Aristotle and Plato on “Appearing”

Published in Mind, 1964.

In part I of this paper my aim is to elucidate a certain argument of Aristotle’s against an account of appearing that can, I think, be attributed to Plato. In part II I proceed to apply this argu­ment to a view of sense impressions recently put forward by D. M. Armstrong in his book, Perception and the Physical World.

I

The Argument (De Anima, III, 3, 428 a24-b9) and its background

Aristotle’s concern in this chapter is to give an account of phantasia, i.e. the state or capacity in virtue of which we say we have an image, or, as I would prefer to say, ” the state or capacity in virtue of which we say we are appeared to in such and such a way”. Translating the word “phantasia” is difficult. In departing from the traditional rendering of it as “imagination”, I follow the precedent set by the Oxford translator of De Memoria, etc., viz. J. I. Beare, who translates “phantasia” as “The Faculty of Presentation” and “phantasma” as “presentation”.1 I feel we must avoid the odd procedure of translating “phantasia” with a word that has very different uses from the Greek one, and then, in order to avoid misunderstanding and nonsense, having to exclude from our minds the ordinary uses of “imagination”, and, especially, of the verb “to imagine”! There are, I think, two advantages in preferring the ‘appearing’ terminology to that of ‘presentation’: (i) We retain in this way the connection, present in the Greek, with the passive forms of the verb “to appear”. So, we can have the awkward but intelligible “the appearing(s)” for “to[ta]phainetai”, “[the] Being appeared to” for “[to]phainesthai” and “appearance” for “phantasma”. (ii) We also have an apt way of expressing Aristotle’s account of appearing, that is given later in this chapter, and which is couched in explicitly causal terms.2

Aristotle has various arguments to show that phantasia, i.e. the state or capacity of being appeared to, is not a state or capacity in virtue of which we judge (krinomen) rightly or wrongly. It is not, therefore, for him sense-perception, opinion (judgement), scientific knowledge or intuition. We have to be careful how we interpret this general thesis of Aristotle’s. We may try the following: Locutions of the form “I am aware of such and such an appearance”, or, less misleadingly, “I am appeared to in such and such a way” need not be construed as expressing a ‘judging’ (a krinein, i.e. an act of judgement). In this respect they contrast with locutions of the form: “I see (that) a …” (I see a grey cat eating a mouse), or “I believe (my opinion, view) is …” (I believe that spicy food is unhealthy), or “A is either X or not X”, or the scientific description of what consti­tutes an eclipse of the moon. (In the last two cases Aristotle would have said that we have intuition and scientific knowledge respectively.) This contrast does not commit Aristotle to the view that the concept, e.g. “being appeared to in a blue way” cannot be exercised in a judgement. For, I do not think Aristotle would, or need, deny that any of the following expressed a judgement: “This piece of cheese will look green tomorrow”, “the wine tasted bitter yesterday”. What he is concerned to deny is that my being appeared to in a green or bitter way is an exercise of judgement.

Before coming to the argument we are interested in, Aristotle points out that while animals are not capable of belief (pistis) they are capable of being appeared to (phantasia). This is not just a statement of empirical observation, for he adds that belief is necessarily involved in judging (doxazein) since it is impossible to judge what one does not believe to be the case. But with the notion of belief is connected the notion of coming to believe, of being convinced, and with that the notion of deliberation or thinking (logos).3 And this last is absent from animals, while the ability of being appeared to is not. The refutation of the view that being appeared to (phantasia) is a sensuously mediated act of judging or a blend of sensation and judging comes at this point. The three phrases Aristotle uses to characterize-this thesis occur in Plato:

(i) Being appeared to is opinion (judgement) plus sensation (doxa met’aisthēseōs)

(ii) Being appeared to is opinion mediated by sensation (doxa di’aisthēses)

(iii) Being appeared to is a blend of opinion and sensation (sumplokē doxas kai aisthēseōs. Plato has summeixis in place of sumplokē).

Let us see now what Plato did say. Since (i) above is not used by Plato to describe phantasia explicitly, I shall concentrate on (ii) and (iii), both of which occur in the Sophist (264, A.B.). Plato’s characterization of phantasia occurs near the end of his investigation of the nature of discourse (logos), judgement (doxa) and appearing (phantasia), to establish the possibility and nature of falsehood. Plato’s aim at this point was to counter the Sophist’s contention that there can be no such thing as falsehood unless a connection is shown between discourse, judgement and appearing on the one hand, with not-being, on the other. Having shown how a piece of discourse, e.g. somebody’s stating or saying “Theaetetus is flying” can be false, viz. in that it combines a real subject (Theaetetus) with a real action (flying) but in a way other than that which is the case, since Theaetetus is now sitting, Plato now asks: “What then of thought, judgement and appear­ing? It is clear, isn’t it, that they may all be found within the soul both as false and as true?” “The thought,” he explains, “and the statement (expressing the thought) differ only in that while the former is a silent converse of the soul with itself (the speaking in one’s heart, as we may say, but not the speaking to oneself, which is quite different), the latter is a spoken flow of thought. But since we affirm and deny statements, similarly then the silent affirmation and denial of a thought is what we call judge­ment. But in the case where the process (including affirmation or denial) occurs through the mediation of sensation we call this ‘appearing’ (phainetai).” He then concludes, naturally, that since thought, judgement and phainetai are the silent analogues of discourse, and since discourse is true or false, the other pro­cesses can be true or false.

The important thing to notice is that what is called phainetai here is the judging not that something, e.g. looks green, but that it is green. Taylor comments: “Imagination, in this passage, thus means more particularly the mental interpretation of an actual sense-perception, which may, of course, often be erroneous, as in an example used in the Philebus…”4 In other words, ‘appearing’ is the silent affirmation or denial (the ‘taking’ or the ‘not taking’) of what my senses tell me. The example in the Philebus, referred to by Taylor, illustrates Plato’s view that perceptual judgement, and the endeavour to form such a judge­ment, always spring from memory and perception in the following way: A man sees an object at a distance not very clearly and so asks himself, “What is that which appears (phantazomenon) to be standing by the rock under the trees?” And he may answer, rightly “A man”, or mistakenly, “A statue”. (As in the Sophist Plato remarks that the statement (logos) is the loud version of the judgement (doxa).) He now offers the following explanation of this mental act: The soul, at such times, is like a book. Memory and perception coincide, and together with the experiences relative to them they seem almost like writing statements in the soul. When this ‘writing experience’ writes truly, then true judgements and true statements are formed within us as the consequence of its work – but when the scribe within us writes falsely, the result is the opposite.

Another, and for our purposes, crucial passage is to be found in the Republic (602, C,D). It is in Book X where the discussion concerns the poet, and he is said to be an imitator not of things as they are but of appearances, i.e. thrice-removed from truth. Socrates then asks what kind of faculty or capacity (dunamin) in man is that to which imitation makes its special appeal? He explains: One and the same body appears to be different sizes when seen near and when seen at a distance. Similarly, the same objects appear straight when out of the water and crooked when in water; and the concave becomes convex, owing to the illusion about colours to which sight is liable. Thus every sort of confusion is revealed within us; and it is this weakness in our nature that is exploited with magical effects by the art of painting in light and shadow, the art of conjuring and many other such devices. But the arts of measuring and numbering and weighing have, gratefully, appeared to help with respect to these illusions, in such a way that the apparent greater or less, or what appears the more, or what appears the weightier no longer rule us but give way to the calculating or measuring part of the soul (logisamenon). And when this faculty has done its measuring and signified that some things are greater than or less than, or equal to, other things, it is often contradicted by the appearances (sc. contradicted by how these things and their relations appear). But it is impossible for the same thing (part of the soul) to hold two contradictory judgements (beliefs) (enantia doxazein) about one and the same thing. So, that part which judges (to para ta metra doxazon tēs psuchēs) against the measurement cannot be the same part as that which judges in accordance with the measure­ment. And the latter (sc. the one that trusts or believes in measures and calculation) is a better and the other, an inferior, part of the soul.

The importance of this passage lies in two things:

(a) That the different parts of the soul talked about are both said to judge. We could express this as the thesis that equates “X’s appearing f to S” with “S’s judging ‘sensibly’ that x is f”.

Hence, (b) ‘Appearings’ can be said to contradict judgements formed by other means, e.g. measurement, precisely because they are or involve judgements; albeit of an inferior sort.5 We can turn now to Aristotle’s argument against Plato’s characterization of phantasia. I shall give it first in full and then comment: “It is obvious that phantasia, cannot be judgement (doxa) with, or through, sensation, or a combination of judgement and sensation for these reasons [sc. those mentioned earlier at 428 a18-23], and also because the judgement, if it exists at all, cannot be of something other than what the sensing is of. What I mean is, that [on this view] my impression (my being appeared to) must be the combination of the judgement that something is white with the sensing (perceiving) of white; it could not be the combination of the judgement that it is good with the sensing of white. [The] Being appeared to ([to]phainesthai), therefore, will have to be [on this view] the judging what is sensed; the two things not being accidental to one another. But, often, the ‘appearings’ of things [i.e. our being appeared to with respect to things] are false [with respect to these things], in cases where we hold true beliefs about these things, e.g. the Sun looks a foot across, but is believed to be larger than the inhabited part of the earth: so, one must either have discarded one’s true opinion, though the fact remains and one has not forgotten or changed conviction by persuasion, or, if one still has the belief, it must be both true and false (a true opinion, however, becomes false only if the facts alter without being noticed). Therefore, phantasia, is neither any one of these things nor a compound of any of them.”

Elucidation of the Argument

Does Aristotle’s argument refute what seems to be Plato’s view? I think so. Aristotle’s remarks bear on the two points (a) and (b) (cf. above), we extracted from the Platonic passages, and in particular the one from the Republic. Against (a) he argues that it will not do to equate being appeared to with judging ‘sensibly’, precisely because there are cases where my being appeared to in a certain way (as of an object a foot across) with respect to a thing (the sun) is quite compatible with my contemporaneously held true belief that the sun is larger than the earth. If my being appeared to was a case of ‘sensibly’ judging, then either (1): I would have to say that my belief that the sun is larger than the earth was lost, even though it is still a fact that the sun is larger than the earth. But in fact in these cases I neither lose my belief, nor forget it, nor am I persuaded against it. For example there are two suitcases A and B, and A is larger than B. I weigh them and find that B is heavier than A. Nevertheless K feels heavier than B. In this situation knowing that B is heavier than A, I am not taken in by appearances. Yet while believing or judging truly that B is heavier than A, it is still true that I am being appeared to with respect to A and B in a certain way, viz. A’s feeling heavier than B. But my being appeared to in this way cannot be a judging, otherwise how could I hold the other true belief in being appeared to in this way.

Or (2), the other impossibility, if I still hold the true belief we would have to say that it was both true and false. There is some difficulty of interpretation here due to compression. The ques­tion is, “Which belief does Aristotle think would be both true and false on the view he is attacking?” Ross6 gives the follow­ing explanation: “Aristotle is arguing that on the view that phantasia is a combination of judgement and sensation, a single appearance (phantasia) will sometimes be both true and false, e.g. if the sun were at the same time looking a foot across and believed to be immensely large, and if phantasia was a combina­tion of a sensation element and a belief element, the same phantasia would be true in respect of the belief-element and false in respect of the sensation element.” I suspect that Ross has got hold of Aristotle’s meaning here, though the way he expresses it might easily mislead his readers cither into thinking that Aristotle is saying something quite different from what he is in fact saying, or in not understanding him at all. For as a result of reading Ross we may think that Aristotle is open to the retort that since the judging and the sensing must be of the same thing, the judge­ment (belief) that the sun is larger than the earth is not part of the phainetai (appearing) in question. For surely, it may be said, the belief or judgement element of the appearing in question could only be that the sun is a foot across. So, Aristotle could not possibly be complaining that the appearing (phainetai) would be true in respect of the belief element, viz. that the sun is larger than the earth, and false in respect of the sensation element, viz. the sun’s looking a foot across. Plato would just insist that the judgement involved in the appearing, viz. the sun is a foot across, is false and that the other belief, independently held, viz. the sun is larger than the earth, is true. Aristotle’s objection is, thus, rendered harmless. And if we do allow two beliefs in the situation, we could certainly not interpret Aristotle as saying that Plato’s characterization of phantasia committed the latter, in such cases as the present one, to the view that a man held at one and the same time two beliefs such that the one entailed the falsehood of the other and vice versa. We could not interpret thus because this is not an impossibility (as Plato would have been quick to point out), and, further, to say this is not to say that a belief is true and false at one and the same time. If Aristotle meant the former he would have said so but would have been saying nothing objectionable to Plato. Now, the important feature of Aristotle’s example is that we have to say that there was only one judgement or belief involved here, viz. that the sun is larger than the earth, in a perceptual situation where the observer is appeared to as of an object a foot across. The view being attacked is committed to the prima facie thesis that there are two beliefs here: the one concerned with or, rather, involved in, the appearing and the other concerned with the relative sizes of the earth and the sun. If the holders of this view still want to say, as the case requires, that only one belief is involved, the one concerned with the relative sizes of the earth and the sun, then this belief would be both true and false: true as a belief (the earth is smaller than the sun) but false as the belief element in the appearing. For the latter to be true, the judgement involved in the appearing would have to be “the sun is a foot across”. The next words in the text Ross finds puzzling, and thinks they may not belong here. They are: “A true opinion, however, becomes false only when the fact alters with­out being noticed.” Ross’s reasons for querying these words are that they refer to a change of facts and no previous hint has been given, and also that the grammar is odd. If my account of what Aristotle is saying is correct, a sense can be given to these words, especially if we connect them with what he says elsewhere (De Somn., III, 461b). Aristotle’s meaning, I think, is this: Since a belief cannot be true and false at the same time, the only thing that can happen is for the belief to become false when unbeknown to the observer, there is a change in the facts. For example, suppose I see an object at the distance which looks a foot across but which I judge to be much larger. Suppose now that un­beknown to me the object is travelling towards me and that, for some reason, it decreases in size in such a way that while I am aware of an object that looks a foot across but which I judge to be huge, by the time it comes much nearer to me the object not only looks but is a foot across. My judgement in this case would become false even though the appearance has not changed.

How would Plato meet Aristotle’s dilemma? Let us imagine the following discussion between them.

P.: With regard to the first horn of your dilemma, surely all we need say about your example is that the thought or judgement involved in the “appearing” is not allowed to become a belief or judgement. It is held back, as it were, by the other contra­dictory belief.7 We need not say, as you suggest we should, that our true belief, viz. that the sun is larger than the earth, is lost. What is lost, in this case, is our inclination to judge that the sun is a foot across. It is an unfortunate fact about human nature that due to ignorance such inclinations are not checked often enough. As to the other horn of the dilemma, I would argue, again, that in the case under discussion, there is only one belief and it is true. There is also, of course, the checked tendency to judge a falsehood which would be the belief element in the appearing. The sensing element in the appearing we can now express in the judgement “the sun looks a foot across” because we know that it is not a foot across. But if we did not know this fact we would have judged that it was a foot across. You must also notice that cases like the one you mention only occur when, as I pointed out in the Republic, judgements formed by means other than our “appearings”, e.g. by measurement, counteract the evidence of the senses. But for this indepen­dent knowledge we would have judged in accordance with our senses.

A.: You avoid my criticism by what seems to me to be a shift in your position; shift detrimental to your view. Your start­ing point is, isn’t it, that the state of being appeared to is a com­bination of judgement and sensation. That is, when a subject S is appeared to f with respect to an object x, this state you describe as consisting of S’s judging (taking) ‘sensibly’ x to be f. But there are cases, like the one I mentioned, where while we would say that S was being appeared to f with respect to x, we could not say that he thereby judged x to be f. Your reply, in effect, seems to deny that in these cases of appearing there is judgement involved. But if you admit this, your general thesis about phantasia falls to the ground. We must separate the appear­ing from the judging both because of cases like the above and also cases where I suspend or delay judgement.8 For example, being appeared to in the same way as in the example, I may say of the sun (being ignorant but cautious), “It looks a foot across but I cannot say whether it is or not. I can’t tell just by looking”. In this case, unlike the other one, the absence of the relevant judging is not due to a previously held independent belief, though my caution may derive from previous mistakes. It follows from this that it is not the case that the absence of belief or judge­ment in perceptual situations only occurs where we have in­formation that contradicts the evidence of the senses. It also follows that from the fact that we did not know the sun to be larger than the earth it is not necessary that we should judge it to be a foot across when it looks a foot across. We may be cautious or uncertain. I said earlier that your reply to my objection seemed to deny your original thesis, because I suspect that you are still not prepared to agree to the separation of the appearing from the judging, though, I confess, I am not at all clear why.

P.: You are quite right. I cannot agree to the separation you propose because I want to say that ‘appearing’ is an in­ferior kind of judging (cf. Republic passage) and also because I want to say that perceptual illusion is a special kind of false judging (cf. the Sophist passage). It seems to me, in any case, that the shift you detect in my position is not as fatal as you think. Let me try and deal more fully with your examples. Though there is a difference between the case where we reject the evidence of our senses and the case where we suspend judgement, what is common to both is lack or absence of assent or affirmation (phasis) to a proposition or thought, viz. e.g. ” the sun’s being a foot across”. Now, in ordinary cases, i.e. in cases where there is no contradictory belief, the thought is affirmed and so becomes a judgement. In the sort of cases you mention, however, the thought is not allowed to become a judgement. Nevertheless, what is disbelieved, or what belief is suspended about, is a thought which is in the soul (is ‘entertained’ as we might say). So, in the one case (the disbelief case) the thought is not affirmed and in the other (the suspension of belief case) the thought is neither affirmed nor denied. I may, perhaps, add at this point that your remark about animals being capable of being appeared to but not capable of thought is not decisive. For, given that men are capable of thought, may it not be the case that their being appeared to takes a different form (i.e. needs a different account) from that of animals? (Perhaps we could put this by saying that our application of the concept to animals involved an analogical extension.)

A.: I think lean show you now how the shift I mentioned earlier is damaging to your original thesis. You are in effect granting the separation between the ‘judging’ and the ‘appear­ing’ I was at pains to establish and you to deny. For if you are not in fact granting it, I can face you once again with my original point about the necessary coincidence of what is judged and what is sensed that your thesis commits you to, and which is precisely what left you open to my objection. Let me explain: You say that in the normal course of events when something looks like a man I take it to be a man if there is no other belief to the contrary. With this I don’t think I disagree. It is your analysis of this situation that I find dubious. You suggest that the thought, e.g. “It is a man standing under the trees near the rock” is ‘entertained’ in the mind, and it is there that it is normally, affirmed or denied. You then argue, that in the cases I mention the thought is still entertained but not affirmed or denied. So far, however, your theory does not make clear the connection between judging (and entertaining a thought) on the one hand, and being appeared to on the other. Your thesis concerning this is that it is a combination of judging with sensing, or, less metaphorically, that the appearing (to phainetai) is the judging through (or by means of) sensation. The implications of this view are crucial, for it seems to me that you avoid my criticism only by switching back and forth between two very different interpretations of your thesis. If your view is that being appeared to is a judging (i.e. the mental utterance of a judgement) made in a context of sense-perception, then you have not avoided the separation between the judging and the appearing, for a judge­ment is a combination of concepts (noēmata, cf. De An., III, 8, 432a 10-14), and so it is general. But appearances, like sensa­tions, are particular. Hence, a judgement’s referring to some (particular) appearances cannot be part of the content of the judgement. If it were so, we would have to say that the judge­ment “A man is standing under the trees by the rock” was a different judgement on each occasion of its utterance (loud or silent) in a sensible context.9 But surely the only thing that may be said to be different is the act of judging, and not the con­tent of the judgement; and it is precisely this act of judging, performed in a particular sensible context, that may be referred to the appearances (phantasmata). But if this is so, my being appeared to in a certain way with respect to a certain object cannot be my judging it to be so, or even my entertaining the thought that it is so, for on this analysis it is not necessary that the thought I entertain or the judgement I make, exactly coincides in content with my sensation. It is because of this that I can say (judge) not only “this is good (pure)” in the sensible context of something appearing white, but also “This is larger than the earth” in the sensible context of an object appearing a foot across. Of course, these are advanced performances, but what they show is the distinction between judging and being appeared to (cf. De Somniis, II, 460b 16-25). Even when I judge “according to appearances” my being appeared to is not my judging, but, rather, the evidence or ground for my judgement. And I think you lean on this account, when you try to deal with the dis­belief cases, because it is the one that makes sense of the fact that in the same context of appearances, e.g. an object’s looking a foot across, we can perform any of the following acts of judge­ment: “The sun looks a foot across but is larger than the earth”, “the sun is a foot across”, “the sun is larger than the earth”, or even “that bright (one-foot) object in the sky is a star larger than the earth” (cf. “that tiny speck on the horizon is a large battleship”).

This brings me to the other way of understanding your thesis which is the one, I think, you want to stick to, and which is open to my objection. According to it, my being appeared to in a certain way is my judging such and such, or, at least, my enter­taining a thought that such and such. It is a necessary con­sequence of this view that the difference between judging a thought by means of appearances and judging it on its own (kath’autēn) as it were, is a difference not merely in the context of judging but in the content of what is judged. It follows, in turn, that the thought I entertain or the judgement I make in (not on) being appeared to in a certain way must coincide exactly in content with the appearing, otherwise you will have the separa­tion of the judging from the appearing that you dread. Given this, I can face you with my objection: Taking the initial formulation of your thesis, viz. that in being appeared to I judge, the dilemma is: either you must say that I judge the sun to be a foot across, in which case what happens to my true opinion, seeing that I haven’t forgotten or changed my mind about the facts, and the facts haven’t changed? Or, my state of being appeared to in this case will consist in my judging something that will have to be both true and false, i.e. if my judgement is, as in our example, that the sun is larger than the earth it will be true vis-à-vis the facts but false as the judgement involved in my being appeared to the way I am. According to the amended thesis what happens is that I merely entertain the thought that the sun is a foot across but judge that it is larger than the earth. My point now would be this: either the amended thesis involves the rejection of the original thesis, or it does not help at all with the objection. For, if, in this case, in being appeared to I do as you say I do, then there is no reason why I should not say that in the normal cases (when I judge according to appearances), what I judge need not be identical in content with the thought I am supposed to entertain in being appeared to in a certain way. In short, if we allow in these cases a separation between the judge­ment made and the thought ‘entertained’, there is no reason why we should not allow it in all cases. The important point being not that the appearing is a thought, in your terminology, but that the judgement made can be different (hence, distinguish­able) in content from the thought ‘entertained’. But this is incompatible with your original thesis.

If on the other hand you want to insist that in normal cases no distinction can be made between judgement made and thought entertained, it should be a mystery how in some cases I can judge contrary to, or even differently from, the appearances. There is no such mystery if we separate the appearing from the judging because we can say that on being appeared to in a certain way I judged so and so. And there is no compulsion here to say that on being appeared to in a certain way I must judge (have judged) so and so, even if normally what an object looks like is what we judge it to be.

P.: I think I can see there is a difficulty. But let me turn critic and ask you two questions: (i) You have objected to my saying that “appearing” is judging. But I am not sure whether you mind my ‘writing’ metaphor in the Philebus. I said there that in a perceptual judgement memory and perception meet … and so on. Now, isn’t it the case that whether I see some­thing that I take to be a man or whether I see something that looks like a man but do not judge it to be so, the fact is that in either case I must be exercising my concept (noema) man, and, surely, this is a sign that judgement occurs? (ii) You seem quite pre­pared to say that “appearings” are true or false, while at the same time you deny that they are, or involve, acts of judgement. How do you reconcile these two contentions? That is, in what way are “appearings” true or false?10

II

I wish now to consider certain arguments put forward recently by D. M. Armstrong in favour of a view of appearing, and its role in perception, which seems to me very similar to that I attribute to Plato.

Armstrong’s view is that when I have a hallucination as, say, of a cat on the mat, I will believe two things: (i) there is a cat on the mat before me; (ii) I am seeing that cat. His suggestion is that the occurrence of these two false beliefs constitutes sensory illusion. In a parallel way the two corresponding beliefs in veridical perception are true.11

He is now faced with the difficulty created by the cases of perception without belief, for, prima facie at least, these cases suggest that there can be sensory illusion without false belief. Armstrong thinks, somewhat in the way I made Plato do, that he needs only amend his analysis and not abandon it altogether. He points out that such cases only occur where we have independ­ent information that runs counter to the senses, and he concludes that in all cases of “perception without belief” (“p.w.b.” from now on), there would be a belief that we are perceiving something but for independently acquired information. “Formally”, he says, “we have a contrary-to-fact conditional statement, ‘If I did not know or believe X, then I would believe I was perceiving a Y’” (p. 86).

Given his view that sensory illusion is nothing but the holding of the false belief that we are perceiving something, he is committed to saying that contrary-to-fact conditional statements of the above form are logically necessary. He cannot say they are empirically true, for if they were it would be conceptually possible to separate in the case of illusion the sense-impression from the belief. Armstrong (p. 86) gives an instance of a logically neces­sary statement, viz. “If he had been a father, he would have been a male parent” and then says that the contrary-to-fact conditionals we are concerned with are logically necessary, e.g. “If I did not know or believe that the sun is larger than the earth I would believe I was perceiving the sun as being a foot across”. For Armstrong, then, “p.w.b.” is essentially belief-inducing; in default of other, contradictory, beliefs it must issue in the belief that we are perceiving something in our environment. The amendment of the original thesis to cover the “p.w.b.” cases is that since these cases are essentially belief-inducing they must involve the thought that we are perceiving something in the world, a thought held back from being a belief by other, contradictory, beliefs. Sensory illusion, then, is nothing but the false belief, or inclination to a fake belief, that we are perceiving some physical object or state of affairs.

There are several points to be made here. To begin with there will be difficulties with the “suspension of belief” cases. In these cases, as we have seen, it need not be that I hold other, contradictory, beliefs that stop me from judging according to the evidence of my senses. For example, “It looks like an air­craft but I can’t say whether it is or not yet. We better hold our fire.” Armstrong might reply that in these cases we do not have cases of perception. Since, ex hypothesi, I have not made a judgement, I could not be subject to an illusion. But this reply is not available to Armstrong if he is to avoid begging the question, for this presupposes that the question of the nature of sensory illusion has been settled in his favour. Our problem is the relation of “appearing” to “judging” in the context of perception. So, in this case too Armstrong would have to say that the object’s looking like an aircraft was the inclination to believe that it was one.

Now, what would in this case be the logically necessary con­trary-to-fact conditional? We may try: “If I was not uncertain or cautious, I would be inclined to believe that I was seeing an aircraft.” Is this statement logically necessary? Hardly, for while I could with perfect sense ask the question “why?” of this statement and receive the informative reply “Because it looks like an aircraft”, it is doubtful whether I could with equal sense, and equal expectations for an informative reply, ask the question “why?” of the statement, “If he had been a father, he would have been a male parent”. To put it differently, while in the latter case the proper reply to the question “why?” is likely to be “because this is what ‘father’ means, viz. ‘male parent’”, in the former case it is more than dubious that the reply would be, “because ‘being uncertain (or cautious) as to whether an object is an aircraft’ means ‘being inclined to believe that one is seeing an aircraft’”.

Perhaps the logically necessary statement should read:

(1) “If I did not think that what I saw might not be an air­craft (supposing we interpret ‘I was uncertain whether it was or not’ in this way) I would be inclined to believe I was seeing an aircraft”. If this is a logically necessary statement it is a very odd one, since its force consists in telling us that a state of affairs in the world would necessarily obtain if certain other empirical conditions were fulfilled. I say it is odd, because we expect a logically necessary statement to show us something about the meaning of the words or expressions contained in the statement, or about the logical relation(s) between the concepts employed in the assertion of such a statement. Armstrong is debarred from saying that (1) is an a priori synthetic because, he informs us (n. 2, p. 89), as an Empiricist he has no truck with the possibility of the synthetic a priori. But now, it seems to me that if a statement is a genuine contrary-to-fact it is unlikely to be logically necessary, and if a logically necessary statement is expressed in a contrary-to-fact form, then this way of expressing it has no special force over and above its expression in a categorical or normal conditional way. The logically necessary statement “If he had been a father, he would have been a male parent” surely has no force other than “If he (anyone) is a father, he is a male parent”. The reference, e.g. to J. Smith’s not being a father does not contribute to the statement’s being logically necessary. Per contra, the reference to Hannibal seems essential to the state­ment’s “If Hannibal had entered Rome at such and such a time, he would have conquered Italy” being a genuine contrary-to-fact.

Now what about (1)? Let us contrast it with what would be a logically necessary statement in this context. Take (1.1): “If I did not think that what I saw might not be an aircraft, I would have thought that it was an aircraft.” This is still not a logically necessary statement, for the question “why?”, to which an ex­plicitly informative reply is expected, is in place even here. For example, “By the way it looks” or “By its steady movement”. (Cf. Armstrong’s own example: “If I did not know you so well, I would have thought you were joking.”) What would be a logically necessary statement is the following:

(2) “If I didn’t refrain from judging that what I saw was an aircraft, I would judge either that it was or that it wasn’t” (i.e. if I did not refrain from judging, I would judge). But here the contrary-to-fact form does not serve any special purpose, i.e. given that “not-judging p” is the contradictory of “judging p”, and that “judging p” means, à la Plato, “affirming or denying p”,12 we could express the above logically necessary statement formally as follows: “Either I not-judge p or I judge p”. Alternatively, “If it is not the case that I not-judge p, then I judge p.” The question “why?” here is out of place. I remarked that our contrary-to-fact (1.1) contrasts with the logically necessary statement (2) in two ways: (i) the question “why?” is significant in relation to (1.1) but not to (2); and (ii) that the contrary-to-fact form is essential to (1.1) but not to (2). Reflecting on these two points we shall see how Aristotle’s argument applies to Armstrong.

(i) Supposing in reply to the question “why?” with regard to (1.1) one said, “Well, what else could I have done (thought, judged)?” This might do as a short-tempered reply to a why-question addressed to (2) but not to (1.1), for the questioner might well retort, “Couldn’t you have thought it was a very large bird?” In other words, there must be a reason why but for your uncertainty you would judge the object to be an aircraft. It does not necessarily follow from your uncertainty as to whether it is an aircraft that, had the uncertainty been absent, you would have taken it for an aircraft, because had your uncertainty been absent nothing follows a priori as to the specific way in which it would have been absent. Armstrong’s thesis, however, does commit him to saying that something did follow a priori as to the specific way in which the uncertainty would be absent. So either Armstrong’s thesis is wrong (the contrary-to-fact (1.1) is not logically necessary), or there is another element in the situa­tion that makes it so. But what is that other element? It cannot be the object’s looking like an aircraft for according to Armstrong’s view an object’s looking like an aircraft is nothing more than the belief (inclination to believe) that one is seeing an aircraft. And on this thesis one cannot say that the object’s looking like an aircraft was the ground for asserting the contrary-to-fact (1.1) since this would be denying the thesis.

If, on the other hand, Armstrong’s ground for asserting (1.1), indeed the ground for asserting it to be a logically necessary statement, is that “a thing’s looking to me like an aircraft” means (is no more than) “I think (I am inclined to think) that I am seeing an aircraft”, then he is open to the retort that a thing’s looking like an aircraft is compatible not only with its being a large bird, but also with its being taken to be a large bird. And nothing Armstrong has said (apart from the bare assertion of his thesis) shows that the situation where something’s being taken to be a large bird, is logically incompatible with its looking like an aircraft.

(ii) But there is also the difficulty of taking the contrary-to-fact (1.1): “If I did not think that what I saw might not be an aircraft, I would have thought I was seeing one” to be logically necessary. For, either the contrary-to-fact form is essential to asserting (1.1) or it is not. There is now an interesting dilemma: To the extent to which we want to consider the statement as a genuine contrary-to-fact conditional, to that extent we do not want to treat it as logically necessary. Hence, we would be resisting the identification of having an impression with the holding of a belief about what we see. For, we feel that the ground for asserting the contrary-to-fact cannot be a statement of logical necessity, but, rather, a particular (and, perhaps, a general) matter of fact, viz. “it looks to me now like an aircraft”. That is, we would be arguing: because it looks to me like an aircraft, had I not been uncertain as to whether it was or not, I would have thought I was seeing one. But this neither supports nor is supported by the identification of impression and belief. Rather, it exemplifies our general belief that normally we judge an object to be what it looks like. This general belief, however, apart from not being equivalent to the identification required by Armstrong’s thesis, demands it denial.

In contrast, to the extent to which we want to treat the con­trary-to-fact as logically necessary, to that extent its form becomes inessential, for we feel that the ground for treating (1.1) as logic­ally necessary must itself be logically necessary, and not a par­ticular (or general) matter of fact, viz. something like “Either it is the case I not-judge ‘my seeing an aircraft’ or it is the case I judge ‘my seeing an aircraft’”. That is, we would be arguing: since it is either the case I not-judge p or I judge p, then had I not not-judged p, I would judge p. But if this is what we are doing we have lost any support we might have had for the identi­fication of sense-impressions and beliefs, for from the above logically necessary statement nothing follows as to what my judgement is or would be and its relation to my sense-impression.

One may wonder, now, how could it ever seem that (1.1) was a logically necessary statement? Only by begging the question, I think. That is, it is not the case that because (1.1) is a logically necessary statement, that Armstrong’s thesis is correct. Rather, only if we do identify impressions and beliefs in the way proposed can we think (1.1) to be logically necessary. That is, it is only if we think that a thing’s looking like an aircraft is nothing more than our believing we are seeing an aircraft, that we might con­strue “If I didn’t think that I may not be seeing an aircraft, I would have thought I was seeing one” as logically necessary. Mistakenly, nevertheless. For it is based on the illusion, re­inforced by the way the “p.w.b.” cases are treated, that this statement is equivalent in force to “If I didn’t doubt (disbelieve) my senses, I would believe them” which equivalence is, of course, false.

We have then been cheated. The alleged logical necessity of (1.1) does not establish Armstrong’s thesis but requires it. But since the alleged logical necessity of (1.1) cannot be established independently of Armstrong’s thesis, we have no reason to accept it. Indeed we have good grounds for thinking that it is anything but logically necessary. Aristotle’s criticism in fact would be that if the identification is granted, it would be a logical mystery how I could come to regard any of the contrary-to-fact condi­tionals, e.g. “If I did not know that my eye-ball was pressed, I would believe I was seeing two candles”, as true. Furthermore, if the identification is allowed, it would be a logical mystery how I could come to formulate the contrary-to-fact at all i.e. if its looking as if there are two candles is identified with my believing I am seeing two candles, how could the statement “I believe I am seeing two candles” justify the contrary-to-fact “If I did know or believe my eye-ball was pressed (and, by implication, that I am not seeing two candles), I would believe I was seeing two candles”. This is precisely what Aristotle meant when he said that this view would involve one in the contradictory position of holding a belief to be both true and false at the same time.

And as we have seen, the amended thesis either involves the abandonment of the view, or is open to similar objections. For, if the inclination to believe something is interpreted as “the thought’s (that we see something) being held back”, what is relevant to the issue is not its being a thought but its being held back. And if we allow a separation here, we allow the conceptual possibility of a separation between impressions and judgements.

It is worth noting how Armstrong prevents himself from seeing this when he says (p. 86):

If I have an hallucination of a cat, and if I have no independent information of any sort which suggests that I am not looking at a cat, then it is logically necessary that I will believe that I am looking at a cat. If I perceive an actual cat (perceiving it as a cat), and if there is nothing to suggest I am being hal­lucinated, then it is logically necessary that I will believe I am looking at a cat.

But it is logically necessary that I will believe that I am looking at a cat only because my “being hallucinated as of X” entails my “believing I perceive X”. This is not the same as “My having the visual impression of a cat, and my not having independent information of any sort which suggests that I am not looking at a cat, entails my believing that I am looking at a cat”. Surely it is this latter entailment that Armstrong needs (and cannot have), but not the former.

University of New South Wales

  1. The expression “being appeared to in such and such a way” comes’ of course, from Chisholm’s, Perceiving. It may be felt that my suggestion will meet with difficulties when we come to Aristotle’s distinction of phantasia into aisthētikē and Logistikē. My reply to this is that “being appeared to” is general enough, as Chisholm points out, to cover both perceptual situations, where we talk of something’s “looking”, “sounding”, “smell­ing”, etc., and to non-perceptual situations where we talk of “after­images”, “ringing in the ears”, “hallucinatory pink rats”, and also mental images. The role of ‘appearing’ in the context of imagining raises issues that are logically akin to those discussed in this paper. They require, of course, separate treatment. []
  2. Chisholm in Perceiving, chap. 10, pp. 143-149, proceeds in a similar way. Chisholm’s attempt to define “appearing” in causal terms is, of course, much more fully developed than Aristotle’s. []
  3. This necessary connection between belief and judgement may be denied. It seems to me, however (though I cannot argue for it here), that such a denial can ultimately be justified only on a ‘behaviouristic’ account of belief; and such accounts are fraught with difficulties. []
  4. Plato, The Sophist and the Statesman, Trans. with intr. by A. E. Taylor, ed. by R. Klibansky and E. Anscombe, 1961, p. 177. []
  5. We could, perhaps, contrast Plato’s and Aristotle’s views here in terms of a distinction Chisholm and others draw (Cf. Perceiving, chap. 4, p. 44) between epistemic and non-epistemic uses of “appear-words”. We may say that while Plato is committed to recognizing only epistemic uses of “appear-words”, Aristotle wants to insist that there are, or, must be, non-epistemic uses as well. []
  6. Aristotle, De Anima, ed. with intr. and comm. by Sir David Boss, 1961, pp. 287-288. []
  7. This way out. which I place in Plato’s mouth, is to be found in Arm­strong. I take it to be quite a natural extension of what Plato says in the Sophist passages considered above. []
  8. Again, it seems to me consistent with Aristotle’s general view to put these words in his mouth. []
  9. Cf. P. T. Geach, Mental Acts, pp. 64-66. []
  10. I hope to deal with what would be Aristotle’s reply to these questions on another occasion. The issues raised by question (i) above are discussed with considerable insight by Geach in Mental Acts, pp. 121-129. For a discussion of points raised by (ii) above, cf. Vesey’s “Seeing and Seeing as”, P.A.S. 1955-56 and, also, Chisholm’s Perceiving, chaps. 9 and 10. []
  11. Perception and the Physical World, chap. 7, p. 83. All subsequent page numbers are to this work. []
  12. Cf. Von Wright’s essay “On Conditionals” in Logical Studies, pp. 127-130, and, also, the Sophist passage considered earlier. []