A dogma of speech act theory

ABSTRACT

In this article I argue that the distinction between illocutionary and perlocutionary acts needs examination, not just in its details but in its philosophical standing. We need to consider whether the distinction is motivated by (sometimes unwittingly) assumed problematic philosophical assumptions concerning the nature of our dependence on the words of others and the rationality of speech reception. Working with an example of the act of telling, I argue against the idea that the distinction is self-evident or easy to draw. By developing an analogy with perception, I argue further that defending the distinction requires one to engage in an argumentative dialectic with powerful alternative positions. I end by suggesting that taking the challenge further would require us to look more closely at how passivity and rationality might be reconciled in the reception of speech.

KEYWORDS:

• Telling
• illocution
• perlocution
• rationality

I, like many of the other contributors to this volume, have been deeply influenced by Stanley Cavell's essay ‘Passionate and Performative Utterance’ (2005). In that essay, Cavell argues that the perlocutionary dimension of speech (the dimension involving the influence speech has on others) is unduly neglected and misconstrued as lacking in systematicity. The tendency to disregard the perlocutionary dimension of speech is for Cavell connected to a more general tendency in philosophy to neglect passion and passivity.

At first, what I took from Cavell's essay was the need to better understand the idea of a perlocutionary act. Having explored that thought for a while, I began to lose my grip on what it was that I was supposed to better understand. I came to think that we might not on a fundamental level know what this alleged dimension of speech is supposed to be. The theoretical criteria introduced for isolating this dimension might themselves be spurious. I started to think that the problem was not only or even primarily the neglect of the passions, but how the very distinction between illocutionary and perlocutionary acts encourages a forced and problematic separation between the speaker's action and the hearer's affection. In this paper, I want to open a case for the idea that it is far from clear both what the distinction amounts to and what its philosophical motivations are. What I argue here is that the distinction needs examination, not merely in its details but in its underlying assumptions and philosophical concerns.

In the limited space I have here, I will not be able to present in full the case I see for re-examining the distinction. I will, for instance, only consider the case of telling in this paper, and not work out the implications of what I say about this case for other kinds of action in speech.¹ I should say, too, that it is not my aim here or elsewhere to propose a new general taxonomy or a competing way of sorting different kinds of action in speech. The argument of the paper goes, rather, as follows. I propose a conception of what it is to tell someone something, which I call the reception-dependent conception. The reception-dependent conception of telling stands at odds with the illocutionary/perlocutionary distinction. With this alternative conception on the table, I examine the idea, recurring in the literature, that the illocutionary/perlocutionary distinction is in some sense natural, self-evident, or easy to draw. The core of this idea is that the distinction itself doesn't require much motivation. By pointing out conflicts among the central authors about where the supposedly natural line is drawn I want to unsettle this sense of obviousness. Further, by using my proposed alternative conception, I want to show that beneath these conflicts in where to draw the line, there is an interesting philosophical dialectic that requires consideration. I show that this dialectic shares a structure with a familiar dialectic in the philosophy of perception. In this way, I establish that the distinction is in need of examination, not just in its details but in its philosophical standing.

This argument is built entirely around the case of telling. It is therefore open to the objection that I can at most establish that telling is an odd case that defies the distinction, not that the distinction itself is unclear and in want of motivation. To this I say that telling is not, and I think rightly so, treated in the literature as an odd case. Rather it is treated as a paradigmatic case of an illocutionary act. If it turns out that telling does indeed defy the distinction, it cannot but cast a shadow on the distinction itself. At least it calls out for an explanation.

I. Telling and being told

What sort of act is telling someone something? Let's think about this question in the first instance through an example:

Sana is hurrying to school, afraid she's running late. She stops a woman, Eva, in the street and asks for the time. Eva pulls up her phone and tells Sana: ‘It's 8.30′. Sana sighs and stops running, realizing she’ll be late for class anyway.

In being on the receiving end of Eva's act of telling the time, Sana is taken from ignorance to knowledge (assuming that conditions are favorable). Just by telling Sana the time, it seems, Eva can affect Sana – give her a new piece of knowledge. On account of this knowledge, Sana orients herself accordingly; she slows her pace down and starts to think about what to do now.

I wish to introduce an understanding of acts of telling that takes the appearance seriously that Eva's act of telling as such changes Sana. On such an understanding, Eva's act of telling includes, as its end point, Sana's acquiring a new piece of knowledge.² On this view, Eva's act of telling the time would have failed, if Sana, say, had begun to suspect that Eva was trying to make her late for school, and therefore had been unable to trust Eva on this matter. Let us call such a notion of telling a reception-dependent notion of telling. Such a notion of telling is not completely new in the philosophical landscape. Here's John McDowell, giving expression to such a view,

One cannot count as having heard from someone that things are thus and so, in the relevant sense, unless, by virtue of understanding what the person says, one is in a position to know that things are indeed thus and so. If it turns out that things are not thus and so, or that although things are thus and so, the person from whom one took oneself to have heard it did not know it, one cannot persist in the claim that one heard from him that things are thus and so. One may retreat to the claim that one heard him say that things are thus and so. (McDowell 1998, 434)

McDowell focuses here on how the act of telling (McDowell uses ‘heard from’ in this passage but he might as well have used ‘told by’) is dependent on the speaker's having knowledge to transmit. But the act might also fail on the receiving end, such as in our imagined case when Sana had some reason for distrusting Eva. In such cases too, the act of telling fails, since the possibility for acquiring knowledge from Eva is foreclosed to Sana (436–437). McDowell's reception-dependent notion of telling emerges from specifically epistemological concerns, but here McDowell's view will stand merely as an alternative for thinking about acts of telling, an alternative which is not in any way obviously wrong-headed.³

If we look, however, at the options for thinking about action in speech in the contemporary philosophical landscape, the reception-dependent notion of telling is entirely off the grid. According to the received view on what it is to do things with words, an act of telling is an illocutionary act, and as such it is what I will call reception-neutral. On this received view, Eva's act of telling the time should be separated from Sana's coming to possess new knowledge. What happens to Sana stands as an effect – intended and foreseeable, perhaps, but still – to Eva's act of telling. The act of telling itself ends somewhere in the process before Sana knows that it's 8.30, and thus might be successful even though Sana doesn't end up knowing the time. On the received view, Eva's act of telling is the illocutionary act – something done in saying something, whereas giving Sana new knowledge is a perlocutionary act – an act done by saying something. On the reception-neutral understanding of telling, the question whether an act of telling is successful is independent of whether the recipient has acquired new knowledge.

Austin (1975) tries out various ways of marking the distinction between illocutionary and perlocutionary acts. Initially, he phrases the distinction in terms what we do in saying something (the illocution) and certain consequential effects this has on the audience (the perlocution) (101). But he soon realizes that the distinction can't be quite so simple and suggests instead that the distinction turns on whether the effect in question is ‘conventional’ or whether it involves ‘the production of real effects’ (103). Later, he tries to clarify by saying that uptake, the hearer's understanding of ‘the meaning and force’ of the utterance, somehow belongs with the illocutionary act (117), as well as certain normative effects (such as commitments and liabilities). Examples of illocutionary acts are informing, ordering, warning, and undertaking (110), and examples of perlocutionary acts are convincing, persuading, deterring, surprising and misleading (110), and getting to obey and getting to believe (119).⁴

The issue of uptake is one of the more controversial points of the distinction, both because Austin leaves the matter unclear and because it threatens the neatness of the distinction. Uptake is, after all, not a mere conventional or normative effect, but is, rather, the production of a real effect. Although many authors follow Austin in taking understanding of the meaning and force to be somehow central to illocutionary acts (e.g. Searle 1969; Langton 1993; Hornsby 1994), we also find authors who simply drop the requirement of uptake for a successful illocutionary act (Alston 2000), or who argues that ‘securing of uptake’ in Austin does not mean that the audience actually understands, but that the speaker makes it possible for the audience to understand (Sbisà 2013). All these authors, however, think that the illocutionary act of telling stops short somewhere before the recipient has acquired a new piece of knowledge. Telling is one act, transmitting knowledge another. Ordering is one act, getting to obey another. This is natural since going further would threaten the very basis of the distinction between illocutionary and perlocutionary act.⁵

How, then, is this reception-neutral notion of telling, this separation between act itself and its effects, motivated? Austin does not think it takes much motivation. In his view, there is a ‘natural break’ between the act itself and its consequences in the case of verbal actions. Austin says,

[T]he vocabulary of names for acts (B) [i.e. illocutionary acts] seems expressly designed to mark a break at a certain regular point between the act (or saying something) and its consequences (which are usually not the saying of anything), or at any rate a great many of them. (112)

Austin thinks that this supposedly natural break warrants him setting aside issues about what philosophers now talk about as ‘basic actions’ – i.e. supposedly minimal expressions of agency such as moving one's fingers or lifting one's arm. Austin takes care to show that the distinction between illocutionary acts and perlocutionary acts is not equivalent to a distinction between what he calls a ‘minimal physical action’ (such as crooking one's finger) and the rest of what happens (such as a trigger moving and a bullet ending up in a donkey's chest). Austin finds the notion of a minimal physical action ‘doubtful’ (112).⁶ Thus, Austin views this separation between act itself and its consequences as a distinction in ordinary language – in his view verbs such as inform, order, and warn come with such a separation inherent in them. The reception-neutral notion of telling is born out of a sense of inevitability. I now move on to challenge this sense.

II. A natural distinction?

Other canonical works in the field reproduce this sense of inevitability and self-evidence. Look at this passage from John Searle's Speech acts (1969):

Let us remind ourselves of a few of the facts we are seeking to explain. […] If I am trying to tell someone something, then (assuming certain conditions are satisfied) as soon as he recognizes that I am trying to tell him something, I have succeeded in telling it to him. Furthermore, unless he 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. In the case of illocutionary acts we succeed in doing what we are trying to do by getting our audience to recognize what we are trying to do. But the ‘effect’ on the hearer is not a belief or response, it consists simply in the hearer understanding the utterance of the speaker. (47)

Hence, Searle finds it natural to draw the line between illocutionary and perlocutionary act in the same place as Austin does. Another major figure in the field, interestingly, finds it equally natural to draw the line before uptake is secured. William Alston (2000) says,

Whether I told you that the dean is coming to dinner or asked you to bring me a towel does not hang on whether you heard or understood me. If you didn't, my communicative purpose has been frustrated. But it doesn't follow that I didn't tell or ask you. In response to a charge that I hadn't told you that the dean was coming, I might reply, “I told you all right; perhaps you didn't hear me”. (24)

In Alston, the passage is meant as an argument for setting aside any further discussion of the distinction between illocutionary and perlocutionary acts. He professes to find this distinction quite unproblematic (23–24). The passage from Searle doesn't so much present an argument as an initial and (purportedly) innocent description of the phenomenon his theory is intended to elucidate. (Searle's transition into the passage is ‘Let us remind ourselves of a few facts … ’.) But where Alston thinks that an act of telling does not as such involve uptake, Searle thinks it does. And both think that their respective position tracks some intuitive view on the concept ‘ … tells … ’.

But is there one intuitive view on the concept ‘ … tells … ’? Alston's form of words, which he adduces in support of excluding uptake from illocutionary acts, ‘I told you all right, perhaps you didn't hear me’, is not immediately suspect. Neither is it difficult to find formulations that appear to support Searle's interpretation, where the act itself stops with uptake, e.g. ‘I told him several times, and he did hear me, but he didn't believe me’. Matters are further complicated when we notice that ordinary language formulations can be adduced in support of a reception-dependent notion of telling, too. Consider, first: ‘I tried to tell him several times, but he refused to believe me’. The use of ‘tried to … ’ indicates a less than successful act, and thus that a refusal to believe on the part of the hearer constitutes a failure of the act of telling. Consider also the following conversation:

Dana:

Did you tell Carl the dean was coming to the meeting?

Peter:

I did.

Dana:

But why did he look so surprised when the dean entered the room?

Peter:

I said I told him, I didn't say he believed me.

Given Peter's unqualified initial affirming response to the question ‘Did you tell … ’, his last response ‘I didn't say … ’ appears not so much as an answer to the challenge, but as a witticism exploiting the flexibility of the concept ‘ … told … ’. What Dana was asking for was clearly whether Peter had managed to inform Carl, not whether Peter had done something that came close to but fell short of Carl's coming to know the dean was attending the meeting. Hence a question of the form, ‘Did you tell … ’ is (in many cases at least) a question as much about what the recipient now knows, as about what the speaker uttered in her presence. Radford (1969, 329) notes precisely this when he writes,

There are, however, contexts in which we can say indifferently either “I told him (e.g. that you would not be coming to his party) but he did not believe it” or “I tried to tell him … but he did not believe it.” In the latter case the implication is that I failed to tell him (just because he would not believe it) or that he at any rate would not be told, though of course we might insist that he jolly well was told. This bit of linguistic schizophrenia does suggest that “told” and “tell” have two different senses – a stronger, which strongly implies tellee-acceptance, and a weaker which does not itself imply or require or even suggest this at all.

Now, there is room for extensive debates about this stock of (seemingly) conflicting evidence on the ordinary notion of telling and being told. There are plenty of distinctions, between e.g. what is said and merely implicated, that philosophers of language would be keen to invoke at this point. It is not my ambition here to stake out a position in such a debate. Instead I will simply note two things. First, a philosopher who wishes to base the distinction between illocutionary and perlocutionary act on a supposedly natural break somewhere in the chain before the hearer is affected encoded in ordinary language had better have a view on how to sort these seemingly conflicting cases. She also needs some explanation for why different philosophers have drawn the allegedly natural line in different places. Second, it seems to me unlikely that any interesting sorting would be independent of a host of quite difficult philosophical considerations. Hence, the idea that the distinction is natural, self-evident, or easy to draw cannot be upheld.

By putting my alternative conception of telling on the table and arguing that it is no more and no less a natural or self-evident conception of telling than the (different and conflicting) ones that build on the illocutionary/perlocutionary distinction, I have shown that the distinction requires more philosophical work for its defense than is often mounted. Next, I will show that beneath the distinction there is a philosophical dialectic familiar from the philosophy of perception. Articulating this parallel gives more substance the idea that the distinction carries problematic philosophical assumptions and needs to be re-examined. But before I go on, I will respond to an objection to the line of reasoning so far.

Could it be, this objection goes, that the best understanding of this distinction is not as a technical distinction underpinned by some general theory, but as a rough and ready tool for helping us reflect on the various dimensions of importance for action and responsibility in language-use? Nancy Bauer (2015, 87–112) argues that we should not see Austin primarily as a theorist but as proposing a shift in perspective in how philosophers approach language. She doesn't contest that there is theory in the book, but argues that Austin's primary contribution lies in how he encourages us to view language-use (including the writing of philosophy) as an ethical matter. Bauer says,

The theory [in How to do things with words], I submit, is a means for us to grasp something philosophically deep about what it is to use words, not a philosophical end-in-itself. That means that even if the theory turns out to be wrong-headed in its particulars, it might still do something important philosophically. (2015, 92)

The illocutionary/perlocutionary distinction should, she could argue in response to me, be taken up in this spirit too. It should be seen as non-technical and context-sensitive tool to consider how different aspects such as intention, convention, or psychological effects matter in particular cases. It points us towards interesting places to look, but we have to use our judgment when we bring the distinction to bear on particular cases.⁷

My response to this objection is first, to contest that we should separate taking up the book as theory and assigning to it a philosophically important ethical function. For such a separation to be meaningful in the way Bauer suggests, the theory can't be too problematic. This is, I think, signaled when Bauer claims that the theory might work as an ethical tool even if it is ‘wrong-headed in its particulars’.⁸ My worry about Austin's theory is not that it is wrong-headed in its particulars, but that it is wrong-headed in a more fundamental, and philosophically interesting, way. As such it requires a theoretical critical engagement. Otherwise it will, I submit, hinder rather than help us realize something philosophically deep about what it is to use words. In particular, my worry is that it might hide the specific ways in which speech-acts as such can change others. This aspect of language-use might turn out, because of the way the distinction sets the issue up, to be hard to think properly about.

Second, if we focus attention on the part of the objection that claims that the distinction is useful taken up as non-technical and context-sensitive, I would argue that it is too unclear and philosophically problematic to function contextually this way. For that to work we need to have enough of a grip on what the terms of the distinction are to be able to decide in a context where to draw the line. I think we do have various resources to decide (often and for specific intents and purposes) where to draw the line between what is done and what is an effect of what is done. This is a perfectly general point about action and effects. But the notion of a perlocutionary effect is, I would have thought, supposed to (1) be specific to action in language, (2) say something interesting about a certain class of effects that speech can have (such as getting to believe, getting to obey). Hence I would agree that we can in a given situation mark a difference between action and effect. But for that we need the notions of action and effect (along with a host of accompanying conceptual resources), not the notions of an illocutionary and perlocutionary act. Indeed, my point is that the distinction actually might distort our capacity to adequately assess particular cases.

With these remarks, I want to turn to elaborating the parallel with perception. Let's return to the dialogue with Dana and Peter above, which ends with Peter saying, ‘I said I told him, I didn't say he believed me’. Compare that dialogue with the following conversation:

Peter:

I saw three horses outside our window just now!

Dana:

But there were no horses there.

Peter:

I only said I saw three horses, I didn't say there were any horses there.

This dialogue is similar to the previous dialogue between Dana and Peter, in that the last response does not meet the challenge as much as constitute a witticism exploiting the flexibility in the concept ‘ … sees … ’. Although we in many cases use ‘ … sees … ’ in such a way that it doesn't leave open whether the purported object is there, there is nothing prima facie odd about someone reporting on a previous hallucinatory experience by saying ‘I saw people floating in space’. Here we seem to be faced with two different uses of ‘ … sees … ’, one that floats free of the existence of the object seen and one that depends on the object being there. Let us call the second use of ‘ … sees … ’ the factive use, and the first the non-factive use. Now, anyone familiar with issues in the philosophy of perception will recognize the beginnings of a familiar dialectic here. Faced with the discrepancy between these uses, we have (at least and roughly) philosophers who take the non-factive use to be primary for understanding perception, and those who give priority to the factive use.

Returning to the case of telling, I propose that the non-factive notion of ‘ … sees … ’ can profitably be compared to the reception-neutral notion of ‘ … tells … ’ where it is independent of whether the audience actually comes to possess a new piece of knowledge. The factive use of ‘ … sees … ’ corresponds to the reception-dependent use of ‘ … tells … ’, where the hearer's acquisition of knowledge is included in the act. Analogously to the people who take the non-factive use of ‘ … sees … ’ to be primary for understanding perception, Austin, Searle, and Alston all think the reception-neutral use is fundamental for understanding telling and being told.

In the case of perception, the dialectic is not usually taken to be resolvable on grounds of ordinary language only. The debate is not (primarily) about what notion of ‘ … sees … ’ is most natural or used most often. Rather, the debate concerns what these uses reveal about perception as a capacity, and perception as a way of acquiring knowledge.⁹ So we might ask whether an argument from naturalness can be at most a first step towards staking out a position with respect to what sort of act we are engaging in when we tell someone something. Why should we privilege the reception-neutral use over the reception-dependent one?

With respect to perception, part of what is at stake is precisely what conclusions we should draw from the possibility of failure. One familiar argumentative move for privileging the non-factive notion of ‘ … sees … ’ is an argument from illusion. Such an argument could go, roughly:

From the perspective of the perceiver, a situation where she is undergoing a perceptual illusion and a situation where she is gaining perceptual knowledge might be indistinguishable. Therefore there is a common element to these situations, and the non-factive use of ‘ … sees … ’ picks out this common element.

This argument purportedly establishes that there is a common ingredient to veridical and non-veridical cases of perceptual experience. Something must then be added in the veridical case, for the situation to amount to a case of latching onto a real object.

Now, we might construe a corresponding argument in the case of telling:

From the speaker's perspective, it might seem as if the audience has heard, understood and been informed, but the audience might have been distracted, temporarily deaf, or secretly incredulous. In such a case, the speaker's action is exactly the same as in the case when the audience takes in what she says. The only difference is in the effect on the audience. Hence the failed case is fundamental, in that it allows us to see the common ingredient in failed and successful cases alike.

The point I want to make by sketching a parallel structure here is that once we have gone down this road of articulating a version of the argument from illusion, it will also become more clear that the reception-neutral notion of telling is, in fact, controversial. The conclusion Austin, Searle, and Alston sought to establish was that the act of telling itself stops short of the audience's coming to possess a new piece of knowledge. But if this conclusion actually depends on mounting a transposed version of the argument from illusion, there are familiar ways of resisting it.

In the case of perception, one way such an argument has been countered is by putting pressure precisely on the idea that the subjective indistinguishability means that there need be an epistemologically salient highest common factor between the failed and the successful cases. To deny such a highest-common-factor view is to embrace some form of disjunctivism about perceptual experience (See e.g. Hinton 1967; Snowdon 1980–81; Martin 2002; McDowell 2009). Although I have never seen the move been made with relation to speech-act theory specifically, there are several versions of disjunctivism about actions and agency out there (e.g. Brewer 1993; Haddock 2005; Hornsby 2008; Stuchlik 2013; Lockhart and Lockhart 2017). Such proposals try to meet the sort of argument outlined above but in a more general version that covers all material action.

A defense of a reception-dependent notion of telling, thus, could argue that the good case – where the speaker has knowledge to transmit, and the hearer is in a position to simply receive this knowledge – is of a distinctive and paradigmatic kind. Bad cases, i.e. cases of telling lies, or cases where the hearer mistrusts the speaker and thus fails to pick up the knowledge made available, are both different in kind and have a derivative status. To spell out in more detail what such a defense would look like is not my aim here. My aim is rather to undermine the argument from naturalness by showing that what Austin, Searle, and Alston all present as a natural distinction is (1) a distinction that isn't natural or self-evident, (2) instead requires an engagement in an argumentative dialectic with significant alternative positions.

Now, someone might argue that what I have done is to propose a new account of illocutionary acts, one that includes the hearer's being affected in a certain way. But I have not, on this line of reasoning, argued that there is no use for the illocutionary/perlocutionary distinction. The first thing to note is that I have not in fact proposed an account of equally general purport as the ones tied to the notion of an illocutionary act. I have not argued that what I say about telling can be simply transposed to all kinds of things we do with words. What we are left with is the suggestion that we need in some (perhaps many) cases be prepared to see the production of response in the hearer as an integral part (not merely an effect) of what we do with words. And the notion of an illocutionary act, tied as it is to the notion of a perlocutionary act, can only work to make this response-dependence obscure.

Second, the notion of an illocutionary act is a technical notion, and it is therefore tied quite intimately to its motivations and accompanying distinction. To say that I merely have redefined the notion of an illocutionary act would be to say that my conception of telling is a recognizable heir to the notion of an illocutionary act as we find it in Austin and beyond. But my account of telling rejects too much in Austin's framework for it to be considered an inheritance of Austin's notion. Not only does my account include a response in the hearer which Austin and others explicitly relegate to the perlocutionary dimension, I also don't place any emphasis on the need or naturalness of (always and everywhere) separating the creation of obligations and commitments (normative effects) from affecting material changes (the production of natural effects).

I would respond in a related manner to the objection that all I have managed to show at this point is that it is a matter of dispute whether telling should be understood as an illocutionary or a perlocutionary act. Hornsby (1994) follows Searle in using the example of telling in her exposition of illocutionary acts while explicitly noting both that the role of speaker acceptance in acts of telling is ‘complicated’ (203), and that the line between illocutionary and perlocutionary acts is not ‘hard and fast’ (205). Why should we not say, then, that telling is a perlocutionary act or a borderline case between illocutionary and perlocutionary act?

In response to this objection I would point out that if telling comes out as a perlocutionary act, then the perlocutionary act turns out to be a far more interesting category than has been generally acknowledged. Telling someone something is as central a case of doing something with words as it gets, and if it cannot serve as an example of an illocutionary act, the common identification of speech-acts with illocutionary acts will start to look like a quite unhelpful stipulative restriction. Second, since telling is so often used as an example of an illocutionary act, this should make us wary of accepting other canonical classifications. Again, the philosopher who wishes to make use of the distinction should be prepared to argue for the usefulness of the distinction and for her classification of a particular kind of speech-act.

But the more important point is that it is difficult to see what becomes of the distinction between illocutionary and perlocutionary act if we think of telling as a perlocutionary act – or even as a borderline case. Doing so cannot preserve Austin's sense that we’re drawing a line between the act itself and its consequences on the audience. The very point of putting the reception-dependent notion of telling on the table is to show that such a notion resists a division in terms of what the agent was doing (or the act itself) and the effects on the audience. If we accept the reception dependent notion of telling, it would be highly misleading to call telling a perlocutionary act or a borderline case. If we do so, we would obscure the point that the reception-dependent notion of telling does not pry the effect on the audience (their coming to possess new knowledge) apart from the speaker’s act.

My conclusion so far is that the distinction is ill-motivated and covers over an unexamined and difficult philosophical dialectic. This is a situation that calls out for diagnosis. What explains its attraction? In the concluding section, I will very briefly sketch one answer to this question (I suspect there might be more than one interesting answer to it).

III. Concluding remarks

I introduced the reception-dependent notion of telling via McDowell's paper on hearsay. For McDowell, it is an important task both in the epistemology of testimony and in the epistemology of perception to arrive at an adequate picture of the relation between receptivity and rationality (McDowell 1994, 1998). In the previous sections I began to articulate the philosophical dialectic that I see underneath the distinction illocutionary/perlocutionary act. It would be a part of a further examination of this dialectic to turn specifically to the question how we should think about the rationality of speech-reception. My suspicion, in short, is that the distinction seems inevitable in part because it is taken to be a necessary prerequisite for preserving the rationality and freedom in taking up and responding to the speech of others.¹⁰ Here I consider McDowell's work on reconciling receptivity and rationality a resource in that it proposes an alternative understanding of rationality, one that doesn't force this distinction upon us.

But this is a topic that awaits another time and place. In this paper I have argued that the distinction between illocutionary and perlocutionary acts is unclear and in want of motivation. What I urge going forward from my conclusion is that the role of affection and response in the hearer should be reconsidered. I am not suggesting that we look for some one way in which hearer-responsiveness might be involved. Austin himself took a special interest in highly formalized, even ritualized, forms of speech, such as christening, marrying, and pronouncing sentence. In such instances, there might be a point in thinking about the acts as primarily setting up a normative space with new entitlements and commitments, and thus as relatively independent of the audience's responses. This is connected, it seems to me, to the fact that such forms of speech are not addressed to any one particular person, but are moves within a legal or (quasi-legal) system.

When we consider, however, acts of speech central to everyday person-to-person communication such as greeting, asking, telling to (ordering), faulting/accusing, complaining, avowing feelings, arguing, giving advice, warning, threatening, congratulating, sympathizing, provoking, joking, and telling a story, my sense is that the question of what (if any) response on the hearer's part is required for the acts to be successful needs to be re-opened. And in thinking about such cases, we should, I urge, keep in mind that there need be no conflict between a response's being passive and it's being rational.

Acknowledgements

I want to thank Åsa Burman, Martin Gustafsson, two anonymous reviewers, and audiences at Åbo Akademi University and at the conference Perlocutoire! (Paris1/Sorbonne), for valuable comments on this paper. I also thank Jennifer Lockhart for asking exactly the right questions when I was in the early stages of writing the paper. Work on this paper has been supported by the Academy of Finland and The Kone Foundation.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Work on this paper has been supported by the Academy of Finland and The Kone Foundation.

Notes

1 See Martin Gustafsson's paper (this issue), ‘On the distinction between uptake and perlocutionary object: The case of issuing and obeying orders’, for considerations about the speech-act of ordering that are of a piece with what I say in this paper. My paper has been greatly improved by discussions with Martin Gustafsson.

2 Someone might wonder why I do not take telling to produce belief rather than knowledge. After all, it might be argued, people can be telling lies or think they know things they do not know. Hence, why not say that telling produces belief, and that some other condition needs to obtain for the situation to amount to a transmission of knowledge? This question leads us into intricate questions in epistemology. I will not resolve these issues, but I will say that I do not think we should simply assume that we should derive our account of telling from bad cases, such as telling lies or speaking from ignorance. Starting our account of telling from the best-case scenario means that we involve knowledge from the start, and we represent speakers and hearers as having the capacity to give and receive knowledge. The worry about starting from the bad case is that this starting point threatens to prevent us from being able to have an adequate hold on this capacity. Thanks to Peter Myrdal for pressing me on this point.

3 I think that a full consideration of the philosophical motivations for the illocutionary/perlocutionary distinction will require us to address anew the issue of what it means to think about speech as a rational transaction. I will make a few remarks about this issue at the end of this paper.

4 How to do things with words is comprised of edited lecture notes, notes that present thought in progress. Hence Austin's conception of this distinction and its difficulties evolves during the lectures. Since my aim here is to criticize a distinction that originates with this work but has been taken up and solidified through the works of others, I present Austin's thought in a way that neither captures its method nor its (intimately connected) texture and style. See Bauer (2015) (chapters 5 and 6) for a reading of Austin's text that emphasizes the philosophical importance of paying attention to Austin's style and method.

5 Rather, some philosophers argue that there are illocutionary acts that have an ‘associated’ or ‘correlated’ perlocutionary goal, i.e., a kind of intended effect that is non-contingently related to that particular act-type. So passing over knowledge or getting to believe could, then, on such a view, be described as the associated or correlated perlocutionary goal of the act of telling (Searle 1969, 46; Cohen 1973). Such a view keeps the separation between the illocutionary act itself and its perlocutionary effects, while trying at the same time to accommodate the fact that certain (supposedly) perlocutionary effects appear to have some role to play in our understanding of at least some types of illocutionary acts, such as telling and warning.

6 Even though Austin expresses doubts about the notion of a minimal physical action, Austin toys with an argument, commonly used as a motivation for the idea of a basic action, for moving the boundaries of the action itself closer and closer to the agent's body (111). The relation between the illocutionary/perlocutionary distinction and the basic/non-basic action distinction is itself an interesting topic to explore. Classical papers on the notion of a basic action are Danto (1965) and Davidson (2001, originally published 1971). See Lavin (2013) and Ford (2014) for critical perspectives on the basic/non-basic distinction.

7 Thanks to an anonymous reviewer for articulating this objection.

8 I am not excluding here that a text that is completely wrong-headed might work as an ethical tool. I take it that it could, in various ways, e.g., by provoking us to think through what it wrong with it or by giving insight by way of making us horrified at its conclusions. But as I understand Bauer, she is not suggesting that Austin's book works as an ethical tool in this way. Its point lies not, as it were, in our coming to reject it wholesale.

9 This is a point emphasized by Martin Gustafsson (2012).

10 The only place I have found this connection articulated explicitly is in the work of Jürgen Habermas (1984). For him the distinction is explicitly tied to a concern with understanding communication as a rational and free transaction.