SOURCE: “Berkeley's Imagination,” in Essays on the Philosophy of George Berkeley, edited by Ernest Sosa, D. Reidel Publishing Company, 1987, pp. 85-102.

[In the following essay, Tipton considers Berkeley's belief about the human imagination and its role in his philosophy of immaterialism.]

In Principles, 22-3, and in a parallel passage in the Dialogues,¹ Berkeley presents an argument that is important in his eyes because he thinks, or appears to think, that it is sufficient to establish his immaterialism. He is, he says, “content to put the whole” upon the issue of whether his reader can “conceive it possible for one extended moveable substance, or in general, for any one idea or any thing like an idea, to exist otherwise than in a mind perceiving it.” The reader, or Hylas in the Dialogues, has only to look into his own thoughts to find the answer. To be sure, one can frame the idea of books or trees without framing the idea of any onlooker, but that does not do the trick. “… it only shows you have the power of imagining or forming ideas in your mind; but it doth not shew that you can conceive it possible, the objects of your thought may exist without the mind: to make out this, it is necessary that you conceive them existing unconceived or unthought of, which is a manifest repugnancy.”

Most commentators, including me, have not been impressed, though some have been more sympathetic than I was in my book, not necessarily arguing that what Berkeley says should convince us, but that it is certainly less silly than is sometimes suggested, or, alternatively, that if he does not establish what he thinks he does, he has at least got on to something of interest.² The debate continues, and it raises complex issues. It emerges that it is far from clear what the argument is.³ It therefore remains unclear how it should be answered. In this paper I want to examine a number of the approaches to it, and I hope we shall learn something on the way. I do not, however, expect to come up with much that is startlingly original. Indeed I would like to think of this paper as defending what some will take to be an orthodox interpretation. If I dwell on commentators with whom I disagree, this is in part to show that the orthodox needs a defence. That is what I shall be doing in my first three sections.

I

There is, first of all, what I think of as the brusque approach, represented by J. O. Urmson who quotes a large chunk of Principles 23, but devotes only eight lines to dismissing the argument as “surely fallacious.”⁴ For him, all Berkeley is entitled to is the proposition that “it is not possible that a body be conceived of and at the same time both exist and not be conceived of,” a proposition that is of no more significance than a claim that I cannot kick a chair that is at the same time unkicked. It does nothing to support the contention that the concept of an unperceived object is logically deficient. Arthur Prior and J. L. Mackie have argued at greater length, bringing formal logic to their aid,⁵ but their conclusion is similar. Putting it simply, Berkeley's mistake will be comparable to that of supposing that because, quite obviously and uncontroversially, I cannot answer the challenge to “think of a number” without thinking of a number (the number 12 perhaps) that is, at that moment, thought of by me, I cannot make coherent sense of the notion of a number that is not thought of by me. Indeed, this parallel points to what both Mackie and Prior see as a further weakness in Berkeley's argument. Not only does it prove too little—that I cannot think of a tree without thinking of it, rather than that I cannot conceive that a tree might be unthought of—it also “proves” too much. As Mackie puts it (and he is rather more impressive here than Prior): “Since the choice of material things, trees and houses, as what Hylas tries to conceive existing unconceived plays no special part in the argument, it would, if sound, show equally that no one can coherently postulate the independent existence, out of his own mind, of other human minds or of God.”⁶

I called this the brusque approach, but not to be insulting. I concede that the line of thought Urmson, Prior and Mackie attack can be abstracted from what Berkeley actually says, and would argue only that there must be more to be said if we are to understand the relevant passages and, in particular, how it is that Berkeley can think that his argument supports the claim that the esse of sensible things is percipi rather than concepi.⁷ This will presumably involve attending to another feature of Berkeley's argument, which stands out in the Principles but is there in the Dialogues too, this being the suggestion that getting at the truth requires “looking into your own thoughts” and considering what goes on there when you think of trees or other sensible objects: an exercise of introspection that would hardly be necessary to come up with the truism that I must conceive what I conceive. The relevant feature stands out in the Principles where it occurs in Berkeley's own presentation of his argument, while in the Dialogues it is presented almost as an afterthought, and by Hylas rather than Philonous. But it needn't worry us that it is the materialist who makes the crucial move. At this point Hylas is yielding, and in fact giving Philonous just what he needs. Indeed it is interesting that while Philonous makes no mention of perceiving during the relevant exchanges, consistently referring to conceiving (or to what is “in the mind”), it is Hylas who brings us back to esse/percipi. He cannot, he allows “conceive a tree as existing unperceived or unthought of.”⁸

Too often, it seems to me, critics of Berkeley's argument ignore this aspect and, what is crucial, the fact that readers of the Principles at least have been alerted to the importance of drawing “the curtain of words” as an aid to reaching truth.⁹ As a result, they sometimes give the impression of talking past Berkeley, who certainly doesn't accept that contradictions and repugnancies are necessarily discoverable at the level of sentences or linguistic utterance. The point isn't new of course: not long before Prior's paper was published in 1955, W. H. Hay had stressed that “if our purpose were to refute Berkeley … we should have to revise his account of language, and not merely ignore it,”¹⁰ and his critical eye was on G. Dawes Hicks's examination of the argument over twenty years earlier. But the point remains important. Nor is there really any excuse for forgetting the strictures Berkeley issued towards the end of his introduction. There he had urged that we should contemplate “bare ideas.” But in the first edition version of Principles, 24, he issued a reminder. “Cou'd men but forbear to amuse themselves with words, we shou'd I believe, soon come to an agreement in this point.”

II

I'd like to think that a natural move here would be towards an approach that did make a clear link with the Introduction to the Principles and the attack on abstraction and which, frankly, took a critical line on that basis. But what I in fact want to do now is to look at a highly original interpretation proposed by Désirée Park¹¹ that conflicts with the sort of approach I would support just as much as it does with that of Urmson, in that while the Urmson approach is abstemious, Park allows Berkeley considerable resources that would in fact allow for a defence of his argument against most of the objections that have been made to it. In my third section I shall try to make a link between her interpretation and what Michael Ayers says about Berkeley's argument in his introduction to the Everyman edition of Berkeley's works.

A feature of the Park approach is her insistence that Berkeley's argument should be taken in its context, an insistence that leads her to lay great stress on the fact that what Berkeley is trying to show is that we cannot conceive any particular sensible thing existing unperceived. Now an obvious consequence of this is that though in the Dialogues (and Park concentrates on the Dialogues version) Philonous argues that it is a contradiction to talk of conceiving “a thing” which is unconceived, thus apparently inviting the Prior-Mackie observation that this would apply to minds as well, it must be assumed that Berkeley still has sensible things in view. He might as well have had Philonous point out that it is “a contradiction to talk of conceiving a sensible thing which is unconceived”; and he might actually have resisted an interpretation that made this simply a consequence of a more general claim about conceiving “things.” It might be something about sensible things that matters, and the relevant consideration might not apply to minds.

This, Park holds, is in fact the case. She believes that if Berkeley did hold that his crucial claim applied to sensible things, he would not accept that the same claim could be applied to minds, this on the ground that, for him, minds and ideas have nothing in common apart from the name “thing.”¹² But she also takes it that Berkeley did have a special reason for applying it to sensible things, for she holds that Berkeley is working with the background assumption that sensible things are ideas “of the peculiarly Berkeleian sort.” Thus:

… an idea for Berkeley must conform to the single condition which he states; that is, any idea must be perceived. This means at least that it is imagible. Indeed a test for a putative Berkeleian idea is whether it can be captured as a mental picture, of whichever sense one pleases. It may be added that each idea is wholly present, momentary and therefore is strictly unrepeatable.

This must not be forgotten when we consider the passage in the Dialogues. So, given the context—that is, given that the discussion concerns sensible things; that sensible things are ideas “of the peculiarly Berkeleian sort”; and indeed that “by definition” no Berkeleian idea can exist unperceived—it is hardly surprising if Philonous takes it for granted that we cannot conceive a sensible thing existing “otherwise than in a mind.”

That at least is Park's view, and though I can't accept it I should perhaps point out that it might seem to have certain advantages. For example, and if I understand it correctly, the interpretation would give Berkeley an easy answer to one objection that I and others have levelled at him in the past, this being the objection that, even when restricted to sensible things, his line of argument must be too strong in that it would lead to the conclusion that no individual can coherently suppose that any object can exist that he does not perceive. Berkeley's answer would be this. Hylas supposes that he can imagine an unperceived tree. That is, he can frame an image of a tree and need include no perceiver in his mental picture. And, on this basis, he concludes that an unperceived tree must be a possibility. But it is Hylas who has made the mistake. All he has offered is an idea, in this case an image, that he himself perceives. This has no tendency to show that a sensible tree can exist unperceived, for on Berkeley's general theory any sensible tree will remain a collection of mind-dependent ideas. There is no solipsism. All Hylas has to recognize is that any mental picture anyone forms is dependent on the mind of whoever it is that forms it, while, any sensed idea remains dependent on some mind, though of course not necessarily on his mind.

Rather similarly, the Park interpretation would enable Berkeley to answer the stock objection that he is guilty of confusing a thought-vehicle with the object of thought,¹³ because now (and again if I have got it right) if anyone is guilty of that confusion it will be Hylas. Not for one second need Berkeley relax a firm grip on the distinction between the mind-dependent ideas that are constitutive of physical reality and those feebler ideas we can form when exercising our imaginations. The thought will merely be that though we can of course form the image of a tree without a perceiver, this carries no implication that trees can exist unperceived. Even the image is perceived, by the imaginer, while for any sensible object, esse is percipi. That is the background assumption.

It seems to me, however, that the Park approach is not acceptable, and this for two different but connected reasons. The first, and perhaps this isn't an objection in itself, is that it downgrades the importance of the argument by making it do much less work than is normally supposed. Thus there is no doubt that philosophers such as Prior, Mackie and Bernard Williams,¹⁴ who would not consider themselves Berkeley scholars, have given the argument the attention they have because, along with most scholars, they have taken it that Berkeley believes himself to be justifying, not presupposing, the esse/percipi thesis. And if that isn't an objection in itself, it seems to become one when joined with the second point which is that it is Berkeley himself who leads us to expect a lot from this argument. Thus it is Philonous who introduces the Dialogues passage by saying that Hylas can “pass by all that hath been hitherto said, and reckon it for nothing, if you will have it so,” claiming that he will “put the whole” on what follows. And of course there is a similar grand claim in the Principles. Commentators have taken the invitations in both works seriously, and I think that they are right to do so.

The objection is a serious one, even if it is not decisive. For even if it is countered that we may not be expected to pass by literally all that has been said but only what has been said on some particular point or issue, Berkeley's wording must mean that some things he would take himself to have established can be disregarded, and that in turn must place some limit on what Park can appeal to and accuse others of overlooking. Thus, in general, it won't do to assume that acceptance of p must be presupposed if p has already been argued for or asserted,¹⁵ given that we are to pass by some things and that p may be one of those things. And, to turn to a concrete case, it won't do to object to Bernard William's reading of the Dialogues passage that it has been “conceded some pages before … that anything which Hylas can see … must be an idea,” given that Philonous is responding to Hylas's failure to take earlier points on board. Philonous's response is, admittedly, to allude to the arguments Hylas has forgotten or misinterpreted, but then he sets them aside. We can “pass by” that material.

Unfortunately Park doesn't tell us what she thinks Hylas is being allowed to pass by,¹⁶ and I do think that that is a serious weakness in her paper. But I think that more must be involved than she can recognize, and certainly that her charge against Bernard Williams is not fair. To see this, we should look at the context.

When the Dialogues open, Hylas assumes that sensible qualities exist “without the mind,” and Philonous sets out to convince him that they are mind-dependent. He does this by running through what are in fact the secondary qualities, arguing that colors, tastes and so on, as Hylas perceives them, are not external to the mind. Then, when Hylas draws a distinction between primary and secondary qualities, Philonous claims that similar arguments apply to the supposed primary qualities. Indeed, there is another point. The ideas of extension and motion cannot be separated off from other sensible qualities, even in thought.¹⁷ Consequently, “where the one exist, there necessarily the other exist likewise.” On p. 194 (in the standard edition) Hylas concedes that “all sensible qualities are alike to be denied existence without the mind,” and this means, though Berkeley doesn't actually use the word here—that they are mind-dependent ideas. Given that this concession comes six pages before the argument we are concerned with, it might seem that Berkeley can take it for granted at the later stage. It must be this that Park thinks Williams has overlooked.¹⁸

What Park has herself overlooked, however, is the fact that though Hylas does make the concession, it is followed by a series of exchanges in which he attempts to retrieve the situation, a quite proper procedure given that his acceptance of the earlier arguments was a qualified one. Thus when he accepted that extension was ideal, he reserved “a right to retract my opinion, in case I shall hereafter discover any false step in my progress to it.” And this qualification is generalized on p. 194:

You need say no more on this head. I am free to own, if there be no secret error or oversight in our proceedings hitherto, that all sensible qualities are alike to be denied existence without the mind. But my fear is, that I have been too liberal in my former concessions, or overlooked some fallacy or other. In short, I did not take time to think.

What follows is a series of attempts by Hylas to locate some “error or oversight” and show that, after all, sensible qualities are not Berkeleian ideas. Whatever he comes up with, Philonous cannot answer him by appealing to the concession that Hylas is now attempting to withdraw.

The argument we are concerned with has to be set in this context. At the bottom of p. 194 Hylas attempts to locate “one great oversight” in their failure to distinguish the object of perception from the mind-dependent act of perception; and then on p. 197 he claims that when he thinks of sensible things as modes or qualities, “I find it necessary to suppose a material substratum, without which they cannot be conceived to exist.” Philonous responds by collapsing the distinction between the perceiving and the object, and by showing that we have no adequate conception of a “support” of qualities. And in each case Hylas concedes. But he is still looking for “some fallacy or other” in the earlier reasoning. Our passage follows.

Locating what he now takes to be “the ground of all our mistakes,” what Hylas points to is their procedure of running through the qualities individually, for, as he now sees it, this leads only to the conclusion that “each quality cannot singly subsist without the mind.” It remains possible that “blended together” they can. Viewed in that way we need not consider them Berkeleian ideas. That is the move, and it exasperates Philonous who reminds Hylas that the earlier arguments were not designed to show that qualities cannot exist singly without the mind, “but that they were not at all without the mind.”¹⁹ But it is at this point that Philonous makes his generous gesture. Hylas can “pass by all that hath been hitherto said, and reckon it for nothing, if you will have it so.” To say the least, it is difficult to avoid the conclusion that this means just what it seems to mean. Hylas is being allowed to withdraw his earlier concession, even though he has found no ground for doing so. It can play no part in what follows.

III

Park attempts to defend Berkeley by allowing him very generous resources. On the face of it, Michael Ayers's treatment of the argument, which is offered within the context of a more general treatment of Berkeley's strategy, looks very different,²⁰ though as already noted, I think the two approaches can be linked. I should state at the outset, though, that, unlike Park, Ayers concentrates on the Principles version.

What Ayers does is to focus attention on what is undoubtedly Berkeley's main target in the Principles, this being the notion of a physical reality conceived in “a sense-independent way”; and Ayers certainly thinks that if we consider what he writes in Principles, 23, in that light we can see that what he is after is the not implausible point that “to think and talk of the table in an unoccupied room is still to think and talk of the potential perception of it, of how it would look, feel and so on.” Again:

if … I “imagine trees … in a park or books existing in a closet, and nobody by to perceive them” [a quotation from Principles, 23, of course], I merely imagine how it would be to perceive them myself. I do not conceive of Locke's “particular constitution which everything has within itself, without any relation to anything without it.”

“The strength of this argument,” Ayers observes, “should be obvious to anyone who has felt the temptations of phenomenalism.” Ayers concedes that we should not applaud too loudly and that the argument as Berkeley actually presents it has a feature that has given it “an unjust notoriety.” But this feature, to which I shall return, is said to be “inessential” to the basic point. It “does not affect the main contention that neither imagining a tree nor anything else can count as conceiving of a tree ‘as it is in itself.’” Ayers's treatment of the argument is brief, and it is perhaps possible that I have misunderstood him. But I think my summary of his position is fair.

If it is, Ayers's interpretation will have the obvious merit that it focuses attention on what nobody could doubt is Berkeley's main target in the Principles: the broadly Lockean notion that our senses do not reveal things as they are in themselves, and indeed (as Berkeley interprets the view) that they do not really acquaint us with the “real” things at all. Against that, Berkeley can plausibly claim that he cannot conceive a tree that is not perceivable. The best I can do in thinking of unperceived objects is to “imagine how it would be to perceive them myself,” and this means that I must, after all, think of them as perceivable, and as having the familiar qualities that I find in them when I do actually perceive them. This would give us something like what is often labelled “common-sense realism.” Add the assumption that to imagine how a tree would look to me is to imagine a possible sensation and we are coming round to phenomenalism. It all begins to look rather impressive.

But it also begins to look rather disappointing. As in the case of the Park approach, it will turn out that the argument is doing less work than is commonly supposed. Helping ourselves to a tag introduced by A. A. Luce,²¹ it seems that the conclusion we might be led to is that esse is percipi aut posse percipi, not esse is percipi. I am reminded of the passage in the Third Dialogue where Philonous appeals to “the common sense of the world” for the truth of his notion, drawing attention to the view that real things are perceivable, only to be met with the understandable protest that there is a gap between “being perceivable” and “being actually perceived.”²² At that stage Philonous is able to help himself to the premise that nothing is perceivable but a mind-dependent idea, and it would seem to be that premise that Berkeley will need on the Ayers interpretation of Principles, 23, given, that is, that Berkeley is to be seen as offering clinching support for idealism, and not just as opposing what is after all, just one possible alternative to it. But if that is correct, Ayers's position will be quite close to that of Park, in that the premise Berkeley will require will be the one that Park gives him as an assumption. Perhaps it is significant, then, that Ayers and Park react to more orthodox critics in much the same way. As already noted, Ayers refers to an “inessential” feature that has given Berkeley's argument “unjust notoriety.” And Park strikes a similar note. She refers to a “notorious” argument that is, she claims, “not Berkeley's argument.”

So far as Ayers is concerned, however, I do have to be a little more conciliatory than I was in the case of Park, and in particular to concede that it is easier to defend this sort of interpretation with our eyes on the Principles version, for the immediate context is different and, in general, Berkeley is readier to help himself to assumptions he shares with his philosopher opponents, including the assumption that we perceive only ideas. It follows that he could be resting on that assumption in Principles, 23, and that his demonstration of his doctrine “in a line or two” does depend upon it.

If I am not convinced, this is partly because the argument does recur in the Dialogues, and, for reasons already given, I find it impossible to accept this sort of interpretation of the later occurrence, but also because the feature of the argument that Ayers takes to be “inessential” to it does seem to me to be so well calculated to support the strong conclusion that Berkeley appears to want. In Ayers's words, that feature is found in the assumption that “imagining or conceiving of a sensible quality [is] perceiving a recalled idea or image.” It is perhaps worth quoting Berkeley's own words:

When we do our utmost to conceive the existence of external bodies, we are all the while only contemplating our own ideas. But the mind taking no notice of itself, is deluded to think it can and doth conceive bodies existing unthought of or without the mind; though at the same time they are apprehended by or exist in it self. A little attention will discover to any one the truth and evidence of what is here said, and make it unnecessary to insist on any other proofs against the existence of material substance.

The context makes it quite clear that, when the object concerned is not present to my senses, the idea will be one I frame and, as Ayers says, an image. But that notion is, surely, highly significant, and it is difficult to believe that it is idling. For if Berkeley does take the notion seriously he will have an argument that will lead him directly to esse/percipi. Putting it simply, if to conceive of a sensible quality (or a tree) is to contemplate or perceive an image I frame, then the only object involved will be mental. That should be uncontroversial. After all, as Berkeley noted in Principles, 3, that “ideas formed by the imagination” do not exist without the mind, “is what every body will allow.”

IV

Now of course there is nothing original in the suggestion that this or something very like it is what lies behind the argument we are considering, and Pitcher for example is one commentator who seems to interpret Berkeley in this way, being as a result very critical of him for confusing a thought-vehicle (or image) with the object of thought (or what is imagined).²³ What Pitcher does not do is to make any attempt to link the disastrous move to its wider historical context, or to Berkeley's strictures in the Introduction to the Principles. In broadest outline, there are two things we need to bear in mind. The first is that Berkeley was strongly influenced by the notion that, to avoid talking nonsense, we should be prepared to “draw the curtain of words” and attend to the ideas we find in our minds. And the second is that he was highly critical of others for not following their own advice. Berkeley's rigor in this area may lead to a strange result, but it is hardly fair to see it as a quirky aberration in his thought. We are on to something central.

Approaching Berkeley in this spirit we get one immediate result, in that we can well understand his assumption that to determine the coherence of the generalized notion of an unperceived object we should be prepared to consider any one object chosen at random—a tree perhaps—and consider whether we can conceive it existing unperceived. That challenge does not seem grossly unfair in itself; but it is backed up by the assumption that to think in general terms is not to have an abstract idea in one's mind, but rather a determinate idea which can represent many particulars. Nor is it surprising that when Berkeley sets the word “tree” aside, all he finds is an image. A test on any first-year student, or on oneself, is likely to produce the same result. And this gets us very quickly to the argument we are concerned with. Look at a tree. Then close your eyes and try supposing that it is unperceived. That involves attending to what is in your mind: the image. Given that you are now allowed to entertain thoughts about that object, there is one thought that you cannot coherently entertain, this being the thought that it is not mind-dependent. It is, after all, your image.

If this line of thought seems just too absurd, that is not the way it looks to Ian Hacking, one writer who is commendably sensitive to the historical context. As he notes:

Descartes' Rules for the Direction of the Mind are famous because of their author, but are entirely typical of the manuals in circulation at the time. The gist of the advice is that to avoid error we must train ourselves to “scrutinize” our ideas “with steadfast mental gaze.” Stop speaking and start looking, looking inside yourself. What is one to look at? The answer is given by that code word “idea.”²⁴

In this tradition, ideas need not be images, but they are, it seems, mental. Indeed, to suppose that one can cast the eye of thought on something that is not mental, that is not, as Hacking puts it, “an object of thought,” can easily appear nonsense. That is the notion that Berkeley needs, so it is significant that Hacking's sensitivity to the context goes along with the uncommon judgment that Berkeley's argument is “very impressive.” As he puts it: “Within the conceptual scheme in which it is formulated, the steps seem to me cogent (not incontestable, of course, but cogent).”²⁵ And one can accept this without suggesting that Pitcher and other less sympathetic commentators have got the story straightforwardly wrong. We do not share the relevant conceptual scheme; nor should we. And the position it leads to is flawed. Hacking shows how Berkeley got where he did. It does not follow that we should take Berkeley's route.

V

It might seem that this paper could end there, and so it could were it not that there is a residual puzzle, a puzzle I can introduce by returning briefly to Ayers, who does note in passing that the feature of Berkeley's argument that he sees as “inessential” can also be seen as “a natural consequence of its historical context.” This hints at the Hacking approach, though it only hints at it, both because Hacking regards the feature as crucial, and because, in elaborating, Ayers refers only to Locke, noting that “[Berkeley's] move may be open to objection, in that Locke … would not treat imagining a tree as perceiving a tree but as perceiving an idea of a tree” (my emphasis).²⁶ This perhaps suggests that Berkeley may have been a little bit careless, but it in fact raises a real difficulty facing the believer in what Hacking describes as “purely mental discourse”: the issue of the “aboutness” of thought.²⁷ The problem may not appear pressing to a philosopher who is either very vague about “ideas,” or who supposes that in some sophisticated fashion he can build “aboutness” into them,²⁸ but I think that it is pressing for Berkeley.

To bring this out it will perhaps help if we start by taking our cue from Hacking and consider the historical context. For if we start from the received notion that to think of X is to have an idea in one's mind, and then ask ourselves how philosophers in this period could live with that notion, I think the answer that emerges, or at least part of the answer, is that they were often very vague about what an idea is. And it does seem that the basic notion often appeared truistic precisely because to say that “A has the idea of X” was to say little more than that A is thinking of X (or even that he has the capacity to think of X). Significantly, this position tended not to go along with an “imagist” view of ideas. For Berkeley, however, an idea formed by the mind is always an image, and this has important consequences. Let me illustrate.

It can appear to be a truism that to think of God is to have in one's mind the idea of God. Clearly that seemed evident to Descartes, but then (as he in fact observed to Hobbes)²⁹ he did not hold that an idea is (at least necessarily) an image. Hobbes, on the other hand, denied that we can have an idea of God, but then, he was an imagist. Berkeley denied it too, for the same reason, but for him this went along with a revision of the received view that to think of X is to have an idea of X. At best this was true of sensible things, things that can be pictured. For other things, Berkeley has to tell a different story. And of course he does. As it happens the point is obscured from Hacking, because he makes the unfortunate mistake of supposing that Berkeley “acknowledges plenty of ideas of which we can form no images, God and the will, for example,”³⁰ but it remains important. For this is not the only area where Berkeley's insistence that any idea we frame must be an image has to lead to a revision of received notions about thought.³¹

With that in mind, we can turn our attention away from the received notion that to think of X (whatever X may be) is to have an idea of X to the more restricted notion that to think of a sensible thing is to perceive an image, for it seems to me, not only that this is an unsatisfactory notion, but that this should have been obvious to Berkeley himself. It should have been obvious precisely because, for wholly understandable reasons, Berkeley is determined to distinguish between the ideas we frame (ideas of the imagination) and those sensed ideas, imprinted from without, which he claims constitute a genuine, though still ideal, “real” world. I suggested earlier that one apparent merit of the Park approach as I understood it was that it would allow Berkeley to retain a firm grip on this distinction. On what I would take to be the correct interpretation, that firm grip is lost. On this interpretation Berkeley does confuse an image and the “real” thing, when he, above all, should not.

In this connection it may help if I draw attention to one entry in the Commentaries which suggests that Berkeley was quite capable of avoiding the confusion. For in entry 657a he writes:

properly speaking Idea is the picture of the Imagination's making this is y^e likeness of & refer'd to the real Idea or (if you will) thing.

I think, incidentally, that Park mishandles this in her Complementary Notions when she sees it (together with the corresponding recto entry) as saying that “perceived ideas are simple and depend for their existence and meaning on being related by spirits or minds to something unlike themselves” (my emphasis);³² but she is quite right to see Berkeley as recognizing that “so-called ‘real’ ideas or ‘things’ are distinguishable from the ideas or pictures referred to them.” But this only highlights my problem. For the entry suggests that Berkeley was indeed quite capable of seeing that thinking of the sun, say, must involve more than attending to an image, and that the “picture” (or thought-vehicle) must, as he puts it, be “refer'd to the real Idea,” which will be what I am thinking about. But of course this is not a problem for Park. It is the rest of us who should be puzzled. For it is the more orthodox commentators who have it that Berkeley does confuse the image with the object of thought; and even Ayers has him holding that conceiving something is perceiving “a recalled idea or image no different from an idea of sense except in its immediate origin, context and vivacity.” Against this, it might be argued that the philosopher who penned PC, 657a, could never have made that mistake.

VI

I see the position differently. For one thing it seems to me that it is very difficult to read Principles, 23, without seeing Berkeley as confusing the image with the object of thought;³³ and, for another, it could hardly be argued that Berkeley did come round to a clear and consistent view on the relevant relationship. I find Principles, 36, particularly interesting in this connection. For it seems very significant that here, where Berkeley is explicitly addressing himself to the distinction between those “faint, weak, and unsteady” ideas we form and those ideas which “are not fictions of the mind perceiving them” but which do constitute “real” beings, he also claims that “the sun that I see by day is the real sun, and that which I imagine by night is the idea of the former” (my emphasis). And what is significant, of course, is the fact that it is supposed here that the sun which I imagine by night is a “faint, weak, and unsteady” image which I form, and that, on the face of it, Berkeley is explicitly committing himself to the view that I cannot imagine “the real sun” at all. Now I am sure that this is a slip, in the sense that Berkeley would not want to commit himself to the view that the very thing that is sensible is not also imaginable, but it is not an unimportant slip. It reveals that Berkeley is confused, and for a quite understandable reason. That is, he is more or less taking it for granted (in the spirit of the “received” view) that a person who has the idea of the sun in his mind is imagining—thinking about—the sun; but this is difficult to reconcile with the independently tempting view that the only object available is a “picture.” And if Principles, 36, provides clear evidence that Berkeley had not achieved a satisfactory reconciliation, the “notorious” argument we have been considering is, it seems, simply a dramatic consequence of that failure. For he faces a dilemma. In short, it is only if the object I imagine is distinct from any image I frame that Berkeley can hold that I am imagining “the sun that I see by day” (or trees); but it is only if the object is the image—“the sun … which I imagine by night”—that the object is obviously mind-dependent. In fact, it seems Berkeley wants both.

Clearly this is a most unsatisfactory position to leave Berkeley in, but I suspect that it may be where we have to leave him if our interpretation is not to go wildly awry. I concede, however, that it would be a service to Berkeley scholarship, and to Berkeley, if someone could show that his apparent confusion is apparent only, and that, underlying even Principles, 23, there is some clear grasp of what the object of the imagination is. But I am not optimistic. I don't believe that Berkeley has a clear and consistent view even on the nature and status of ideas of sense;³⁴ and it will hardly be surprising if his attitude to the ideas of the imagination was equally Janus-faced. And, again, the problem of the “aboutness” of thought has been perennially troublesome. It will hardly be surprising if Berkeley was unsound on it. Finally, it will not be surprising if Berkeley's mishandling of the issue led to one argument that continues to puzzle us.

Notes

1. Works, ii. 200.

2. Christopher Peacocke, for example, opens his ‘Imagination, Experience, and Possibility: A Berkeleian View Defended’ (in John Foster and Howard Robinson, ed., Essays on Berkeley, Oxford University Press, 1985) by observing: “Bishop Berkeley had a gift for making clear, crisp statements of doctrines that are also unbelievable. His doctrine that it is impossible to imagine an unperceived tree is often taken as a prime example of this gift. I will argue here that, on the contrary, there is an important sense in which the doctrine is true.” It is significant, however, that Peacocke offers his own argument for the conclusion, and makes it quite clear that he does not accept Berkeley's. And Peacocke supports Berkeley's doctrine only in so far as it concerns the imagination. He does not agree that we cannot conceive of an unperceived tree.

3. As just noted, Peacocke is concerned to support a Berkeleian view, not the Berkeleian argument, which he deals with only in a footnote. That footnote is interesting, however, in that it confirms that there is some haziness over what Berkeley is up to. Thus: “Mackie … regards Berkeley as having misused a self-refutation argument … There is some support for Mackie's reading … But it may be that an even cruder, and totally indefensible, argument was influencing Berkeley …” Textual support is given for both readings.

4. J. O. Urmson, Berkeley (Oxford University Press, 1982), p. 45.

5. Arthur Prior, ‘Berkeley in Logical Form’, Theoria 21 (1955); and J. L. Mackie, Problems from Locke (Oxford University Press, 1976), pp. 52-54.

6. What Mackie brings out rather more clearly than Prior is the tendency towards solipsism. Hence his observation that the line of argument should show that nobody can conceive of anything at all existing “out of his own mind.” Prior is content to point out that the argument would make everything, including minds, dependent on thought. The trouble is that, put in that way, Berkeley might find the supposedly absurd conclusion broadly acceptable.

7. It has even been thought that Berkeley changes tack in Principles, 23. J. O. Wisdom takes it that the crucial argument is contained in the thirteen words he quotes—“you conceive them existing unconceived or unthought of, which is a manifest repugnancy”—and dismisses the point as “entirely specious and not worth a moment's academic discussion.” He does however observe that “Berkeley has shifted his ground, and replaced his doctrine of Esse percipi by Esse concepi.” (J. O. Wisdom, The Unconscious Origin of Berkeley's Philosophy, The Hogarth Press, 1953, pp. 8-9.) Prior, Mackie and Urmson don't actually make that criticism, but their reading of Berkeley would seem to justify it.

8. It could I suppose be argued that the wording here might allow the suggestion to be that we can only conceive an object as existing when perceived or, if not perceived, then thought of. That reading would at any rate be implausible, but a comparison with Principles, 22-23, virtually rules it out.

9. Principles, Introduction, sects. 21-25.

10. W. H. Hay, ‘Berkeley's Argument from Nominalism,’ Revue internationale de philosophie, VII (1953): p. 25.

11. Désirée Park, ‘Prior and Williams on Berkeley,’ Philosophy, 56 (1981).

12. The reference is to Principles, 89.

13. See for example George Pitcher, Berkeley (Routledge and Kegan Paul, 1977), pp. 113-5; my Berkeley: The Philosophy of Immaterialism (Methuen, 1974), pp. 165-66; and (though not in his discussion of the argument we are concerned with) Urmson's Berkeley, pp. 30-31.

14. Bernard Williams, ‘Imagination and the Self,’ a British Academy lecture reprinted in his Problems of the Self (Cambridge University Press, 1973).

15. It does strike me as odd both that Park nowhere hints at any limit to what can be appealed to, and that she refers so readily to what Berkeley has stated in the Dialogues, and indeed elsewhere. For example: “… in Berkeley's account there are no grounds for supposing that a thing which one does not perceive is like a thing which one does perceive. In fact he expressly states that they must be different.” The reference she gives here is to Principles, 8, and entries in the Commentaries.

16. It can't be much. In evaluating Berkeley's argument we are to bear in mind that: “It is simply inconsistent to speak of a Berkeleian idea as if it were absent for the moment, but none the less somehow poised in the wings of future quietly awaiting its cue. In short, it cannot be insisted on too much that an absent idea is not an idea in absentia, but no idea whatever.” Even more clearly, Park tells us that in order to give force to his argument: “[Berkeley] needs to bring into play some of his claims about the relations between minds and ideas … Not least among these relations is the rule [!] that esse is percipi, with its full complement of implications.”

17. The thought-experiment Hylas is challenged to perform here is of course comparable to the one that is central to the argument we are concerned with.

18. Park does in fact refer to p. 194, though she quotes only Philonous's statement that: “… thus much seems manifest, that sensible things are only to be perceived by sense, or represented by the imagination” (see her notes 7 and 21). I am far from clear that conceding this would amount to conceding that any sensible thing is “an idea of the peculiarly Berkeleian sort.”

19. This point is made in relation to the supposed secondary qualities in particular. Philonous notes that in discussing figure and motion it was indeed argued that these are mind-dependent because “it was impossible even in thought to separate them from all secondary qualities, so as to conceive them existing by themselves.” “But then,” Philonous adds, “this was not the only argument made use of upon that occasion.”

20. See his introduction to George Berkeley: Philosophical Works (Dent, 1975), pp. xii-xiii.

21. A. A. Luce, ‘Berkeley's New Principle Completed’ in Warren E. Steinkraus (ed.), New Studies in Berkeley's Philosophy (Holt, Rinehart and Winston, 1966). I examine Luce's use of this tag in my Berkeley, pp. 101-2 and 114-20.

22. Works, ii. 234.

23. See the reference in note 13 above. What we have here is, basically, the “even cruder, and totally indefensible” notion that we found Peacocke suggesting “may” have been influencing Berkeley (note 3, above).

24. Ian Hacking, Why Does Language Matter to Philosophy? (Cambridge University Press, 1975), pp. 17-18.

25. Ibid., p. 41.

26. This formulation is in fact not unproblematic, though no doubt it captures a vagueness in Locke's own mind over whether ideas have “of-ness” built into them. In Essay, IV.xi.1, he urges that “the having the idea of anything in our mind no more proves the existence of that thing, than the picture of a man evidences his being in the world.” It is the analogy that poses the problem. If we take it seriously (more seriously than perhaps Locke intended) we find ourselves wondering, not simply how we can know that there is something corresponding to the picture, but how perceiving the picture can amount to thinking of the thing.

27. This is the term used by Anthony Kenny in his ‘Aquinas: Intentionality,’ in Ted Honderich (ed.), Philosophy Through Its Past (Penguin Books, 1984). Kenny shows how Aquinas tried to account for the sense our thoughts have in terms of the presence to the mind of abstract “forms”; but he sees his attempt to account for reference to individuals as marred by an erroneous understanding of sensory imagination and the role of “phantasmata” in thought. For all Berkeley's attack on the abstract notions of the schoolmen, I don't see that he has a superior understanding of the imagination.

28. Thus, if a philosopher is vague about “ideas” he can take it as just obvious that when I think of John and Jim I must have two different ideas in my mind, and it needn't worry him if John and Jim are identical twins and there happens to be no difference in my images. Again, the champion of abstract ideas or “forms” will hold that universality is “built in” to at least some ideas.

29. Hobbes, of course, wrote the third set of objections to the Meditations. Descartes's observation comes in his reply to the fifth objection there.

30. Hacking, op. cit., p. 40.

31. I have in mind, in particular, Berkeley's account of general ideas. Thus, once it is clearly recognized that an idea is an image, it becomes clear that having an idea in one's mind can't be thinking in general terms of, say, man. For Berkeley, it is the mind that contributes generality by using the idea in a certain way, so that it is “rendered” universal. I don't think that Berkeley is at all clear on how this happens, but obviously this is one context where what I am thinking about cannot be identified with what I am thinking with.

32. Désirée Park, Complementary Notions (Martinus Nijhoff, 1972), p. 34.

33. Note in particular the suggestion that it is the idea we frame that we “call” books or trees, for it does seem that it is only on the assumption that the referent of the word in each case is the image that the argument goes through. The trouble is that the idea we frame should only be a “copy” of a supposed “real” or sensible thing (cf. Principles, 33, for example), and yet Berkeley clearly thinks that he is demonstrating something, not of images, but of things.

34. Thus it is certainly not clear whether a Berkeleian idea of sense is “private” or “public.” It is not particularly controversial to suggest that, in different contests, it is both.

Abbreviations

The present volume follows John Foster and Howard Robinson, eds., Essays on Berkeley: a tercentennial celebration (Oxford: Clarendon Press, 1985), for most of the abbreviations here listed.

Works: The Works of George Berkeley Bishop of Cloyne, ed. A. A. Luce and T. E. Jessop, 9 vols. (Edinburgh: Thomas Nelson, 1948-57). References, except to the works listed below, are by volume and page number.

PC: Philosophical Commentaries (Works, vol. i). References by entry numbers (Notebook B = entries 1-399; Notebook A = entries from 400 onwards).

NTV: An Essay Towards a New Theory of Vision (Works, vol. i). References by section numbers.

TVV: Theory of Vision Vindicated and Explained (Works, vol. i). References by section numbers.

Principles: The Principles of Human Knowledge, Part I (Works, ii). Referred to by section number.

Principles, Intro.: Introduction to The Principles of Human Knowledge (Works, ii). Referred to by section number.

Dialogues: Three Dialogues between Hylas and Philonous (Works, ii). References are to pages numbers.

Alc.: Alciphron or the Minute Philosopher (Works, iii). Referred to by page number.