Buddies,Dudes,and Bros of Ontario: Trends and Patterns of Vocative Change

2.1. Familiarizers in Sociolinguistic Research

In this study, the focus on familiarizers highlights the main social correlates of vocative usage and facilitates comparison with earlier sociolinguistic research (see section 2.4). Familiarizers show several characteristics of particular interest to sociolinguistic research. First, familiarizers are a highly versatile and flexible type of vocative. They may be used across many different relationships: for kin, peers, strangers, pets, etc. Familiarizers may be less constrained by formalized social structures (compare doctor versus mom) or conventionalized relationships (e.g., darling, cabbage); that is, dude can be used in a wider range of interpersonal contexts than darling. In addition, familiarizers are highly interchangeable: individuals alternate between dude with buddy or bro, while it is only sometimes possible to do this with other vocative categories as described above (e.g., possible for honey/darling/baby, not possible for mom/pop/gramps). While we do not claim each token has identical meaning (indeed, we do not believe they do), our data demonstrate intra-speaker variation such that a speaker can conceivably use any of man, buddy, dude, bro in most instances where one of them was used.

Second, familiarizers tend to be steeped in locally-grounded social meaning, in relative contrast to more neutral vocatives such as sir, miss, or someone; offering an impetus for sociolinguistic scrutiny (see section 2.4). Familiarizers also have practical methodological advantages: they are common enough for analysis, and because they are known to change, they are a good subject for studying change over time, especially among young people. Familiarizers have been shown in the literature to play a unique role in the encoding of complex relationships and identity between people, and their patterning regarding gender, age, and generation provides evidence of important overarching trends in usage. Taken together, this suggests that the choice of familiarizers in contemporary vernacular dialects of English in North America will provide important information about their nature and, by extension, new insights for language variation and change.

2.2. Discourse Functions

In an ongoing conversation between two people, there would seem to be little need to address or call attention; one effect of doing so is to emphasize the utterance. In (3) VF repeatedly punctuates her responses with dude across a sequence of turns in which she refutes the interviewer’s statements. A similar function is shown for man in (4): AE addresses the interviewer even though he already has the interviewer’s attention, to strengthen his response to the question.

(3) VF: I really hope we can actually go out in high school. But then the thing is I don’t know how to ask people. And dude he won’t even ask me because he doesn’t like me anymore.

  Interviewer: You never know.

  VF: Dude he doesn’t.

  Interviewer: In my experience it’s always been Valentine’s Day.

  VF: Valentine’s Day. Dude, no, no come on. (Violet Fong, 13, W, Toronto)

(4) Interviewer: Who was your favourite that day?

  AE: It was nuts. Don’t even ask man. ACDC stole that whole place, stole it.

  Interviewer: I was going to say, you’ve got to see ACDC. (Andy Eaton, 18, M, Toronto)

The instance of dude in (5) is used for topic management (i.e., ER is attempting to end the discussion). Here, dude may also be used to soften ER’s effective refusal to answer the interviewer’s question, which could have negative social consequences).

(5) ER: I know. I already decided. I’m going to graphic design or media studies which is a course.

  Interviewer: Okay so yeah, why- why that then?

  ER: Because it’s what I’m into, dude. (Eliot Rousseau, 20, M, Kirkland Lake)

2.3. Pragmatic Functions

While vocatives are involved in discourse functions like turn and topic management, they can also serve to enact interpersonal relationships, depending on context (McCarthy & O’Keeffe 2003): compare Mother, what are you doing? to Mum, please help! They may claim or instantiate social relationships, intimacy, and distance. As we have seen in (1), buddy is used with a child to express concern and empathy. The same term activates a familiar friendly relationship in (6). Vocatives may also convey a speaker’s stance, tone, or framing, as in (7) and (8), where dude contributes to an authoritative and chastising tone, and (2), where the speaker uses dude to frame her response as a rebuttal.

(6) FF: I bumped into Milo at Jane and Bloor restaurant the other day. I went for breakfast with my buddy Andy. Then I turned around and he was sitting right behind me. I was like “Oh buddy, what’s up?” (Fred Flynn, 25, M, Toronto)

(7) CF: That’s what I told him. And he has like anti bacterial wipes at home. I’m like “Dude, you’re going to get sick.” (Clara Felipe, 24, M, Toronto)

(8) AS: [. . .] rather than dealing with the situation himself, he’s always like, “Go talk to the pool manager”, and I thought, “Dude, you could have done this yourself.” (Alyssa Sailner, 29, W, Toronto)

Examples (6-8) demonstrate that the vocatives rely on the interlocutors’ interpretation of the cultural and interpersonal context. Choosing a vocative for these pragmatic effects comes with a risk of failure in terms of inappropriateness, social miscommunication, or rejection of an attempt at use, and may have other ramifications. In an example of a rejected vocative use in Shakespeare’s Hamlet, Busse (2006:283) notes that Claudius attempts to enact a relationship by addressing his nephew Hamlet as “my son,” which Hamlet refuses with a pun: “No, I am too much in the sun.”

2.4. Sociolinguistic Meanings

Research on vocatives has shown that they are used in rich systems of social meaning, including stance, relationships, and social-indexical meaning. Because of their loose syntactic integration and strong sensitivity to social context, vocative terms are often reflective of gender, community, or other identities. These meanings may be complex and contingent: for example, the social meaning of the same vocative form can be interpreted differently by men versus women because of their different social experiences and positionality (Rendle-Short 2010; Alimoradian 2014) . Formentelli (2014) finds that vocative use is highly indexical, expressing social meaning and embodying relationship work; and that intended pragmatic meanings must always be interpreted in their context, to which they are intimately connected. We note that the identity of both the speaker and the addressee are critical aspects of these intricate social meanings (see e.g., Kiesling 2004). In parallel with their social embedding, vocatives exhibit extensive variation by region, ethnicity, and age, as well as cross-linguistically and over time.

Vocatives have also been shown to participate in both rapid language changes and long-term historical shifts (Shiina 2003). Much of recent sociolinguistic research on vocatives focuses on the familiarizer subcategory (Kiesling 2004; Rendle-Short 2010; Alimoradian 2014; Heyd 2014; Palacios Martínez 2018; Pastorino 2022; Flesch 2023), which often derive from historically masculine terms, such as dude, mate, man, buddy. While familiarizers in English reflect a robust history of masculine variants and usage by men, there is also evidence of a shift in these correlations (see Kiesling 2004; Rendle-Short 2010; Alimoradian 2014; Pastorino 2022).

Hill (1994) and Kiesling (2004) documented the development of the form dude from common American English usage in the late nineteenth and early twentieth centuries, into use by members of young and hip social groups during the twentieth century (e.g., pachucos, zoot-suiters, surfers, and others), and into more general use again. Over this time frame, dude ranges from a referential noun, a discourse-pragmatic marker, and a vocative. Across these functions, dude can carry various social meanings, including masculinity, ‘coolness’ and solidarity with other men (Kiesling 2004). Kiesling describes the stance of cool solidarity as “effortlessness (or laziness, depending on the perspective of the hearer),” which is linked to its embedding in the “surfer” and “druggie” subcultures where this type of stance is highly regarded (Kiesling 2004:282). The form and this social meaning then diffused to youth of the 1980s to express “dissatisfaction with careerism,” and in the 1990s to express “nihilism” (Kiesling 2004:21). Kiesling also describes how the predominantly masculine dude is also used by and to women, and that women and men can employ dude in a gradient pattern for quite different pragmatic uses. For example, while men are more likely to use dude in greetings, for example, Hey dude, how are you?, women are more likely use it when they are attempting to “ameliorate a confrontational and/or hierarchical stance” (Kiesling 2004:285). Kiesling’s example, “Dude, that’s just not cool” (286), is comparable to examples (2) and (8), both from women born in the 1980s. Similar gender-related patterns are shown for dude in British English by Pastorino (2022): dude is “now used by female and non-binary speakers as much as by male speakers” and “to address individuals of all genders equally” (27).

Rendle-Short’s (2010) analysis of mate in Australian English presents a variety of context-dependent social meanings: along with solidarity and familiarity, mate can carry meanings of masculinity versus gender-neutrality, and friendliness versus hostility. As with dude, there are contrasts here between the ways men and women use this familiarizer. The early uses signaling male solidarity have changed among younger women, who “interpret mate as being a fun, friendly term of address, almost like a term of endearment” (Rendle-Short 2010:1202). Such changes also implicate generational shifts and are diagnostic of linguistic change in progress. Further nuance is shown by Alimoradian’s (2014) study of mate in Australian English, which focuses on Australians of a non-English speaking background and reports similar patterns to Rendle-Short: mate is associated with men (and Anglo-Australian culture), but women are increasingly using the term, and with a wider variety of addressees; an effect of ethnic orientation is also reported.

Effects for familiarizers related to ethnicity/community, age, and gender appear to be common and separable, with similar developments being highlighted in Palacios Martínez’ (2018) study of London English (UK), which demonstrates a key effect of age group in familiarizer choices: adults use man and brother (including other variants such as bruv, blud), but teenagers also use other forms such as mate and boy. Urichuk and Loureiro-Rodríguez (2019) also find that bro is more common for younger speakers (especially men) in Manitoba, Canada; they further report that related variants like bruh and brah are also associated with younger speakers, particularly younger non-Caucasian speakers.

In summary, familiarizers are a proven site for rich sociolinguistic patterns, particularly with respect to gender and age. The extant research on familiarizers demonstrates all the hallmarks of change in progress. Based on the social trends reported in the literature on familiarizers—that is, dude (Kiesling 2004; Heyd 2014; Pastorino 2022; Flesch 2023), mate (Rendle-Short 2010; Alimoradian 2014), and others (Palacios Martínez 2018)—we can expect their distributions to be patterned by social factors. We may also expect change over time, either: (1) shifting types as the popularity of particular forms changes across generations, as evident in the metalinguistic commentaries in (16-20) and in the literature; or (2) a shift toward certain familiarizers as documented in the history of English from the Early Modern period into the present day (Shiina 2003); or (3) a pattern of geographic diffusion such that incoming forms are more frequent in urban centers like Toronto, while others may be more typical of outlying localities. Consistent with studies of other lexical categories in the ODP, such as adjectives (Tagliamonte & Pabst 2020) and certain swear words (Tagliamonte & Jankowski 2019), there may also be different rates of variants across time. Further, frequent variants with long-term presence in English may be fairly stable, for example, man, while those of recent vintage, for example, dude, may evidence short-lived popularity (Kiesling 2004; Tagliamonte & Brooke 2014).

4.1. Circumscribing the Variable Context

We refer to the baseline extraction of forms as ‘potential vocatives’ because they could be instances of other categories, such as referential nouns (I saw my sweetie.), discourse-pragmatic markers (Man, that’s hot!), etc. Consistent with the nature of layering described in section 1, not all tokens had any vocative function; and many tokens had unambiguous functions that were not primarily vocative. The data contain a vast array of these forms in their use as nouns or discourse-pragmatic markers. Indeed, the majority of uses of buddy, dude, and man—all used across Ontario—were referential nouns, as in (9-11).

(9) DD: ‘Cause my buddy has a fish shack there that he built. Put four-thousand-dollars into it. [. . .] And then sometimes my buddy Jason and I, him and I’ll take a hike into Algonquin-Park to go for some ah speckled-trout. (David Daniels, 57, M, North Bay)

(10) SC: ‘Cause like, he gets so much stigma for being this redneck from the South, but he’s actually like this really really awesome dude [. . .]. (Sam Carr, 21, M, Parry Sound)

(11) EA: The first schoolteacher was a man. And he was an old man. (Ethel Ashford, 90, W, Ottawa Valley)

Discourse-pragmatic uses of many of the same forms are also evident, for example, dude in (12) and man (13-15). The data were coded for these functional distinctions; that is, whether the instances of each form were primarily vocative uses (e.g., How are you, man?) or non-vocative, such as a referential noun (I saw a man.), a discourse-pragmatic marker (Man that’s good!). ⁷ To be included as a vocative, we required a token to unambiguously refer to an interlocutor or person in the immediate situation; for example, in (6) buddy is immediately followed by “What’s up?” In the absence of an obvious addressee, a non-referential-noun token is likely to be discourse-pragmatic; for example, in (14) “Man!” does not refer to an addressee, but expresses intensity or other affective nuance and is involved in the structure of the discourse. In cases of ambiguity, we erred on the side of rejecting the token.

(12) KM: Those were awesome. I saw those, I was like “Dude! That’s awesome!” Like I couldn’t believe someone actually wore that, because like it’s the Olympics. (Kaitlyn McAngus, 15, W, South Porcupine)

(13) SW: My uncle always has ah- he always has like a fish-fry there in the summer. And he’ll go catch his pickerel, all that and he’ll deep-fry it and that’s just amazing, man. (Sam Wysocki, 17, M, Kapuskasing)

(14) CF: He had like, soup to eat. Man! I didn’t know what to do. (Clara Felipe, 16, W, Toronto)

(15) RN: I went to Waterloo instead of Toronto. Oh, man! I could have taken all my courses in Waterloo. I was stuck in Wycliffe, Hoskin Street, in the summer, in the basement. (Richard Northman, 67, M, Kapuskasing)

We also distinguished and excluded tokens of direct reported speech or metalinguistic commentary, as in (16-20). These tokens were excluded on the grounds that they are not in the same sense uses of the familiarizers but mentions. However, the many comments of this type highlight the degree of salience these forms hold in the population, as well as individuals’ awareness of their uses over different social groups (16, 18) and generations (17). In particular, the commentary in (19) describes the contextualized meanings of buddy, dude, and hun (‘honey’) across different social dyads.

(16) JC: I’ve been finding myself saying ‘bud’ or ‘buddy’ quite a bit.

  JC: I think that stems from my cousin too, like you see someone in the hall, “Hey buddy!” You know? And it’s a little less Canadian than the, (affected Northern accent) “Hey bud! How’s it going there, bud?”

  (John Courchaine, 18, M, Kapuskasing)

(17) DS: I don’t really say that much but some of my friends say ‘dude’ a lot and ah, mom and dad don’t really say that. (Dwight Shooter, 14, M, Thunder Bay)

(18) CM: [. . .] I think North Bay has a real like surfer kind of slang too. Like I say “dude” and like I used to say “gnarly” [. . .].

  CM: [. . .] we say “like” and “dude” and you-know ‘cause it’s just really laid-back slang I guess, yeah. I don’t know. (Chloe McDougal, 23, W, North Bay)

(19) BS: Queer people don’t use the word buddy. Buddy doesn’t exist. Or if it does, it’s like, you’re talking to a straight guy to prove that you don’t like him.

  BS: Um, within lesbians, dude is a big word because if you call- if you’re a dyke and you call your friend dude, that means you don’t like them, but you won’t call a femme dude, because it just doesn’t work like that.

  BS: [. . .] you can have like the dykiest person going, “Hey hun, come here.” But it won’t be to another dykey person. It’ll be to a femme. (Brigitte Santino, 20, W, Toronto)

(20) NJ: We used to say dude all the time. Like, “Dude!” I still hear people say that and I laugh at them, but. “Oh dude, that was so lame. Yeah, dude.” (Neha Jairath, 27, W, Toronto)

4.2. Coding Information for Addressee Analysis

For the addressee analysis, age of the individual and assumed binary gender come from the metalinguistic information recorded at the time of interview. Addressee gender was derived from the corpus when the addressee was an interviewer or another interlocutor present in the interview, or estimated by the research team from close reading; addressee gender was labeled ‘indeterminate’ when no reasonable estimation was possible. For addressee analysis, we coded age group as ‘young’ (less than twenty-five years old) and ‘older’ (at least twenty-five years old). When the addressee is present in the interview (i.e., the interviewer or another interlocutor), age information comes from the field reports completed by the interviewers, in the same way that individual metadata was gathered. For non-present addressees (e.g., in a recounting of a conversation), age coding was based on close reading by the research team. We chose a binary scale with a threshold of twenty-five years old due to the difficulty of estimating non-present addressee age; using twenty-five years old corresponds to a distinct generational divide. This coarse approach is a necessary compromise between the limited data available and the importance of addressee in the literature. Addressee age group is coded ‘indeterminate’ when no reasonable estimation was possible.

4.3. Random Forest Modeling

In order to establish whether the trends in the data according to individual usage are statistically significant, we employ Random Forest modeling (see Tagliamonte & Baayen 2012). In keeping with the distributional trends shown in Figures 1 and 2, we include coarse-grained community (Toronto versus other communities) and assumed binary speaker gender (men and women), as well as occupation (student, blue collar, or white collar), education (some secondary education, or none), and year of birth (continuous). A random forest model is constructed out of a large number of randomized Conditional Inference Trees (CITs; Hothorn, Hornik & Zeileis 2006). In brief, CITs divide the data by recursive binary splits, each of which uses the single strongest predictor for that split, for parsimony. Each split is evaluated by significance testing and non-significant splits are not used. The tree stops growing when no significant splits exist, to avoid overfitting. This means that the contrasts presented in the analyses below are statistically significant when all the factors in the analysis are considered simultaneously.

OPEN IN VIEWER

Figure 1.

Proportion of Familiarizer Type by Decade of Birth ⁸ and Assumed Binary Gender, Outside of Toronto (W Denotes Women, M Denotes Men)

OPEN IN VIEWER

Figure 2.

Proportion of Familiarizer Type by Decade of Birth and Assumed Binary Gender, in Toronto (W Denotes Women, M Denotes Men)

This modeling approach is better suited to these data than common alternatives such as mixed-effects regression for several reasons. First, instead of the binary nature of regression modeling, random forest modeling allows the analyst to construe the choice of familiarizer as a multinomial dependent variable: given an opportunity to use man, dude, or buddy, an individual may choose any one of these options, and these variants are not ranked relative to each other. Note that this model relies on the assumption that these three familiarizers are effectively interchangeable, which we believe holds for this application (see section 2.1); a comparable regression approach would require a similar assumption. Indeed, it is specifically the multiple-variant mixture of familiarizers within a given group (e.g., people born after 1980, Toronto women, men in non-urban communities, etc.) that bears sociolinguistic investigation given the literature, and confirmed in the addressee/speaker distributional analysis. Therefore, a binomial or ordinal regression could not straightforwardly address this question. We also consider random forest modeling to be well-suited deal with the data imbalances present in the ODP, which are typical of sociolinguistic corpora compared to experimental data (see Sarda-Espinosa, Subbiah & Bartz-Beielstein 2017). Finally, random forest models are flexible, readily interpretable, and considered robust against some kinds of data bias and overfitting.

5.1. Distributional Analysis

The analyses that follow expose the general social and areal patterns in the data to characterize the nature of familiarizers in Ontario across the twentieth century from an overarching perspective. Table 3 shows the frequency of the target forms, the frequency of vocative use of each form, and the proportion of vocative use. The table reveals the extent to which the majority of uses for these forms are non-vocative and excluded: that is, referential nouns, discourse-pragmatic markers, reported speech, or metalinguistic commentary. Overall, only 22.5 percent of the data comprises a primarily vocative function. We note that the vocative share of tokens of the different lexical items shows wide variation: dude tokens are 36.9 percent vocatives, while the proportion of vocative uses for man and bud/buddy are lower (22.4 percent and 14.5 percent). It appears that the vocative share for a given term is most strongly influenced by the popularity of its non-vocative functions (e.g., noun, ‘a buddy of mine’), which varies widely between terms. This exhaustive procedure led us to establish that the main familiarizers in Ontario in the twentieth century are man, buddy, dude; in contrast, bud, fella, and mate are rare, and bro appears later. Notably, vocative uses of these forms are relatively rare in comparison to non-vocative uses, providing support for Cheshire, Adger, and Fox’s (2013) contention that vocatives arise from earlier nominal usage. This finding provides a foundation for understanding the trajectory of development of familiarizer vocatives in Ontario English, and may be used as a comparative base for studies in other corpora.

OPEN IN VIEWER

Table 3.

Frequency of Target Forms and Vocative Uses

Variant	Frequency of form	Frequency of vocative use within form	% vocative
bud	38	8	21.1
buddy	218	29	13.3
dude	195	72	36.9
fella	6	0	0.0
man	1589	356	22.4
mate	29	2	6.9
Total	2075	467	22.5

5.2. Social Patterns

We begin our analysis of the Ontario familiarizers by considering the distribution of forms in time and space. As outlined in section 3, data sparsity does not allow us to compare each of the ODP communities individually; instead, we contrast the smaller communities of Ontario (Figure 1) and the large urban community of Toronto (Figure 2). Another key social factor is assumed binary gender, indicated by W (women) and M (men).

Figures 1 and 2 expose dramatic patterns across time and space for familiarizers, and between men and women in Ontario. In terms of apparent time, the familiarizer man—the dashed line—is both the most frequent and the most stable form across generations. However, there is evidence of a decline in its use over the course of the twentieth century. This decline is particularly visible among men (right panels). Another notable contrast by region is that buddy is a strong early variant outside of Toronto, consistent with the North American associations cited in the OED (OED Online 2021), but it is virtually non-existent in Toronto. In contrast, dude is more prevalent in Toronto, and appears earlier in the apparent time trajectory. Both women and men born after the 1960s use dude, but there is a surprising peak of usage among women born in the 1970s and 1980s; see related examples (2) and (17-19). In contrast, the use of dude among men shows an incremental rise and much lower frequency.

These comparative trajectories highlight the notable gender differences, generational changes, and community contrasts in familiarizer use. Women use more of the novel forms: in the city of Toronto they use dude, and in the smaller communities outside of Toronto, they use buddy. In contrast, men use more of the older and more conservative form man regardless of community. The differences across the twentieth century also reveal shifts over intervals of ten to twenty years, with individuals born in the 1970s ushering in the most marked adjustments. Notably, Toronto speakers adopt dude earlier than the other communities, which instead conserve buddy.

5.3. Addressee Effects for Age and Gender

The literature notes that one of the strongest effects on vocative use is the addressee within social context. For example, Kiesling (2004) showed that women and men used dude at very similar rates when speaking to women, and that the relationship with the addressee could be more decisive than the influence of gender. In an effort to account for the effects of addressee, we coded all the familiarizers for the key attributes associated with familiarizer use—age group and gender of both speaker and addressee (see section 4.2 for coding details). The following section summarizes the results from a proportional analysis of these factors and their combination, focusing our discussion on each of the main forms. Although we cannot include these factors in our statistical modeling due to scarcity and exaggerated imbalances in the distribution of the data, we will discuss their relevance to the subsequent analysis. We refer to Table 4 (distributions by gender) and Table 5 (distributions by age group).

OPEN IN VIEWER

Table 4.

Distribution of Familiarizers by Speaker and Addressee Assumed Gender

Speaker gender	Addressee gender	Familiarizer	Tokens	% Familiarizer by dyad type
man	indeterminate	buddy	9	11.4
man	indeterminate	dude	11	13.9
man	indeterminate	man	59	74.7
man	man	buddy	14	6.3
man	man	dude	23	10.3
man	man	man	186	83.4
man	woman	buddy	0	0.0
man	woman	dude	3	3.4
man	woman	man	86	96.6
woman	indeterminate	buddy	9	23.7
woman	indeterminate	dude	15	39.5
woman	indeterminate	man	14	36.8
woman	man	buddy	4	12.5
woman	man	dude	15	46.9
woman	man	man	13	40.6
woman	woman	buddy	1	4.3
woman	woman	dude	5	21.7
woman	woman	man	17	73.9

OPEN IN VIEWER

Table 5.

Distribution of Familiarizers by Speaker and Addressee Age Group

Speaker age group	Addressee age group	Familiarizer	Tokens	% Familiarizer by dyad type
older	indeterminate	buddy	7	16.7
older	indeterminate	dude	3	7.1
older	indeterminate	man	32	76.2
older	older	buddy	3	13.6
older	older	dude	1	4.5
older	older	man	18	81.8
older	young	buddy	1	0.9
older	young	dude	0	0.0
older	young	man	110	99.1
young	indeterminate	buddy	14	18.7
young	indeterminate	dude	24	32.0
young	indeterminate	man	37	49.3
young	older	buddy	0	0.0
young	older	dude	7	46.7
young	older	man	8	53.3
young	young	buddy	12	5.5
young	young	dude	37	16.9
young	young	man	170	77.6

As reported earlier, man is the most frequent and diffused form. The distributional results show that is it overwhelmingly used when the speaker is a man, regardless of addressee gender (75-97 percent of familiarizers from men, depending on addressee gender; see Table 4). Women use man more rarely than men, and less than men in proportion to buddy or dude: 37 percent when addressee gender is indeterminate, 41 percent to men, and 74 percent to women. This distribution may suggest that speaker identity and not addressee gender drives some avoidance of man for women speakers. Additionally, man is dominant among older speakers, and largely directed at younger individuals: compare 99 percent usage for older-to-younger interactions versus 82 percent for older-to-older ones; and 78 percent for younger-to-younger versus 53 percent for younger-to-older (though, younger-to-older tokens are rare in the data). In summary, man is the dominant baseline familiarizer, especially for older speakers and for men. It is particularly common when men address younger men and when women address women.

The use of buddy is strongly directed at men, with almost no tokens for man-to-woman or woman-to-woman dyads. Many of the addressees for buddy are indeterminate for gender and especially for age group. Aside from these trends, buddy is largely used in younger-to-younger interactions (5 percent versus 0 percent for younger-to-older) and older-to-older interactions (13 percent versus 1 percent for older-to-younger). In sum, use of buddy is correlated with addressing men, and for age group peers.

The use of dude is also strongly directed at men, for both man-to-man and woman-to-man interactions, with many indeterminate addressees. For older speakers, dude is rarely used; but for younger speakers, it is more frequent outside of younger-to-younger dyads (32 percent younger-to-indeterminate, 47 percent younger-to-older, but 17 percent younger-to-younger). Overall, the use of dude is limited to younger speakers, is predominant among women, and largely man-directed; it is most frequently used with older addressees.

While bro is not included in the analysis, we note striking patterns in its use. As mentioned earlier, it is restricted to individuals born after 1999; the following proportions are calculated by including bro tokens with the study data. For these younger speakers, bro is used in 25 percent of cases for older addressees, and 26 percent for younger ones. It is used much more for same-gender dyads (9 percent versus 5 percent for man-to-man and man-to-woman; 67 percent versus 9 percent for woman-to-woman and woman-to-man). The relative dominance of bro for women speakers exceeds other familiarizers, including man.

In summary, there are pronounced distributional patterns associated with the use of familiarizers by addressee gender and age group. Man is diffused and generalized, with a bias toward younger addressees, but more neutral with respect to addressee gender. In contrast, buddy and dude pattern similarly with addressees that are men. Use of bro is innovative and is not strongly associated with a particular addressee gender. As we shall show, these results converge with and support the speaker-centric patterns we discuss in the next section.

5.4. Random Forest Modeling

We present the random forest model by variable importance analysis, which indicates the degree to which each predictor has explanatory power across the large number of randomized trees which comprise the model. Importance values are interpreted relative to each other, and values near zero indicate that a predictor is no more useful than noise (i.e., not significant). Figure 3 summarizes the random forest model by showing relative variable importance of factors, with the dashed gray line indicating the absolute magnitude of the least important predictor. This value is used as a rule of thumb threshold for considering the importance of variables in the model.

OPEN IN VIEWER

Figure 3.

Random Forest Model Summary Showing Relative Variable Importance. Dashed Gray Line Indicates the Absolute Magnitude of the Least Important Predictor

The analysis confirms statistically significant effects on familiarizer use, which are Gender, BirthYear, and Community. Gender and BirthYear are the most important, each contributing a strong and comparable influence. The Community difference between Toronto and the non-urban communities is also significant. In contrast, the effects of occupation and education are not significant in this model. To illustrate the intersecting effects of Gender, BirthYear, and Community, we present a single conditional inference tree (Figure 4) that has been fitted on the same predictors as the random forest model.

OPEN IN VIEWER

Figure 4.

Conditional Inference Tree Showing Variable Structure. Gender, Community, and BirthYear Drive Significant Splits

Figure 4 shows that the main split (node 1) is Gender, with women (W) on the left and men (M) on the right. Men show more exclusive use of man than women (light gray bars). While women in the data also use man, they employ a wider range of familiarizers which includes buddy (dark bars) and dude (gray bars). The subsequent splits contrast the effects for men and women. For women (node 2), familiarizer use is further split by community: women outside of Toronto sometimes use buddy, but Toronto women do not (instead they use dude). For men (node 5), familiarizer use is split by year of birth in 1980: for men born after 1980, dude is more frequent. These patterns generally match those shown in Figures 1 and 2, though the difference in buddy use between men in Toronto and men in other communities is not significant in the tree model.

6.1. Sociolinguistic Patterns of Ontario Familiarizers

The most popular familiarizers of the late twentieth century in Ontario, Canada, are man, buddy, and dude. Importantly, the system of familiarizers has been changing incrementally across generations, with statistically significant correlations related to the individual’s gender and the place they live. The dominant and longstanding form is man for individuals across the province. However, buddy is a persistent option for communities outside of Toronto, especially among women. The form dude becomes popular among individuals born after the 1960s, consistent with previous research on American English (Kiesling 2004; Flesch 2023). Interestingly, the earliest Ontario users of dude are women, notably in Toronto. The strong effect of an individual’s (binary) gender on familiarizer use is a recognizable pattern in sociolinguistics. Women are known to lead in linguistic innovation (Labov 1990 et seq.) and to be at the forefront in changes from above (Principle 2, Labov 2001:274) and below (Principle 3, Labov 2001:266). In this study, women are ahead of men in timing and extent of their use of dude or buddy. That women lead in the use of these familiarizers suggests familiarizers (and perhaps vocatives more generally) are yet another area of grammar for which women “perceive and react to prestige or stigma more strongly than men do, and when change begins, women are quicker and more forceful to employing the new social symbolism, whatever it might be” (Labov 2001:291).

In addition to the well-known sociolinguistic pattern of women leading what becomes a whole-community change, we suggest that the gender differences in familiarizer use in Ontario may indicate that men and women are exercising distinct strategies in social meaning for their familiarizers, and are using familiarizers in different contexts for different pragmatic functions. Kiesling (2004) notes that women use dude more when speaking to other women in the Dude Corpus of University of Pittsburgh college students in 2001 and 2002; and that in such cases, women used dude at a higher rate than men outside of simple greetings (i.e., in contexts of commiseration, confrontation, or directives). Following Kiesling (2004), Flesch (2023) finds that “the frequency of ‘dude’ more than doubles between 1995 and 2010” in the international TV Corpus (29), and that dude is used increasingly by women with a greater variety of addressees; and Pastorino (2022) concludes that the term is perceived and used as gender-neutral. This parallels Rendle-Short’s (2010:1202) observations on the familiarizer mate in Australian English: young women use mate gender-neutrally for fun and intimacy, rejecting the “traditional meaning of male solidarity, camaraderie and equality.” Rendle-Short notes that young women specifically reject the use of mate in some contexts, especially hostile ones. These possibilities of divergence in context and social meaning can be observed in example (19): a twenty-year-old Toronto woman describes strong and specialized usage of buddy and dude within her queer community, suggesting that a more fulsome study of gender fluidity and familiarizer use is warranted.

Given the gendered patterns we have observed, a possible explanation for the overarching result that women use man to a lesser extent than men do (in favor of buddy and dude) may be that women are avoiding the dominant variant man. Specifically, women may be avoiding the masculine meanings associated with man: Urichuk and Loureiro-Rodríguez (2019) conclude that “[man] indexes a higher degree of masculinity than dude in Canadian English.” Since man has the strongest statistical and semantic connections with masculinity, women may choose the other familiarizer options in order to avoid those associations. This retreat from a prevailing usage in relation to a gender contrast has been observed in earlier research; for example, Trudgill (1972) reported that men in Norwich, England align with local men’s norms by retreating from the prestige or prescriptive uses of women. Similarly gendered patterns have been noted by Kroch and Small (1978) and Labov (2001), as well as in studies by Laitinen (2008, 2018) and Nevalainen and Raumolin-Brunberg (2003); we expand on this concept and its connection to grammatical change in section 6.2. In an analogous pattern of disassociation, we suggest that Ontario women use buddy and dude as options less associated with men than man, depending on the norms of their localities: dude in Toronto; buddy in the non-urban communities.

The contrast of individuals in Toronto versus those in smaller, outlying communities, especially with respect to gender, raises additional questions. One of the strengths and goals of the ODP materials is the ability to compare language patterns across regions and—where possible—individual communities; however, due to the rarity of vocative tokens, the current study could not pursue finer contrasts either with respect to gender or community. These coarse-grained groupings undoubtedly miss minor variants and locally-situated uses that can be studied in future research. Nevertheless, we have uncovered a striking difference in use of buddy: Toronto women avoid buddy entirely, while it is a major choice for women outside of Toronto. Moreover, this pattern is a significant choice for women outside of Toronto within each community for which adequate tokens are available, particularly smaller communities where buddy is frequent in both men and women.

The gender pattern for buddy is distinct from the rise of dude for women inside and outside of Toronto. In the United States, dude is reported to be a relatively recent innovative form (Kiesling 2004). That it is also ‘cool’ and trendy is consistent with general sociolinguistic patterns in which women favor positively-regarded incoming forms. Women outside of Toronto do not pick up dude until the generation born in the 1980s, but even here their usage is split between buddy and dude. This pattern suggests that Toronto women had contact with dude decades earlier than individuals in the outlying communities: buddy was not used in Toronto, while women outside of Toronto already used buddy as an alternative to man. Outside of Toronto, the circumstances are reversed: both the lack of contact with dude and the prior presence of buddy are consistent with the delayed and reduced adoption of dude. A further possibility is that speakers from different communities maintain oppositional identities such that buddy has associations with being positively associated with outlying regions and dude has (or had) associations with being ‘urban’ or ‘southern’ (i.e., from Toronto). Further investigation of these possibilities would require more anthropological investigation on site in both Toronto and the other communities, as well as qualitative analysis of the contexts of use.

6.2. Systems of Familiarizer Change

Conversational speech data from individuals born from 1950 to 2004 in Ontario, Canada has provided insight into the rise and fall of familiarizers across the latter half of the twentieth century. We have documented a vigorous generational shift for these terms, with roughly twenty-year spacing between peaks for buddy and dude; illustrated in example (16). As discussed in section 2, buddy and dude are each innovations compared to the much older familiarizer man. Their advance is localized to this telescoped time frame, concordant with the idea that vocative forms are quicker to change than other grammatical systems. As Kiesling (2004:300) says, “What was cool in 1982 is not necessarily cool in 2002 but may become cool again in 2005.” This pattern of relatively rapid generational change can be corroborated by the familiarizer bro, a rising new form for Toronto teenagers. Unfortunately, we were not able to include bro in the statistical analyses because it occurs only in Toronto and among speakers born after 2000. However, the reason for its exclusion is precisely what makes bro relevant here: tokens of bro from teenagers in Toronto at the time of writing straightforwardly indicate a recent innovation. Moreover, bro is used by both young men and women, and with great frequency. In fact, these speakers overall use bro more than man or any other familiarizer. Consistent with the notion that familiarizers synchronize with prevailing social trends, this pattern converges with the significant rise in the same familiarizers, for example, bro, blud, bruv, reported by Palacios Martínez (2018:39), collected in London, England between the early 1990s and 2010 in the neighborhoods of Hackney and Havering (Cheshire, Kerswill, Fox & Torgersen 2011); a similar pattern of increasing bro is observed by Flesch (2023), and Urichuk and Loureiro-Rodríguez (2019). Study of these forms among individuals born in the early decades of the twenty-first century is critical to tap these incipient trends at a later phase in their development and into the future.

At the same time, we note that there are also indicators that individuals may change their familiarizer use over their lifetime: example (19) shows metalinguistic commentary by a twenty-seven-year-old Toronto woman who claims that she used to use dude, but now does not. However, because our data permit apparent-time analysis only, we are not able to disentangle age-grading patterns, another pathway for further investigation.

The literature for familiarizers describes their particular character as tightly bound in discourse context and group identity; accordingly, our study demonstrates that they are systematic and patterned at the broad social level, visible evidence comes from the importance of year of birth across generations in our results. We suggest that this reflects a tight association of familiarizers with generational identities and that, as with many linguistic changes, younger speakers drive the adoption of new forms which distinguish them from their elders, so that familiarizers shift in a tight temporally-spaced trajectory across these decades. Whether they recycle as has been found with intensifying adverbs is not apparent in the present study, but is something to keep in mind in future research.

A key characteristic of the full set of types with the category vocatives is that their use has been documented to change over time to accommodate changing social proprieties. Shiina (2003) reports a historical shift in vocative usage between Early Modern English (EME, c. 1640-1760) and Present Day British English (PBDE, c. 1999), which she argues reflects a change in social structures. The most prevalent forms in EME are honorifics and titles, for example, your ladyship, sir, etc. (see also Busse 2006:230). In PBDE, they are replaced by kinship terms (e.g., father, mother), first names, and familiarizers (Shiina 2003; see also Murray 2002).

We argued above that vocatives (and especially familiarizers) are of particular interest to sociolinguistic because they are so strongly embedded in social context and thereby may be among the most rapidly changing systems of grammar. This includes changes in types (e.g., the rise of dude and the fall of previous terms) and changes in vocative categories (e.g., the fall of honorifics and the rise of familiarizers as a whole). Vocative terms also play an important role in other large-scale grammatical changes, so an understanding of these interactions can in turn inform our sociolinguistic inquiries into familiarizers. Studies by Palacios Martínez (2018), Kleinknecht and Souza (2017), and Cheshire, Adger, and Fox (2013) describe a pathway of grammatical change for words such as man, which start out as referential nouns and then evolve into vocatives, pragmatic markers, and pronouns. Cheshire argues this process is crucially driven by the social needs of speakers, and states that “The sociolinguistic setting is all-important” (Cheshire, Adger & Fox 2013:627). Palacios Martínez notes, however, that the shifts in this pathway (i.e., noun > vocative > pragmatic marker > pronoun) may vary and “need not be regarded as necessarily sequential” (Palacios Martínez 2018:42). Further, Kiesling (2004:284) has suggested a similar dynamic for dude, which is developing into a discourse-pragmatic marker conveying a “cool solidarity” stance. The referential noun dude (‘a dandy, fashionable man’) became used as an in-group term of address (a familiarizer), and then a discourse-pragmatic marker, and then a pronoun; for example, John just arrived. Dude is hungry. Our results corroborate these trends.

The role of familiarizers in this process provides an insight into the gender patterns found for vocatives: masculine-derived familiarizers may not all retain the same masculine meaning of their source nouns, at least not for the time frame studied here. As terms shift from one grammatical function to another, their meanings also evolve and go in different directions. Cheshire, Adger, and Fox (2013) discuss the process underlying loss of meaning features, desemanticization, which occurs coincident with grammatical shifts. For example, the use of man as a discourse-pragmatic marker has lost the [+male] feature (keeping in mind that layering means that uses may not be categorical and may co-exist at different points in time). This again parallels discourse-pragmatic use of dude, which has carried forward a cool solidarity meaning and not the masculine feature of its referential source. We note that desemanticization steps may be divergent or incomplete, such that vocative man might be simultaneously used by and to women, yet also carry a measure of masculine meaning, as evident in the distributional results for familiarizers by speaker/addressee age group and gender. Historical sociolinguistic evidence shows that women avoided the use of masculine-associated pronouns in the past, which suggests incomplete desemanticization of that feature. As highlighted by Cheshire, Adger, and Fox (2013:624-625):

Laitinen’s (2008) analysis of the development of he, they and those as indeterminate pronouns in Early and Late Modern English found that women strongly disfavoured the use of masculine he as an indeterminate pronoun, presumably, he suggests, because they were avoiding a form that for them indexed masculine gender. Similar gendered patterns are found in studies focusing on the loss of –man in indefinite pronouns. Nevalainen and Raumolin-Brunberg (2003) point out that gender is the only social factor that correlates with the reorganisation of the indefinites in Early Modern English; and Laitinen’s ([2018]) study of individuals who lag behind and polarise language change shows that male writers continue using the –man indefinites several generations after female writers.

We note that vocatives and their homophone terms have varying associations in the lexicon: for example, man frequently appears in adjectival collocations (I saw a tall man) while buddy does not (*I saw a tall buddy). These differences in usage, which are affected by syntactic, semantic, and social contexts, lead to variation in the frequency and flexibility of these terms in the lexicon. Cheshire, Adger, and Fox (2013) argue that this kind of variation affects the developmental trajectories of the terms as vocatives, pragmatic markers, and pronouns. Thus, despite their commonalities (e.g., as masculine-derived familiarizers), there is significant variation across forms in their developmental circumstances. Vocative man has had a continuous presence in English since the time of Old English (OED Online 2021) and is thought to be ubiquitous; buddy is attested much later, since the nineteenth century (OED Online 2021) and is reportedly limited to North America and the Caribbean. The familiarizer dude is most recent, accelerating in use in the US during the 1960s and afterward (Hill 1994; Kiesling 2004; OED Online 2021; Flesch 2023). Perhaps because of its long usage and diffusion, man has more general meanings (referential and social) than buddy and dude. Here too, further study of these factors and how they impact vocative change is ripe for future study.

To conclude: we have demonstrated that familiarizers are an exceptionally rich area of grammar for study, especially for the analysis of language variation and change. Their variety and availability for speaker choice, along with their close ties to generation, place, human relationship, and gender, give them high value for insight into the interface of external influences on language change and social change. In this study, we showed that familiarizers, while exhibiting strong associations by speaker/addressee age and gender, also reflect broad scale social patterns that are highly sensitive to Labov’s principles of linguistic change (Labov 1994, 2001) and follow known patterns of grammatical change, involving layering, desemanticization, and other linguistic correlates.

Further, our findings suggest that familiarizers are ripe for furthering the understanding of diachronic pathways of change (cf. Cheshire, Adger & Fox 2013). In the vocative system, lexical nouns are tapped for address functions while bringing along their prior associations and acquiring new ones. At the same time, we have every reason to think that adjustments within the inventory of familiarizers are taking place within in a long-and-ongoing move away from linguistically-marked formal social structures in different spheres of life—at work, at school, at home. To take a mundane example, historical norms of address with title and last-name, for example, Missus Donna Taylor, could continue to move toward increasingly familiar modes from abbreviated last names or first name only, to more intimate forms: for example, Missus T, Donna, lady, girl (Murray 2002; Shiina 2003). Further, while generational cycles of linguistic change do not necessarily require concomitant social or cultural shifts, the dramatic changes evident in certain familiarizer uses by birth year since the 1980s (Figure 4) suggests a shock point in time that is known to correlate with other innovations (see Tagliamonte 2016:256). Given the increasing gender fluidity in contemporary society and progressivism, we suggest that the next stage in the development of familiarizers may prove critical for understanding the interplay between language variation and cultural change.

Moreover, the evidence for changes in usage that transcend geography—that is, timing and uptake across North America in the rise of dude and between Canada and the UK for bro (compare Kiesling 2004; Palacios Martínez 2018; Flesch 2023)—means that this area of grammar may be a decisive test for ongoing global patterns of change (see Tagliamonte & Hudson 1999:168; Tagliamonte, D’Arcy & Rodríguez-Louro 2016). It will be of great interest to assess how the social, geographic, and generational patterns we have uncovered here compare to familiarizer use in other countries, regions, social groups and social networks, across ethnicity and other societal characteristics, and how these are embedded within a longer trajectory of change reaching back into the history of English and forward into the twenty-first century.