The first sociolinguistic study in San Francisco was DeCamp’s (1953) dialectological analysis that formed part of the Linguistics Atlas of the Pacific Coast (Reed and Metcalf 1979). Out of 25 speakers in the total sample, DeCamp included three speakers labeled ‘Jewish’ and three labeled ‘Negro’, one for each of three ‘educational types’ (less than high school, high school, and more than high school), the other 18 speakers labeled ‘white’. Rather than comprising a sample that was statistically representative of the population’s ethnic distribution, DeCamp included these speakers in the linguistic atlas tradition, which approximated representativeness through the inclusion of token speakers. Based on this sample, and the analysis of a wide range of linguistic variables typical of linguistic atlas work of the time (e.g., the entire vowel system, the entire consonant system, and several morphosyntactic features), DeCamp noted a linguistic pattern in line with the migration trends of the time, namely, that the African American speakers were more likely to exhibit speech patterns similar to the South or South Midland dialect areas, in contrast to the European American (and Jewish) speakers, who displayed patterns of linguistic production that better approximated northern dialects. DeCamp’s study was conducted after the end of World War II, but prior to any major change in San Francisco’s ethnic distribution, and his speaker sample, in part, reflects this. For example, Chinese Americans were not included in DeCamp’s study, which was a descriptive account of English dialect variation, because despite generations of residency in the United States, many of Chinese San Franciscans would not have been English dominant. Although only including participants who were native to the city of San Francisco, DeCamp’s analysis sought to draw connections between particular speakers’ pronunciations and dialect areas to the East, rather than to describe particularly Californian or San Franciscan ways of speaking.
The second variationist study of San Francisco English by Hinton et al. (1987), 35 years after DeCamp (1953), focused on recruiting young speakers from across the Bay Area, in an effort to analyze the most advanced forms of new sound changes they were observing across California and in popular, California-based media: the fronting of the goat and goose vowels (where those vowels are pronounced with the tongue body further front in the mouth), the lowering of the kit and dress vowels (where the tongue body is lower), the lowering and backing of the trap vowel, and the merger of the lot and thought vowels (where those vowels are pronounced in the same way).6 As a result, their analysis focused more on speakers’ age as a relevant sociolinguistic variable, rather than speakers’ ethnicity. The non-white participants who appeared in their final sample consisted of four African Americans, two Asian Americans (where ‘Asian’ was not specified, although one of the two was of Singaporean descent), and one Latina, out of 22 total speakers (so, in combination with 15 white speakers). Hinton et al. (1987), found that, overall, the two Asian Americans matched the vowel production patterns of the 15 European Americans, whereas the other five speakers of color were discussed as not producing these sound changes, with production patterns more similar to the older white speakers than the white speakers of their same age cohort. That said, the data (unsurprisingly) show quite a bit of variation between the individual speakers, and the numbers are simply too small to make any conclusions from this study with respect to ethnicity. At the time of the study, in the mid-1980s, demographic shift in San Francisco with respect to ethnicity was proceeding rapidly, and adequate coverage of ethnic variation with respect to linguistic variation would have been a herculean task.
The third analysis of San Francisco English, by Moonwomon (1992), focused specifically on the low vowel system, analyzing phonetic shift in the vowels of trap, lot, and thought. This study, however, eliminated ethnicity as part of the analysis through its methodological design. By confining her speaker sample to white women, Moonwomon investigated research questions particularly relevant to that demographic, considering variation in socioeconomic class as it is realized between speakers of comparable sex and ethnicity. At the same time, Moonwomon’s choice to define her participant sample in this way implied that ethnic variation (and gender or sex variation as well) has such a potentially strong impact on the analysis of San Francisco English that it was beyond the scope of her study. Her analysis did confirm that the sound changes found among younger speakers in the Hinton et al. (1987) study (of which she was a part) were in progress within this sample, specifically that trap was raising and fronting before nasals, lowering and backing elsewhere, and that lot and thought were moving toward merger. Furthermore, Moonwomon found differences between middle-class and working class women with respect to attitudes toward ethnic shift in San Francisco, with some of the working class participants expressing greater resistance to the increased Asian and specifically Chinese presence in their neighborhood. However, Moonwomon did not draw specific connections between the low vowel production patterns of these women and their orientations toward ethnicity, although she did find more advanced (i.e., more raised before nasals) productions of the trap vowel for the middle-class speakers. As Bucholtz’ (2007, inter alia) work in a Bay Area high school suggests, discourse about race often stands in for discourse about class, and analyses of class ideologies in the San Francisco area are inseparable from the local construction of ethnicity, particularly since the early 1990s.
Linguistic research across the San Francisco Bay Area, outside the city itself, has been much more actively engaged in investigating the role of ethnicity and ethnic identities, particularly since the 1990s. Rickford (1992, inter alia) analyzed morphosyntactic variation among African Americans from East Palo Alto, comparing usage patterns of several variables across three age groups, and contrasting the results with those found in studies among African Americans in other parts of the United States. Rickford and McNair-Knox (1994) and Alim (2004) have also analyzed style-shifting with respect to morphosyntactic variation in African American English in the Bay Area, joining a larger canon of research on language and ethnicity that argues for the crucial role of an interlocutor’s race and ethnic identity in affecting patterns of linguistic production. With respect to African American English, in particular, the role of ethnicity in linguistic production in the San Francisco Bay Area became a national media firestorm in 1996 with the passing of the Oakland School Board ‘Ebonics’ resolution and the resulting media controversy, an event addressed by Rickford (1997) and Rickford and Rickford (2000), among others.
Bucholtz’s (1999, 2001, 2004, 2007) work in another part of the Bay Area analyzed the role of language in identity construction among students at an ethnically diverse high school. She found that several linguistic variables, often considered to be features of African American English (including various phonological, syntactic, lexical, and discourse features), were employed by whites (Bucholtz 1999) and South Asian Americans (Bucholtz 2004) in the construction of their own ethnic identities, and were specifically avoided by others in the construction of other, for example, ‘nerd’, identities (Bucholtz 2001). The nerd identity, in particular, was in turn represented as ‘hyperwhite’ through the employment of linguistic resources marked as ‘superstandard’ (e.g., ‘lexical formality, carefully articulated phonological forms, and prescriptively standard grammar’; Bucholtz 2001:88). Bucholtz’s work is perfectly situated in the ethnic landscape of the present-day San Francisco Bay Area, in that it highlights the fluid construction of ethnic meaning, which, although rooted in a speaker’s ethnic identity, is highly negotiated in a community through linguistic and other social practices.
Similarly, Shankar’s (2008) analysis of South Asian American teens in Silicon Valley (located directly south of San Francisco) argues for the creation of ethnically marked, ‘FOB styles’ comprised of linguistic features from ‘Punjabi, Desi Accented English, California slang, and hip-hop lexicon’ (268). These linguistic styles encode ethnic identity and socioeconomic class membership, and Shankar argues that middle-class, Silicon Valley language ideologies are complexly linked to linguistic features that indirectly index any number of multiple ethnic meanings at any given time (South Asian, African American, and Mexican American among them), whereas upper middle-class speakers avoid all such ethnically marked linguistic strategies. Shankar’s study reflects that important shift in the social history of the San Francisco Bay Area since World War II whereby non-white ethnic groups, especially Asians, have become much more socioeconomically stratified.
Mendoza-Denton (1997, 2007) examine the production of phonetic variables among Chicana teenagers in another San Francisco suburb. She found that the raising of the kit vowel before velar nasals was produced by those speakers with the greatest investment in the creation and maintenance of a Latina identity. Mendoza-Denton and Iwai (1993) and Mendoza-Denton (1995) analyzed Asian ethnicity in the San Francisco Bay area by comparing four Japanese Americans and four white Americans. Their results showed that the pronunciation of the vowel in face and the vowel in goat differed between Japanese Americans and whites, as well as between younger and older speakers of both groups. Their analysis for the Japanese Americans, in particular, argued that experience with the World War II internment camps marked a crucial turning point in Japanese American language ideologies, and highlighted the important intersection between ethnicity and immigrant generation with respect to linguistic practice.
Eckert (2008a) has also examined the use of linguistic resources in the construction of identity among Anglos and Chicanos in the San Francisco Bay Area, this time in San Jose (the focal city in the ‘South Bay’) and among younger adolescents, at two different schools. Her analysis of the trap vowel shows that speakers produce their vowels in relation to the particular crowds that define the social order at the two respective schools. Although one school’s crowd orients to the raising of the trap vowel before nasals, which is the Anglo norm, the other school’s crowd orients to the Chicano norm, where productions of the trap vowel are not conditioned by the following consonant. Crucially, both Anglo and Chicano kids from both schools can participate in both these ethnically correlated vowel patterns, because rather than directly indexing ethnicity, the variable production of trap indirectly indexes a field of available social meanings particular to each community. Eckert argues that kids of all ethnicities actively construct the meanings that are tied to ethnicity, so that the relation between variable production and ethnicity emerges through kids’ construction of their peer-based social order. Thus, for any multiethnic context, an informed analysis of patterns of phonetic variation benefits from viewing each moment of linguistic production as indirectly indexical of a field of social meanings that are linked to ethnicity through iterative practice (see Eckert 2008b).
The only known research within the city of San Francisco (as opposed to locations around the Bay Area, more generally) that explicitly considers ethnicity as a factor intersecting English variation and change is Hall-Lew (2009). This, along with Ong (1993), Williams (2006), and Starr (to appear), is also one of the few sociolinguistic studies of Chinese Americans in the San Francisco Bay Area, despite the group’s importance in shaping the local social history. Hall-Lew (2009) analyzed variable production of the vowels in lot, thought, goat, and goose across a sample of 30 residents of one of the city’s New Chinatowns (the Sunset District) – 16 Asian Americans (mostly of Chinese descent) and 14 European Americans. The study found statistically equal production of sound changes in progress across both groups, and no difference between them according to ethnicity. For both Asian and European Americans, the vowels of lot and thought were more likely to be merged among younger than older speakers, and the vowels in goat and goose were more likely to be fronted for younger and older speakers. This finding, which supports that observed by Hinton et al. (1987), contrasts with separate claims within sociolinguistics that regional sound changes, such as these, are more robust among white than among non-white speakers (Labov 2001). Based on an ethnographic analysis of language attitudes in the neighborhood, Hall-Lew (2009) further argued that, on a whole, older San Franciscans orient to a variety of English associated with East Coast urban centers and particular European American migration histories, whereas younger San Franciscans participate in western United States and specifically Californian social practices and ways of speaking which indirectly index ethnicity and specifically Asian American identities. In San Francisco, Chinese identities and practices have a particular historical prominence and present-day influence on the city’s ethnic landscape – Chinese practices are part of local practices. The linguistic markets in present-day San Francisco have thus emerged as a consequence of the city’s social history, in this case, the long-standing, indeed, foundational presence of Chinese ethnicity and culture, and the socioeconomic diversification of the Chinese American community, has made the acquisition of local authenticity possible. In this way, the social history of the city has important implications for understanding the present-day language attitudes, language ideologies, and patterns of variation and change in San Francisco English.