Listening as Self‐Positioning: Revisiting Inoue's Psychoanalytic Approach to the Listening Subject

In 2015, while pursuing my master's degree in linguistics at the National University of Singapore (NUS), I signed up for a three-day workshop by postcolonial theorist Raka Shome, who was visiting NUS at the time. On the very first day, we discussed Spivak's (1992) seminal essay Can the Subaltern Speak? Confronted with this challenging text, I found Shome's explanation to be immediately lucid and incisive—especially for a linguistics student like me. She explained that the difference between talk and speak lies in the fact that when we speak, we speak a language, and language is a structured system; to speak is thus necessarily to engage in an act that links the individual to a structure. Talk, however, does not inherently involve this linkage. From Spivak's perspective, the subaltern can talk, but what they say can never enter—let alone unsettle—the very structures that dominate them. Their words must always be mediated, fragmentarily and with alterations, through the authoritative perspectives of elites before they can be transmitted. Smalls's (2018) ethnography, in this virtual issue, on discourses of “black-on-black” violence in the United States offers a contemporary illustration of this dynamic. In her analysis, anti-Black modes of listening construct Black people—especially African Americans—as inherently violent, enabling the criminalization of Black youth and shaping how African transnational youth present themselves in contrast to their African American peers. These listening modes sustain socio-racial hierarchies through “discursive violence” that casts groups or persons into a subaltern position (p. 361). A similar mechanism is visible in my study of Taiwanese netizens’ backlash against Vietnamese migrants’ antiracist petition (Wan 2025). The petition protested a multinational supermarket's collaboration with a Taiwanese male YouTuber who profits from a fictional Vietnamese woman speaking a parodic accent. In response, many Taiwanese commenters hailed themselves as the listening subject by recasting the migrants’ antiracist action as itself “racist” to uphold the existing racial order where the Taiwanese majority defines how multiculturalism should be manifested. These cases produce a dynamic in which the margins’ self-representations are refracted through the dominant's listening practices. From this perspective, the reason the subaltern cannot become a speaking subject is that they cannot truly be heard. Here, I want to bring in something that frequently surfaced in my past sociolinguistic interviews on identity negotiation among deaf individuals in Taiwan: interviewees would often repeat a saying from their speech-language pathologists— “Being able to hear doesn't mean being able to hear clearly; being able to hear clearly doesn't mean being able to understand.” This originally emerged as a clinical discourse in audiological rehabilitation: patients were instructed to progress from hearing sounds as meaningless noise, to distinguishing their natural types (thunder, impact, and human voices), and finally to understanding their semiotic meanings, including linguistic ones. Yet during the interviews, what I found myself reflecting on more often was: Why isn't this same discourse directed toward those who cannot understand the pronunciation of deaf speakers? Listening, too, is imbued with power relations. The main idea here is that it is the listener—the listening subject—who determines the terms on which the speaker appears within the power structure. In recent years, a growing number of sociolinguists, including both linguistic anthropologists and variationists, have shifted their focus from the speaker to the listener (e.g., Campbell-Kibler 2021; Cushing and Snell 2023; D'Onofrio 2018; Jessee and Calder 2025; Reyes 2017; Wong and Babel 2017). One motivation, especially from a liberal perspective, is the pursuit of social justice: through the lens offered by Spivak above, it becomes clear that the question is not about the speaker, nor about interrogating what the speaker has done “wrong” to remain unheard or to fail to become a speaking subject. Rather, it is the listener who must be interrogated—why do they fail to listen? Campbell-Kibler's (2021) work makes this especially evident: some listeners explicitly reported that they consciously avoided linking foreign-accented speech to non-white faces so as not to appear racially biased. This illustrates that, at least to some extent, listeners can deploy ideological packages that yield varying evaluations of the same linguistic signals. When listeners occupy a dominant position, their listening practices have the structural capacity to reproduce or alter the social order of power—they are the subjects who must be interrogated. And Inoue's conceptualization of the listening subject offers, in my view, another important reason we need to take this turn: beyond the listener's power to shape social structures by virtue of occupying a dominant position, their listening (and interpretation) of voices from the margins has no a priori or “natural” drive to maintain those structures. In Inoue's account, a crucial point is that listening as a social practice is, to a great extent, a response to the subject's anxiety—in the psychoanalytic sense: the anxiety experienced when the subject's sense of self-completeness feels threatened. I argue that one way to further mobilize Inoue's (2003) concept of the “listening subject” is to return to her Lacanian reading of how Meiji-period Japanese male intellectuals tried to enregister the speech style of schoolgirls. Inoue demonstrates that, no matter how insistently these men linked the girls’ way of speaking to negative social meanings—low socioeconomic status, vulgarity, and so on—the labels they produced never fully absorbed their sense of unease. That unease, she argues, arises from the voice as a “psychic object” (p. 176), whose hailing force exceeds any act of metapragmatics. What they target as “unpleasant to the ear” are sentence-final particles teyo and dawa, the uptick in kango (Chinese words), or English words among the girls. Yet even the most exhaustive catalog of such linguistic features could not capture the ineffable affect that unsettled the men—the anxiety that surfaces when the masculine order of modernity feels destabilized. In this context, anxiety should be interpreted psychoanalytically. As what is cited in Inoue's discussion of Lacan's objet petit a (p. 177), the symbolically irreducible “little a” embodies the subject's void or lack that they unconsciously pursue yet never reach. And, when the subject encounters the “little a” in the symbolic stage, the very architecture of the psyche threatens to collapse, for the subject sees their own incompleteness. As Inoue reminds us, “there is always a residue or excess that is irreducible to language and meaning, inconvertible into the signified, and not necessarily linguistically present and presentable” (p. 176). That is, anxiety emerges when the subject sees themselves from the Other, and this disconnect, however, cannot be symbolized. Likewise, Reyes (2017) shows how Philippine middle-class elites hear the conyo elite—their English–Tagalog linguistic hybridity—as inappropriately modern, even betraying the nation, although semiotically English–Tagalog mixing can have an equal potential for indexing modernity. Reyes notes that “this anxiety is often produced from having glimpsed at a modernity that is out of reach to most” (p. 217). Confronted with conyo elites, the middle-class elites seem to hear the conyo elite as dramatizing a modernity that cannot be fully symbolized. As Reyes argues (p. 228), these elites encounter an “internal other”: a mirror of their own deficiency and unfulfilled desire for modern legitimacy. Again, the deeper anxiety surrounding modernity cannot be verbally located, for the gap between the subject and a part of the subject that is projected onto the Other has formed when the subject enters the symbolic world. From this perspective, Inoue's listening subject does more than redirect indexicalization from speakers’ self-styling to listeners’ interpretation. It highlights the instant when those male intellectuals, hearing the voices of schoolgirls, are hailed by the voices’ residue or excess and thereby are constituted as listening subjects. For Inoue, the payoff is not in logging which speech features are linked to which negative meanings, but in tracing the process of subjectification through which interpreting the Other's voice becomes a strategy for coping with an ineffable, modernity-induced unease. Put differently, Inoue's concept foregrounds how listeners negotiate their own position in the social world: they situate pre-sign, pre-logical affect—the raw emotions triggered by another's voice—within the power relations that bind hearer and speaker. “Listening,” in this sense, is the work of translating that affect into discourse, and any overt commentary on the heard speech can only ever partially embody the listening subject who articulates it. Variationist research on sociolinguistic perception has taken a different path. Using the speaker evaluation paradigm, researchers play listeners carefully manipulated speech variants and ask them to deliver social evaluations of the speaker, or present listeners with social information and invite them to identify the sounds they hear. Recent studies foreground how listeners’ own social biographies shape those evaluations. For instance, Wong and Babel (2017), in this virtual issue, show that prior exposure to ethnolinguistic variation improves listeners’ accuracy in identifying the social group indexed by an accent. Other studies may further foreground the role of ideology in mediating indexical interpretation. For example, Jessee and Calder (2025) develop the notion of the cisgender listening subject to highlight how cisgender participants in their linguistic perception experiment are ideologically predisposed to expect transgender speakers to be more hyper-normative in their speech than cisgender speakers (i.e., trans women sounding even more feminine than cis women, and trans men sounding even more masculine than cis men). Such expectations, in turn, shape how these participants categorize the phonemes of transgender speakers’ speech. Here, the concept of the cisgender listening subject underscores that listening is neither neutral nor passive, but is shaped by the listener's gender ideologies. Collectively, this line of work foregrounds listening as a social practice of interpreting linguistic variation into a social semiotic system. If we treat Inoue's listening subject as a theoretical construct that sees the listener not as a passive recipient of a ready-made sign–object link but as someone who exerts a socioculturally mediated capacity (Ahearn 2001) to interpret that link—then linguistic anthropology and variationist sociolinguistics do, in this sense, converge. Yet at its core, the concept of the listening subject concerns the processes of subjectification that make “listening” possible in the first place. In sociolinguistic perception experiments, however, such processes are often technically harder to capture, since listeners are typically restricted to a limited set of researcher-designed options—whether a forced choice between phonemes, a slider on a Likert scale, or eye-tracking of competing visual stimuli. Even when accompanied by an open-ended question intended to generate a word cloud, the method itself leaves the subjectification process underexplored, as what is observable is largely the end product of listening—a tip of the iceberg in relation to the full scope of the listening subject. Even when researchers document how sociolinguistic evaluation shifts with listeners’ ideological orientation to, say, the speaker's ethnic group—the data do not reveal how the listener becomes the listening subject, hailed by a voice to interpret it in relation to themselves. By contrast, linguistic anthropological work often relies on spontaneous discourse, as in Inoue's data on Meiji-era male intellectuals’ unsolicited metapragmatic commentary. Similarly, Reyes (2017) shows how elite Philippine netizens position themselves as a listening subject that claims intimacy with its object of critique, labels the targets as the “wrong” kind of elite Filipino, and brands themselves as the “right” kind. The nature of spontaneous discourse allows researchers to probe the nuances of how listeners position themselves vis-à-vis speakers when making indexical interpretations. We make a similar point in Su et al. (2025), which examines how Taiwanese netizens interpret a government official's pronunciation of a high-falling tone as a mid-falling tone as indexing the People's Republic of China's encroachment on Taiwan's democratic polity—and how others, in turn, meta-interpret such linguistic policing as a betrayal of Taiwan's political value of liberalism. These metapragmatic moves vividly reveal how listeners imagine themselves, and by extension, the imagined community of Taiwanese, when engaging in indexical interpretation. The concept of the listening subject is not always useful, especially for power relations where the dominant systematically erase the existence of the dominated. An example is disabled people. Deaf accents are routinely misrecognized (cf. Hall and Parish 2024). In perception experiments, phonetic features of deaf speech are judged as index of illiteracy or cognitive impairment (Cowie and Douglas-Cowie 1992); in Taiwanese online comments to a live streamer whose deafness was not publicly disclosed, his pronunciation was mocked as “fake” or “stupid”; in everyday encounters, deaf speakers are asked whether they are foreigners (Wan 2024b). A variationist instinct would be to run, for example, a matched-guise study: isolate particular phonetic features of a “deaf accent,” combine them with visual cues such as skin tone or clothing, and test whether hearing listeners map those cues onto specific nationalities. Or a variationist approach that highlights listener agency can probe into the subjective bias against deaf people among hearing individuals. During my PhD exploration, I ultimately set that perception study aside. Although such experiments highlight listener role—reminding us that communicative responsibility is shared, not speaker-bound—they do little to expose the ableist structures that crip linguistics urges us to confront (Henner and Robinson 2023). As Wong and Babel (2017) demonstrate, exposure to diverse ethnolinguistic groups improves indexical identification; yet most hearing people who stigmatize deaf speech have never met a single deaf person and assume deaf people simply cannot talk. To them, “deaf speakers” do not exist. Irvine (2002) notes a parallel in Wolof: the absence of a label for “slave style” reflects an ideological erasure of enslaved voices. Likewise, before Japan produced the “unprecedented category” of educated schoolgirls, Meiji men had no reason to comment on women's speech; only when women's education threatened male privilege did those men become listening subjects, summoned by teyo–dawa and compelled to police it. In this light, hearing a deaf person's speech rarely turns a hearing listener into a listening subject. From the listener's standpoint, deaf people remain outside their perceptual horizon and thus never hail them. Lulu—a deaf Taiwanese woman I interviewed—illustrates the point. While visiting a rural town, she was asked by an elderly woman what country she came from. Lulu replied, “I am Taiwanese,” but the woman insisted she must be a foreigner. When Lulu explained that she was deaf, the woman found the claim absurd—she said nothing and walked away. The encounter convinced Lulu that most hearing people neither realize deaf people can speak nor know how to speak with them. Due to socially systematic forces, hearing individuals have been deprived of the capacity to perceive any potential indexical link between deafness and speech (cf. Wan 2024a). Hence, empirically, deaf speech is present; ideologically, they are And what is ideologically cannot become the object that a hearer into the position of a listening subject. In this virtual issue, in her of the at shows that a language is not simply a matter of one of with another (e.g., visual to are by social and which make some natural and not by how to linguistic but by in and the so that could From this perspective, the of deaf speakers as foreigners can be as the of a In most hearing deaf speech has never been as a This that hearing people have never become listening subjects to the voices of deaf but to an imagined this as in the would first for and the social so that deaf speech becomes an within the listening can we to the question of what of subjectification are when hearing people to (and make comments deaf speech. In the variationist of sociolinguistic perception and the anthropological of the listening subject can Variationist have to a between and modes of sociolinguistic and (2025) show that even when listeners are explicitly the they is to the voice they their in a matched-guise is from that of a group no such Although their listeners were not asked for social but to identify an a the that a about who is linked to which linguistic does little to the of indexical Campbell-Kibler (2021) a similar listeners remain by cues even being instructed that those cues are to the these perception experiments that sociolinguistic that cannot fully the of the sociolinguistic perception studies show that and processes do not map onto one Using eye-tracking to capture D'Onofrio (2018) found that but did not a yet in a more evaluation it was the than the A similar appears in and the link between and they that in a matched-guise a of to even discourse labels can be of an indexical link at the while largely to it when speech more Inoue's “listening subject” into sociolinguistic perception work a method of the pre-logical affect that arises before listeners any social As Campbell-Kibler reminds us, existing research on person perception in can be such a in the which bias by how a a that one of which has been in anxiety research (e.g., et al. In studies of racial the a participants who are more socially not to racial bias toward a Black at of about but when the for and If we could an experiment to that with the Meiji male elites, translating this into the of language, we ask whether those men would a toward a Japanese woman a Japanese woman when teyo–dawa orientation that could an anxiety modernity. In the negative metapragmatic commentary to a extent, be about the men themselves. An for this question is to listeners’ during speaker evaluation In a matched-guise experiment from my (Wan to the of participants were asked to how they their own speech to This with the between social evaluation and the more a listener their own the more they of a by Taiwanese As and such to shifts reflects sociolinguistic are not passive whose mirror their by they position themselves in relation to the they not that kind of Taiwanese That is, not only speakers may linguistic variation as a social semiotic in speech to construct of but listeners make of linguistic variation in their listening to themselves within the semiotic To construct a perception study by the concept of the listening subject, we must by that, at its core, the listening subject is shaped by that are and thus cannot be fully by studies can only this core, it to the kind of psychoanalytic that Inoue on Meiji male on this it is to and like metapragmatic not as but as and through which listeners engage with the listening subject does not interpret language they interpret the voices of as through which they negotiate their own social position in the world. we must approach to listening as a social practice to capture the process of that the into In we can further understanding of how the socially fail to speak of the the dominant I am to and for their on the first of this commentary. The no of not to this as no were or during the