2. The Dynamics of Socially Indexed Linguistic Variables
The capacity of speakers to employ social meaning in the construction of identities, stances, and personas relies on two fundamental pillars. Firstly, there is the necessity for the linguistic variable to index social meaning, thereby implying the social category along with any other semantic contributions (e.g., context-free truth-conditional meaning). Secondly, individuals must, across a speech community of any given size, possess an awareness of the indexed social meaning for the variable to successfully serve as a tool for identity, persona, or stance construction. If either of these two fundamental pillars is problematic or non-existent, individuals could still produce the form as a result of imitative social conditioning, but the intended social information would be opaque. For instance, consider the English quotative
be like. This variable has been shown to correlate with social categories, such as gender, age, and socioeconomic status (
Dailey-O’Cain 2000). If these categories are indeed socially indexed onto the variable in addition to its quotative meaning, it implies that individuals possess an awareness of this additional meaning, be it unconscious or conscious knowledge. Explicit and conscious knowledge would suggest that speakers can utilize
be like to construct specific identities related to the indexed social meaning. However, such sociolinguistic performance becomes ineffective if individuals lack awareness of the potential indexed meaning. In
Buchstaller (
2006), British listeners were shown to be able to determine the age and gender of a speaker from quotative
be like use, suggesting that individuals in Britain employed
be like to construct identities in relation to age and gender. However, while socioeconomic status was also shown to be a correlating social category with the variable in production, listeners were unable to deduce this meaning from quotative
be like use, at least not in an overt manner, and were thus unable to employ
be like to construct identities in relation to socioeconomic status.
The process of indexicalization is consistent with usage-based approaches to language learning, notably exemplar models. Within this framework, exemplar models propose that individual speech utterances accumulate in memory as exemplar representations, incorporating comprehensive linguistic and non-linguistic information (
Bybee 2001;
Foulkes and Docherty 2006;
Goldinger 1997,
1998;
Johnson 1997,
2006;
Pierrehumbert 2001,
2002). This aggregation allows for the mapping of relevant social categories onto each exemplar, which has been observed in studies of psychology and, more recently, those in the field of linguistic inquiry (
Drager 2005;
Foulkes and Docherty 2006;
Hay et al. 2006a;
Johnson et al. 1999). Exemplar representations are theorized to be stored in an individual’s mind and can be accessed during both speech production and perception, a phenomenon which has also been supported by linguistic research (
Hay et al. 2006a,
2006b;
Johnson 1997;
Pierrehumbert 2001). Thus, according to usage-based approaches to language learning, speakers can produce forms that index corresponding social categories and perceive the associated social categories. Returning to the example given above of Orthodox Jewish men’s released /t/ patterns, exposure to such correlations in speech production creates a mental representation of the released /t/ and its associated social categories: religion, education, and gender. In accordance with exemplar theory, it is expected that speakers could then employ this knowledge as a stylistic device to craft a specific social persona.
Experiments in regional dialect labeling have provided evidence to demonstrate individuals’ overt awareness of social categories that correlate with variables (
Baker et al. 2009;
Cramer 2011;
Fuchs 2015;
Kirtley 2011;
Purnell et al. 1999;
Suárez-Budenbender 2009).
Clopper and Pisoni (
2004) focused on the ability of Indiana college students to accurately categorize six North American regional dialects. Although general identification accuracy among listeners was relatively low, the participants’ responses surpassed chance levels statistically. Notably, speakers who had lived in at least three different states exhibited higher accuracy compared to those limited to Indiana. Moreover, individuals who had lived in a particular region demonstrated a more precise categorization of talkers from that region than those without such residency experience. Furthermore, social evaluation studies have shown that listeners possess a capacity to recover social meanings from linguistic variables (
Campbell-Kibler 2007,
2008,
2011). Campbell-Kibler specifically investigated the impact of the sociolinguistic variable (ING) (e.g.,
walkin’ vs.
walking) on listeners’ attitudes towards speakers by manipulating the realization of the final nasals in (ING). This manipulation influenced listeners’ judgments about the speaker, though results diverged from prior studies on the social stratification of (ING). While Campbell-Kibler identified the associated social categories, including education and intelligence, previous research also linked (ING) to gender, socioeconomic status, dialect, age, and race (
Fischer 1958;
W. Labov 1966;
Shopen 1978;
Shuy et al. 1967;
Trudgill 1974). In addition to (ING), this inconsistency between production patterns and individuals’ awareness extends to other linguistic variables such as
t/
d deletion in English (
Baugh 1979;
Campbell-Kibler 2005;
G. R. Guy and Boyd 1990;
W. Labov 1972c;
Rickford 1986;
Staum Casasanto 2010;
Wolfram 1969), quotative and focuser
like (
Buchstaller 2006;
Dailey-O’Cain 2000), fundamental frequency (
Kirtley 2011;
Linville 1998;
Smyth et al. 2003), and /ay/ monophthongization (
Kirtley 2011;
Plichta and Preston 2015;
Rahman 2008). It is this inconsistency that prompts questions about why individuals exhibit awareness of the association between some linguistic variants and social categories but not all.
The context of a variable has been investigated as an explanation for individuals’ inability to overtly deduce social meaning that is expected to be indexed onto a variable. Speech, being inherently social, unfolds between speakers and interlocutors, whose interpretations of meaning hinge on their experiences, social positions, and goals. Campbell-Kibler’s research (
Campbell-Kibler 2007,
2008,
2011) exemplifies this, demonstrating that listeners rate speakers who use the alveolar variant of (ING) as compassionate when they perceive the speaker to be Southern but condescending when they perceive the speaker to be from elsewhere. Social information about the speaker also influences how listeners perceive the speech of an individual (
Hay et al. 2006a,
2006b;
Hay and Drager 2010;
Koops et al. 2008;
Niedzielski 1999;
Strand 1999). In
Hay and Drager (
2010), New Zealand English speakers were exposed to either stuffed toys associated with Australia (kangaroos and koalas) or toys associated with New Zealand (stuffed kiwis) during a vowel perception task. Participants showed a shift in their perception boundaries according to which set of toys they were exposed to, i.e., participants matched natural vowels with more Australian-like synthesized vowels when they were in the Australian “kangaroo” condition. Recently,
Sherwood et al. (
2023) furthered the examination of context as a factor which influences individuals’ perception of speech and indexed social meaning to the situational context, i.e., the place where the utterance was spoken. While they found evidence to support that individuals form associations between the gender of a speaker and prescriptive variables in Japanese speech communities, knowledge of the context of the variables had no significant effect on individuals’ judgements.
Crucially, however, one of the most striking findings in the
Sherwood et al. (
2023) study was the ease by which prescriptive forms in the speech community were perceived by participants. This finding is in line with
W. Labov’s (
1972b) proposed model of social salience which delineates three variable types, demarcated by speakers’ awareness of their existence. The first level are
indicators, which show zero degree of social awareness and are therefore difficult to detect for both linguists and native speakers.
Markers are usually socially stigmatized forms characterized by sharp social stratification across groups and styles. The highest level of social awareness for variables is the
stereotype category.
Stereotyped forms display both social and stylistic stratification and are subject to explicit meta-commentary due to their overt level of social awareness in the speech community. The salience of a variable in the speech community therefore offers a clear explanation for why a certain variable’s expected social meaning may not be detectible by listeners. That is, if the variable is non-salient at the
indicator level, its associated social meaning(s) may not be learned by the listener. What remains to be investigated along this line of inquiry, and what was not within the scope of
Sherwood et al. (
2023), is the extent to which stereotyped attitudes and beliefs about a certain variable may influence listeners’ evaluative judgements.
In
Levon (
2014), stereotyped attitudes and beliefs about groups of speakers were examined as factors that could influence listeners’ evaluative judgements. The attitudinal and cognitive factors targeted in this study were in reference to listener endorsements of normative stereotypes pertaining to male gender roles. Listeners who endorsed normative stereotypes of masculinity and male gender roles used pitch and sibilance as salient cues which signaled ”nonmasculinity” and ”gayness”; however, listeners who did not identify with these stereotypes showed no effect for pitch and sibilance. This finding suggests that when individuals desire to conform to socially accepted norms, they are more likely to interpret linguistic features in ways that align with these norms, thereby reinforcing stereotypes. Individuals have also been shown to be aware of stereotyped linguistic features to the extent that they claim to use these features differently than they actually do in their speech. In
W. Labov (
1966), speakers from New York showed a tendency to report a higher usage of standardized forms than their actual usage. The opposite effect was observed by
Trudgill (
1972), where men from Norwich reported a higher use of non-standardized forms than their actual usage. Given the findings pertaining to social stereotyping, both of groups of speakers and of variables themselves, the role of social desirability presents an interesting line of enquiry. Individuals’ endorsement or positioning towards a linguistic variable, where some may identify as users of the variable and others may not, could offer an explanation as to why some correlations between variants and categories are detectable by speaker–listeners, and thus are available for use for the purpose of identity construction, while other correlations are not. Therefore, this study aims to determine how individuals’ positioning to a variable mediates the awareness of social meaning within a language community where not only the correlating social categories are salient but also the linguistic variable itself.
3. Life-Stage, Gender, and Yeah-No
To investigate the role of social positioning on judgements of social meanings, our study design incorporated specific constraints essential for hypothesis testing. The first constraint was that, to gauge an individual’s positioning effectively with respect to a variable, it is imperative that the variable in question holds salience within the community, either as a stereotype or a socially significant marker. Consequently, we opted for the marked and stereotyped discourse marker in Australian English, yeah-no, as our chosen variable for examination. The second constraint pertains to the social categories in focus. While it is acknowledged that linguistic variables can index multiple complex, dynamic, and contextually dependent social categories, we deliberately limited our analysis to two potentially indexed meanings of the Australian English discourse marker: age and gender. This study’s emphasis was on exploring the role of social positioning rather than delving into the subtle nuances of the chosen variable. Thus, we underscore that the chosen design and methods align with the research aim’s scope and subsequently encourage future investigations to delve deeper into this line of inquiry by exploring yeah-no and other variables with consideration for styles and their respective indexical fields.
As continuous variation in the phonetic realization of vowel allophones is the most heavily employed resource for the construction of social meaning in English (
Eckert and Labov 2017), it is no surprise that the majority of the research that examines the communication of social identities has largely focused on vowel allophones. There are, of course, many studies which examine variables of different linguistic levels and their correlated social meaning(s), for example, but not limited to, consonant allophones (
Benor 2001,
2004;
Campbell-Kibler 2007,
2008,
2011;
Podesva et al. 2015), quotatives (
Buchstaller 2006;
Dailey-O’Cain 2000), intensifiers (
Bauer and Bauer 2002;
Stenström 1999;
Stenström et al. 2002;
Tagliamonte 2005), and discourse markers (
Andersen 2001;
Erman 1997,
2001;
Macaulay 2002;
Tagliamonte 2005). This paper contributes to the investigation of social meaning by examining the understudied discourse marker
yeah-no in Australian English.
To date,
yeah-no, like other discourse features, has received very little attention in linguistic research, despite its salient reputation in the speech community. Labeled as “speech junk” (
Campbell 2004) and a “verbal crutch - an epidemic from which no strata of society is immune” (
The Age 2004),
yeah-no had a very negative reputation in the early 2000s. More recently, however, it has received more positive attention as the punchline of a road safety campaign in which the advertisements would conclude with “Say ‘Yeah… NAH’ to taking risks” (
Kelly 2018), and the variable was considered to be the ABC’s second greatest Aussie slang word or phrase in a 2021 listener poll (
Hughes et al. 2021). In studies that have sought to better understand the variable in terms of its linguistic distribution,
yeah-no has been shown to serve a number of functions in speech, including discourse cohesion, the pragmatic functions of hedging and face-saving, and assent and dissent (
Burridge and Florey 2002). Burridge and Florey performed a corpus analysis of formal conversation and interviews, examining the interplay between intonation and turn-taking, along with the usage of
yeah-no in relation to topic, conversational genre, and the age and gender of the speaker. In their study, Burridge and Florey also examined the variant forms of
yeah-no which include
yep-no,
yes-no, and
yeah-nah, and occurrences of
yeah-no with additional markers of agreement—for example,
yeah yeah yep-no. The stratification of their results demonstrated a higher frequency of
yeah-no use in speakers between the ages of 18–49 years of age (25% of speakers produced the variable), with a slight preference for the 35–49 age range (25.6%) compared to the 18–34 range (23.5%). In terms of gender, Burridge and Florey speculated that, given the interactive politeness phenomena of
yeah-no, a gender difference would be apparent in the distribution of the results. However, no differences were apparent at the conclusion of the study. It is also worth noting here that the size of the corpus in the Burridge and Florey study was limited with respect to investigating a potential age stratification. Specifically, there was a difference in
yeah-no use between formal and informal interactions among younger speakers (18–34 years); however, only five tokens of
yeah-no were produced and only in a television and film setting.
Moore’s (
2007) study, incorporating data from radio and television broadcasts, the Australian International Corpus of English corpus, and the Monash University Australian English Corpus, showed similar findings, particularly regarding the influence of age. Notably, however, a higher frequency of
yeah-no usage was observed among male speakers, constituting 85% of tokens, compared to female speakers. While other social categories have not yet been explored in relation to
yeah-no, both Burridge and Florey and Moore speculate that a potential socioeconomic and style stratification may be present in the distribution of
yeah-no. Ultimately, due to its salience in the speech community, either as a stereotype or a socially significant marker,
yeah-no satisfies the first constraint required by this study. The results of
Burridge and Florey (
2002) and Moore likewise demonstrate that
yeah-no has the potential to socially index the categories of age and gender due to their correlations with the variable in natural spoken speech, satisfying the second constraint employed in the study.
Wolfram’s study of African American Vernacular English in Detroit, specifically, found that the adolescent age group (14–17 years) in the study used fewer phonological, morphological, and syntactic variables than both the preadolescent group (10–12 years) and the adult group. Across the eleven variables that were examined according to age group and socioeconomic class, word-final
t and
d, 3rd singular -
z, and double negatives were all examples of variables that did not show social stratification in the adolescent group compared to the preadolescent and adult groups. This discrepancy has been attributed to socioeconomic mobility, with
Eckert (
1988) suggesting that the lack of variables used in the adolescent sample reflects a relatively smaller relevance of parental socioeconomic class to adolescent social identity. Therefore, the use of a continuous analysis could present problems for participants, as they may struggle to distinguish between seemingly arbitrary landmarks according to age in years. Age as a categorical variable, however, offers an interpretive lens that can both aid participants in making their judgements and provide normative landmarks which may contribute towards a better understanding of other social categories indexed by
yeah-no, such as the predicted socioeconomic and style stratification correlates proposed by
Burridge and Florey (
2002) and
Moore (
2007).
Gender, like age and life-stage, has also been widely researched in the domain of sociolinguistics. We would like to note the distinction here between gender, a constructed ideology that depends on perception, and sex which is a biological category. Both have been examined in the variationist literature; however, while some studies specifically distinguish between biological characteristics and social factors (e.g.,
Chambers 1992,
1995), the majority often have an overlap between the two. Where possible, we distinguish between ”sex” when discussing research that relies on a simplistic classification of speakers into males and females, and ”gender” when describing research that takes at least some account of relevant social and cultural factors. Gender, unlike sex, abstracts over a range of globally and locally constructed speaker–listener practices (
Eckert and Labov 2017) and is claimed to be as impactful to the constructions of identity as the dimensions of region and age (
Podesva and Kajino 2014). In one of the foundational studies to examine variations with speaker sex in production,
Fischer (
1958) found that girls consistently used more of the perceived standard form of the (ING) variable [ɪŋ] than boys, a pattern that was later discussed by
W. Labov (
2001) as a preference for women to use more standard variants than men. In addition to prestige, a number of sociolinguistic variables have been studied in connection with sex and gender, for example, the Northern Cities Chain Shift (
Eckert 1989), high-rising terminals in Australian English (
G. Guy et al. 1986) and in New Zealand English (
Britain 1992), and glottal stops in British English (
Milroy et al. 1994). Sex and gender have also been studied in other languages within a sociolinguistic framework. Some examples include phonological, morphological, and lexical differences between male and female speakers of Koasati (
Haas 1944), monophthongs and diphthongs in the speech of women from Tunis (
Trabelsi 1991), and patterns of non-palatized [l] in Crete (
Mansfield and Trudgill 1994). Listeners’ evaluative judgements of speech with reference to sex and gender have also been investigated (
Owren et al. 2007;
Traunmüller 1997;
Whiteside 1998), as has the perception of sexual orientation (
Clopper et al. 2006;
Levon 2007;
Smyth et al. 2003;
Squires 2011). For example,
Smyth et al. (
2003) examined listener judgements pertaining to gender for varying discourse types and associated stylistic features. The results showed a main effect of discourse type, where more formal speaking styles were judged as more “homosexual” sounding, which had an interaction with the speaker’s sexual orientation. That is, straight speakers were judged to be more homosexual sounding when reading a scientific passage. Ultimately, as with age and life-stage, the category of gender can thus be found as a correlate in the production of linguistic variables, as well as a social meaning that is detectable from exposure to the variable in speech, making it an ideal category of study in the present research.
Therefore, this study investigates how individuals’ positioning to a variable mediates the awareness of social meaning by examining social evaluations of gender and life-stage in
yeah-no users and non-users. In two evaluation experiments, we examine whether native Australian speaker–listeners associate life-stage (Experiment A) and gender (Experiment B) with the discourse marker
yeah-no in Australian English and the role that positioning plays in these associations. We use individuals’ self-reported use of the variable as evidence of their social positioning towards the variable. It is worth noting here that we do not consider reported use as actual production. As discussed above, speakers have been shown to report higher usage of standardized forms than their actual usage (
W. Labov 1966). Given that
yeah-no is a stigmatized variable in the speech community, it is expected that it will pattern according to what is considered to be socially desirable in the community, whether that be using (or not using)
yeah-no in speech. If speaker–listeners show a difference in their evaluations of the social meaning with regards to their positioning to the linguistic variable, either as a user or a non-user of
yeah-no, it would suggest that positioning is a contributing factor to the attitudinal and cognitive factors that mediate the awareness of social meaning.
6. Discussion
In two experiments, we investigated how individuals’ positioning to a variable mediates the awareness of social meaning by examining perceptions of gender and life-stage in
yeah-no users and non-users in Australian English. We found that, at least for this paradigm, the positioning of an individual towards a linguistic variable influences their awareness of the variable’s social meanings. In Experiment A, the discourse marker
yeah-no in Australian English was judged as more likely to be uttered by a student than an employee, especially among those who did not self-report as
yeah-no users. This finding aligns with previous studies on
yeah-no (
Burridge and Florey 2002;
Moore 2007), albeit with a deviation in the age effect, which was more pronounced for the 35–49 age range in Burridge and Florey. Experiment B, interestingly, revealed no overall effect of form, contrary to
Moore’s (
2007) observation of a higher frequency of variable use among males. Notably, participants who self-reported as non-
yeah-no users exhibited a significant form effect, perceiving
yeah-no sentences as more likely to be spoken by a male. It is important to consider here the sample demographics when interpreting the results of Experiment B. Our sample included a higher proportion of young participants (18–25 years old) compared to older participants. This age distribution may have influenced the findings, as younger individuals might have different linguistic perceptions and usage patterns compared to older individuals. This was not the case with Experiment A, however, though we recommend that future work using this framework includes a more balanced age distribution to avoid possible conflicts across the different age groups. Ultimately, the outcomes from both experiments in this study suggest that individuals’ positioning mediates the evaluation of social meaning.
Our discovery regarding the role of positioning contributes significantly to ongoing research that seeks to understand speakers’ motivations to use one linguistic variant over another. Context plays a significant role in individuals’ inability to deduce the social meaning expected to be indexed onto a variable. Stereotyping—both of groups of speakers and of variables themselves—also contributes to establishing a connection between individuals’ awareness of social meanings and their use, whether overt or covert, as tools for identity construction. Our study extends this line of inquiry, which investigates the attitudinal and cognitive factors that influence individuals’ awareness of social meaning by demonstrating that not only do an individual’s attitudes toward a speech community and normative stereotypes mediate judgments, but their positioning—how they position themselves through speech choices—also holds substantial significance. Those who self-report as speakers of a linguistic variable within a community appear less sensitive to the social meaning associated with that variable, while non-users exhibit heightened awareness of social meanings. This suggests that linguistic insecurity among non-users may contribute to their increased sensitivity to the social connotations of the linguistic variable. Thus, an individual’s positioning, akin to their beliefs and endorsement of stereotypes, emerges as a factor influencing cognitive processes in the evaluation of social meaning.
Examining the mismatches between the production of linguistic variables and evaluations of their potential social meaning(s), particularly when listeners are unaware of these correlating social meanings, raises questions about individuals’ positioning. For instance, if the listeners in these previous studies on (ING) self-reported as users of the alveolar form of the variable, they may not have been sensitive to the additional social meanings that were not perceived by listeners, such as the social categories of gender, socioeconomic status, dialect, age, and race that were also shown to correlate with (ING) in production (
Fischer 1958;
W. Labov 1966;
Shopen 1978;
Shuy et al. 1967;
Trudgill 1974). On the other hand, if some listeners did show a sensitivity to the additional social meanings that were indexed onto the variable, they may not have been users of the alveolar form of the variable themselves. Ultimately, the ratio of those who show a sensitivity may be smaller than that of the listeners who do not have a heightened sensitivity, but the effect was unable to be identified without investigating individuals’ positioning to the variable. Approaching this line of reasoning from an alternative angle, it is also possible that speakers who self-report as users of a variable do not create associations between the linguistic variants and social categories of their community. That is, their variable use is natural and automatic, compared to explicitly learned, conscious language choices. Thus, users of a given variant may have implicit knowledge of speech patterns in their community but show no awareness as the relationship between the variant and its social categories is meaningless for the purpose of their communication.
An interesting point pertaining to the discourse marker
yeah-no specifically is the overt nature of the variable in the community. The variable is marked, if not stereotyped, and the media attention surrounding the variable suggests it is salient in the speech community.
Yeah-no’s status in the community as “speech junk” and a “verbal crutch” could be considered as negative, certainly a vernacular speech variant, and would thus be expected to impact individuals’ positioning. As discussed earlier,
W. Labov’s (
1966) study, whereby speakers from New York showed a tendency to report higher usage of standardized forms than their actual usage, differed significantly from
Trudgill’s (
1972) findings in the opposite direction which showed a tendency for speakers to report higher usage of non-standardized forms than their actual usage. Given the status of
yeah-no, it appears that individuals are positioning in a similar way to the findings in Trudgill, suggesting that
yeah-no has a non-standard social desirability bias. This possibility, however, will require further investigation, as the data used in this study are self-reported and thus cannot be directly compared to W. Labov or Trudgill’s work which investigates actual usage. Additionally, future work comparing variables that have standard or positive connotations compared to vernacular or negative connotations, in line with Trudgill’s distinction between overt and covert prestige, would be a very interesting line of further inquiry for understanding the role of positioning. Since we expect stronger reactions regarding positioning to a variable that has a marked status in the community compared to variables that are considered indicators in a speech community, a study comparing variables with different levels of social salience is highly encouraged to further unpack the investigation of positioning with respect to the awareness and control of social meaning.
Further to the association between
yeah-no and its correlating social categories, we have found a production- and evaluation-based match between the stratification of
yeah-no and the social category of age. For speakers who did not self-report as users of
yeah-no, we also found a match between the stratification of
yeah-no and the social category of gender. Both findings suggest that an association exists between the discourse marker and the social categories of age and gender, and this finding can be interpreted as the variable indexing the categories as social meanings. Given that age and gender are the only categories to have been investigated within a sociolinguistic framework on the discourse marker
yeah-no, we encourage further investigation of the variable and other potentially relevant categories, especially since it has been demonstrated that variables can index multiple social categories. With respect to the Australian road safety campaign, which uses
yeah-no as their punch line, the categories of region and socioeconomic status appear relevant, the latter having been previously noted by both
Burridge and Florey (
2002) and
Moore (
2007). These categories have also been discussed in research regarding Australian English, specifically, the divide between Australian English accents (
Cox and Palethorpe 2010;
Harrington et al. 1997;
Mitchell and Delbridge 1965). We advocate for further examination into
yeah-no in the hopes of improving our understanding of the current study’s results and, more broadly, our understanding of sociolinguistics in Australian English.
The final point we wish to raise here pertains to the incorporation of self-reports in sociolinguistic research. Researchers, quite rightly, cite the risks of using self-reports in linguistic research, as they do not reflect natural language in use. We do not contest this; however, we can confirm from the results of this study that when examining an individual’s awareness of social meaning, self-reports offer a unique insight into how individuals position themselves toward normative stereotypes. The results showed, through a combined method of evaluation tasks and self-reporting, that the positioning of the individual plays a role in the evaluation and awareness of social meaning. As such, our methodology builds upon research that has found that the association between linguistic variables and social categories can be mediated by both attitudinal and cognitive factors, such as the speaker’s normative endorsements and beliefs. In the future, a more robust examination of self-reports, especially those examined in combination with production-based research methods, would aid this line of research. Both cross-sectional and longitudinal investigations into individuals’ positioning to a variable may offer finer-grained and subtler nuances that could reveal more about how we perceive social meaning and how we manipulate our speech for the purpose of communicating social meaning.