The object of this review is the videoconference titled The sociolinguistic perception of ‘sounding gay’, presented by Dr. Ronald Beline Mendes[1], promoted by ABRALIN (Brazilian Association of Linguistics) and broadcast to the internet on the platform YouTube.com, on June 23^rd, 2020, at 5 pm EST. The conference was part of the Abralin ao vivo activities. It had the duration of two hours and was mediated by graduate student Dany Thomaz Gonçalves, from Universidade Federal do Rio de Janeiro.

Dr. Mendes is a professor of the department of Philosophy and Human Sciences at Universidade Estadual de São Paulo (2018). His research if focused on the following topics: variation, change, sociolinguistic perception and correlation between social categories and linguistic usage.

The conference was divided in three parts: (i) a few theoretical assumptions about the relation between linguistic elements and their potential social meanings (indexical relation), (ii) metadiscursive information from a research developed by the lecturer in 2007 on what it is to “sound gay”, and (iii) results from an experiment about sociolinguistic perception, where the associations between linguistic forms and social meanings are made implicitly – most of the results from this research is available in his 2018 thesis.

There are two major approaches to a perception experiment in Sociolinguistics. The first, roughly speaking, examines how a linguistic variant has influence over the social information ascribed to a speaker, while the second is dedicated to assessing how the social information about a speaker may have influence over the perception of a given linguistic variant (DRAGER, 2014[2]). In this conference, Mendes reports, as main result, an experiment of the first approach, once his interest is to investigate the social meanings to which a certain linguistic element can point.

The investigation of the bonds, through which linguistic forms summon many aspects of the social structure, allows us to unveil the most varied subjective assessments. An example is the one that Mendes presents, productively, when he exposes, in the third part of the conference, the implicit assessments experiment, which aims to assess the effect of the variable standard agreement of nouns (SAN henceforward) on the perception of masculinity. The desire to investigate and try to associate linguistic elements with the perception of gender and sexuality is not something new. Papers published in English have already built a certain tradition, with research done by Gaudio (1994[3]), Smyth e Rogers (2001[4]), Symth et al. (2003[5]) and many others. In Brazil, however, papers on the sociolinguistic perception of gender and sexuality started to gain power with Dr. Mendes’ prolific production (2007; 2011; 2015; 2018[6-9]), and have motivated others such as Sene (2018; 2019[10,11]) and Gonçalaves (2018[12]).

In 2006, Labov mentions, in the second edition of his book on New Yorker English, that experimental approaches, comparatively to the production studies, had not thrived in Sociolinguistics (LABOV, 2006 [1966][13]). Almost fifteen years later, the scenario has changed. An example to highlight, at least in Brazil, is the sociolinguistic perception experiment displayed by Mendes. From the stereotypical constructions shown in the second part of the conference, especially the one about the notion that “gay men speak properly”, the lecturer, in a punctual way, proposes to investigate the following thesis: Would a man be perceived as less masculine due to using the SAN?

Even though the stereotypical constructions are general reactions presented spontaneously, and apparently without any correlation with the objective linguistic data, what the population, in general, believes regarding language and its usage has been more and more acknowledged as a legitimate and important object to empirical studies. In this direction, Labov (2006 [1966][13]) states that:

On implicitly eliciting an evaluative behavior, without the participants knowing what was being assessed, Mendes provides us with his experiment, which translates, at least for now, into the most modern approach to perception studies, since he works with spontaneous recordings, a sophisticated survey platform and a refined statistical tool. From two excerpts provided by four speakers, taken from the SP20101 tool, the professor, aided by the software PRAAT, creates paired stimuli that contain, for instance, a disguise in the SAN variant and another in the NSAN (non-standard agreement of nouns) variant, as shown in the table below.

In spite of the stimuli developed not having been presented to the attendants of the conference, Mendes explains that the results from Carlos and Robson, for example, came across as less effeminate, while Jaime and Lucas were perceived as more effeminate. When the professor analyses the answers from the stimuli that the participants heard (whether SAN or NSAN), it is possible to verify that all the four participants in the survey, when listened to in the SAN disguise, were perceived as more effeminate than when listened to in the NSAN disguise. In other words, what Mendes summarizes with these results is the fact that the AN variant has an effect on the perception of how a male individual would sound, and hence the NSAN, besides indicating low education level and socioeconomic class (as other sociolinguistic studies (SCHERRE (1988; 1994)[14,15]) show), also (indirectly) index masculinity.

That illustrates, in the indexical field of agreement of nouns, that besides education level and social class, masculinity is also a potential social meaning. These results may not have been easily understood by everybody who attended the conference, since analyses that emphasize the relation between linguistic elements and their social meanings, such as the ones developed by Mendes, are a novelty in Brazil, in a way. However, the professor didactically clarifies that indicating masculinity, for instance, does not imply that if one wishes to sound more masculine, then one must use the NSAN; the fact is that the indexical relation is neither essential nor necessary.

The professor exemplifies this discussion when he brings up the case of the word “gosta” ([one] likes) as pronounced by some cariocas (Rio de Janeiro locals): “goysh-tuh”, where “oysh” is pronounced as a diphthonged open-mid back rounded vowel [ɔj] followed by the voiceless palato-alveolar sibilant consonant [ʃ]. The linguistic form [ɔjʃ] may indicate the social meaning of “being carioca”, but not because there is something essentially carioca in the linguistic form in question; to the contrary: this is only possible because “so frequent did the co-occurrence of said linguistic form and its regional identity became, that one turned into a ‘specialty’ of the other.” (MENDES, 2018, p. 6[9])

With this discussion, the lecturer brings important contributions to a higher questioning, in which other researchers are also interested: Which linguistic elements make a listener think that a certain speaker sounds gay? It is worth mentioning that the question distances itself from an essentialist approach, once it is not suitable to ask or seek to explicitly identify whether “being gay means to speak in a certain way” – there is no such thing as a “gay way” of speaking – also, a male individual may “sound gay” but not be gay, and vice-versa. It is in this sense that Mendes, quite wisely, opts for “sounding gay” as a title for his conference.

In summary, the conference is productive and systematic, for it invites everybody to reflect on the advances in sociolinguistic research, in particular the papers on the perception of gender and sexuality. Besides that, Mendes shows that, from a linguistic perspective, perceptions related to stereotypical intuitions must be tested through adequate methodologies, once it is commonly acknowledged that language does not only convey purely informational content, but also reflects social information.