Anscombe vs. J. L. Austin on Pretending:
Was There a Victor?

by Steve Bayne
Nov. 31. 2003

What I intend to do in the following pages is; first, state Austin's position as carefully as is required for an understanding of Anscombe's reply; and second, state Anscombe's position, but as a running criticism of some of her central ideas. I will not take a side as to which had the greater insights, although my effort with respect to Anscombe rightly suggests that, while I disagree with a number of her various opinions on the subject of pretending, I believe it was she who demonstrated the greater understanding of the complexity of what it is to pretend. One can't help but be left with the impression of a light hearted Austin and a tough minded Anscombe; and this is the way I believe both would have preferred it. They shared the belief that philosophy is not logic but has a great deal to do with words. The years in which they wrote were near the high water mark of this conception of philosophy, and it is instructive to read both with care if for no other reason than to see how things have changed, perhaps not necessarily for the better.

In 1958, when my world was young, Elizabeth Anscombe and J. L. Austin published their remarks on pretending, surely this was one of the greatest exchanges of ideas ever to appear in The Proceedings of the Aristotelian Society (referred to here as 'PAS': The Proceedings of the Aristotelian Society Supplementary Volume xxxii. 1958. pp. 261-294). Both philosophers were at the height of their powers and the ensuing discussion exemplified best the styles each had perfected. Austin writes with a sense of humor, but one that often conceals the details of what he is seriously setting forth. Anscombe style is copious, methodical and deliberate; a style that would remain characteristic throughout her productive philosophical career. If there is a debate here, it is one-way, as Anscombe is replying to remarks by Austin - who in turn is reacting in large measure to some ideas Errol Bedford had presented on another occasion ("Emotions." PAS. Volume lvii. 1956-57; reprinted in Essays in Philosophical Psychology edited by Donald Gustafson. Anchor Books. Doubleday. 1964. pp. 30-41). Bedford's puzzle all too frequently gets lost in the exchange, and neither Austin nor Anscombe to this writer's satisfaction address some of the central points he raises. In part, Bedford's interest, like Anscombe's, was in arguing against an appeal to introspection in the identification of emotions, leading him to raise the central question of how to draw the line between real and pretend anger. Let us examine very briefly how this comes about.

If I exhibit pain behavior in the absence of a "raw feeling" of pain, then there is the likelihood that I am only pretending to be in pain; but if I exhibit behavior typical of anger, the absence of such a feeling does not with equal assurance, at the very least, imply the absence of anger. At least part of the reason for this difference is that, unlike pain, anger is not something we associate with a "raw feeling" or sensation. But this raises a problem for the philosopher: if anger is not a sensation or feeling then we are left, so it seems, with its being a behavioral phenomenon; and, if this is the case, then it becomes difficult or impossible to distinguish pretending to be angry and actually being angry - should we behave angrily in a compelling way. It is difficult, however, to imagine that there is nothing non-behavioral separating the pretenders and the truly angry.

This is the central problem raised by Errol Bedford during the halcyon years of philosophical behaviorism, and the problem which J. L. Austin takes as the basis of his remarks on pretending. Let us begin by discussing some of what J. L. Austin has to say on the matter, concentrating in particular on those aspects of his remarks which are taken up by Anscombe in her carefully composed reply.

Pain is a sensation; anger an emotion. One can easily come to feel certain that there are profound differences, if for no other reason than that while there are reliable pain producing mechanisms, there are no similarly reliable anger producing mechanisms. If we wish to distinguish the angry man from the man who merely pretends to be angry in those cases where it is difficult to tell the difference on the basis of "more evidence of the same kind," that is, more behavioral evidence. Pretending to be angry as Austin observes  (PAS. p. 266) differs from my pretending to bite your calve, say, inasmuch as more evidence of the "same kind" - although the evidence here may be more than behavioral - appears sufficient to establish reality over pretense. Nevertheless, in most cases we are not seeking knowledge of an internal sensation that will serve as a sufficient basis for making the distinction between pretender and non-pretender, or perhaps more interestingly between pretending and pretending to pretend. Even if pretending is not acting, there may, indeed, be some difference between asking an actor to act the role of actor and merely asking him to act, one which has something corresponding to it in the case of pretending. In acting we are not trying to get people to believe but to get them to believe we are believable. Austin's remarkable linguistic sensitivity and grammatical judgments afford us a number of distinctions that will serve us well in discussing Anscombe's own views on the role of sensation (or lack of it) in anger, as well as introduce greater precision into the discussion of how we are to conceptualize such subtle differences. One such distinction is that between pretending 'to do' and pretending 'to be' (PAS. 265):

However, we have to consider also the construction in which "pretend" is followed by "to A" or "to be A-ing", especially in cases where the verb "A" is one which describes the doing of some *deed* (e.g., "bite" as opposed to, e.g., "believe") and more particularly when that deed is of a pretty "physical" kind (e.g., biting as opposed to e.g., giving).
While Austin is correct that it is impossible to pretend to do what it is impossible for the body to do, e.g., to "curl your trunk" (since you don't have one), he neglects to mention cases which are least impossible: e.g., I can pretend to be trying to make a fist, but I can't in any non-controversial sense "pretend" to make a fist. Austin having first mentioned cases of deeds in connection with pretending, we might raise another point, one inspired by Anscombe's notion of acting 'under a description' ("Under a Description." Collected Philosophical Papers of G. E. M. Anscombe. Vol. 2. Blackwell. 1981. 208-219, and Intention. Oxford. 1957. p. 11). As she points out, I may saw a piece of wood under the description "plank" but not under "oak," when my intention is to saw wood but not oak wood. In this connection, one might ask if, since pretending is done intentionally, I may pretend to act under one description but not another. Can I pretend to murder John under the description 'my broker' without pretending to murder John under the description 'my long lost uncle who bequeathed me a million dollars'? Suppose I can. But now by what behavioral criterion can this behavior be identified as such? Future behavior may apply towards arriving at some decision in the sawing case, but matters are not so clear in the pretend case. Perhaps a more telling difference worth examining is that between pretending to try to move my finger and pretending to try to move my hand. Here there is a contrast, as 'under a description' does not apply but there is behaviorally no criterial difference to speak of in terms of what is observable. Besides the difficulty presented by ambiguity in behavioral criteria for discerning pretending to one thing from doing another. There is the further problem of distinguishing pretending to act from acting itself. Austin cites (PAS 268) three instances where it would seem that we are in fact doing what we only pretend to be doing.

1. Where two men pretend to be sawing a tree down by allowing the teeth of the saw to bite into the bark.

2. Where a magician pretends to be sawing a girl in half by allowing the teeth to bite in, but only briefly.

3. Where a man pretends to be cleaning windows, and is doing so, but is actually spying on people in the room.

While these cases are special there are three things one can say of the typical act of pretending to perform some deed. First, there is the actual pretending behavior, such as the back and forth motion associated with sawing. Second, the "reality-dissembled," such as the reality of my true feelings in such cases as where I'm pretending to be angry; and third, "real behavior-dissembled," such as the real behavior of biting when I pretend to kiss a piece of carpet. These features he contrasts with the usual ones invoked in philosophical discussion, those of the actual public behavior, the behavior which is being "simulated," and the "genuinity" of the simulating behavior. There are some special considerations attaching to the case of the man pretending to clean windows. In the case of the tree sawing, as well as that of the magician, while the actual behavior is the very doing of what it is said the agent is only pretending to be doing the action is never completed. The situation here is in its semantic detail related to what David Dowty called the "imperfective paradox" (Word Meaning and Montague Grammar. Klewer 1991 [1979]. p. 133). The problem is that in such cases as sawing the tree it is difficult to give an account of how sawing the tree down does not entail bringing it about that the tree is sawed down. What I want to suggest is that the "imperfectivity paradox" does not apply to verbs when the action described is pretend behavior of this sort, rather than the window washing sort. That is, an account of pretending to V entails not perfecting the action, suggesting that the semantics of progressives is not context free, and that giving it an account must not involve reliance on the semantics of the perfect form of the verb, at least in any obvious way. This is not a central concern, but it does serve, I believe, to show that commonplace linguistic behavior often rides the surface of very deep grammatical waters. We shall have occasion to address these puzzling cases shortly when we arrive at Anscombe's response to Austin. An important conclusion Austin draws from his examination of the general features of pretending in cases such as the window washing case is that there is a contrast between the "real behavior dissembled" (the spying) and the "mere pretense behavior" (the window washing itself). Yet, this is not entirely convincing, since there seems to be a similar contrast between the "real behavior dissembled" and any of a number of activities the spy is engaged in while washing the windows. Consider the situation where the spy is not really washing the window at all, but merely going through the motions - in mime-like fashion. Do we have here two cases of pretending? Or pretending with the intention of pretending not to spy? The agent might want to convey the following: "Don't suspect me of spying; I am just a mime pretending to wash windows." Perhaps the best we can do is to say that washing the windows in the original case is done with an intention of dissembling the "real" behavior, but that more than real behavior can be dissembled. It is to be noted in passing that an equally defensible way of describing the situation is to say that in neither case is the performance the actual spying; it merely conceals it, whereas sawing the tree does not merely conceal cutting the tree down it is in fact part of what cutting the tree down consists in doing. According to Austin's account of pretending what is essential is merely that the window washing conceal something else.

The essence of the situation in pretending is... that my public behaviour is meant to disguise some reality, often some real behavior. (ibid.)
Suppose I pretend to be trying very hard to move my hand. I groan and stare at my hand with mouth twisted in feigned agony. What am I trying to disguise? That I can in fact move my hand? But how do I know I can, and besides, perhaps I can't and know it. What all may I be trying to disguise? Here there seems to be no contrast such as Austin requires between "real behaviour dissembled" and "mere pretense behaviour." I am not pretending not to be able to move my arm; what is being concealed is one's intention. One can pretend, therefore, to have a certain intention and in such a case contra Austin it is "real intention dissembled" and not behavior. Pretending to try is not like pretending to be in pain. Even more certainly than in the cases of fear or anger, there is no special sensation one readily associates with trying - nor do I cease, for example, to have a sensation (such as dizziness) by trying (e.g. to regain my balance). There is behavior closely tethered to pain; there is no such behavior so closely related to trying. As we shall soon have reason to believe, pretending to try to move one's hand is more like acting than pretending, as acting and pretending need to be distinguished. A final word on 'pretending to try' will serve to illuminate Austin's distinction between 'pretending to A' and 'pretending to be A-ing'.

Austin gives two further examples of pretending; one where someone in the next room is trying to get us to believe he is playing chess by loudly uttering chess like remarks and clacking pieces of wood together, and another where a child is making movements of his arms, pretending to be driving a race car. Austin raises the question in regards to the second as to why we would not say that the child is 'pretending to drive' rather than "pretending to be driving." He maintains that to do the latter would require a real car (PAS 273). He might have said that 'pretending to A' is in a sense de re and pretending to be A-ing' de dicto, but more important is that there is no corresponding distinction to be made between pretending to try to move one's hand and pretending to be trying to move one's hand. Although Austin has critiqued Bedford for not seeing that there is more to pretending to be angry than acting angrily, no matter to what extreme, he is insistent that there is more to pretending than acting a certain way while leaving something out, or where something is left out. (PAS. 274)

I may, says Austin, act angrily and not be angry and still not be pretending. Here he cites, among other things, the case of acting. He doesn't make much of this, but a brief statement of the difference is worth making. The pretender disguises in some sense the reality of what he is up to. At least this much is true of the typical cases in ordinary experience. To make use of one of Austin's examples, in an attempt at deceiving someone in the next room I may pretend to be playing chess - making the noises of a chess player. Suppose I do this to conceal the fact that I am keeping an eye on his car outside my window. Here one may speak of disguising a certain reality. But acting is something somewhat different. Instead of disguising reality I may be said in a manner of speaking to be "guising unreality." I am taking something unreal and trying to make its reality believable; I am not taking some real behavior and dissembling it. I am not taking the reality of being a man and making that reality over so that it is believable that I am, say, an owl. We may judge a pretense as good or bad, convincing or unconvincing; we judge acting. In the latter case the script may not even be something with which we are familiar, but to judge a pretence I must know or have some idea of what sort of thing the pretender is pretending to be. How might I make the attempt to pretend to be someone who does not exist? Is it possible to pretend to be a character from Dostoevsky without acting the part? Do we go about our daily business in such a case in a Karamazovian way? Is it that in pretending and not acting we can add to the actions of the subject of our pretence? Answers here do not come easily, nor shall I attempt them here. We must consider that such questions call into question not only the border separating pretending from other forms of behavior but create, as well, a lingering doubt whether even within the category of pretendings there is a common defining aspect - a point Anscombe will make in a strong way.

While Austin does not identify a defining element of pretending, he does consider what he calls the "basic case."

To be pretending, in the basic case, I must be trying to make others believe, or to give them the impression, by means of a current personal performance in their presence, that I am (really, only, etc.) abc, in order to disguise the fact that I am really xyz. To neglect to notice this is to put in the bathwater without the baby. (PAS. 275)
Anscombe will argue, as I've said, against this. What she will not make entirely clear is that it is by no means obvious that Austin's comment was intended to capture in one fell swoop the meaning of 'pretend', indeed as Austin notes, even at this stage, "we are far from having a full account." While lamenting the nuances of 'pretending' he has failed to capture, Austin alludes to what is, perhaps, a category of its own. Austin speaks of pretending which "masks our intention," (PAS. 276) and thereby introduces a theme of  Anscombe will pursue. Unexpectedly, perhaps, Austin remarks, merely, that; "we do not call for precision" (ibid) and skirts the matter. Those familiar with Anscombe's work know that she rarely skirts such matters.

Austin having already made careful use of the distinction between 'pretending to A' and 'pretending to be A-ing' proceeds to inquire into the character of 'pretend that', noting that it provides some "emancipation" from 'pretending to'.

...when pretending-to I can deceive only as to my states or activities, and contemporary ones at that, but surely when I "pretend that it was in the garage yesterday" I deceive as to something other than my own states or activities, and something non-contemporary at that. (PAS. 276)
No sooner than he introduces the notion he begins to express doubts as to the viability of 'pretending that' as possessing a meaning independent of 'pretending to be': how am I, for example, to distinguish pretending to be on my way to Antarctica from pretending that I am on my way to Antarctica? Still, there does appear to be - at least within his surmise - a use for 'pretending that' in "creating impressions about our cognitive states," as when by pretending that the car was in the garage yesterday I dissemble my present awareness that it was not. (PAS. 277) He adds, what is arguably, his most forceful reason for distinguishing an independent meaning for the expression, noting that 'pretend that' is especially useful in describing where we 'pretend to oneself'. To the very end of his discussion, Austin is unsure of the difference between 'pretending that' and 'pretending to'.

He tries opposing pretending that he is playing chess with pretending to play chess. What he comes up with is that in pretending to play chess I am "more...going through the motions," but the uncertainty remains. Let's carry this in a slightly different direction. Suppose you and I play chess without a board, merely pretending that there is a board. Is this different from imagining a board or using an imaginary board in order to play a game of chess, or are we pretending that something is the case? Keep in mind, also, that were we to use something besides a board we might have what children call a 'pretend' board. Supposing that I actually move my king to a place where if there were a board the appropriate place would there awaiting the move. Here I should rather say that I am pretending that there is a board, whereas if all my moves are imaginary, then I am inclined to say that I am merely imagining a board, or that the board is imaginary. If pretending implies deception this cannot be, for as anyone can see there is no board before us. Notice that in order for us to play such a game in any case, we imagine we are playing with the same board. We could imagine different boards but then we would have difficulty imagining that we are playing with the same pieces. Nor can we merely pretend to play chess on an imaginary board. If 'pretend that' pertains to cognitive states, then properly understood as such, we cannot pretend that we are playing chess on such a board but we might imagine doing so, but we cannot even pretend to be doing so.

Anscombe (PAS Vol. xxxii. 1958 279-294) begins her discussion with a distinction that serves to illustrate the inadequacy of Austin's idea that there is a "basic case" description of the meaning of 'pretend'. Suppose I am behaving as I would were I playing tennis, only I don't have a ball. I'm not trying to fool anyone. I am simply acting as though I'm playing tennis, only without a ball. This sort of behavior is a "mock performance." Anscombe may be overstating the degree of reality there is to a mock game of tennis, however. When after remarking that more than being a mock game of tennis the point of the example (and one might justifiably ask, "Example of what?") is not to illustrate mock tennis playing only "just that it is a tennis-game without a ball." (281). The problem is that she has elsewhere expressed considerable agreement with Davidson that we perform certain actions which are in a sense basic, such as arm raisings etc., while "nature does the rest"; but in this instance of a pretend game of tennis there are only "basic actions" and nature has been deprived of any effect on either the outcome of such actions within the context of the "game" or in the determination of its outcome.

Real pretending may involve at attempt to deceive, such as in Austin's case of the window washer who is actually a spy. But even though the intention to disguise one's motives is present this does not (according to Anscombe) mean, contrary to Austin, that there need be a "current personal performance.". So, to use Anscombe's examples, I may pretend to be a meat eater by having large deliveries of meat to my door, or I may pretend to be angry in a letter. In neither case is there a "current personal performance." I think what Austin might say is this: in none of the cases produced is the person doing the pretending actually doing anything that is easily identified as pretending. In the one case where this may be argued - where x is pretending to marry y but it is actually z marrying y using x as proxy - the intent to deceive is absent. Further, marriage by proxy would appear to be neither mock behavior, since a marriage is taking place, nor "real" pretending. What concerns Anscombe most, and for reasons much like those Wittgenstein produced in his remarks on 'game', is that Austin speaks of 'essential features' of pretending.

Supposing there are, as Austin suggests, "essential features" of pretending, one might suppose that some sort of definition of 'pretend' is in principle producible. The difficulty is that if we affirm that real pretending involves some mental occurrence removed from behavior, then we are back to the idea of a "private ostensive definition" (282).

 The problem is that we can only assume that "we have correctly identified something within ourselves." This is a particularly strong argument, especially in light of Wittgenstein's arguments. What might Austin say? Austin might say that the central question is whether if we reject feelings as a necessary condition for being angry can there be said to be a difference between real anger and pretend anger? There may be certain problems of knowing on what basis a correct identification of a feeling related to anger can be made, but, then again, we have yet to come up with a suitable criterion even if we resist construing anger as something essentially "inner." Moreover, while Anscombe is critical of Austin for invoking essences and "basic cases" she does something similar when, for example, she speaks of the "best general account" of pretending, which comes down to 'trying to appear what you are not'. Austin, playing Wittgenstein's "game," might introduce for consideration cases where we are trying to create a mere impression of an appearance that does not exist in order to disguise the real appearances such as they may be, as for example when a judge tries to appear a judge, knowing he does not appear to be a judge. But there is something else Austin might cite as a problem.

We require a certain amount of angry behavior in order for a performance to qualify as pretence of anger. Anscombe in commenting on the view that there is something that distinguishes real anger from angry behavior remarks that on this view of the relevant behavior: 'if there is enough of it' then we might have a bona fide case of pretending to be angry. (282) The basic problem is this: in the case of pain there seems to be something that determines the fact of the matter (whether there really is pain). But what about anger? If there is no sensation or inner "fact," then what determines the fact of the matter of whether someone is angry and not just pretending? Perhaps more importantly we might ask whether there is a fact of the matter. If there is no fact of the matter- and here I do not mean something criterial for warranting the belief in such a fact of the matter - then Anscombe's "general account" is useless for purposes of determining whether there is "really" pretending going on. Framing the dispute in these terms places Bedford and Austin observations in a very different light, but there are reasons for believing, which I shall not elaborate, that Anscombe would challenge my use of 'fact of the matter'.

Austin never comes out and says that there is an inner sensation that is essential in order to account for the difference between anger and pretend anger; nor does Anscombe make this attribution. If we ask, "How exactly does Anscombe come down on the matter of the dispute between Austin and Bedford?" no answer is clearly given. One thing can be said, however, with certainty, viz. that Anscombe is interested in applying Wittgenstein's critique of inner "criteria" as the basis of the application of public language expressions, such as 'pain'. This she takes in concert with his critique of essences in, for example, his discussion of 'game' (PI 71). In shifting our attention to 'facts of the matter' it must be admitted that we may have shifted the nature of the issue, but in what I would argue is the right direction.

Anscombe resists the notion that determining what anger, not pretend anger, is relies essentially on inner mental episodes. 

The words, or thoughts, are themselves an angry reaction, and there is no need to postulate, indeed no sense in postulating, another reaction, not the words or the thoughts which is the ground or the thought which is the felt anger itself. The fact that the verbal reaction may be a sham does not prove such a need (287).

The first thing she says is reminiscent of Wittgenstein, on that occasion where he says he says that when we teach a child language which one is inclined, at first, to characterize as describing pain what in fact occurs is that we "teach the child new pain-behavior" (Philosophical Investigations 244). But in the case of anger what Anscombe appears to be suggesting is that there are only angry reactions. Whether this is in fact the case is difficult to decide, but it is worth considering that in this rather "Wittgensteinian moment" she is saying something about 'anger' highly suggestive of what Wittgenstein says about statements of fear: that we ought to question whether they are always "a description of a state of mind" (PI. p. 189e). Those who view Wittgenstein as some sort of behaviorist may believe he is hedging in the case of 'fear' - in that he does not repudiate states of mind - but it is quite possible Anscombe is not hedging in the case of 'anger', and that what she intends is that anger is never descriptive of a state of mind. This is at the heart of the position she takes with respect to the Austin/Bedford dispute per se. This much seems certain. Anscombe  is interested in what it is to feel angry. But explaining what anger is and giving an account of what determines the fact of whether one is angry or merely pretending is something different. Supposing that there is a feeling that determines the "fact of the matter" of whether someone is pretending or really angry is not to imply the need for some feeling to serve as the basis for the meaning of the term 'anger'. That a feeling should be essential to anger is one thing; that it's presence may be a requirement for the truth of 'I am angry' is another. To suppose otherwise is to confuse essences with meanings. One possibility I will not pursue is that there are private facts that have public explanations and that a variety of private facts may factor in the explanation of public behavior. It is not a necessary truth that private facts are states of mind, particularly if we seek out the most commonsensical notions of 'private'. These matters raise "metaphilosophical" issues we must resist, at this juncture.

 If we had an efficient anger simulating performance, and we wanted to know what makes it a simulation, that is, a pretence, it does little good to do what Anscombe does and merely state the conditions for the "diagnosis" of pretence based on facts such as knowing that the agent does not really mind what he seems to be expressing anger at. It might be argued that the "diagnosis" of behavior as pretending is not the important issue; the important issue is what makes it a fact that a person is pretending and not really angry.

Anscombe rejects the position that inner states figure in determining whether a person is angry or not; whether they are, therefore, inessential to anger is not - we have argued - a necessary consequence. It is only natural to ask how, then, her position differs from Bedford's - his view being that beyond a certain point behavior is no longer pretence, but the real thing. What, then, is the difference between pain and anger, if any, for Anscombe? Correspondingly we might ask: "What for Anscombe is the difference between sensations and emotions?" We have already said the answer to the former question is not entirely clear. In a sense, Anscombe's behaviorism, what there is of it, is more radical than Bedford's. Anscombe maintains that there is in fact a "great" (288) difference between anger and pain. At first, she appears to suggest that unlike anger in the case of pain there may be pretendings that are impossible to know as such; but, then, she says that 'Cases can be constructed for anger too...' (289) and the reader is most likely going to be confused, for now we are back to the question as to what this "great" difference is. The difference is shown in the fact that an actor who professes "I am really angry" in describing his state of mind while acting may be regarded as entirely sincere,  but even if we believe him we may question whether his anger is real or not. Is the actor really mad at the character portrayed as at the receiving end of his anger. At least a question exists. But in the case of the actor playing in King Lear who professes to feel pain when acting the part of one having his eyes plucked out no similarly warranted question arises. However, as long as this does not nullify what she says earlier about there being cases of simulated anger, just as for pain, the conclusions to be drawn from this example cannot be expressed with any degree of certitude. But this introduces larger issues, having to do with "verificationism" which cannot be explored at present. It involves, as we have said, other "metaphilosophical" issues.

When Austin attempts to explore "basic features" he is merely trying to probe cases where there is a fact of the matter whether there is merely pretend or real anger. Imagine the following game: I observe a behavior I've never observed before and must pretend to be doing it myself. In a manner of speaking I "translate" observed real behavior into pretend behavior. The game is somewhat "radical" (compare here Quine on "radical translation") as it involves a game entirely unfamiliar to me and familiarity only with how to go about pretending to do certain other things. Two points: if I succeed in disguising the fact I am not really performing the behavior is the behavior, alone, what qualifies my actions as pretending? Secondly, is there not a fact of the matter whether I am pretending to perform this behavior, but if so what is the fact? Like translation, pretending may be something that is done better or worse with no criteria of its having been performed. What Anscombe resists, as have so many others under the influence of the later Wittgenstein, is that there is something that determines the fact of the matter of whether behavior is indicative of real pain, a something we can know only by introspection. It is this which, she says, "is the one great locus of the discussion of pretending" (p. 282).

There is, however, an important distinction to be made between claiming that there is something which insures the "fact of the matter" that pretending to be angry is what is going on and the claim that what this something is is the "essence of the situation in pretending" (p. 283) to be in anger. In arguing against there being such an "essential feature" Anscombe appears to assume that such an "essential feature" is what Austin is arguing determines the fact of the matter, rather than being, say, a necessary condition of there being a fact of the matter in the first place. It is one thing to deny that there is any single feature that determines whether a set of behaviors is a game and another to deny that there is any necessary condition to be satisfied in determining whether a game is being played. Arguing that the rules do not constitute the essence of chess is not to argue that we cannot know whether we are playing chess or merely pretending. While Austin does speak of "essential features" he does not specifically introduce a required inner criteria.

Anscombe entertains what she calls "central features" of pretending (p. 284) while rejecting Austin's more or less untechnical reliance on "essential features." Indeed, it would not be cynical to assert the possibility that Anscombe's rejection of essences and her rejection of the need to introduce inner states in distinguishing real and pretend anger amount to being one and the same thing. For once having rejected this role for inner states, she not only introduces "central features" but in addition "four main features" of anger and doing this invites the same sorts of criticism frequently lodged against essences, in particular their proneness to disproof by counter example, when one imagines the need for speaking of "main" or "central features" rather than essences. The point is whether there is any more significance to be culled from her discussion of pretending than a rejection of the particular role of inner states other philosophers have reserved for them. The answer is clearly yes, which can be seen from her discussion of the four features she associates with anger.

The four features Anscombe identifies are: 1. its object, 2. its expression, 3. feelings, and 4. its aims (285). Notice that nowhere does Anscombe reject feelings but only the role suggested for them as deciding the fact of the matter as to whether what we have is pretend or real anger. Consider the object of anger. She echoes Aristotle's claim that anger is more rational than lust. At least anger "often" has grounds, and here the use of "often" suggests the possibility that none of these main features are required of anger. At first she speaks of someone as if they could be angry at a chair. (286) But how could a man's anger at a chair be more rational than lust? She appears to equivocate, saying that "we do not yet know what he was angry at." But why not the chair? But if so, he is no more irrational than the man of lust. Notice this much: one can pretend to be angry at a chair for good reason, and one can recognize what we are doing; but one cannot be angry at a chair, or so it would seem, with good reason. Furthermore, suppose someone says, "I am in an angry mood." Where is the object? "Ah, but this is not anger," one might reply. But consider this: What does it mean to be in an angry mood? Is it to be disposed to anger or being made angry? No answer here can be certain. But an angry mood needs no object, nor can the mood behaviorally be distinguished from anger itself, or is it possible that anger is no feeling at all and that I may be merely in an angry mood and never show it. But if an angry mood is neither then how are we to distinguish between pretending to be angry and pretending to be in angry mood? Perhaps, contra Anscombe, it is that I am pretending to have a feeling, a feeling that I am prone to anger without reason. One is inclined here to look for more than two alternatives, that is, more than either making a feeling a necessary condition for anger or making it a matter of crossing some imaginary threshold of behavior. Having briefly discussed "main features" let's take a look at those "central features" in more detail, all the while avoiding essence.

The first such central feature is that "the pretender should figure as a principal and in that by which it is pretended" (284). The first conjunct is sufficiently obvious - the pretender must figure in an important way in pretending - the second is more subtle. To illustrate its application, Anscombe cites the case of a King who creates the appearance that one of his subjects is planning to kill him. If we consider this situation as it would be were it not a pretense, the King would play no role; it would be the subject who would be the agent. The second "central" feature is what she calls a "rule of sequences of tenses," which she illustrates, as follows, without stating it in general terms. If we imagine someone breaking some crockery in order to convince someone else that he was angry (when in fact he is not), it cannot be said that later, when he points to the alleged evidence of his anger, that at this time he is pretending to have been angry at that earlier time. Attention to time is characteristically at the center of many of the things Anscombe has said about human action, such as in treatment of objections to her work on intention by Goldman and Judith Thomson ("Under a Description" in Collected Papers of G. E. M. Anscombe Vol. 2. Oxford. 1981.). We have seen that Anscombe prefers to speak of "central" rather than "essential" features of pretending, and she is careful to note that her concern is with merely two, but there is more to be said, perhaps of greater significance than about the two she discusses, particularly given the distinction both she and Austin draw between 'pretending to' and 'pretending that'.

In cases where I pretend that there appear to be few constraints. I can walk a short walk and pretend that I'm on a long walk to St. Petersburg from Istanbul, with little to persuade others that I am not; although persuading them that I am may be difficult, should there arise a skeptical doubt. But, now, pretend to stand under a tree while standing in the desert; try standing under a tree falling under the description 'Oak', now 'Elm'. On the other extreme, try pretending to dance without actually dancing. My point is this: there seems to be a fact of the matter whether on these occasions I am pretending or not pretending. It may make sense to try to pretend to stand under an Elm as opposed to an Oak in the desert, but to correct one's effort as long as one is standing may be made difficult when this is my intention. It is easy enough to imagine pretending that I am standing under an Elm or dancing when I am not. But next consider this: it seems that there are certain minimal facts about performance that must obtain if there is to be an outward criterion of my intention to pretend. Pretending for Anscombe always carries with it an intention (284). This is an important claim.

I may do x with the intention of doing y, even though I do not do x intentionally. For example, I may spill the cream with the intention of passing it. But I cannot pretend to be ill with the intention of having you call a doctor without intending to pretend to be ill. Notice in connection with this that it does make sense to say: 'I went to the party with the intention of pretending to be a bear', although going to the party is not part of what I do in order to pretend to be a bear. What all must be done in order to pretend to do something is not always clear. One might ask, for example: what all must I do to pretend to wash the windows?

In one sense - in the case of the window washer who is actually spying - one might say that more is involved in pretending to wash the windows than merely washing the windows. Commenting on Austin's case, Anscombe recalls his description of what makes the activity a pretence.

...we must allow that he is indeed actually cleaning the windows, from start to finish and throughout the whole time he is pretending to be cleaning them. But it is still a pretence, because what he's really doing all the time is something different, namely noting the variables: he is only cleaning the windows to disguise and promote this other activity...(270)
Anscombe engages Austin on the matter of what is meant by "actually cleaning the windows."
Professor Austin explains "It is still a pretence [i.e., though the windows are being cleaned], because what he is really doing is something quite different." But the point of the expression "What he is really doing is something quite different" is that 'what he is really doing' falsifies the appearance he presents by cleaning the windows. (290)
The difficulty Austin would quite likely have with this is that looked at in what may first appear the most compelling way the idea that somehow actually doing something falsifies the appearance created by doing that very thing is misleading at best. How can actually washing windows falsify the appearance created by washing windows? What is the 'appearance' actually being falsified, if not the very activity performed? Anscombe has an answer:
The appearance presented by cleaning the windows is that, in cleaning the windows, he is doing something in some ordinary proper course of things. (ibid)
Yet, all Austin needs to say in order to caste a shadow of a doubt on this way of viewing the matter is to say that the spy washing the windows is doing something in some "ordinary proper course of things." Indeed, this is why he is washing windows and not, say, clicking pictures in the face of the one upon whom he is spying with, say, a large camera. If we are to get clear on this matter we must understand how it is that an appearance becomes "falsified." Part of the difficulty is that Anscombe appears to believe that what does the falsifying is the false appearance that the window washer is "doing something in some ordinary and proper course of things." What is problematic is that this is not a false appearance at all; it is this very appearance that makes the pretence so successful, maybe, but it is not a false appearance. The window washer is washing the windows in a very ordinary way. While Anscombe is entirely correct in pointing out that there are other things that the washer may "really" be doing, such as composing verse, she has not given any basis for singling out the window washing as falsifying. What is it, then, about the real window washing that makes it a pretence, whereas should the culprit be composing verse at the same time this would not count as falsifying the appearance. To get around this difficulty we need a clearer statement of what the relevant appearance actually is.

If the appearance of actually washing the windows were actually false, as Anscombe suggests, it would be correct to say, "He is merely pretending to be cleaning the windows." But since it is not cleaning the windows that he is pretending to do (he is not a mime) a more correct formulation is to say "He is pretending merely to be cleaning the windows." This is the sort of scopal consideration of adverb behavior ("merely") that Austin was sensitive to, in particular in his treatment of the case of the child who is 'pretending to be driving a car' without 'pretending to drive a car'. (272) Once we describe the window washer as 'pretending merely' to be washing windows as opposed to 'merely pretending', we get a better grasp of what distinguishes the other things he may "really" be doing, such as composing verse, and the activity of window washing. It can be put this way: The difference between composing verse and window washing in the case of the spy is that 'ordinary and proper' actions have objectives which are ordinarily unexpected, precisely, because the activity is 'ordinary and proper'. Composing verse may not be unexpected; earning a wage may not be unexpected. But if spying is expected (the victim having been tipped off) still the windows are getting washed and the washer is still pretending. There is no firm reason to believe that either Anscombe or Austin would object to this, given what they have said. What is important is that Anscombe has raised the issue: what constraints and conditions are there on pretending without being sufficiently precise.

Anscombe asks, "Why cannot a baby six months old pretend to be in pain?" (290). The question harks back to Wittgenstein's, "Why can't a dog simulate pain?" (PI. 250). Wittgenstein notes that "...the surroundings which are necessary for this behaviour to be real simulation are missing." Our distinction allows us to set aside comparison of the baby and the window washer, for what is being asked of the baby is whether it can merely pretend to be in pain and not whether it can pretend merely to be in pain. An important difference between pretending to be in pain and pretending to be washing windows surfaces: I can pretend merely to be washing windows, but even though I wash windows, that activity may be a pretence - I am not a window washer; but where I create my own pain in order to deceive, the pain itself is no pretence - I am a pain-feeler. Although on a number of occasions the situation may correctly be described as one in which I am pretending merely to be in pain). In the former case, I may say that I am not merely pretending to be a window washer; in the latter I cannot say I am not merely pretending to be a pain-feeler, or "one who feels pain." In discussing the unpretentiousness of babies, Anscombe contrasts our inclination to attribute pretence to animals with our disinclination to do the same in the case of babies. Her position is stated succinctly and straightforwardly.

Anscombe may have had a reason for rejecting such a contrast, but she never gives it. In order to see what it might have been consider the case of the window washing example. What is falsified is said to be the appearance presented by the behavior of washing the windows. Now if it should turn out that the appearance created by being in pain is falsified, then some symmetry between the pain case and the window washing case may be maintained. However, while the lack of characteristic appearances may be associated with window washing, such is not the case with pain; and it is for very similar reasons that Anscombe alleges the difference between a baby's pretending to be in pain and an animal's pretending to be in pain. Just as the characteristic appearance of limping may cause us to believe, when voluntary, that the behavior is pretence, voluntary behavior such as a baby might display, while thought to be voluntary, will not cause us to suspect pretence; so, too, the behavior of a man in pain may, if we suspect it is voluntary, causes us to attribute some pretence. The issue, while not entirely wide open, will be closed in these remarks on Anscombe and Austin by noting that there does appear to be what I would call an "asymmetry" between pain and window washing which is recapitulated in the semantical difference between 'pretending merely to X' and 'merely pretending to X'. As Anscombe observes: we may wish to appear something we are not with no further end in mind (292). In this instance we are merely pretending to be such and such, even though we might very well be doing the same thing we are doing here but with some further end in mind; however, in the case of pain there is little point or inclination to merely pretend we are in pain and this for the reason that unlike limping there is no characteristic appearance, but also because even when the spy is washing windows there is a sense in which he isn't washing windows, such a sense does not apply to the man who pretends merely to be in pain, for such a man there is no sense in which he is not "really" in pain. Before concluding this critical examination of Anscombe's views, a mention of one further distinction she draws is well worth considering for its broader implications.

Anscombe retains a special category for cases where the pretender "unreflectively knows he is pretending." (292) This she calls "plain" pretending. She gives as an example the following dialogue between a teacher and the parent of a pupil:

Did James tell you I had to beat him today? Yes, he said he got beaten. Oh, did he tell you what it was for? Hm! He told me it was for something he had written in his notebook. I don't know - what he said was that he wrote "Casson is a sod." I gather Mr. Casson is one of the masters. Oh!...Well, that's not very nice is it? Well, I understand your beating him, but all the same, surely this is quite an ordinary thing for a boy to do? No, in my experience, not at all normal. Let the parent's reply to him be unspoken, since it is: "Stop pretending".
In this discussion of Anscombe's reply to Austin we have reserved most of our criticisms for Anscombe. Austin is masterful at drawing from subtle semantical differences in our ordinary usage of 'pretend'. Here, however, in this notion of "plain" pretending we have in Anscombe an ingenious observation with applications yet to be fully explored.

Donald Davidson has made very good use of the notion of self-deception in connection with the irrationality of the incontinent man ("Deception and Division." in Essays on Davidson: Actions and Events. edited by Vermazen, Bruce and Merrill Hintikka. Clarendon Press. Oxford. 1985). Throughout Davidson's extensive work on the subject of incontinence he has availed himself of the idea that the incontinent man - the man who believes there is a better course of action but proceeds intentionally do something else - is in some sense irrational. Late in the development of his ideas on the subject he introduced the notion of self-deception. In some concluding remarks he at one point says of the incontinently irrational man,

...the irrational step is therefore the step that makes this possible, the drawing of the boundary that keeps the incontinent man's beliefs apart. In the case where self-deception consists of self-induced weakness of warrant what must be walled off from the rest of the mind is the requirement of total evidence. (op. cit. p. 148)
What Anscombe tells us about "plain" pretending is, I believe, of considerable value in providing a greater understanding of the psychological component to incontinence of philosophical significance. The incontinent man is like the school teacher of Anscombe's example, the one just mentioned. Like this plain pretender who, as Anscombe remarks, "would take some reflection, in the realize that he knew the contradictory of what he said"; so too, the incontinent individual believes there is a better alternative to the one he takes, but fails to "reflect" on the comparative value of the alternatives before him. The incontinent man may, then, be viewed as a pretender of sorts, in particular a pretender who is deceiving himself. The question of irrationality, while important, does not reach to the core of the agent's circumstances. An enrichment of Davidson's view with this insight is well worth the doing, but regrettably not part of our task here and now.

We have explored what Austin said; we have gleaned a significant portion of Anscombe's response. One may ask: who best captured the meaning of 'pretend'? There can be no final word. Austin's distinctions are clear, but not always as firm as he no doubt supposed. He pursues the problem Bedford addresses doggedly and with some success. By contrast Bedford's problem as seen by Anscombe is not free from some of the very misconceptions one can understand her to be attributing to Austin and yet she never convincingly answers the following question: What is the difference between pain and anger which is such that in the case of pain there seems to be some inner state or event that provides the ground for distinguishing pretend pain and real pain, whereas in the case of anger no such thing comes immediately to mind? At times Anscombe conflates essences and inner states. We have seen that this may be an error. We see also, however, that the "essentialist" line apparent in Austin is open to what are now commonplace criticisms. J. L. Austin's remarks are penetrating; given this and the fertile creative mind of Elizabeth Anscombe there are conceptions yet unborn. Hopefully what has been said here will allow this to become a fact should others, inspired by their example and the standards they set, decide to make the necessary investment of energy and time.