Bending as Counterspeech

Abstract

In this paper, we identify and examine an overlooked strategy to counter bigoted speech on the spot. Such a strategy we call ‘bending’. To ‘bend’, in our sense, is to deliberately give a distorted response to a speaker’s harmful move – precisely, an ameliorative response, which may turn that move into a different, less harmful, contribution. To substantiate our proposal, we distinguish two ideas of uptake – interpretation and response – and argue for the general claim that a distorted response on the hearer’s part may end up transforming a speaker’s contribution. Patterns of distortion have been analyzed in the literature as unjustly undermining speakers’ agency and exacerbating oppression. Our analysis shows that, under certain circumstances, distortion can be employed to derail bigoted speech and thus serve the purposes of social justice. We close by discussing the virtues and limits of bending vis-à-vis a different, much-discussed, counterspeech strategy, i.e. ‘blocking’ (Langton 2018).

1 Bending: A Foretaste

Over the past two decades, social philosophers of language and political philosophers have been increasingly interested in counterspeech, i.e. the practice of responding to bigoted speech in ways that defuse or moderate its harmful effects.¹ In this paper, we identify and examine an overlooked counterspeech strategy. We call it ‘bending’. To ‘bend’, in our sense, is to deliberately give a distorted behavioral response to a speaker’s harmful move – precisely, an ameliorative response, which may turn that move into a different, less harmful, contribution.² By way of illustration, consider the following scenario.

Example 1. Exclusionary Roundtable.

Anna, Jason, and a few colleagues of theirs are brainstorming about potential invitees for a graduate roundtable. Jason says, “My supervisor would avoid inviting, you know, affirmative action students”. Anna perfectly gets that Jason is implicitly suggesting that they do the same, and yet she goes on replying, “I know. It’s terrible how racist some professors are around here”. After a moment of bewilderment, Jason mumbles a ‘yeah’. The organization process goes on with no regard for the bigoted suggestion to exclude students from underrepresented groups.

In responding to Jason’s move as if it were an expression of indignation over bigotry rather than a bigoted suggestion, Anna engages in bending. She deliberately gives Jason’s move an ameliorative response – and this, we will argue, may sabotage its potential for oppression. Provided that Jason doesn’t rejoin by making his original suggestion explicit, the discriminatory line of conduct he hints at fails to enter the set of options under discussion. The conversation does not get drawn into whether certain students should be excluded, and discriminatory treatment is never talked over as a viable option.

Bending can operate on an utterance’s content, force, or both. In Example 1, bending touches upon both force and content: Jason’s implicit suggestion to do as his supervisor would do (roughly, “We should exclude students from underrepresented groups”) is bent into an expression of indignation (“It’s awful how racist my supervisor is”). Examples 2 and 3 illustrate how bending can target only one of force and content.

Example 2. Coffee Break Microaggression.

At a conference coffee break, Brown, Zheng, and Miller are chitchatting. They don’t know each other. All three have an American accent. At some point, Brown asks Zheng, “Where are you from?”. Zheng perfectly gets that Brown is inquiring about her ancestry, and yet she goes on replying, “I’m from D.C. Aren’t you, too?”. While Brown says he is, Miller jumps in to gather practical info in view of her upcoming trip to Washington. The conversation turns to gluten free restaurants and must-see spots in Washington.

Zheng responds as if Brown’s question concerned her hometown rather than her ancestry. Thanks to her bending maneuver, and provided that Brown doesn’t follow up by asking about her descent, the conversation can easily veer towards topics – restaurants, touristic spots – that do not trigger the bigoted assumption associated with Brown’s intended question (i.e. that Asian people do not belong in a Western country). Zheng does not touch on the force of Brown’s move, which is both intended and replied to as a question. Rather, she twists its content: roughly, “Where are your ancestors from?” is turned into “What’s your hometown?”. Zheng deliberately gives Brown’s utterance a distorted response – and this allows her to avoid the harm caused by her ethnic origin becoming the question under discussion.³

Example 3. All-Male Panel.

Professor Smith, Professor Murphy, and a few colleagues of theirs are organizing a panel on climate justice. After drawing up a list of potential invitees, Smith exclaims, “Whoa, a men-only panel!”. Murphy perfectly gets Smith’s gloat, and yet he goes on replying, “You’re right, it’s awful. Let me think… Got it! What about Tara Williams? She just published an amazing paper on environmental responsibility”. Following suit with Murphy’s proposal, a couple of colleagues add a few more names of outstanding women scholars. Smith acquiescently nods. They end up with a gender-balanced list of invitations.

Murphy responds as if Smith were calling attention to a problem rather than gloating over political incorrectness. If Smith does not clarify that he actually likes the idea of a men-only event, what he says (“Whoa, [we’ve put together] a men-only panel!”) gets associated with a different force than the one he meant. He intended to gloat over a certain state of affairs, but his act ends up counting as a warning about that very state of affairs. Murphy bends the force of Smith’s utterance, while leaving its literal content intact. Note that, once its force is bent, the sexist assumptions associated with Smith’s gloating utterance fade away. Particularly, the assumption that it is OK – perhaps, even ‘brilliantly un-PC’ – to exclude women from academic events. Murphy’s bending move thus defuses the sexist potential of Smith’s comment.

These three scenarios should provide an intuitive picture of the phenomenon we are after. They suggest that acting as if your interlocutor’s move were less prejudiced than it actually was may transform it into a different, less prejudiced, contribution. We characterize bending as a deliberate ameliorative maneuver. Of course, one can genuinely misinterpret a bigoted move as non-bigoted or less bigoted than it was, and ameliorate it unintentionally. Since our focus is on discursive activism and counterspeech strategies, we here leave “accidental bending” (as it were) out of the picture.

To bend is to engage in a form of pretense – to act as if. To ‘act as if’ encompasses two dimensions: one’s interpretation of and behavioral response to a speaker’s move. In the next section (§ 2), we suggest that these dimensions capture two different ideas of uptake, and argue for the general claim that a distorted response on the hearer’s part may have the power to transform a speaker’s speech act. The philosophical literature has hitherto focused on distortion as a force for evil – a mechanism that tracks and exacerbates social injustices and disadvantages (Kukla 2014). The one who bends, we submit, deploys distortion in the service of social justice (§ 3). We close by comparing bending to a different, much-discussed, counterspeech strategy, i.e. ‘blocking’ (Langton 2018),⁴ with the aim to detail, by way of contrast, its distinctive benefits and costs (§ 4).

2 The Power of Response Distortion

As already stated, bending involves a form of pretense: the counterspeaker interprets the speaker’s move as a speech act A but responds as if they interpreted it as a speech act B. ‘Interpretation’ and ‘response’ capture two different ideas of uptake.

2.1 Two Ideas of Uptake

In speech act theory, uptake is traditionally couched as the hearer’s interpretation of the content and force of the speaker’s speech move. Call this idea of uptake ‘Interpretation’.⁵ Content and force have been variously read in the literature. One prominent reading casts content in terms of the speaker’s locutionary intention, and force in terms of the speaker’s illocutionary intention.⁶

Interpretation is widely⁷ taken to influence illocutionary success. John L. Austin famously claims,

Unless a certain effect is achieved, the illocutionary act will not have been happily, successfully performed […]. Generally the effect amounts to bringing about the understanding of the meaning and of the force of the locution (Austin 1962: 115 − 16).

In a similar spirit, John Searle underlines that,

unless he [the hearer] recognizes that I am trying to tell him something and what I am trying to tell him, I do not fully succeed in telling it to him (Searle 1969: 47).

That is to say that an illocutionary act is fully successful only if the hearer gives it a compliant interpretation. A hearer provides a compliant interpretation when they correctly grasp the speaker’s intentions, and a distorted interpretation when they misread the speaker’s intentions. If I try to promise you something, but you think I am joking, or misinterpret what I am trying to commit myself to, my illocutionary act will not count as a fully successful promise.

Uptake-as-interpretation has to do with the hearer thinking that the speaker is performing a speech act with a certain force and content. As such, it is independent from any behavioral response on the part of the hearer. But uptake need not be just a take on the speaker’s intentions. And indeed, in conversation analysis, ‘uptake’ is often given quite a different meaning. Herbert Clark, for one, writes,

What am I doing by asking you to sit down – by performing an illocutionary act? I am proposing […] a project for us to carry out jointly – namely, that I get you to sit down […]. Joint projects have two parts. In my terminology, the speaker proposes a joint project, and the addressees take it up. I propose that you sit down, and you take up my proposal by sitting down or by agreeing to sit down. A proposal is expected to be followed by its uptake (Clark 1996: 150).

‘Uptake’, in Clark’s terminology, stands for the hearer’s behavioral response to the speaker’s speech act. Call this second idea of uptake ‘Response’. Just as interpretation, response can be either compliant or distorted. A hearer provides a compliant response when their subsequent behavior is consistent with the speaker’s intentions, and a distorted response when their subsequent behavior is inconsistent with the speaker’s intentions. I ask you to sit down and you sit down: your response is clearly compliant. I ask you to sit down and you reply, “I’d rather not. My back hurts”: your response is still a compliant one – although it does not satisfy my perlocutionary goal, it is consistent with my intention to ask you to sit down. But if I ask you to sit down and, say, you pass me the salt or turn off the TV, you will be giving my words a distorted response.

Bending is an ameliorative response distortion. Consider again our opening scenarios. In Example 1, Jason intends to suggest that certain students be excluded, but Anna gives a distorted, ameliorative response to his speech act – she responds as if he were expressing indignation over bigotry. Similarly, Zheng responds as if Brown’s question revolved around her hometown rather than her ancestry (Example 2), and Murphy responds as if Smith were flagging a problem rather than gloating over political incorrectness (Example 3). In all three cases, the hearer’s response diverges from their interpretation: they interpret the speaker’s move one way but respond as if they interpreted it another way.

It is to be underlined at this point that uptake-as-interpretation and uptake-as-response stand in a proxy relation: response usually reveals interpretation. Put otherwise, the hearer’s response provides clues to figure out how they interpreted the speaker’s act. For example, if following an order of mine, you drop everything you were doing to act as I told you, your response suggests, although it is no guarantee, that you interpreted my words as an order. More generally, the only way we have to ascertain or determine the hearer’s interpretation of a given speech act is to look at their subsequent behavior. We are not the first to note this. Marina Sbisà points out that

the response (verbal or non-verbal) which follows the illocutionary act under examination […] makes manifest how the hearer has taken the speaker’s illocutionary act (Sbisà 1992: 101).

And, in discussing turn-taking dynamics, William Turnbull suggests,

In responding to S’s turn, A both produces an action and manifestly displays how he treated S’s turn (Turnbull 2003: 161).

The addressee’s response makes their interpretation manifest – it displays how they construed the speaker’s move. Interestingly, Turnbull adds,

Of course, S in the next turn also has the opportunity either to accept or contest A’s manifest interpretation of her prior turn (ibid.).

Conversations do not end in two turns. They typically proceed further, and the speaker can have their say, in subsequent turns, about what they were trying to do with words in past turns.⁸ This is crucial to what we are going to claim, namely, that despite being a causal or perlocutionary fallout of the speaker’s utterance, the hearer’s response may play a role in its illocutionary success. To unpack this claim, let us first map what a speaker can do after a hearer responds to a move they have just made.⁹

Firstly, the speaker can contest the hearer’s manifest interpretation. Faced with Anna’s distorted response, Jason could have clarified that he was suggesting that the organizing team excluded certain students (Example 1). In doing so, he would have preserved the success of his original speech act. Similar considerations apply to Examples 2 and 3: Brown and Smith could have contested their addressees’ responses to their moves. Had they done so, their original speech acts would have succeeded and Zheng’s and Murphy’s bending maneuvers would have failed.

Secondly, the speaker can simply do nothing. Brown, in Example 2, picks this option. He does not contest Zheng’s response, nor does he openly endorse it. Rather, he does nothing about it, until Miller jumps in and the conversation is veered towards different topics.¹⁰ In contexts of this sort, doing nothing often amounts to a tacit acceptance of the hearer’s manifest interpretation.¹¹

Thirdly, and finally, the speaker can endorse the hearer’s manifest interpretation by openly signaling that they accept it. This is what both Jason (Example 1) and Smith (Example 3) do, the former by mumbling a ‘yeah’ and the latter by nodding as his colleagues suggest potential female invitees.

We claim that, when the speaker accepts a hearer’s manifest interpretation, either tacitly or openly, they end up performing the speech act that the hearer has manifestly ascribed to them. How the hearer responds matters for illocutionary success: a distorted response has the potential not only to make a speech act fail, but to transform it into a different one. To avoid any misunderstandings, we are not claiming that the hearer’s response is sufficient for the speaker’s intended move to turn into something different. For this to occur, the speaker needs to play along, thus ratifying or validating the hearer’s manifest interpretation.¹² It is this ratification that makes bending succeed.¹³

Importantly, conversation participants need not believe that the bending maneuver is consistent with the speaker’s intentions. Nor do they need to take the counterspeaker to be in good faith. They may have perfectly understood that the counterspeaker is deliberately giving a distorted response, and yet follow suit for all sorts of reasons: they may genuinely share their values, or play along for conformism, personal interest, lack of initiative, etc. Token instances of bending will diverge in terms of how credible the counterspeaker, and their distorting move, are taken to be.¹⁴ Participants may find it more or less credible that the counterspeaker’s response is consistent with the speaker’s intentions. And they may find it more or less credible that the counterspeaker’s manifest interpretation matches their actual interpretation (i.e. that their response is sincere and not manipulative).¹⁵ Arguably, the more credible a distorting move is taken to be, the smoother bending goes. Yet, whether bending is successful in ameliorating a speaker’s act is fairly independent from the extent to which it comes across as credible. Success, to reiterate, depends on whether the counterspeaker’s manifest interpretation of the speaker’s act is ratified.

2.2 On/Off Record

Bending, we have claimed, does not depend on how credible conversants take it to be. To this, one might object that, when everyone got the speaker’s intentions right, bending cannot transform the speaker’s act.¹⁶ Suppose that, in Example 1, Jason’s intention to suggest that certain students be excluded is clear to all. Suppose also that everyone goes along with Anna’s manifest interpretation, but as soon as Jason has left, people start commenting on how racist his suggestion was. How can we explain it, if – in our view – Jason’s words do not count as a suggestion at all? Or imagine that, as the election for the Equality, Diversity, and Inclusion committee approaches, people do not even remotely consider Jason as a candidate, because they think he is a bigot. How can we explain it, if – in our view – Jason counts as having expressed indignation over bigotry?

We explain it by emphasizing that there is a distinction between what officially occurs in a conversation (the conversational record) and what participants think has occurred (their mental scoreboards). We interpret the conversational record as the official register of what speakers said and did. The record of a conversation tracks all and only those conversational facts to which speakers are publicly committed, in such a way that “if a transcription or video footage [of the conversation] were available then they would be undeniably liable for them” (Camp 2018: 59). Since Jason accepts Anna’s distorted response and the conversation proceeds as if he did not do anything bigoted, what goes on record is that he expressed indignation over bigotry. And what speakers count as having performed depends on what goes on record. This does not mean that Jason’s attempt to make a racist suggestion disappears entirely. As long as conversation participants get Jason’s original intentions, his attempted suggestion enters their mental scoreboards. For each participant P in a conversation C, P’s mental scoreboard incorporates P’s mental representation of what occurred in C. With the notion of mental scoreboard in place, we can easily explain how people can later talk about Jason’s bigoted suggestion. It is because the conversants’ mental scoreboards registered Jason’s intended move that they can comment on it once he leaves the scene. A similar story can be told as to why, in discussing potential candidates for the EDI committee, Jason is not even remotely considered.

The ‘gap’ between the official record of a conversation and conversants’ mental scoreboards creates the possibility of bending. The one who bends steers the conversation away from what their mental scoreboard tracks, in an effort to make the record overlook the speaker’s intended move and register an ameliorated move instead. Provided that everyone plays along, one can succeed in bending a speech act even if one’s fellow conversants realize what one is doing. By playing along, conversants (more or less wittingly) resist the problematic direction that the conversation would have otherwise taken: excluding students from underrepresented groups is never talked over as a viable path; all-male panels are never considered an option (let alone one to proudly endorse); etc.

So, participants’ mental scoreboards may well track a speaker’s bigoted move, but as long as everyone acts as if the speaker performed something different, their original attempt stays off record and the conversation may proceed as if that attempt had never been made.¹⁷

In this section, we maintained that hearers can deliberately provide distorted responses to thwart speakers’ speech acts. While we are interested in how this mechanism can be exploited in the pursuit of equality, patterns of distortion are more familiar in the philosophical literature in the negative guise of discursive injustice.

3 Evil Distortions: Discursive Injustice

Quill Kukla (writing as Rebecca Kukla) (2014) has introduced the notion of discursive injustice to capture the idea that a speaker’s social identity, and in particular their belonging to a relatively disempowered social group, can derail the process from speaking to uptake. Kukla’s conception of uptake holds together interpretation and response. They write,

as I am using the term, “uptake” does not mean mere recognition of the intentions of the speaker […]. The uptake of a speech act is how it gets recognized and responded to in practice (Kukla 2014: 453).

Discursive injustice occurs when a speaker intends to perform a speech act A; they are entitled to do so; they use the appropriate conventions to do so in the context of utterance; and yet, because of some stigma attached to their social identity, they are taken up in practice as having performed a speech act B. This alternative uptake, Kukla argues, will constitute their performance as a speech act B. To illustrate, let us bring in Kukla’s central case of discursive injustice.

Example 4. Floor Manager

Celia is a floor manager at a heavy machinery factory where most workers are men. It is part of her job to tell the workers on her floor what to do. “Put that pile over there, Steve”; “Unplug the drill press, Pete”. Compliance, however, is low. Although Celia is entitled to give orders in that context and follows the standard conventions for ordering, because of her gender, her workers interpret and respond to her acts as mere requests – which explains why they feel free not to comply.¹⁸

Perhaps, Celia’s workers are so unaccustomed to having female co-workers, let alone bosses, that they cannot even conceive that a woman could give orders in their workplace; and so, the most reasonable interpretation for them is that Celia is making requests instead. Unlike orders, requests do not introduce any obligation, but leave the requestee free to grant them (and do the requester a favor) or refuse them. The obedience rate among Celia’s workers is low because they read her acts as requests and respond accordingly. And this, in Kukla’s view, seems to be enough for those acts to count as requests.¹⁹ Celia is a victim of discursive injustice: her utterances receive a distorted uptake, which constitutes them as different acts than the ones she was trying to perform. This distortion both depends on and upholds unjust gender imbalances. Were a man in the same position as Celia to utter the same words she utters, those words would have most likely been taken as orders.

Kukla’s proposal differs from ours in at least one significant respect. As Celia’s scenario illustrates, Kukla takes the hearer’s interpretation-cum-response as constitutive of what speech act a speaker turns out to perform. We agree that the hearer’s response has the potential to transform a speech act, but for the transformation to occur, the speaker must go along with the hearer’s manifest interpretation. How Celia reacts to her workers’ noncompliance matters for the force of her acts to be fixed in one sense or the other. She could openly state, “That was an order, Pete” – or, should that not be enough, issue a formal reprimand to inform him that further disciplinary action may occur if the noncompliant pattern continues. Reactions of this sort would fix her initial utterances as orders.²⁰

Note also that, in Example 4 (and similar others from Kukla), the hearer’s response manifests how they truly interpreted the speaker’s words. Kukla is interested in cases in which the hearer genuinely misinterprets the speaker. By contrast, we are interested in cases in which interpretation and response come apart – in which the hearer understands the speaker’s intentions and yet behaves as if they got them wrong. What if Celia’s workers correctly interpreted her imperative utterances as orders, but responded to them as if she were merely requesting? More broadly, can a hearer’s deliberate distortion of a speaker’s act reshape its pragmatic structure? To reiterate, our answer is yes, but only insofar as the speaker does not retort to make it clear that they intended to perform a different act than the one the hearer ascribed to them.

Discursive injustice and bending involve, with due specifications, discursive distortion. While the former employs the mechanism to worsen unjust social imbalances, bending puts it in the service of equality. Discursive injustice and bending thus play opposite roles in the economy of social justice. Discursive injustice is defined in negative terms. Cases in which distortion does not exacerbate preexisting social disadvantages do not instantiate discursive injustice. In a specular way, bending is defined in positive terms, i.e. it amounts to the hearer deliberately providing an ameliorative response to a speaker’s move in order to reduce or defuse its oppressive potential. Of course, this does not preclude the possibility that sometimes hearers deliberately give speakers’ utterances a pejorative response, overall contributing to social injustice. Such distortions, however, would not be cases of bending in our sense.²¹

By characterizing bending as a positive phenomenon, we do not mean to conceal the controversial character of any discursive distortion. Bending is, in a sense, ‘unjust’, too – for distorting a speech contribution is unfair to the speaker. By engaging in bending, one disregards the discursive agency of one’s interlocutor, whose attempt to do things with words gets thwarted and twisted. One might worry that, in the long run, this lack of regard for others-qua-agents may lead to discursive alienation and aggravate political polarization.

Although bending is unfair to the speaker in this sense, its distinctive features make it a promising tool to counter bigoted speech on the spot. We fully detail such features in the next section.

4 Bending as Counterspeech

Bending provides a subtle way to counter bigoted speech on site. We here analyze its distinctive features vis-à-vis prototypical instances of a different, and much-discussed, counterspeech strategy, i.e. ‘blocking’ (Langton 2018). Let us return to Example 1. Here is the case, once again.

Example 1. Exclusionary Roundtable.

Anna, Jason, and a few colleagues of theirs are brainstorming about potential invitees for a graduate roundtable. Jason says, “My supervisor would avoid inviting, you know, affirmative action students”. Anna perfectly gets that Jason is implicitly suggesting that they do the same, and yet she goes on replying, “I know. It’s terrible how racist some professors are around here”. After a moment of bewilderment, Jason mumbles a ‘yeah’. The organization process goes on with no regard for the bigoted suggestion to exclude students from underrepresented groups.

Anna engages in bending: she ameliorates Jason’s bigoted suggestion into an expression of indignation over bigotry. Imagine now a slightly revised case.

Example 1*. Exclusionary Roundtable.

Giulia, Jason, and a few colleagues of theirs are brainstorming about potential invitees for a graduate roundtable. Jason says, “My supervisor would avoid inviting, you know, affirmative action students”. Giulia, who perfectly gets what Jason is implicitly doing, rebuts, “Are you suggesting we do the same? That’s off the table, man. We won’t exclude students who are just as valuable”.

Giulia engages in blocking: she explicitates²² (“Are you suggesting we do the same?”) and rejects (“That’s off the table, man”) Jason’s bigoted suggestion. By bringing what was implicit to the surface, and no matter whether people take her side or not, Giulia prevents Jason’s suggestion from entering the conversation through the ‘back door’ (Langton 2018: 146). Her intervention forces Jason either to explicitly defend the suggestion he was implicitly making or to renege on it.

Thus, blocking and bending counter implicit harmful moves in different ways. Blocking explicitates and openly challenges them. Bending, in contrast, distorts them into different, less harmful, moves.

A terminological note before moving forward. Blocking (broadly construed) can take many forms: the blocker may explicitate the objectionable content without openly rejecting it; they may raise a query and urge the speaker to explicitate that content themself (Brandom 1994: 191ff; Tirrell 2018: 135); they may resort to irony; etc. Since we are here interested in how bending differs from paradigmatic instances of blocking, we will use ‘blocking’ as a shorthand for what Cepollaro (ms) calls ‘full-fledged blocking’, which consists in explicitating and rejecting implicit harmful contents.²³

4.1 The Virtues of Bending

The distinctive features of bending and blocking make their effects, costs, and benefits importantly different.

For one thing, (full-fledged) blocking is confrontational in nature.²⁴ In Example 1*, Giulia voices a view that openly conflicts with Jason’s. Her maneuver threatens Jason’s face by calling him out for bigotry. More generally, the blocker takes an argumentative or disputatious stance towards a certain bigoted move. This threatens the bigoted speaker’s face and exposes the blocker to the risks and costs of verbal disputes. Open disagreement comes with social perils. This is so in peer-to-peer exchanges (as in Example 1*), and even more so in asymmetrical contexts: it can be especially risky for subordinated speakers to express disagreement with or openly confront their superiors.²⁵

Unlike blocking, bending is not confrontational in character. In Example 1, Anna makes it as if the conversational context were so obviously egalitarian that it would be almost unconceivable that her interlocutors could voice a deeply bigoted view. In doing so, she gives Jason a chance to tacitly disavow his prejudiced suggestion. Not only is he given the opportunity to step back from it; he can do so without publicly admitting that he intended to make it. Anna’s bending maneuver is not face threatening, and to an extent, safeguards Jason’s reputation. Thus, bending may be a preferable counterspeech strategy to blocking when taking a confrontational attitude towards one’s interlocutor would be too risky or otherwise unwise.

Of course, as already said, Jason could stick to his original suggestion by bringing what he meant into the open (“What? No, I meant that we should not invite affirmative action students!”). Yet, by assuming that the local context is an egalitarian one, Anna makes it socially costly for any subsequent speaker, Jason included, to act bigoted. Her maneuver does not only discourage bigoted behavior; it also contributes to enact egalitarian norms for the ongoing conversation.²⁶

Secondly, the blocker may be expected to provide reasons against the view they challenge. After all, if you reject a certain move, you must be prepared to spell out what was wrong with it. Giulia (Example 1*) may thus be pressured into articulating why excluding students from underrepresented groups is not a viable option. Bending lifts the burden of arguing for or against a view off the counterspeaker’s shoulders. Anna (Example 1) need not argue against excluding underrepresented students – in fact, thanks to her distorted response, that suggestion is never taken into consideration. The counterspeaker is spared from the fatigue associated with giving reasons. This is, we think, a general advantage of bending over blocking, and an especially valuable virtue when the counterspeaker belongs to an oppressed group (as in Example 2) and is thus vulnerable to epistemic exploitation – i.e. the phenomenon whereby “privileged persons compel marginalized persons to educate them about the nature of their oppression” (Berenstain 2016: 569).

In addition, it has been suggested that blocking may backfire in a distinctive way, i.e. by making harmful assumptions contextually salient.²⁷ Bending is less susceptible to this worry. To see why, it is helpful to distinguish two senses in which something may be made salient. First, speakers can make something salient by turning it into the question under discussion (QUD).²⁸ Second, something is made salient in a conversation when it is rendered highly cognitively available to its participants. Merely evoking something, without turning it into the QUD, may suffice to make it salient in this second sense: it may render conversation participants more prone to represent it. In turn, this may consciously or unconsciously influence their subsequent decisions and conduct.²⁹

Blocking risks making harmful assumptions salient in both these senses. Recall Example 1*. Giulia replies to Jason, “Are you suggesting we do the same? That’s off the table, man. We won’t exclude students who are just as valuable”. Her reply makes the question whether certain students should be excluded the main topic of conversation. The blocker turns the bigoted content they target into the QUD. This is problematic because questions like ‘Are students from underrepresented groups as valuable as any other student?’, ‘Who belongs in the US?’, ‘Are women less suited than men for scholarly research?’ have been raised and (painfully) addressed over and over in the past, and their answers ought not to be argued for anymore. To reiterate such questions is to force new exhausting discussions of past settled issues (Maitra, ms). Furthermore, by articulating and challenging prejudiced associations, blockers render them highly cognitively available to participants. And this can bias them towards bigoted choices and behaviors.

Bending is less likely to make harmful assumptions salient in the first sense. By replying to Jason, “I know. It’s terrible how racist some professors are around here”, Anna (Example 1) does not turn the question whether certain students should be excluded into the QUD. Rather, she takes it for granted that they should not. Notice, however, that Anna’s words highlight the persistence of racism in education. Merely mentioning racism may suffice to evoke (i.e. make it cognitively available) a host of disparaging stereotypes against racial minorities. Anna’s bending maneuver may thus inadvertently make harmful associations salient in the second sense.

That said, we should not go as far as to conclude that bending in general falls prey to salience-raising effects. This ultimately depends on how a given bending maneuver is devised. Consider Example 2. Unlike Anna’s, Zheng’s reply (“I’m from D.C. Aren’t you, too?”) is devised in such a way as to distort Brown’s question without evoking any ethnic-related associations. In cases like this, bending sidesteps the risks associated with salience raising.

Thus far, we have focused on the relative advantages of bending over blocking. However, we are by no means suggesting that bending is always preferable to blocking,³⁰ nor that bending is always beneficial. To the limits of bending we now turn.

4.2 The Limits of Bending

We have characterized bending as less face threatening than blocking, but this need not be an asset across contexts. In bending, one ends up protecting the bigoted speaker’s face.³¹ In certain circumstances, though, we may want to (and perhaps should) call people out for bigotry. We may want bigoted speakers to publicly commit to the things they were trying to smuggle in through the back door, and draw everyone’s attention to the harmful contents that certain conversational moves convey. When this is the case, blocking may be a more adequate counterspeech strategy than bending.

This is not the only potential limit of bending. For starters, bending might foreclose opportunities for fruitful dialogue. Acting as if a speaker’s contribution were less oppressive than it was is a way to evade debate. Bending is a form of gentle bullying:³² it pushes conversation participants to conform to a worldview that is taken for granted, rather than argued for. The blocker, on the other hand, confronts the speaker – and confrontation may lead to a debate between opposing views. Such a debate may, in principle, be transformative for the bigoted speaker. So, while bending makes it harder for the speaker to stick to bigotry in that context, and it may even rule out oppressive courses of action, it has little chance of turning the bigoted speaker into a less bigoted person.

In confronting their interlocutor, the blocker recognizes and respects them as a fellow agent who is doing things with their words. As we saw in § 3, by engaging in bending, a counterspeaker disregards their interlocutor’s discursive agency. Bending is a strategic (i.e. non-collaborative) move, which, one might worry, runs the risk of exacerbating discursive alienation and political polarization. Bending, so the worry goes, might win a battle but lose the war.³³

Notice also that bending is not always a viable option: outright, blatantly bigoted utterances may leave very little room for ameliorative responses. It may thus be particularly hard to bend them.

Finally, when bending is available, it is rather cognitively demanding. One needs to grasp one’s interlocutor’s intended move (i.e. give it a compliant interpretation); sense its harmful potential; come up with an ameliorative construal of that move; and behave in such a way as to make this construal manifest (i.e. provide a distorted response). All this requires a swift, skillful, and highly ingenious counterspeaker.

5 Conclusion

Overall, this paper adds to the philosophical literature on counterspeech by identifying a neglected strategy – i.e. ‘bending’ – to counter bigoted speech on the spot. We examined such a strategy vis-à-vis blocking and spelt out, by way of contrast, its distinctive features. While blocking amounts to challenging a speaker’s harmful move, bending is to give it a distorted, ameliorative response – a response that may turn that move into a different, less harmful, contribution. To substantiate our proposal, we distinguished two ideas of uptake – interpretation and response – and argued for the general claim that a distorted response on the hearer’s part may end up transforming a speaker’s speech act. While distortion has so far been analyzed in the negative guise of discursive injustice, our investigation shows its potential to derail bigoted speech and thus contribute to social justice.