2. Acts

‘She intends to write John an email’. ‘She knows how to make a cup of tea’. Looking at these examples, it seems that that which is intended by someone who intends to do something is expressed with an infinitival phrase (‘to write John an email’), and that which someone knows how to do can likewise be so expressed. That is how it seems. And since infinitives, and infinitival phrases are not sentences, prima facie that to which a person relates in intending to do something or in knowing how to do something is not expressed using a truth-evaluable sentence. When belief, or any other attitude, is said to be propositional, the assumption is that the place of ‘p’ in an instance of the schema ‘She atts that p’ is to be filled by a sentence ‘having truth-conditions’. A sentence has a main verb which is inflected; it is inflected for tense, and how it is inflected depends upon its subject. But the infinitive of a verb is tenseless, and not bound to a particular subject. Given that infinitives are subjectless, intend and know how appear not to be propositional attitudes.

If ‘know how’ and ‘intend’ are not propositional attitudes, then what are they? Well, if someone actually did something that she knew how to do, there is something which she knew how to do and which she did. If someone actually will do something that she intends to do, then there is something she intends which she will have done and might be doing even now. So it seems that the objects of knowledge how, and of intention, for which infinitives are used, are things that someone (anyone) may have done, or be doing, or will do, in the role of intentional agent. I shall sometimes use the word ‘acts’ for such things – for do-able things, one might say. It seems that in English, they are not only infinitival phrases that may specify acts, but also phrases containing so-called zero, or bare, infinitives (which lack the ‘to’ of the official infinitive) and gerunds. Consider: ‘I intend to take warm clothing [infinitive]; ‘You had better take warm clothing’ [zero infinitive]; ‘For him, taking warm clothing is recommended’ [gerund].

If acts are what we do, then it could introduce confusion to speak of things we do as actions. That is one terminological option, of course: many would say ‘actions’ where I say acts, and that can be fine. But it is not an option for someone who tells what has been called the standard story of action, who means by ‘actions’ events and takes events to be concrete unrepeatable particulars (such as Meyer’s climbing of Kibo).Footnote ¹ What I am calling acts are sometimes called act types, and contrasted with actions, now sometimes said to be act tokens. But that can’t be right – not at least if the type/token relation is so understood that which type some type is depends upon what condition something must satisfy to be an instance of it. An event-denoting nominal, being derived from a whole sentence (such as ‘Meyer climbed Kibo’), will not only specify an act (such as climb Kibo) but also make reference to an agent (such as Meyer). If there is an action in the sense of the standard story, a particular agent must be related to an act in a certain way – sc. by being one who did it.

To say that the actions of the standard story do not stand to acts as tokens do to types need not be to reject every sort of type/token distinction in respect of acts or to deny that ‘particular act’ must lack any definite sense.Footnote ² An agent may intend something quite particular: she might, for instance, intend to make an online donation of £100 to Oxfam using her computer at work when she arrives there. The act she then intends is of the type giving to charity. But to think of the particular act, even in its connection with a particular agent, is not yet to think of a particular event in its own right, such as an action has been supposed to be. Suppose that you now intend to make a donation in a few minutes’ time. There is not now any particular which is your making a donation; and we don’t have to wait until there is an event denoted by ‘your making of a donation’ before we can speak of the particular act that would come to have been done in a few minutes’ time if you should do what you intend.

When I give the label acts to things such as agents may relate to in one or another way, an ontological claim might seem to be made. But I mean to introduce nothing that could be contentious. I don’t suppose that acts are properties: verbs, when predicated are tensed, and acts are given by infinitives of verbs.Footnote ³ If I say that an act is something that a person may stand to in some relation, then I mean something analogous to what is said when a proposition is said to be something that a person may stand to in some relation – a relation such as belief. Debates about propositions are plentiful, of course; and insofar as different so-called attitude words seem to introduce objects of different sorts, it may seem that no word such as ‘proposition’ could really be an apt catch-all term for what a ‘that’-clause may introduce. But I don’t pretend to sort out any metaphysical issues here.Footnote ⁴

3. Intention: sources of the propositional view

Many have treated intention as propositional in consequence of taking a certain view about how the mind is to be accommodated in a physical world. According to them, the attitudes are states of mind which have a role to play in an account of a sort which they suppose that a philosopher of mind must give – an account in which such states are sometimes causes, sometimes effects, where the cause/effect relations are, as one writer puts it not ‘foreign to science’. They say that in explaining interactions between any subject of the attitudes and the world beyond her body, the attitudes must be attributed contents, and that explanations of the subject’s interactions with the world must advert to the relations among those contents. If a systematic account is to be reached of the cause/effect relations which the attitudes (supposedly) enter into, then uniformity of objects of the states of mind will seem to be key. And then if ‘propositions’ is the word for what are believed, known, hoped and feared, what are desired and intended must also be propositions, expressible with truth-evaluable content. (When desired propositions are in the picture, the ingredients are all in place for the so-called ‘desire-belief theory’, whence the standard story of action.Footnote ⁵)

Another source of the supposition that an intention’s object has the truth-evaluable content of a proposition is a line of thinking about practical reasoning – about reasoning concluding with an intention. John Broome and David Velleman both make something out of the point that an agent may express her intention by saying ‘I will Φ’. In Broome’s hands, the claim that intention has a propositional content is then got from a quick argument, and an assumption. Broome points out that one who says ‘I will Φ’ can always be named. Call her Chris. The object, then, of Chris’s intention, expressed when she says ‘I will Φ’ is the proposition that Chris will Φ’. And now comes Broome’s assumption: ‘an infinitive and a ‘that’ clause are alternative ways of expressing a proposition’. Once the argument and assumption are in place, if we find ourselves saying that Chris intends to Φ’, we may take ourselves to express the very proposition that Chris expresses in saying ‘I will Φ’ (see Broome Reference Broome, Bermúdez and Millar2002).

Velleman agrees with Broome in basing a claim about intention’s contents on a conception of practical reasoning. This is in spite of a disagreement between the two which runs deep – a disagreement about intention’s direction of fit. Whereas in Broome’s book, a person stands to the content of her intention by being set to make it true, Velleman says that a person stands to it by ‘representing it … as true with the aim of doing so only if it … really is true’ (Reference Velleman and Hume2007, xix). It seems that when the contents of intention are treated as propositions, something other than a view about the nature of intention is at work.

Some philosophers have no recourse to argument when they treat intention as propositional. They may enquire, for instance, into ‘the involvement of belief in intention’, and then ask whether ‘if you believe you intend X then you believe X’ (see e.g. Bratman Reference Bratman and Robertson2009, 31). They don’t stop to ask what sort of schematic letter ‘X’ might be supposed to be: they don’t raise the question whether the same words could always be written after ‘intend’ and after ‘believe’. To that question, however, the answer would seem to be a simple No.

4. Intentions: reasons for the contrary (‘Act’) view

The word ‘intend’, like ‘believe’, can be followed by the word ‘that’. But I shall argue now both that ‘intend’ is not propositional even when followed by ‘that so-and-so’, and that ‘intend’ followed by an infinitive must come first in the order of understanding.

One can distinguish between two sorts of ‘intends that’ sentences – according as the one to whom an intention is ascribed isn’t, or is, mentioned in the scope of that which is intended. On the one hand, there are examples such as ‘He intends that the lecture start on time’ and ‘She intends that Johnnie go to the local school’. And on the other hand, there are examples such as ‘She intends that she pay the bill’ and ‘He intends that he give up smoking’. Noticing these two different sorts, someone might think that ‘intends to’ must be understood in terms of ‘intends that’. The thought would be that all the occurrences of ‘intend’ will conform to a single pattern if ‘intend to Φ’ is everywhere glossed with ‘intends that she Φ’.

The first thing to notice is a feature of what comes after ‘intends that’ in all these sentences. ‘The lecture start on time’; ‘Johnnie go to the local school’; ‘she pay the bill’, ‘he stop smoking’. The verbs here are subjunctives. But then the ‘that’-clauses that contain them are not propositional clauses – not at least on the usual view according to which propositions are truth-evaluable.

We should ask how sentences containing ‘intend’ followed by a ‘that’-clause are understood. In cases such as ‘He intends that the lecture start on time’, one is likely to think of someone who intends that it should be that p as someone who, even if she cannot herself simply bring it about that p, nonetheless has it in her power that it come to be that p. (Perhaps she can’t herself start the lecture on time, but one way or another can see to it that it should start on time.) If this is applied to the case in which the one who intends is mentioned in saying what’s intended, then the result will be that someone who intends that she pay the bill has an intention which is fulfilled if she sees to it that the bill will have been paid by her. Well, one might have thought that someone who intends to pay the bill can fulfil her intention more straightforwardly: she can pay the bill. True, someone who is going to pay the bill might say ‘I intend that I pay the bill’: she might have a reason to say this if she wants to remove any suggestion that she intends to pay directly and forthwith. (‘I intend that I’ll pay, but am not going to until the deficiencies are remedied.) ‘She intends that she Φ’ appears to introduce a sort of distance between intention and its fulfilment – a distance never heard when ‘She intends to Φ’ is said. The same sort of distance may be heard even when A’s intention that so-and-so is not an intention that A herself act. Thus, one is more likely to think it still to be in question how A might bring it about that Johnnie should be going to the local school if one is told that A intends that Johnnie go to the local school than if one is told that A intends Johnnie to go to the local school. In the latter case, one is likely to take it that Johnnie’s going to the local school is immediately in A’s power.

Suppose that Aelfric is a very powerful man with a finger in many pies. Aelfric intends that the Chair of the Board of Management speak to the Regent. But Aelfric has forgotten who the Chairs are of all the Boards. Luckily (as he thinks), Aelfric’s office wall is adorned with labelled photographs of all the important people, and he can see a photograph labelled ‘Chair of the Board of Management’. So Aelfric now decrees ‘HE, that man, shall speak to the Regent’. Aelfric, however, fails to realize that the photograph he sees is actually a photograph of himself. So now Aelfric intends that HE should speak to the Regent, and actually HE is Aelfric, but he does not intend to speak to the Regent. The example shows that Broome is wrong to think that an intention of Chris’s to act in some way is an intention on Chris’s part that Chris act in that way.

What emerges here about the difference between ‘intend to Φ’ and ‘intend that she Φ’ not only stands in the way of assimilating the use we make of the two: it helps to show why ‘intends that’ cannot substitute for ‘intends to’. Intend to and intend that have it in common that the one who intends takes the fulfilment of her intention to be in her power. But someone who has it in her power to Φ may be someone who can Φ. And no-one can take anything to be in her power unless there are some things she can do. It would be impossible, then, that one should ever do anything unless one was capable of intending to do things of a kind one can do; and it appears that one could not know what it would be to intend anything unless one knew what it is to intend to do something. ‘Intends to ––’ is irreducible, then. It must be allowed that someone who intends to Φ intends just that (whatever it is).

5. Knowing how

In the case of intending, there were two locutions ––’intend to Φ’ and ‘intend that so-and-so’. In arguing against the eliminability of ‘intend to Φ’, I have meant to show that there is much more than a prima facie case for treating ‘intend’ as introducing acts. When it comes to knowing how, there are not two locutions, but a single one, and controversy about how it is understood. The question might crudely be put as a choice between: (1) A [knows how] [to —–]’ and (2) A [knows] [how to —–]. In instances of (1), a genuinely infinitival phrase is a genuine component of the sentence, and A is joined to an act by ‘knows how’. In instances of (2), ‘how to Φ’ is to be glossed so that propositional knowledge is attributed to A: it might be replaced by ‘of some way (or ways) that it (or they) are ways of Φ-ing’. A gloss of some such sort is given by those who defend propositionalism about ‘know how’ by reference to semantic theory. They say that to know how to do something is to know ‘a truth about the world’.Footnote ⁶

When it comes to ‘know how’, the propositionalists are intellectualists in a sense that Gilbert Ryle gave to that word ([Reference Ryle1949] 2009). In opposing intellectualism, Ryle set himself against philosophers who thought that ‘the defining property of a mind … was the capacity to attain knowledge of truths’ (15). Ryle’s anti-intellectualism was hardly a thesis in syntactic and semantic theory. What was he driving at?

Consider what someone may want to know if he asks someone else whether she knows how to do something. A says to B ‘Do you know how to such & such?’ Two rather different things might be going on. On the one hand, A might seek information about how the act in question can be done. Perhaps he wants to do it himself, and if B knows how to, she can tell A what he needs to know. On the other hand, it might be that A wants to learn whether she has the knowledge in question. There is probably a case of the first sort, in which facts about how something can be done are sought, when he asks her ‘Do you know how to get to the station?’ But if he asks her ‘Do you know how to swim’, it is more likely that he wants to find out whether she herself knows how to swim; he need have no interest in how swimming can or might be done. The distinction between cases of the two sorts points up a difference that one can mark between knowing how something can be done, and knowing how to do the thing oneself. Intellectualists for their part do not deny that there is some such distinction as these examples are meant to illustrate.Footnote ⁷ But Ryle, I imagine, under the head of ‘knowing how’ was interested in knowledge of how to do something oneself.Footnote ⁸

An example may help to bring out the idea of knowledge a person has which is knowledge specifically of how to do something herself. Someone asks you how to get to the refectory. If the person who asks this can easily walk there, you will tell her a walkable route, up the stairs perhaps. But if the person who asks is in a wheelchair, you will tell her a route a wheelchair user could adopt, with ramps and elevators perhaps. In either case you hope to convey facts knowledge of which combines with capacities she already possesses (to walk, to manoeuvre her wheelchair) which would enable her to get to the refectory herself. You take her perspective. You know that she is not interested simply in how the refectory can be got to; you want to ensure that she knows what she needs to in order to come to know how to get to the refectory herself.

Ryle surely was concerned with what a person knows by virtue of which she can do something herself – knowledge which she can employ in practice. Ryle spoke of ‘the exercises of knowing how’ as ‘deeds, overt or covert’ (34). And he spoke of our activities as ‘displaying qualities of our minds directly’ (15). In Ryle’s view, it can be that when someone is seen to be doing something, it is manifest that she knows how to do it. So if A had seen B swimming in the pool, he would have had no need to ask her whether she knew how to swim. But still it can be that when A asked B whether she knew how to get to the station, his interest was in learning some of the facts known by someone who knows how to get to the station. For it is not as if a person’s knowledge how to act is to be conceived as a stock of knowledge separable from her knowledge of facts.

Ryle’s own examples of acts someone might know how to do included play a musical instrument, cook a meal, swim, play chess, speak grammatically, write, sew. Such examples are relatively specialist and relatively generic. They are specialist, inasmuch as some people might know how to do them, others not, so that it can be a real question whether someone can do the thing herself. But they are generic inasmuch as one’s knowing how to do any of them ensures one knows how to do plenty else, given only that one knows a multitude of facts. Assume that we all know masses. Then a compendious potential for action will be attributed to someone if she is said to know how to do just about any generic thing, even if it is somewhat specialist. Even where an instance of Φ is just some single verb phrase, someone’s knowing how to Φ, given how much else they know, ensures that they will know how to do a very great deal themselves. If you know how to cook a meal, for instance, it is likely that you know and can use some of the methods of separating egg yolks from whites; and someone who didn’t know how to cook a meal, would be less likely to know so much. And knowing, as you do, how to cook a meal, there will be plenty of other things you know how to do on one or another occasion. This is because you both know plenty of general facts, and you readily learn facts about situations in which you act. You learn that the local delicatessen sells sumac; and, knowing already how to get to the delicatessen, you now know how to assemble all the ingredients for raqaq u addas. Here factual knowledge impinges on, and expands, your knowledge how. At any time, some of that which a person knows how to do at any time will be dependent upon what she knows about the situation in which she may then act. As she moves about, her knowledge how to do this or that particular act will be entangled with her shifting stock of knowledge perceptually acquired.

6. Knowing how and intending

I have suggested that one may see a person’s having masses of factual knowledge as entangled with her having capacities to get plenty of stuff done. No-one will deny that there is some connection between knowledge how and action. Indeed Stanley and Williamson, the authors who together started the recent intellectualist campaign against Ryle, said ‘We do find it very plausible that intentional actions are employments of knowledge-how’ (2001, 442). What they find very plausible is formulated with a schema.Footnote ⁹

(‖) If A intentionally Φ-s, then A knows how to Φ.

I’m going to suggest that the very plausible (‖) actually gives Ryle, the anti-intellectualist, the advantage. I start with two points about it.

The first concerns the ‘s’ which marks an inflection on the (schematic) verb ‘Φ’. The ‘s’ of ‘Φ-s’ in the antecedent introduces the simple present tense for the instance of ‘Φ’: ‘She intentionally sweeps the floor’, for example. But the most likely reading of such a sentence is habitual (perhaps she intentionally sweeps the floor every Tuesday). A habitual reading of ‘Φ-s’ is presumably not what is meant. Presumably the ‘s’ is to be understood generically somehow – as a sort of dummy inflection. So someone who is Φ-ing knows how to Φ, and someone who Φ-d or was Φ-ing knew how to Φ, and someone who will Φ will (if she doesn’t already) know how to Φ.

The second point relates to what makes (‖) compelling. If A could not know how to Φ, then she could never intentionally Φ. We cannot digest our food intentionally, for instance; and the reason would seem to be that digesting one’s food is something that one doesn’t know how to do.

What these two points bring home is that (‖) registers something universal about the domain of intentional agency, and about precisely that domain. Save for examples of things not done intentionally, ‘Φ’, suitably tensed, can be replaced with verb phrases of lesser or greater complexity and of greater or lesser contextual dependence. Instances of ‘Φ’ might be ‘swam’, or ‘is travelling from an address on one side of the world to an address on the other using seven modes of transport’, or ‘will place your cup down on that table’. (‖), then, captures the thought that the acts any person ever is or was or will be doing are properly included among those she knows how to do herself. What (‖) makes vivid is that in doing something oneself, one counts on knowing how to do it. Notice that one doesn’t count on knowing how to produce the unintended side effects of things one does – not even if de facto one does know how to produce them.

Of course no-one could count on knowing how to do anything unless she knew plenty else. But if one endorses (‖), that would seem to be because one allows that the knowledge attributed to A in its consequent is caught up with A’s acting in a way that it could not be if it were knowledge just of how Φ-ing can be done by someone or other. This is why (‖) plays into Ryle’s hands: it is small wonder that someone who is doing something intentionally should know how to do it if she is then exercising her knowledge of how to do it.

My question was about intending, rather than about intentionally doing. But the connection with knowing how remains. One immediate connection shows up if it is allowed that one who is Φ-ing intentionally intends to be Φ-ing. But something more general can be said. It is true that someone may intend to do something in consequence of some delusion; and true that the undeluded may never do something they intend to – whether through forgetfulness, or weakness, or a change of mind about its desirability. It is true also that someone may intend to do something but not know yet exactly how to. But someone who intends but does not know how to do something can be expected to come to know. One does not intend to do something if one thinks that one could never know how to. And it is no sort of accident that the acts it is possible to intend are those that one can know how to do oneself, given that they are those acts which knowledge how to do oneself is necessary for doing (intentionallyFootnote ¹⁰) – acts that may figure in actual instances of (‖).

I think that in seeing how ‘intend’ and ‘know how’ relate one to another, one starts to expose the errors both of the propositionalists about ‘intend’ and of the intellectualists about ‘know how’.

7. Infinitives and ‘first personal’ cognition

My claim was that if ‘intend’ always related a person to a proposition, then no-one would do things they intend to. I took this to suggest that it is not a mere appearance that verbs following ‘intend’ are infinitives, and thus subjectless. Here I take issue with Stanley Reference Stanley2011, in which not only anti-Rylean intellectualism but also a much more general propositionalism is defended on linguistic, theoretical grounds. Stanley thinks that verbs which occur in the infinitive when spoken or written must be treated as having subjects for the purposes of syntactic and semantic theory. According to him and like-minded semanticists, there is an unpronounced pronoun in infinitival clauses – unpronounced in speech, but to be written in the theory using the word ‘PRO’. Someone who sides with Stanley about the syntax and semantics will say that, in order to reveal its underlying character, a sentence such as (E) (chosen for simplicity’s sake) should be re-written as (S).Footnote ¹¹

An infinitive remains in (S). And ‘PRO’ cannot be a genuine subject expression unless and until the infinitive ‘to go home’ is somehow transmuted into something predicable of the reference of ‘PRO’. If ‘PRO’ is rendered with ‘herself’, and the infinitive is left to stand, then one reaches something intelligible: ‘Hannah intends herself to go home’. But a symptom of the fact that even here an infinitive still follows ‘intends’ is that ‘herself’ is now most naturally understood as introducing a sort of emphasis. (‘Hannah intends herself to go home. Don’t think that there’s someone other than Hannah whom Hannah intends to go home.’) Of course the pronoun ‘PRO’ is not supposed to add emphasis: it is meant to introduce a subject for a proposition constructed somehow out of ‘PRO’ and ‘to go home’. So perhaps (S) should give way to ‘Hannah intends PRO go home’, which rendered in English is presumably ‘Hannah intends [that] she herself go home’. Well, as we saw, if this were actually said, it would carry a suggestion that something might intervene between Hannah’s intention and Hannah’s going home. But in any case, this won’t help with propositionalism: there is a subjunctive here: ‘PRO go home’ is not truth-evaluable. (It is easy to be confused here: the ‘go home’ of ‘PRO go home’ could be a zero infinitive [‘to go home’ without the ‘to’]. But if ‘go home’ is in construction with ‘PRO’, then it has to be the [equiform] subjunctive.Footnote ¹²)

Stanley says that when ‘PRO’ occurs in a sentence such as (S), it ‘has an obligatory de se reading – a reading that involves a first person way of thinking’ (2011, 77). At first blush, this might seem to help with the problem for Broome which showed up in the example of Aelfric who lacked an intention to speak to the Regent. Broome based his claim of an equivalence between ‘A intends to Φ’ and ‘A intends that she Φ’ using an argument which assumed that the ‘she’ of ‘A intends that she ….’ is co-referential with ‘A’; and Aelfric exposed the error of Broome’s assumption. So it might now be thought that the problem for Broome can be solved by making him a gift of ‘PRO’. Broome could then say that if A intends to Φ, the proposition intended by A is not a regular singular proposition concerning A, but one in which reference is made with ‘PRO’ read de se. As the example was presented Aelfric did not intend to speak to the Regent. But it is possible at least that he would come so to intend if it should dawn on him that he is himself the Chair of the Board. Broome and any other proponent of propositionalism could say, then, that what is intended by one who intends to Φ is to be given using a sentence whose subject is a pronoun understood de se.

Too much has been written about ‘the first-person way of thinking’ for it to be possible rapidly to dismiss the idea that the needed treatment of ‘PRO’ can be given. But one can start to appreciate the difficulty about treating it by thinking about poor Hannah who still intends to go home. If Hannah’s intending to go home were a matter of Hannah’s standing in relation to a proposition, then it would be a proposition Hannah might express in saying ‘I intend to go home’. But if the content of Hannah’s intention is given with a sentence containing a term whose sense Hannah employs when she uses the word ‘I’, it will apparently contain a term which no-one except Hannah herself can use in making a reference to Hannah. Yet the propositionalists must provide a proposition accessible to anyone who tells or is told what Hannah intends. It can be difficult to see how to provide one. And of course it is only when ‘PRO to go home’ is supposed to be forced into the shape of a proposition that any difficulty about accounting for ‘PRO’ presents itself.

Someone on the side of the propositionalist will probably say that whatever difficulty might be avoided by allowing intentions to be intentions to act, it has to be dealt with because it is a difficulty encountered elsewhere. They will be quick to point out that if it dawns on Aelfric that he himself is Chairman of the Board of Management, this can be comparable to its dawning on John Perry, in his famous supermarket example, that he himself is the man making a mess (Perry Reference Perry1979). In that example, Perry followed a trail of sugar on a supermarket floor, seeking the shopper with the torn sack in order to tell him that he was making a mess. Perry came to believe, of a certain man, that he was the messy shopper. The man of whom Perry believed this actually was Perry. Only later did it dawn on him that he himself was the shopper he sought. Perry’s case apparently concerns a belief he had when it dawned on him that he was making the mess, and belief is a propositional attitude. So it may be said that some account of de se propositions is needed to deal with such cases, and that whatever account is given, it can be pressed into service to treat such propositions as (it is alleged) anyone who intends to do something, intends.

But here must one remember that there was not only a difficulty about finding a subject for a proposition intended by someone who intended to do something (to go home, as it was for Hannah), but a difficulty also about finding anything predicable of such a putative subject. One might want to turn the tables now. Rather than use a claim about how Perry must be dealt with in order to prove the need for de se pronouns, one might find something different to say about someone who intends to do something. We know very well what it is that Perry takes to be predicable of himself when he comes to believe that he himself is the messy shopper – being the messy shopper. So instead of saying that Perry believes a new proposition when it dawns on him that he is making a mess, we could say that Perry comes to stand in a certain relation to this property. Perry then will be in one way comparable to someone who intends to do something and stands to an act in a certain relation. Hannah’s intention (for instance) is shared with anyone who intends to go home. Perry’s state of mind is shared with anyone who thinks that he himself is making the mess. (There is nothing peculiar to Perry’s case that leads to this. If there is a number of people each of whom believes that she herself is F, there is not any proposition [not as ‘proposition’ is ordinarily understood] belief in which is attributed to them all. What they have in common is that each ascribes F-ness to herself. Perry for his part came to ascribe to himself the property of being the mess maker.Footnote ¹³)

Aelfric’s case must not be assimilated to Perry’s, however. Even when it dawns on Aelfric that he himself is the F, and has intended that the F speak to the Regent, Aelfric may very well not intend to speak to the Regent. (It might be a question for Aelfric whether he has any reason not to speak to the Regent; or it might be that Aelfric takes it to be out of the question that he should speak to the Regent.) The manner in which someone relates to an act they intend is markedly different from the manner in which someone relates to a property they ascribe to themselves as a result of a discovery. An intention to do something is not reached by finding out that one is some way, but by making up one’s mind what to do. This is one sign of the fact that intending-to is more primitive than intending-that. And it draws attention to the fact that one who reasons practically is conscious of herself. Evidently there are questions here, about self-knowledge, which are questions for philosophy and not to be answered by studies in a branch of linguistics.

It may be that phenomena that ‘involve first-person thinking’ are not best illuminated by way of introducing a de se ‘PRO’ into language. Not that I have illuminated these phenomena here. Perry has made an appearance only because the propositionalists might have wanted to adduce examples like his in their support. I’ve argued, however, that there is something to say about such examples which makes no use of ‘PRO’. I’ve also argued that even if a semantic treatment of ‘PRO’ were vindicated, we should need to be told what predicates they are which follow ‘PRO’ when ‘PRO’ comes into sentences attributing intentions. And even if we were somehow persuaded that subject-predicate sentences always follow ‘intend’, still we could not understand those sentences as expressing propositions. I’ve wanted to show that standard syntactic theory (or at least theory which Stanley takes to be standard) can lend no support to propositionalism.Footnote ¹⁴

8. Philosophy of action

But we saw that there are strains in philosophy proper which are supposed to lend support to propositionalism (Section 3 above). I shall finish by responding to those philosophers of mind who seek a systematic account of the cause/effect relations which they think the attitudes enter into, and who advocate the standard story of action.

The standard story trades in desires and means-end beliefs. An action, it is said, is a bodily movement caused by a desire (‘that the world be a certain way’), along with a means-end belief (‘that the thing done is a way of making the world that way’). In some tellings of the story, intention is set aside. But in others, the desire and belief are said to cause the action by way of causing an intention that causes it: an intention is supposed to cause a movement of whoever’s body it is whose suitably related desire and belief combine to cause it (the intention). Should we not balk at the idea that the role of intentions, whatever they might be intentions to do, is to cause just the movements of the body of the person whose intentions they are? Are we unable to do such things as we may intend, and know how, to do? (See end of Section 4.) An agent acts – does things that she intends, knowing what she intends, and knowing how to do what she intends to.

Michael Smith’s version of the standard story sticks closely to the belief-desire theory, so that intention has no place in the story as he tells it.Footnote ¹⁵ Still Smith acknowledges that the causal roles filled by desires and means-end beliefs are not adequate to account for the causation of an action. Smith makes use of the idea that agents have the ‘capacity to put their desires and beliefs together so as to produce a bodily movement’ (Reference Smith2012, 399). One wonders whether we are supposed to know that we have such a capacity before we construct theories. It isn’t said. But if this capacity needs to be attributed to us in a theory about ourselves, then presumably it is proscribed that the theory should mention such capacities as we know ourselves to possess – capacities, for instance, to do such things as we intend to. I suppose that the proscription might be defended by appeal to a requirement that mental states to which causal roles are assigned should be propositional. But such a defence wears thin as soon as it is noticed that ‘desire’, like ‘intend’, may be followed by an infinitive, and that it too may take the subjunctive when followed by a ‘that’-clause (‘that the world be a certain way’).

It must be a good question why an agent in the hands of a philosopher should be allowed to possess only certain sorts of mental state and only strictly bodily capacities. Perhaps the standard story’s widespread appeal is founded in a misconceived and ill-considered propositionalism. Or perhaps propositionalism as such has little to do with it. It might be that the story derives from a view about the limits of the contents of the mind of a subject who makes use of reasoning. Or perhaps the standard story and its understanding of human capacities rely upon a view about what causal truths there can be. Certainly, a conception of causality that some will consider ‘foreign to science’ will be needed if an agent’s intentions, including those that she enacts, are not all of them intentions to move parts of her body. But unless it is to be denied that we can often enough do what we intend to, a relation cause which obtains between two items at some time can hardly be the only causal notion that has a place in an account of agency.

Once it is allowed that we can stand in cognitive relations not only to propositions but also to acts, I think that we are bound to acknowledge that an account of ourselves as the thinking and moving beings that we are is not to be given by the semanticist or the scientist.