Photo by ragz13/Getty

Philosophy of language

Photo by ragz13/Getty

The ethics of speech acts

It’s one thing to say something. It’s quite another for a person to do (or not do) something because of what you’ve said

Guy Longworth

Photo by ragz13/Getty

Guy Longworth

is associate professor of philosophy at the University of Warwick in Coventry, UK. His latest book, co-authored with Jennifer Hornsby, is Reading Philosophy of Language: Selected Texts With Interactive Commentary (2005). 

3,600 words

Edited by Nigel Warburton

Syndicate this Essay

Aeon for Friends

Find out more

Total autonomy is a myth. Much of what each of us does depends on others playing their part. I can’t close the suitcase without your help. You can’t find your way to the station without mine. Neither of us will drive the train that gets us to Blyth. And so on, and so on. However, some of what I do seems closer to being solely up to me. I type, and then delete, the word ‘expression’. You suggest ‘word’. I scowl. Where should we draw the line, or lines, between the things we do that depend on others playing their part, and the things we can do freely, without depending on others? And what turns on the answer to that question?

One key battleground here concerns some of the specifically linguistic things we do. Your uttering some words seems to be solely down to you. Your persuading me to replace one word with another, by contrast, depends on my cooperation: your attempt to persuade me will be successful only if I accept what you propose. What about your act of proposing that I make the change, sandwiched as it is between your merely uttering some words and your ultimate end of persuading me? Is your proposing autonomous, or is it like persuading, in that it depends on others playing their role?

J L Austin, a British philosopher working around the mid-20th century, made important progress in trying to classify the linguistic things we do as acts of speech or speech acts. In his book How to Do Things with Words (1962), Austin called the act of uttering meaningful sentences – eg, ‘You should write “word” there’ – a locutionary act. Acts of that sort might be performed with any of a variety of further goals, from practising one’s diction or performing in a play, to commanding or persuading. And he called the act of persuading someone to replace one word with another, or offending someone by calling into question their literary abilities, a perlocutionary act. These are acts that depend on consequences beyond the merely linguistic: I will have persuaded you only if you accept the course of action I proposed; I will have offended you only if you take offence.

Austin distinguished locutionary and perlocutionary acts from a crucially important third sort of thing we do with words. Acts such as proposing, or stating, or telling, or asking, or warning, or refusing he called illocutionary acts. As can be seen from the list of examples, this sort of act is of central importance. Illocutionary acts are our basic moves in the game of linguistic communication. We rarely perform locutionary acts without doing so in order to perform illocutionary acts. We utter meaningful sentences in order to tell people things, or ask people things, or perform some other illocutionary acts. And although we also perform illocutionary acts with further ends in view – in this case, perlocutionary ends – we will have achieved our minimal communicative ends in performing an illocutionary act even if our further ends are frustrated. For example, it might be important that I warned you even if I failed to dissuade you.

Austin’s aim was to clarify the distinction between illocutionary acts and acts of the other two sorts, the locutionary and the perlocutionary. His thought was that we perform illocutionary acts in uttering meaningful sentences (hence the Latin il); and we perform perlocutionary acts by performing illocutionary acts (hence the Latin per). His idea was that the performance of a locutionary act in propitious circumstances constitutes the performance of illocutionary acts. And the performance of an illocutionary act can have among its further consequences the performance of perlocutionary acts. A useful intuitive marker of illocutionary acts is that it makes sense to try to do them by making explicit what one is up to. For example, it makes sense to try to warn someone that the ice is thin by saying: ‘I hereby warn you that the ice is thin.’ By contrast, it would be senseless to attempt the same trick with a perlocutionary act: ‘I hereby convince you not to skate,’ is hopeless.

So, you perform the illocutionary act of proposing that I replace one word with another in performing the locutionary act of saying: ‘You should write “word” there’; you perform the perlocutionary act of persuading me to rewrite by so proposing. We utter words in order to propose, or tell, etc; and we propose, or tell, etc, in order to persuade, inform, etc. We perform locutionary acts in order to perform illocutionary acts; and we perform illocutionary acts in order to perform perlocutionary acts. Illocutionary acts are therefore sandwiched between our merely uttering meaningful words and extralinguistic consequences.

From Austin’s perspective, our question about illocutionary acts is this. Can performing an illocutionary act depend on others playing their part? Can others figure in the propitious circumstances that enable us to perform such acts in performing locutionary acts? Austin seems to have thought so:

I cannot be said to have warned an audience unless it hears what I say and takes what I say in a certain sense. An effect must be achieved on the audience if the illocutionary act is to be carried out.

Austin’s plausible suggestion caught on. It is now common to hold that illocutionary acts are distinctively subject to an uptake condition of the sort that Austin suggested – to hold that you will not have proposed, or warned, or told, etc, unless and until an audience recognises that what you are trying to do is proposing, or warning, or telling. On this sort of view, illocutionary acts fall on the side of the things we cannot do unless others play their part.

The question whether Austin’s suggestion is right is important in thinking about freedom of speech. For, on one hand, we don’t think of this as concerning only our freedom to utter meaningful words – to perform merely locutionary acts. And, on the other hand, we don’t usually think that people’s freedom of speech is confounded just because they are unable to persuade other people, or to bring about any other such further consequences. So, we don’t think of freedom of speech as just a matter of freedom to perform perlocutionary acts. If that’s right, then in thinking about freedom of speech we are mainly thinking about freedom to perform specifically illocutionary acts.

Obviously enough, such freedom depends on freedom to perform locutionary acts, since it is typically only in performing such acts that we are able to perform illocutionary acts. One way of undermining freedom of speech is therefore to block the performance of locutionary acts – for example, by violence or threat. We can think of the effect here as a form of locutionary silencing. However, Austin’s distinction between locutionary and illocutionary acts makes space for the possibility that freedom of speech might be limited in another way: by disabling their capacities to perform illocutionary acts. And Austin’s suggestion that illocutionary acts can be subject to an uptake condition indicates a route to such disability. According to that suggestion, my performing an illocutionary act, such as warning or telling, depends on others being willing and able to recognise what I am trying to do. It is this that enables me to perform such acts. So, if others are unable or unwilling to recognise that I am trying to tell them something or warn them of something, then I won’t be able to tell or warn. That aspect of my freedom of speech will be distinctively undercut. We can think of such an effect as a form of illocutionary silencing.

Asking whether performing illocutionary acts depends on others, and so whether such illocutionary silencing is possible, has potential moral and political significance. One important arena in which the question has come to the fore concerns the standing of pornography. The right to produce pornography has sometimes been defended by appeal to the idea that it is a special case of a right to freedom of speech. However, that defence is effective only insofar as the exercise of that right doesn’t at the same time undermine others’ freedom of speech. And it has been argued – for example, by Catharine MacKinnon in Feminism Unmodified (1987) – that the production and consumption of (at least some forms of) pornography can have precisely that effect on women’s freedom of speech.

One form this argument takes is the following. According to Austin’s suggested uptake condition, performing an illocutionary act depends on being recognised to have attempted to perform it. If one’s audience doesn’t see what one is up to, then one will have failed to warn them that it’s late, tell them the time, refuse another drink, etc. Being recognised as having attempted to perform an illocutionary act depends in turn on our attempt taking a form that is recognisable as such by our audience. In ordinary circumstances, we can rely on a sort of match between, on the one hand, our abilities to make such attempts and, on the other hand, our audience’s ability to recognise such attempts. In those circumstances, it is typically straightforward to bring our audience to recognise our attempts. We are able simply to do what it takes to get our audience to see what we’re up to – for example, by uttering the words: ‘It’s time we left.’ However, being able to do this depends on our audience having appropriate abilities, and in particular on their having abilities to recognise attempts that match our abilities to make such attempts. For example, if I tried to tell a monolingual German speaker that it’s time we left by uttering the words ‘It’s time we left,’ then it is unlikely that they would be in a position to see what I was trying to do. Although I have an ability to make recognisable to fellow English speakers what I am up to, and although the German speaker has the ability to recognise such attempts by fellow German speakers, my ability and the German speaker’s ability don’t appropriately match. If that is right, then to meet Austin’s uptake condition depends on our, and our audience, having appropriately matching abilities.

Pornography trains consumers to see attempts to refuse sex as attempts only to pretend to refuse sex

The difficulty now is that such matching abilities can be lost. Suppose, for example, that I regularly attempt humour by pretending to warn you that there is a fire, and that you know this about me. In pretending to warn you that there is a fire, I exploit my ordinary abilities to make recognisable that I am trying to warn you that there is a fire. To begin with, you might have been fooled by my pretence, via the operation of your ability to recognise attempts to warn you. Eventually, however, it is likely that your abilities will change: you will come to see my attempts to warn you as, instead, attempts to pretend to warn you. You acquire an ability to recognise attempts to pretend, rather than attempts to warn. Furthermore, it is likely that you will lose your ability to recognise my attempts to warn you that there is a fire. A fire breaks out, and I try to warn you, but you lack the ability to recognise what I’m attempting: all you can see is an attempt to pretend. As a consequence, uptake fails. If Austin’s suggestion is right, then I am unable to warn you.

Now suppose that the cause of such a shift in your abilities is not my undertaking foolish attempts at humour, but rather malicious and misleading gossip about me by some third party. ‘He’s always pretending that there’s a fire,’ they say. Trusting them, you come to see my attempts at warning not as what they are, but rather as attempts at pretending to warn. Match between our abilities is lost. I try to warn you about the fire. Uptake fails. According to Austin’s suggestion, I am unable to warn you. Malicious gossip has led to my illocutionary silencing.

The argument that pornography can undermine women’s freedom of speech is based on the idea that men’s consumption of pornography can have similar effects on their abilities to recognise which illocutionary acts women are trying to perform. Crucially, the central claim is that some forms of pornography train their consumers to see women’s attempts to refuse sexual intercourse as attempts only to pretend to refuse sexual intercourse. Such pornography either presents women as always willing, in a way that can undercut evidence to the contrary provided by what they say, or it presents women as regularly only pretending to be unwilling. In either case, its uncritical consumption can undermine men’s abilities to recognise women’s attempted refusals of sexual intercourse as such. In that way, pornography can lead to failures of uptake. And according to Austin’s suggestion, that can make women unable to perform acts of refusal. If that is right, then the consumption of pornography can lead to a distinctive form of illocutionary silencing.

Before proceeding, it is very important to notice two things about this argument. The first, and most important, is that the advocate of the argument is not claiming that illocutionary silencing is the only negative potential consequence of men’s inability to recognise acts of refusal, or that it is the worst such consequence. (Or, indeed, that this is the only negative consequence of the consumption of pornography.) Obviously, there can be other, and far worse, consequences. What the advocate wants to emphasise is that, in addition to other potential consequences, the consumption of pornography can have specific consequences for women’s freedom of speech. The second thing it is important to notice is connected with the first. Failure to refuse sexual intercourse is not the same as consenting to sexual intercourse. So, even if the argument succeeds in showing that women can be prevented from refusing sexual intercourse through a man’s inability to recognise them as so attempting, it does not support the claim that the man is thereby permitted to undertake sexual intercourse.

If the argument that pornography can undermine women’s freedom of speech is successful, then it presents a challenge to the free-speech-based defence of pornography. That defence relied on the idea that censoring pornography would have only a negative impact on the distribution of freedom of speech. However, if pornography itself has a negative impact on the distribution of freedom of speech, then censoring it might have a net positive impact on that distribution. Deciding whether or not censoring pornography will have an overall effect on the distribution of freedom of speech that is positive or negative will depend on comparing the negative effects of pornography on women’s freedom of speech with the negative effects of censorship on pornographers’ freedom of speech. (In coming to such a decision, we would need to consider the potential effects of the consumption of pornography on women’s freedom to perform illocutionary acts other than acts of refusal. Although those acts are obviously of great importance, there is no obvious reason to think that they are the only potential casualties of the sorts of illocutionary silencing liable to be brought about by the consumption of pornography.)

Is the argument that pornography can undermine freedom of speech successful? One key premise in the argument is that the consumption of pornography can cause men to lose their abilities to recognise women’s attempts to refuse sexual intercourse. Whether that is so is a large, and delicate, empirical question. However, the premise is independently plausible, and there is some empirical evidence that consumption of pornography can have relevant effects on men’s attitudes and abilities. (An even more difficult empirical question, bearing on the comparison discussed in the previous paragraph, concerns the extent and strength of such effects.) However, even if we accept that premise, the argument seems to depend on the further claim that Austin’s suggested uptake condition applies to illocutionary acts in general, and to acts of refusal in particular. Should we accept that further claim?

The claim in question is that you cannot perform an illocutionary act, and in particular an act of refusal, unless your audience recognises that you are attempting to perform such an act, that you are trying to refuse. Is that true?

One thing that makes it difficult to address the general question is that it is not entirely clear what the criteria are by which acts of speech count as illocutionary, as opposed to locutionary or perlocutionary acts. Consider, for example, the act of cursing. It seems possible to curse when alone. However, that would show only that there are illocutionary acts that can be performed alone if cursing is an illocutionary act rather than a locutionary act. Consider, for another example, the act of informing. In this case, successful performance seems to require that one’s audience becomes informed, and that seems to depend, in turn, on uptake. However, that would show only that there are illocutionary acts that cannot be performed alone if informing is an illocutionary act, rather than a perlocutionary act such as persuading. So, the question whether illocutionary acts depend on uptake is hard to answer without more information about Austin’s three-way distinction among speech acts.

In the face of that difficulty, we might turn instead to specific examples, assuming for the time being that they count as illocutionary. However, that course raises difficulties of its own. Consider, for example, the act of telling someone something. In cases in which there is a failure of uptake, we sometimes speak of having tried, but failed, to tell someone something: ‘I tried to tell you, but you were too engrossed in the news to listen.’ But equally often, we stick to our guns: ‘I did tell you, but you were obviously distracted and didn’t take it in.’ So, our ordinary thought about telling doesn’t seem decisively to support Austin’s suggestion over its denial. Similarly, we sometimes speak of having refused without uptake: ‘You must not have heard me refuse your offer of another glass.’ Again, ordinary thought seems indecisive.

Illocutionary acts depend not on attempts to perform them being recognised, only on their being recognisable

A third approach would be to try to treat Austin’s suggested uptake condition as itself marking off the illocutionary acts, or one special range of illocutionary acts, from the other sorts of speech acts. Consider what is involved in your persuading someone of something. There is your uttering some words. There is your recognisably trying to propose something to your audience. There is, perhaps, your proposing something to them, whether or not they recognise it. There is your audience recognising that you are trying to propose something to them and, thereby, your opening up a channel of communication with them. And there is your exploiting that channel by persuading your audience and their thereby becoming persuaded. Bringing about your audience’s recognition of what you are trying to do is surely something that you do, albeit with their cooperation. Whether or not this act is captured by the ordinary notion of proposing, it still involves a distinctive sort of achievement involving the opening up of a channel of communication, without yet exploiting that channel in order to bring about further, perlocutionary consequences.

Acts of this sort are clearly subject to Austin’s uptake condition. Can they figure in the argument in place of either a less clearly delineated class of illocutionary acts or specific examples of acts, such as warning or refusing? The difficulty facing this suggestion is to connect acts of this sort – acts that depend on recognition – with our concern to protect freedom of speech. In advance of further discussion, it is not clear that acts that depend on uptake are the sorts of acts that we are concerned to protect when we are concerned to protect freedom of speech.

There is more to say about the three approaches we’ve considered to this point, but suppose that they all fail. Would that mean that the argument that pornography can undermine freedom of speech also fails? Not immediately. Although that argument is often presented as depending on Austin’s uptake condition, it is possible to develop a version of the argument that depends only on a weaker condition. The uptake condition, recall, has it that illocutionary acts depend on recognition. A weaker condition would be that the performance of illocutionary acts depends not on attempts to perform them being recognised by their audience, but only on their being recognisable.

The weaker condition is more plausible than the uptake condition, at least when we consider specific examples. ‘I told you, but failed to do so in a way that you were able to recognise’ seems immediately bad in a way that ‘I told you, but you were obviously distracted and didn’t take it in’ isn’t. Furthermore, the weaker condition can sustain a version of the argument. According to this version of the argument, the problematic effect of pornography isn’t preventing recognition of attempts to refuse. It isn’t supposed to serve, for example, as a momentary distraction. Rather, its alleged effect is undermining men’s abilities to recognise such attempts, and so rendering those attempts unrecognisable by their audience.

If that is right, then the argument surrounding pornography provides one example of how questions about the autonomy of speech acts can have potential moral and political significance. But it is plausibly only one example of a wider class. Acknowledging our dependence on others, even with respect to what we can do with words, is essential to understanding the proper extent of our freedom.

Guy Longworth

is associate professor of philosophy at the University of Warwick in Coventry, UK. His latest book, co-authored with Jennifer Hornsby, is Reading Philosophy of Language: Selected Texts With Interactive Commentary (2005).
Syndicate this Essay
Aeon is not-for-profit
and free for everyone
Make a donation
Get Aeon straight
to your inbox
Join our newsletter

Fremont Street, Las Vegas, 1982. Photo by Harry Gruyaert/Magnum

Mathematics for gamblers

If philosophers and mathematicians struggle with probability, can gamblers really hope to grasp their losing game?

Catalin Barboianu