the socioeconomic and urban continuum.This urban-suburban axis was embedded in each school, where there was an opposition between the school-oriented "Jocks" (not necessarily athletes) and the urban-oriented "Burnouts."The hegemony of this opposition was reflected in the fact that the rest of the student body referred to themselves as "In-Betweens," defining themselves in relation to the two polar categories.
The Jocks came primarily from the upper half of the local socioeconomic hierarchy and the Burnouts primarily from the lower half, constituting two class-based and ideologically opposed communities of practice.The opposition between Jocks and Burnouts yielded a recursivity that built social geography into each town -and into each school.The Jocks were a middle class community of practice based in the corporate culture of the school.They limited their main associations to members of their graduating class in their school, competing with other graduating classes and other schools, and forming a hierarchical social network based on roles in the school's extracurricular sphere.The Burnouts, on the other hand, rejected the school as a locus of social life and based their activities in the local area, and particularly urban parks and cruising strips.The kids in every high school were acutely aware of their place in the social geography of the conurbation.A Jock from an upper middle class family in one of the urban schools told me that she felt outclassed when she went to regional student government events, because Jocks at more affluent schools had more political and administrative experience.She was concerned about not being sufficiently knowledgeable to succeed at similar activities in college.It was also not unusual for a Jock moving from a more urban school to become a Burnout on arrival in a less urban one.One of the Burnouts at Belten High recounted this very experience, describing his arrival at Belten: "… all these short haired kids.My hair was long, it was really long, you know, and these people were, "well get your hair cut," you know.And they all had these Nike tennis shoes on.And that's what I remember.Nike tennis shoes.So I went home and said, "Mom, screw these Trax tennis shoes, I got to get some Nikes" you know."We're moving up in the world."So I had to get Nike tennis shoes like the rest of them.You know, that's about the thing they all dressed like way nicer than in Garden City.Garden City was strictly jeans and tee shirts, you know."Shoes were one aspect of clothing style, and the Jocks and Burnouts were easily distinguishable on the basis of their clothing, makeup, hair -and just about any other thing that could be incorporated into a style.Smoking was a key cultural sign for Burnouts, indexing their claim on adult prerogatives.And to cement this indexicality, one of the most popu-4 This work was funded by a pilot grant from the Spencer Foundation and NSF # BNS-8023291.lar extracurricular activities for Jocks was an anti-smoking committee.Burnouts indexed their urban and working class orientation by wearing Detroit jackets or better even, auto factory jackets; while the Jocks wore school, and where possible varsity, jackets.Perhaps most salient was the cut of jeans (Eckert 1980).Bell-bottomed jeans had given way to the more fashionable straight legs, but their association with the drug culture of the sixties and seventies made them still attractive to Burnouts.Furthermore, the Burnouts prided themselves on being the have-nots in the school, and saw keeping up with global fashion as a sign of Jock privilege.The width of the bottoms of one's jeans, therefore, was central to the semiotic landscape.Quantification of jean styles during lunch hours showed the average width of jean bottoms increasingly gradually and significantly from the jock territory outside the cafeteria to the burnout territory in the courtyard.The jean width semiotic was extended to suburban geography as one Belten student, describing a predominantly working class, hence "Burnout" neighboring school, spread her arms wide and said "that school has bells this wide." Most salient in the urban-suburban continuum was the view of urban kids as more independent, street-wise, and tough.Burnouts were eager for contact with more urban kids, and sought out and welcomed kids moving out from Detroit.A girl who moved from Detroit in elementary school found the kids in Westtown sheltered: "I'd -you know, like I used to tell [my friends] that I used to go out, you know, and walk, you know, across Seven Mile and everything.And they couldn't even cross the street and stuff, and like I'm crossing these big main streets, and, you know …. it was like we were years ahead of these people, it seemed… we were much more sophisticated because we, you know, we were all into all this stuff -like we, we had the weirdest ideas.Now this is when we were like little kids, like ten years old.We would sit around and talk about sex and everything, you know …" Across the schools of the suburban area, Jock and Burnout styles were highly elaborated, ranging from territory to eating choices to language -and contained elements associated with a range of terms that differentiated them.Their speech played the same role, ranging from the kinds of speech activities they engaged in to their vowels.The Detroit area is in the center of the Northern Cities, and the white dialect is characterized by the Northern Cities Vowel Shift, a clockwise rotation of the mid and low vowels (Labov, Yaeger & Steiner 1972;Eckert 2000).The older components of the shift, the raising of trap and fronting of lot and thought, show little geographic variability across the conurbation and seem to have gone to completion.More recent and still ongoing are the backing of strut and dress.In addition, while not strictly part of the chain, the nucleus of price is raising.Comparison of speech in schools at the urban edge with schools deeper in the suburbs shows that these more recent changes are currently spreading outward from the urban periphery.The urban-oriented Burnouts at Belten High led in the use of all of these newer changes.The older changes, which were no longer on the move, correlated primarily with gender (with girls leading in use).And the intermediate change, the fronting of thought, showed an intermediate distribution, led both by girls and by Burnouts.Although the Jocks and Burnouts came from predominantly middle class and working class homes respectively, there was enough class mobility to establish that these patterns of variation did not correlate with parents' socioeconomic status or with the neighborhood the speaker grew up in.All of the recent changes correlated with the speakers' participation in the peer social order.In other words, these variables had taken on social meaning, and were used as stylistic resources in the construction of the adolescent social order.While these late sound changes correlate with social category, one cannot say that they are simply markers of category affiliation.A regression that includes the in-Betweens shows that the urban sound changes correlate robustly with the practice of urban cruising, overwhelming the significance of social category.This suggests that these sound changes are associated with particular activities, stances or characteristics independent of categories.The Jocks and Burnouts constitute the two extremes in the school's culture, and just as the rest of the school's population falls between them ideologically, there are differences within each of the two categories.This is particularly clear among the Burnout girls, who fall into two network clusters joined only loosely by a couple of neighborhood ties.The smaller of the two clusters, sometimes referred to as the "Burned-out Burnouts," pride themselves on being wilder than their peers, on being, in their own words, the "biggest Burnouts."The majority of the Burnout girls may occasionally get in trouble for smoking weed or skipping school, but they identify as Burnouts primarily on ideological grounds rather than from a desire to be wild.The Burned-out Burnouts make particularly dramatic use of the urban changes, standing out significantly from the other Burnouts who, in turn, lead Jocks and In-Betweens in the use of urban variants.This is completely consonant not only with the extremity of their actions and activities, but with their stance in relation to the school and their fellow students, and their consumption patterns more generallyfrom their hair, makeup and clothing to their drug consumption.
In other words, the social world of the school constitutes a social-semiotic landscape that is in turn placed within the wider landscape of the Detroit conurbation.The current changes in the Northern Cities Shift spread outward from the urban periphery as urbanoriented suburban kids construct personae consonant with the qualities that they associate with urban life.

Conclusion
It is common to acknowledge the social meaning of variation but to treat it as epiphenomenal.What I propose here is that while the origin of a sound change may be purely phonological, it must take on meaning to spread.And that meaning emerges as part of a semiotic system that expresses everything from macro-social membership to affect -a system that is essential to human society, hence to language.Stylistic practice involves a variety of linguistic variables that are not changes in progress, but sound changes inevitably become sociolinguistic variables.Any change will become part of a stylistic construction, and it will be by virtue of its role in this construction that it spreads through the landscape.One could say that every individual is always playing a role in sound change simply by participating in the semiotic landscape.But the question is whether any particular individual can have a greater influence than any other or whether the spread of change is a completely collective endeavor.Those who have been found to be leaders in sound change (e.g.Labov 2001) -women with moral authority in their neighborhood and with diverse contacts beyond -do not determine the direction of change, but accelerate it in their extreme stylistic activity.But they may well also play an important role in the social construal of a change, inasmuch as their use of that change is likely to be more noticed and carefully construed, and by a more diverse and mobile audience.But any construal requires legibility, and no individual innovation will catch on unless it is already legible -unless it is skillfully used with reference to the existing indexical system.Needless to say, those who are less sensitive to social or linguistic nuance will play a minimal role in the structuring of the landscape, and the leaders in change will be those who are both sensitive to the landscape and sufficiently influential to motivate others.But their influence itself will lie in their ability to construe the landscape and to innovate in a way that is already meaningful to those around them.The perspective I am offering is not substantially different from Labov's perspective, except inasmuch as it focuses on the collective construction of meaning that underlies the larger patterns of community norms.I acknowledge, furthermore, that the claim that sound change spreads by virtue of social meaning is a somewhat radical hypothesis, but presumably a testable one that merits serious exploration.

3A
slam book is initiated by writing a question (e.g."What's your favorite body part?") in a notebook which then gets passed around for individuals to write their responses.Both access to the slam book (i.e. the invitation to write in it) and knowledge of people's responses are important cultural capital.