Catcalls and Unwanted Conversations

Abstract

Catcalls have been said to insult, intimidate, and silence their targets. The harms that catcalls inflict on individuals are reason enough to condemn them. This paper argues that they also inflict a type of structural harm by subordinating their targets. Catcalling initiates an unwanted conversation where none should exist. This brings the rules and norms governing conversations to bear in such a way that the catcall assigns their target a “subordinate discourse role.” This not only constrains the behavior of the target here and now, but also influences the norms governing future conversations. Catcalls are then not only bad because of the effects on their target, but also because of their pernicious contribution to the wider normative landscape.

Men behaving badly

Men say all sorts of things to women in public. Some comment on body parts (“Nice ass!”), some on attitude (“No need to be sad!”), some on the weather (“Well now it's a nice day!”). Some make assertions (“Looking good!”), some ask questions (“You like what you see?”), some give instructions (“Smile, beautiful!”). Some intend to compliment their targets (Lennox and Jurdi-Hage 2017); others do not (Gardner 1989). Many are far less polite than these examples.

And this does not only happen on the street—men say these things in shops, in schools, waiting for the bus, on the bus, after getting off the bus … Some are loud for all to hear, others whispered sotto voce. This behavior is, to put it mildly, rather unpleasant.

These are catcalls, and as these examples illustrate, catcalls have no uniform shape. This makes a uniform analysis difficult (Vera-Gray 2016). Reactions to catcalls are similarly varied. While many targets feel objectified, humiliated, intimidated, or degraded, some are able to ignore them. A few even find such comments affirming (and women are often told to take this harassment as a compliment).

A uniform analysis of the harm of catcalling is thus also difficult. The most keenly felt harms of catcalling are those inflicted on the target—emotional, psychological, and dignitary. But these harms may seem contingent on the target's reaction. I think that token catcalls are bad regardless of how their target reacts, so in this paper I identify a wider structural harm of catcalling. I argue that catcalls subordinate their targets by assigning them a “subordinate discourse role” derived from a sexist social hierarchy (borrowing from Popa-Wyatt and Wyatt's 2018 analysis of slurs). If token catcalls strengthen gender-oppressive norms and expectations, then they constitute acts of oppression (invoking Langton 1993; Langton and West 1999; McGowan 2009).

In the next section, I describe some central features of paradigmatic gendered catcalls. After that I discuss how these flout Goffman's (1963) “norm of civil inattention” and so visit further individual harms upon their targets. Then I explain how a wider structural harm—conversational subordination—is inflicted, and finally show how this infects future conversations.

What catcalls are (and what they are not)

Catcalls come in all manner of forms—my focus is the paradigmatic gendered catcall (sociologists often use broader terms such as “street remark” or “street harassment,” see Gardner 1980, 1989; Bailey 2017). McDonald (2022) defines catcalls as “unsolicited remarks or gestures made in public, usually by a man to a woman or LGBTQ person he doesn't know, often concerning the latter's appearance” (208). This outlines the phenomenon under discussion, but some further clarification will be helpful.

Catcalls are unsolicited; while sexist utterances that occur during an existing conversation may similarly intimidate, humiliate, and degrade, these are subject to (some) different conversational norms and so require a different analysis. The unexpected, unsanctioned intrusion of a catcall is part of what makes these distinctive (Gardner 1989; Bailey 2017; McDonald 2022). Similarly, private sexist utterances will interact with social norms in different ways to public utterances, although there will be some overlap.

The gendered catcalls I consider here are those delivered through spoken comments, although similar effects can be achieved with gestures, non-verbal noises (such as grunts or kissing sounds), and even directed gaze (Gardner 1980). These non-verbal cues are still meaningful—the target is supposed to understand what is meant by a kissing sound, a sexualized grunt, or a lingering stare. While comments on the target's appearance are common, catcalls cannot be defined by their content alone (McDonald 2022).

So, some of the central features of gendered catcalls are that they are unsolicited, public, and meaningful. But a little more should be said about how they are gendered.

There is a difference between a man catcalling a woman, and a woman saying the same thing to a man (for this example, all are heterosexual). They might say the same words—“Hey, sexy, nice ass!” Both targets might be in the same situation—walking from their home to the bus stop. Both utterances might have the same effects on the target, who could feel objectified, humiliated, and threatened (or, plausibly, pleased with their appearance). But in a patriarchal society, with an ever-lurking threat of male violence, there is something worse about the man's utterance. McDonald (2022) argues that men's catcalls silence and exploit their targets, who are unable to respond in the way they would prefer because of this threat. I will argue that the man's utterance leverages the communicative norms of a sexist society to subordinate his target.

I focus on gendered catcalls because they are not simply unsolicited sexualized public utterances—if they were, the identities of speaker and target would not matter so much. And it is not just that men's catcalls of women are more harmful, but that they can enact a different kind of harm—one derived from the sexist context in which they are uttered. What matters for such catcalls, I think, is that they are relationally gendered. Gendered catcalls make apparent not only the target's gender, but also the speaker's gender and the unequal relationship between genders (and seem to implicitly endorse that unequal relationship).

For example, a woman might say to another “You'd better watch out, dressed like that!”, admonishing the target for her attire and making her gender salient. But this is not really a catcall. If a man were to say the same thing, he invokes the possibility that it is he who the target must watch out for—from his mouth, I think this is a catcall. It is not the content that is distinctively harmful (even if it is victim-blaming regardless of who utters it), but the relationship between the comment, commenter, and the edifice of gender oppression. This is why utterances like “Give us a smile!” or “Cheer up!”, when used by men to police women's facial expressions in public, should be considered catcalls, even though the same words said to an unhappy child may be innocuous. In public, in a sexist context, when said by a man to an unfamiliar woman, such comments call on the target to look happy for the speaker, highlighting the patterns and practices of patriarchy.

So, gendered catcalls foreground not only the genders of speaker and target but also the relationship between them: that the speaker is a member of the dominant group in a subordinating social hierarchy, and the target is not. Gendered catcalls then have something important in common with transphobic and homophobic catcalls (and racist street remarks) that they do not share with in-group gay “catcalls” or those from women to men. I expect that the analysis in this paper will generalize more easily to the former than the latter. “Intersectional” catcalls, where speaker and target may belong to multiple relevant groups, will interact with similar social structures but in subtly different ways. I hope to show that gendered catcalls inflict the sort of structural, subordinating harm outlined in the introduction (from here on, unless stated otherwise, by “catcall” I mean these gendered catcalls).

But, many catcallers will say, catcalling is not harmful—it is complimentary (Bailey 2017). Compliments are nice and, so the inference goes, nice things are not harmful. These men are just complimenting women—where could be the harm in that?

There are lots of ways that compliments might be harmful. “You and your boyfriend looked beautiful last night” might be a lovely thing to say in some contexts, but in others, might cause considerable harm—such as by outing a gay man. As McDonald (2022) says, not all compliments are benign. Even when a compliment is intended, expressing admiration does not get the speaker off the hook for causing harm, nor for being inattentive to the likely effects of their words. But this concedes to the catcaller that they do, in fact, pay their target a compliment. McDonald (2022) bites this bullet, accepting that (some) catcalls are compliments, while arguing that they are nonetheless harmful.

I think we can leave this bullet unbitten.¹ Even catcalls with the appearance of a compliment might not function as a compliment; the illocutionary force of an utterance is not strictly determined by its “surface grammar” (Kukla 2014).² Something seems missing from a catcall that genuine compliments should have. Holmes (1986) suggests as a preliminary definition that compliments are “a speech act which explicitly or implicitly attributes credit to someone other than the speaker, usually the person addressed, for some ‘good’ (possession, characteristic, skill, etc.) which is positively valued by the speaker and the hearer” (485).

Many purportedly complimentary catcalls appear to fit this definition—at least until closer inspection. A catcall like “Hey sexy!” might seem to compliment the target's sexiness, and both speaker and hearer might value the “good” characteristic of sexiness. But we tend to value sexiness from a perspective; a teacher might want to be considered sexy by their spouse, but not their students. If the target of a catcall does not value sexiness-judgments of male strangers in public, then an utterance that might in a more intimate situation count as a compliment would misfire (to use Austin's 1962 term) when shouted across the street. The characteristic cited (sexiness in the eyes of the speaker) is only valued by the speaker, and not the target—and so the catcaller fails to properly perform the speech act of complimenting. This could be developed further (for example, to account for those who find catcalls affirming), but we might not need to concede to the catcaller that they have complimented their target. There is more to compliments than saying the right words, and whatever this is, catcalls may not have it.

Catcalls are unsolicited public remarks that draw attention to the unequal relationship between the genders of speaker and target. This can be intimidating, humiliating, enraging, and upsetting. But these effects will not occur after all catcalls, and not always to the same extent. Perhaps, though, there are other ways that catcalls inflict harm that are not contingent on their target's resilience.

Catcalls as silencing and exploitative

In one of the most extended philosophical analyses of catcalls (sociologists have taken the lead here), McDonald (2022) argues that catcalls exploit and silence their targets by taking advantage of unjust social norms and sexist contexts. These harms are supposed to occur with every catcall, even those that seem complimentary, regardless of the psychological effect on the target (my own account later uses some of the same concepts to explain wider social harms).

McDonald's argument draws on a longstanding recognition that catcalling is distinctive in part because it transgresses what Goffman (1963) calls the “norm of civil inattention” (Gardner 1980, 1989, 1995; see also Bailey 2017; Hesni 2018). In public, we are normally expected to leave each other alone, and only engage with strangers if they are of “special curiosity or design” (Goffman 1963, 84). Gardner (1980) suggests, for example, someone wearing a costume, hopping on one foot, or carrying a sofa. Paying attention to strangers in those circumstances would not flout the norm of civil inattention as they are doing something especially curious. But otherwise, we are supposed to observe the norm of civil inattention and refrain from initiating conversations with strangers, treating them as “participants in the gathering and not in terms of other social characteristics” (Goffman 1963, 86). Goffman notes that, in some contexts, such as social clubs or sporting crowds, this norm may not apply.

But catcalling does not fit these exceptions. Women in public are not “of special curiosity or design,” nor are they in a special context like a social club; they are simply women daring to traverse a patriarchal public space. The norm of civil inattention should apply. Catcalling is then a kind of discourtesy, treating the target not as equal participant in the public space but in terms of another “social characteristic”: being a woman.

The norm of civil inattention has certain advantages in protecting people from unwanted interference and reducing the exposure of being in public places. But, as Gardner (1980) argues, many people have (historically) not enjoyed (full) protection from this norm; for example, people of color being treated as “open” for comment. It also lets us ignore those who might need our help (unhoused people, those asking for money, people in distress)—we can tell ourselves that we are right to ignore them. We are simply minding our own business, observing the norm of civil inattention. But whether or not it is all-things-considered a good norm, it does seem to inform behavior as Goffman describes—at least in many urban contexts.

McDonald argues that catcalling presupposes that the speaker has the authority required to ignore this norm. A catcall is a call, and so demands acknowledgment from the target (Lance and Kukla 2013). To demand acknowledgment, one should have a certain kind of authority for the call to be felicitous. We might call to our friends, and we gain the required authority to do so from the conventions governing the friendship (i.e., a reciprocal right to call to each other). Teachers may make calls of students, police officers of people “loitering,” and so on—in having the right to demand someone's attention, the speaker is supposed to have some kind of authority that empowers them to over-rule the norm of civil inattention.

Catcalling then presupposes that the speaker does, in fact, have the authority required to make the call. The catcaller acts as though they are authorized to flout the norm of civil inattention and demand the attention of their target. Much has been said about how the presupposition of authority might be accommodated (Lewis 1979; Langton and West 1999; Witek 2013; McGowan 2019). To rehearse an example adapted from Austin (1962), a group of people are shipwrecked on an island; one orders the others to gather supplies (Maitra 2012; Witek 2013). The speaker had no authority before speaking, so they should not be able to issue an order. But if the others go along with it (following the order), the speaker seems to acquire authority by presupposing that they have it. If no one objects, it becomes (conversationally) appropriate for the speaker to continue issuing orders and for others to follow them.

But of course, people can object. McDonald notes that ordinarily (following Langton 2018), we can block this kind of presupposition, preventing it from becoming part of the common ground. In the shipwreck example, someone could have spoken up, perhaps saying “Hang on, you have no right to order us around!” This would block the presupposition of authority. It does not become (conversationally) appropriate to follow the orders; instead, it becomes mutually recognized and salient information that the speaker does not have authority. Their utterance no longer functions as an order.

So, we might be tempted to say that the targets of catcalls could similarly block the presupposition of authority by the catcaller. They might say “You have no right to call out to me!”, appealing to the norm of civil inattention and making it clear that the unauthorized catcall is violating that norm. But, McDonald cautions, this would ignore the violently misogynistic context in which catcalls occur. The targets of catcalls rightly fear reprisal should they respond in the “wrong” way. Ignoring the catcaller, rebuking them, mocking them, playing along … depending on the catcaller, any one of these might dangerously escalate the situation. Blocking a catcall is often an unjustly risky proposition. It is a mark of the injustice of patriarchy that women risk worse consequences for their response to a catcall than the man does for uttering it.

The targets of catcalls then cannot block them, and instead have to respond to them as a call—responding as though the speaker actually had authority (i.e., not telling the catcaller to “Piss off, dickhead!”). This exploits the target, as the possibility of violence forces a response that seems to grant the speaker the authority they had presupposed—leaving the target in a “degrading” position (McDonald 2022). It is not that the speaker gains real authority this way, but they get away with acting as though they had it. And as it is too risky for the target to respond in the way they would prefer, they are silenced—unable to perform a blocking conversational move because of the misogynistic context they inhabit. Catcalls then exploit and silence their targets—two harms unrelated to the catcall's content.

This might often be a fruitful way to analyse the harm of catcalls. But there is a worry that it might overgeneralize. A child says to a parent about to serve dinner “I order you to give me dinner!”; the parent gives the child dinner. This is not because of the child's utterance, but because even cheeky children should be fed. It would be wrong to say that the child successfully ordered the parent—the child lacks the authority, or standing, to issue orders (Hesni 2018; McGowan 2019). But in responding this way, the parent appears to act as though the child had the authority to order (and is forced to act in this way because leaving children hungry is frowned upon). On McDonald's analysis, it may seem that the child has exploited the parent. But here the power dynamics between speaker and target are entirely unlike the catcalling case. There is no coercive threat behind the parent's choice to act in this way—they would have provided food regardless of anything the child might have said. If this is also a case of exploitation, then it is less clear that exploitation is distinctively harmful.

But even if this worry was resolved, these harms are individual. The target of the catcall is exploited and silenced here and now (unless they ignore the risk and respond as though the speaker does not have authority—“Get lost, creep!”); the psychological effects linger but are limited to those who heard the catcall itself. I think that McDonald's overall project is right. The harm of catcalls is often independent of their content. But token catcalls are not only bad for their individual targets. To fully explain the harms of catcalling, we need to recognize both the individual and structural harms they inflict (McDonald's focus is the former; mine is the latter).

In the rest of this paper, I argue that the norms and structures of a patriarchal society enable catcallers to not only offend, humiliate, and exploit their targets, but also to subordinate them (the norms of other oppressive hierarchies would similarly empower homophobic and transphobic catcalls). This reinforces those unjust social structures themselves. Catcalls then inflict diffuse and ongoing harm, in addition to their (sharper and more keenly felt) individual and immediate harms. The explanation of this wider harm of catcalling begins, as for Gardner (1980), Bailey (2017), and McDonald (2022), by examining how catcalls violate the norm of civil inattention.

Unwanted conversations and discourse roles

In flouting the norm of civil inattention, catcalls initiate an unwanted conversation. This brings to bear the rules, norms, and niceties that govern our conversations (Hesni 2018 explores the difficulties of ending these unwanted conversations). In turn, this enables the speaker to inflict uniquely conversational harms.

In this section, I argue that in initiating a conversation with a catcall the speaker assigns their target a “subordinate discourse role” (drawing on Popa-Wyatt and Wyatt's 2018 analysis of slurs). This subjects the target to conversational norms restricting their behavior and assigns the speaker a related dominant discourse role. Popa-Wyatt and Wyatt (2018, 2895) call this an “unjust … power grab”, as the speaker maneuvers themselves into an undeserved position of dominance. Catcalls are then not only moves in the new (unwanted) conversation, but also in what McGowan (2009, 402) calls the “rule-governed activity of gender oppression,” changing the rules governing the behavior of their targets. This means that in addition to their psychological effects, catcalls constitute acts of subordination.

Oppression is the unjust systematic subordination of a social group (Matsuda 1993; Frye 2000; Haslanger 2004; McGowan 2009). Frye (2000) describes it like a birdcage. No individual wire by itself constrains the bird, but once enough are arranged in the right way, the bird is trapped. Similarly, no individual rule or norm itself makes it the case that a group is oppressed, but once enough rules and norms are placed in the right way around a social group, we can say that the group is oppressed. There are lots of ways that subordinating rules or norms could be enacted, such as through the passing of legislation or public policy unjustly constraining the options available to members of the target group. Laws prohibiting abortion would, on this view, subordinate women and thus be aptly described as oppressive.

Norm-enactment can also happen at the interpersonal level—and this means that subordination can happen interpersonally too. Many of the norms connected with catcalling are conversational; my argument is then that catcalls conversationally subordinate their targets. Langton (1993) describes some of the ways that speech can subordinate: by ranking targets as inferior, legitimating discriminatory behavior towards targets, and depriving targets of certain powers (303). This confers to the target an “inferior civil status” (Langton 1993, 297).

Subordinating norm-enactment can also be sneakier. McGowan (2009, 2019) argues that small shifts in conversational permissibility also change the rules governing a speech situation in ways that unjustly constrain the target. Hate speech, for example, makes it more permissible to say other nasty things about the target of the hate speech. This is, of course, a much smaller act of subordination than depriving the target group of the right to vote—it is a bit like one small piece of wire that is part of the much larger birdcage. Catcalls, I will argue, subordinate in a similar way; in the next section, I explain how they help to sustain the “rule-governed activity of gender oppression.”

Conversations are complex activities, governed by a vast array of norms and rules determining who speaks, how they speak, and what they say. I should not, upon meeting the Queen of Denmark, exclaim “Oh my god, I love Bohemian Rhapsody!” This utterance would, on the other hand, be appropriate when meeting a member of the band Queen. Ayala-López (2018) argues that the norms governing conversation afford situated interlocutors a limited range of possible actions. Not all participants in a conversation are permitted, according to these conversational norms, to do the same things. So, what are the conversational norms that catcalls interact with to constrain their targets and thus constitute acts of subordination?

Goffman (1967) notes ways in which conversation proceeds according to “interaction pairs”, with each utterance requiring a response. Conversational norms often require that we take it in turns to speak, lest conversation become unruly. A greeting requires reciprocation. A question requires an answer. A compliment requires thanks:

Hello!

Hi! How are you?

I'm well. Nice umbrella!

Thanks!

Many archetypal catcalls, at least on the surface, are paired in this way (Bailey 2017). “What's up, sexy?” asks a question; conversational norms dictate that questions are supposed be answered. “Nice ass!” is phrased as a compliment, and those complimented are supposed to respond with thanks. This demands something more specific from the target than mere acknowledgment of the catcall. The phrasing matches the first half of an interaction pair to make it the case that the target is supposed to give a certain kind of response. By contrast, if I pass someone who exclaims “There's that coin I dropped!,” I am not supposed to respond in any particular way, even if they have made it clear that their utterance is directed at me (to explain why they are scrabbling around on the pavement). The utterance is not part of an interaction pair.

Bailey (2017) explains that the initial utterance in a catcalling interaction pair is not correctly delivered, as it is uttered in a context where the norm of civil inattention is supposed to be observed. Nonetheless, the listener faces a “powerful normative pressure” to deliver the paired response—an answer to a question, a thanks for a compliment, and so on. Targets of catcalls are in a double, or even triple, bind (see also McGowan 2018). If they deliver the paired response, they engage with the catcaller and tacitly endorse his behavior. If they respond in a different way, especially if they confront the bad behavior of the catcaller (“Shut up, jerk!”), they risk violent reprisal. And if they say nothing at all, they might either be taken to endorse the behavior or risk violent reprisal for ignoring the catcaller.

So, catcalls initiate a conversation, often with an utterance that requires a paired response. This means that norms governing conversation and structuring behavior (such as rules about whose turn it is to speak) are now in play. Some actions are now appropriate (replying), whereas other actions are now inappropriate (ignoring the speaker). But the same is true of many conversation-starters. If someone approaches me on the street and says “Do you know the way to the station?,” they likewise begin a conversation, flouting the norm of civil inattention with an utterance that requires a paired response. But this is hardly structurally harmful. Why, then, are the unwanted conversations started by catcalls so bad? This can be explained with Popa-Wyatt and Wyatt's (2018) notion of discourse roles.

Much of our conversational behavior is governed, at least to some degree, by the social roles of conversational participants. A student is supposed to defer to a professor, to use certain honorifics, to speak politely, and so on (although precise rules in play can be negotiated—the professor might suspend the need for honorifics by saying “Just call me Whitney”). But even though these norms will often govern conversation between a student and professor, this will not always be the case. Perhaps they play for the same local cricket team. During training, or cricket matches, the student would not need to defer, or use honorifics, in the same way that they would in the classroom. This is because the rules governing conversational behavior are not determined precisely by general social roles, but by more specific discourse roles that are enacted during the conversation. The people in this example have the social roles of student/professor and teammates, but at a single point in time, typically only one of those will be active as a discourse role. Sometimes, they converse as student and professor, other times as teammates.³

Popa-Wyatt and Wyatt argue that which role speakers are supposed to observe is tracked by the conversational score (Lewis 1979). The permissible (i.e., intelligible) conversational moves available to participants are determined not only by general rules governing conversation, but also by specific information about what has happened in this conversation so far. For Lewis (1979), this is like a game of baseball. “Correct play” depends not only on the rules of the game but also on the score. A batter in baseball retires after three strikes, not two—so to know if they are supposed to retire, they need to know the score (how many strikes have they made?). In the same way, a speaker who responds to “Let's meet back at the house!” needs to know not only the conversational rules, but also the conversational score—which includes information about which house is most salient right now and is therefore referred to by “the house.”

The student and professor will sometimes need to follow the conversational rules befitting their academic roles, and at other times, their sporting roles. Which roles are “in play” according to the conversational score can change during a conversation. For example, they might meet in the supermarket—neutral ground that does not obviously require either academic or sporting roles. By saying “Don't forget to finish your assignment!,” the professor could make their academic social roles salient. In doing so, they assign themselves the discourse role of professor, and the student the discourse role of student. Alternatively, they could say “Well played on the weekend!” to assign their sporting roles. The social roles are always in the background; the discourse roles let conversational participants know how they are supposed to interact right now.

Once the academic roles are “in play,” it would be inappropriate (i.e., incorrect play) for the student to use the kind of informal language to address the professor that they might use during a cricket match. But once the conversation ends, so does the discourse role assignment. When the student and professor next speak at their cricket training, the sporting roles would be more salient and would determine how formal that conversation should be. Unlike background social roles, discourse roles are short-lived, ending with the conversation.

Popa-Wyatt and Wyatt deploy this framework to explain how slurs subordinate their targets. Two speakers might begin a conversation as equals (although real-world conversations will almost never be perfectly balanced). If one is from a subordinated social group, there will be background social roles that could be made salient to activate them as discourse roles, changing the rules governing how speakers are supposed to behave. Uttering a slur is a very straightforward way to make those social roles salient and thus assigns the related discourse roles. Before the utterance of the slur, conversational participants may have had equal (conversational) status; after the slur, the target has a subordinate discourse role and the speaker a related dominant role. The conversational score is updated—now, according to the norms of conversation and information about what has happened in this conversation so far, the participants should no longer be treated as equals (perhaps the subordinated speaker should defer to the other or use overly polite language when disagreeing with them). Popa-Wyatt and Wyatt call this an “unjust power grab”—the slurring speaker gifts themselves a position of (relative) dominance in a racist milieu and forces their target into a subordinate position.

Speakers do not need to begin conversation with equal status for slurs to allow this kind of “power grab” (Tirrell 2018, 25 argues that there are no such universally neutral situations, as identity factors are always salient and the role that they play is highly dependent on context). A white professor could use a racist slur in conversation with a Black student, and in doing so exacerbate the existing status imbalance. A white student could use a racist slur in conversation with a Black professor, and in doing so reduce or reverse an existing (justified) status imbalance by making racial discourse roles more salient than the academic roles that should be in play. This is further complicated by the intersectional identities of real-world speakers (who are not hypotheticals in philosophy papers). For example, there will be different power imbalances in play if one (or both) are women, or gay, or trans. Context will also have an effect; background social roles indexed to race, gender, and sexuality are unlikely to be precisely identical in Toronto, Tehran, and Tokyo.

Plausibly, a feature of oppressive contexts is that some speakers are “automatically” given an unjustly higher status than others. For example, Tirrell (2018) argues that in patriarchal contexts, entering a conversation as a woman limits the available moves they can make by undermining their discursive authority (see also Ayala-López 2018). Different sets of limitations would be afforded to different intersectional identities; entering a conversation as a Black women will often not be identical to entering it as a white woman, and the extent to which discursive authority is undermined will differ accordingly.

So, when a white person uses a racist slur in conversation with a Black person, they may already (unjustly) be enjoying a dominant discourse status that is assigned by “default” in racist contexts—the slur would exacerbate this, making racial roles more salient than they were at the (unequal) beginning of the conversation (more discussion of “default” role assignment follows in the next section). Slurs need not only shift conversational norms from “not-oppressive” to “oppressive”; they might also move them from “oppressive” to “more-oppressive.” They can thus constitute acts of subordination even in already-oppressive contexts, rather than reflecting oppressive norms already in place. I have argued previously that reinforcing oppressive norms makes them stronger and so itself constitutes a type of subordination (Cousens 2020).

I think that gendered catcalls work similarly in relation to the “activity” of gender oppression (and that transphobic and homophobic catcalls work similarly in related activities of oppression). Slurs and catcalls are not the only ways to assign subordinate discourse roles, but they do so particularly effectively. So, with a catcall, the speaker does not merely initiate an unwanted conversation in a patriarchal context. In flouting the norm of civil inattention and beginning a conversation in a relationally gendered way (“Hey sexy!,” “Nice ass!,” “Give us a smile!”), the speaker makes the (apparent) genders of the speaker and target salient, in a similar way to a racial slur making the race of the target salient. What was previously background information (a social role) is now front and centre, and part of the conversational score (a discourse role). By making the target's gender salient in a context where that gender is subordinated, the speaker assigns the target a subordinate discourse role, which brings with it norms governing their participation in the conversation (deference, etc.). By also making the speaker's gender salient (such as by sexualizing their catcall) they “grab” the related dominant discourse role.

Not just any gendered comment brings these discourse roles to bear. If a man says to their colleague (a woman) “I heard that Jones kept calling you ‘Miss’ instead of ‘Doctor’,” they make the woman's gender very salient. But they do not do this in the same way as a catcall, which invokes gender relationally. The catcall makes not only the (apparent) genders of speaker and target salient but also the unequal social statuses of those genders and the fact that the target belongs to the subordinated group.⁴

This does not happen when a woman calls out “Hey, sexy!” to a man. She might similarly objectify, belittle, and demean her target—but she does not make it apparent that she belongs to the dominant gender in a sexist society (because she does not). And she does not subordinate her target by activating a subordinate background social role as a discourse role; men as a social group are not subordinated in patriarchal societies.⁵ This does not mean that her utterance is harmless—objectifying, belittling, and demeaning people seems bad whether or not it also subordinates them—but it does mean that her utterance does not play the same role in the wider practice of gender oppression that it would if said by a man. Her utterance does not inflict the same sort of structural harm (reinforcing the metaphorical birdcage of gender oppression) as his.

A similar analysis would, I think, explain transphobic catcalls. The caller and target's statuses within a social context where trans people are subordinated is made salient in a way that highlights the caller's membership of the dominant group and the target's membership of the subordinated group. A cisgender person who shouts out “What's under the dress?” to a trans woman makes relative group membership salient, activating background social roles as discourse roles. They make social roles salient in a way that makes it clear who is more powerful—the speaker who flouts the norm of civil inattention and reminds their target that such norms do not protect them from unwanted interaction in public places.

So, catcalls initiate an unwanted conversation. The way that this is done also assigns a subordinate discourse role to the target (and a related dominant role to the speaker). These roles change the rules governing how participants are supposed to act for the rest of the conversation. What counts as “correct play” for each will be different. Women might be expected to defer to men (at least on some “manly” topics), to wait for men to speak, or to qualify their assertions. This is a very truncated description of the norms and expectations assigned by actual discourse roles, but hopefully it illustrates the kind of constraints that a subordinate discourse role might place upon the target of a catcall. The focus here is on the mechanism used to assign these roles, rather than the full description the rules and norms that make up the roles themselves.

This might help to explain why men catcall without the expectation of obtaining sex or intimacy. Catcalling to initiate an amorous encounter has, anecdotally, a very low rate of success. But the catcaller gets something else—the power grab—that comes with the newly assigned dominant discourse role. It might be satisfying to feel powerful, even though such power is undeserved. Berdahl (2007) demonstrates that in a similar way, the majority of sexual harassment is driven not by sexual desire, but to punish women who “step out of line” (see also Manne 2017).

Many catcalls are never responded to. I have suggested that catcalls initiate a conversation, but a lot of street harassment is a single comment “left hanging.” A person walks past a construction site (in the most cliched of examples); a worker yells out something, but the target keeps walking. The paired response is never delivered—although Hesni (2018) shows how difficult it can be to end these kind of conversations, and McDonald (2022) outlines the potentially violent response of men whose catcalls are ignored. I think that, even when no response is given, the same kind of discourse role assignment occurs.

While conversations paradigmatically involve back-and-forth between speakers, some end abruptly. If I begin speaking on a telephone call, but the call cuts out before the recipient can answer, I still began a conversation. If what I say assigns a discourse role (such as by mentioning that the listener, a student, needs to complete a report), that role assignment still happens even if the student never has a chance to reply. In flouting the norm of civil inattention, the catcaller begins a conversation and assigns discourse roles even when this is the only utterance that occurs in the (abruptly ended) conversation. And, as the next section argues, this can influence the norms governing future conversations as well.

If this is right, then one of the harms inflicted by catcalling is subordination. Catcalls create a conversation, and once it exists, conversational rules (including discourse roles) apply, governing “correct” behavior for participants. By gendering their catcall, whether through sexualized content or the way that their utterance flouts the norm of civil inattention, the speaker makes the genders of the target and speaker salient, as well as the fact that the speaker belongs to the dominant social group. This activates background social roles as discourse roles, assigning the target a subordinate discourse role and the speaker a dominant one (as these are the social roles in a sexist society that are activated by the gendered catcall). This occurs regardless of the reaction of the target or the psychological states of anyone involved. By introducing unjustly restrictive conversational rules, the catcall structures the normative world in a way that subordinates the target. It adds a very tiny piece of wire to the vast birdcage of gender oppression.

This, I hope, avoids the potential overgeneralization described in the previous section. A child might exploit their parent by ordering them to provide dinner but would certainly not subordinate in doing so. The catcall subordinates, and this should be recognized as part of its overall harm. As Gardner (1995), Hesni (2018), and McDonald (2022) note, catcalling is backed up by the threat of male violence, making a response that might challenge the assignment of the subordinate discourse role unjustifiably costly.

This means that there is no way for a man to catcall a woman without enacting this kind of harm. No formulation of words, benign intent, or proclamation of ignorance can get them off the hook. But how, a catcaller might wonder, might they tell a passing stranger how attractive they find them? The answer is that they should not. They should stay quiet and stop staring.

Future harms

Conversations initiated by catcalls tend to end shortly afterwards. Targets often attempt to leave as soon as possible, unsettled, angry, disgusted (Hesni 2018). And discourse roles are supposed to last only as long as the conversation, because they are made up of a bundle of rules and norms governing conversational behavior. It might then seem that the subordinating effects of catcalls should also end with the conversation. And yet they linger. The normative effects of subordinate discourse role assignment can seep out of the conversation and influence the wider (misogynistic) normative landscape. A catcall does not only subordinate here and now—it also contributes to subordination in the future.

The most obvious ongoing effects of a catcall are psychological. Targets feel offended, threatened, angry, or unsafe, even once they are far away from the catcaller (Major et al. 2002). Sexist behavior can take time to process. Catcalls might also affect the future beliefs and emotions of the target, the speaker, and bystanders. The speaker feels powerful, the target feels demeaned, and bystanders might come to think that such behavior is acceptable. Catcalls also make public space less accessible to its targets, as they risk being catcalled (and violent retribution for “incorrect” responses) when they enter that public space (Crouch 2009). This requires increased “mental vigilance” and avoidance tactics to use public space as emotionally and physically safely as possible (Condon et al. 2007; see also Matsuda 1993 on similar effects of hate speech).

The normative effects of catcalls also persist. Changing conversational rules for one (unwanted) conversation can also change (sometimes in very small ways) the rules governing future conversations. In this section, I argue that token catcalls reinforce pernicious norms and undermine positive norms, and so constitute moves in the “rule-governed activity of gender oppression” (McGowan 2009, 402). Even if catcalling was so ubiquitous that it was the norm (rather than the norm of civil inattention), instances of catcalling would reinforce that norm, making it stronger and thus harder to change for the better (see Cousens 2020 for similar analysis of slurs). Catcalling in those contexts would then still inflict the same kind of structural harm described in this paper, although someone could plausibly judge that strengthening existing subordinating norms is less awful than undermining egalitarian norms (although I make no such determination here).⁶ But how can a single catcall influence the vast set of sexist norms and rules at play in a patriarchal context?

Discourse roles (as tracked by the conversational score) make certain behavior inappropriate, or impermissible. Utterances that assign such roles thereby change what is conversationally appropriate. The student might be supposed to defer to the professor, but should the professor say something to make their sporting roles salient, such deference is no longer appropriate. However, conversational permissibility is not only determined by what has been previously said in a conversation. When conversations begin, there are lots of norms already in play. Do not swear. Speak clearly. Take turns. Avoid terms of endearment with strangers. While some of these permissibility conditions will apply to most conversations, others will be more specific. Even the norm against using terms of endearment with strangers seems to be absent during some conversations—for example, bar staff who might call a customer “love,” “pet,” or “dear.”

Together, these norms that are already in play when a conversation begins constitute the “initial permissibility conditions” of a conversation (Cousens 2020). For example, when the student and professor begin a conversation just outside of the lecture theatre, their academic roles will already be salient, and active as discourse roles. Nothing needs to be said to make this the case; the context is one in which those roles are going to influence “correct” behavior by default. Popa-Wyatt and Wyatt (2018) adapt Lewis’ (1979) example of a master and slave to explain how default role assignment works. The master does not need to say anything to assign themselves a dominant discourse role in conversation with the slave. When the master speaks, they receive the dominant role by default. An adult who enters a room of children will likely be similarly assigned a dominant role by default; children are (defeasibly) expected to follow instructions from an adult. The adult does not need to say “As the adult here”—context and visual cues assign the dominant role automatically.⁷

Speakers can also use utterances to remove a discourse role, even one assigned by default. The master could say something to the slave releasing them (temporarily) from a subordinate discourse role (“Tell me what you really think!”). But when their next conversation begins, the default assignment would be back in play.

The initial permissibility conditions of oppressive contexts could be described using this kind of terminology. It may well be the case that, in relatively progressive contexts, an utterance (like a slur) is required to assign subordinate discourse roles, as Popa-Wyatt and Wyatt argue. But this need not be the case in more starkly identity-oppressive contexts—Ayala-López (2018) argues that there are lots of unjust social norms governing speech interactions that people tend to follow automatically. Similarly, Tirrell (2018) argues that women cannot enter most language games as equals (a “neutral” entrance move); the permissibility conditions in place at the beginning of the conversation already limit the conversational moves that women may make more than they do for men. So, in some contexts, conversations will begin with subordinate discourse roles already in place. In Jim Crow-era USA, Apartheid South Africa, or early colonial Australia, it may be that a Black person in conversation with a white person was assigned a subordinate discourse role by default. Certain utterances would be impermissible for Black speakers even before any conversational contributions have changed the “score.”

My hypothesis is that the default assignment of subordinate discourse roles changes incrementally. In 1923, women may have been assigned subordinate discourse roles by default. Over the next hundred years, the social roles (which get activated as discourse roles) will have changed. Now, an active conversational contribution (like a catcall) may be required to bring subordinate discourse roles to bear. But the more catcalling that occurs, the more often women are subordinated in this way, the closer this becomes to a default role assignment. In some contexts, such as male-dominated workplaces, this might already be the case (see Files et al. 2017 on the incorrect use of honorifics for women). And (more optimistically) vice versa; the less catcalling, and more criticism of catcallers, the less easily or automatically such assignment would occur.

The initial permissibility conditions of a conversation are both prescriptive, in that they tell participants how to act, and descriptive, in that they are derived from patterns of behavior. When I arrive at a restaurant, I expect dining to be governed by a norm prohibiting plate-licking. I would expect patrons to refrain from licking their plate, no matter how delicious the meal, and I would expect other diners to think poorly of those who do (Bicchieri 2017 distinguishes between these types of empirical and normative expectations). If I see one patron lick their plate, I think that they are rude. Should I see several more follow suit, I might still think poorly of them. But if more and more diners lick their plates, at every restaurant I visit, I would be foolish to continue to think that there is a general norm prohibiting plate-licking. Such a norm would not accurately describe people's behavior, as many people lick their plates, and it would not prescribe behavior, as there is little normative pressure to conform to a norm so frequently flouted.

While restaurant dining might usually be governed by a norm against plate-licking, norms against catcalling seem less pervasive—or at least, the strength of such norms changes across different contexts (both in time and location). But similarly, the more catcalling that occurs, the more norms against it are undermined. It would be nice to think that women and men are, by default, equals when a new conversation begins. In some contexts, this may seem to be the case. But the more often that women are assigned subordinate discourse roles (such as via catcalling), the more descriptively accurate it might be to say that this role assignment has become “default,” and the more prescriptive such roles become in setting unequal initial permissibility conditions for conversations between women and men (Bicchieri 2017 and Ayala-López 2018 both argue that such shifts in norms need not be derived from conscious processes).

As more catcalling occurs, the equal status of women is further undermined, and this likely enables further catcalling as the norms against such behavior are also undermined (just as more and more restaurant patrons licking their plates undermines norms against plate-licking). Discrimination begets discrimination. At some point, it might become “correct play” for conversational participants to assume that women are assigned subordinate discourse roles by default (and that an utterance is required to restore equal footing, such as a pointed mention of a woman's credentials).

I am not sure that there is a threshold at which we should say that women are assigned a subordinate discourse role by default. The point is that a token catcall shifts the needle of oppression in the wrong direction. It assigns a subordinate discourse role here and now, and also contributes to the pattern of behavior that (descriptively and prescriptively) forms some of the norms governing the activity of gendered oppression. These norms then provide the initial permissibility conditions governing future conversations, which could assign subordinate discourse roles as default. As Popa-Wyatt and Wyatt (2018) argue, the discourse role is not only influenced by the social roles of participants. Broader social roles and expectations are also influenced by the discourse.

It might be tempting to say that, while women used to be assigned a subordinate discourse role by default, this is no longer the case. Society has progressed. But progress is not always permanent. Hard-won respect, regard, and rights can be stripped away—immediately, by judicial or legislative decree, or gradually, through an accumulation of pernicious behavior (like catcalling).

Norms do not simply spring into existence, then vanish altogether. There is not some magical number of plate-licking patrons that makes a norm against plate-licking suddenly disappear. Norms can grow stronger or weaker. The stronger the norm, the more influence it has over people's behavior. As I have argued in this section, catcalling strengthens pernicious social norms (making assigning women subordinate discourse roles easier) and weakens egalitarian norms (undermining the status of women in the initial permissibility conditions of conversation). A single catcall does not, by itself, noticeably shift norms beyond the unwanted conversation it has started. But if many others are also catcalling, the weight of their collective discourse role assignments alters permissibility conditions here and now as well as contributing to changing the initial permissibility conditions governing future conversations.

Catcalls can have this effect even where the norm of civil inattention is not in play (or at least, not as strongly). In some small towns, people often greet each other on the street (while, e.g., Gardner 1980 discusses specifically urban contexts). This might be thought of as one of Goffman's (1963) exceptional circumstances; just as people in a sports crowd might expect each other to excitedly start conversations about their team's recent victory, so might people in a small town expect to be greeted by others because they inhabit the same small community. Here, inattention would not be civil, as different (greeting) norms apply.

Some utterances that (in urban contexts governed by the norm of civil inattention) might seem like catcalling may not function that way in the friendly small-town context (“Hey, that's a nice hat!” perhaps). But other utterances would still count as catcalls in that context (“Nice ass!”)—while inattention may not be the norm, certain kinds of attention (such as overtly sexualized comments) would flout the norms that are in place. While greeting strangers might lead to wanted conversations, those initiated with catcalls would still be unwanted. In doing so, they will assign subordinate discourse roles to their targets and shift (slightly) the norms that govern future conversations. The more of this that happens (I conjecture), the less likely that “Hey, that's a nice hat!” would have an innocent, non-gendered, or non-threatening character.

Token catcalls thus erode egalitarian norms, a bit like a wave crashing against the shore. Each wave may only wear away a few tiny grains of sand. But slowly, perhaps imperceptibly, the dunes and cliffs that once seemed immovable may disappear into the sea. Catcalling reshapes the normative landscape—and not for the better.

It's not what catcalls say, it's what they do

Catcalls are not really compliments, despite appearances. Sometimes, what looks like a duck and quacks like a duck is in fact an elaborate decoy. But even if they were compliments, they are still bad. Catcalls are both immediately and structurally harmful. They assign their targets subordinate discourse roles in unwanted conversations, and in doing so change the initial permissibility conditions of future conversations in oppressive ways.

This shows that the harm of catcalls is not just in what they say, but also in what they do. It might seem uncomfortable to interpret an utterance like “You look pretty today!” as an act of subordination. But this is part of the problem with patriarchy. We are trained to minimize and overlook the harmfulness of commonplace behavior. As I hope to have shown, oppressive norms work together, reinforcing each other. If we are to properly assess the harmfulness of catcalls, we need to look not only at their sharpest and most immediate harms, but also at the subordinating normative changes that lurk beneath.