Berkeley - Arguments for Idealism Tom Stoneham Professor of Philosophy University of York, UK Keywords: Berkeley, Realism, Idealism, Ideas, Objects, Perception, Existence, Secondary Qualities, Mind-dependence, Scientific Realism. Berkeley’s idealism, which he called immaterialism, has two fundamental theses, which we can call the ontological and the metaphysical. (ONT) Everything which exists is either a mind or an object of perception. (MET) Objects of perception exist when, only when, and in virtue of being perceived by some mind. (ONT) has some anti-realist consequences all on its own, ruling out unobserved particulars (e.g. a small rock on a distant planet with no sentient life) and unobservable kinds (e.g. quarks). (MET) is common to Berkeley and Indirect Realists (see below) so does not immediately have anti-realist consequences. But the combination of the two theses is a distinctive and radical view of the world, characterized, or perhaps caricatured, by the consequence that things pop in and out of existence depending on whether they are perceived or not.

© Steven Appleby 2002 (ONT) and (MET) are indeterminate in a few ways. (ONT) does not specify what kind of thing a mind is and what kind or kinds of thing might be objects of perception, while (MET) says nothing about how something could exist ‘in virtue of being perceived’. Berkeley says very little in his published works about what kind of things minds are and we will follow him in that. He calls the objects of perception ‘ideas’ and this leads many to think there is an easy answer to the question of how they could exist in virtue of being perceived: they are mental items, feelings or sensations like pains and tickles. If that is right, immaterialism is even more radical, for it says that everything that exists is mental, that there is no physical world, just minds and what happens to them.

1

But notice that someone who held (ONT) + (MET) and yet thought that the objects of perception were not mental, would be saying something much more amenable to commonsense – so long as they could persuade us that (MET) might be true of those non-mental objects of perception. And this is exactly what Berkeley intended, despite his misleading use of ‘idea’ for the objects of perception: the objects of perception are not states of our own minds but things in the world. It is just that they only exist when perceived. As he put it (PHK 38, see also DHP3 251): If you agree with me that we eat and drink, and are clad with the immediate objects of sense which cannot exist unperceived or without the mind: I shall readily grant it is more proper or conformable to custom, that they should be called things rather than ideas. And more bluntly (DHP3 244): I am not for changing things into ideas, but rather ideas into things. Furthermore, Berkeley was confident that ordinary, non-philosophical folk cared little or none about unperceived objects, and thus that consequences of holding (MET) for the ordinary physical objects we take ourselves to perceive would not be too much in conflict with commonsense (PHK 45, DHP3 249). The Dialectic The opponents of Berkeley’s immaterialism can be divided into three camps. Two have a common assumption, namely that ordinary objects like apples and houses are material: (MAT) Ordinary physical objects (OPOs from now on) can exist unperceived. The Indirect or Representative Realists accept (MET), which combined with (MAT) entails that OPOs are not among the objects of perception. So if such physical objects exist, (ONT) is false. The Direct Realists accept that OPOs are amongst the objects of perception and thus, given (MAT), deny (MET). When considering Berkeley’s arguments, it seems that the Principles is primarily addressed to Indirect Realists and more attention is paid to Direct Realists in the Three Dialogues. But dialectically speaking it looks as if he needs first to argue against the Direct Realist to establish (MET) and then against the Indirect Realist to establish (ONT). Though Berkeley has some things to say about the merits of (ONT) itself, he prefers to argue indirectly: the Indirect Realists’ conjunction of (MET) and (MAT) forces them to deny that OPOs are amongst the objects of perception, which Berkeley takes to be unreasonable. (The conjunction of (MET), (MAT) and (ONT) is worse, for it forces one to deny that there are any ordinary physical objects. That view is sometimes erroneously attributed to Berkeley.) So if (MET) has been established then (MAT) must go, and once we deny (MAT) we are a long way towards establishing (ONT). The third opponent of immaterialism appears at this point in the dialectic. This philosopher accepts the arguments for (MET) and against (MAT) but still denies (ONT), for he thinks that even if ordinary physical objects are not material, they do have material counterparts which cause or occasion our experiences. A version of 2

this view can be found in Berkeley’s contemporaries Malebranche and Norris, but it may be more familiar to contemporary readers as a form of scientific realism: (MET) is true, and thus the table exists only when perceived, but the swarm of particles which physics finds in its place (quite literally: in the location where we take the table to be) exists unperceived.

As the diagram above makes clear, to argue for idealism Berkeley needs at least three distinct arguments: one for (MET), one against (MAT) and one against there being other, unperceivable, matter. In fact, Berkeley offers dozens of arguments and it is not always clear exactly which version of materialism is the intended target of each. Furthermore, there are about as many different interpretations of his arguments as there are interpreters. To avoid scholarly bickering, the arguments I am going to discuss are Berkeleian in spirit and based in the texts, though they go beyond what can be indisputably found there. For (MET) Berkeley does not have a single argument for (MET), but a great variety of arguments which work in many different ways. Some are variants of traditional arguments from 3

illusions and perceptual variations across individuals and species, others seem original to Berkeley, such as his argument that a great heat is indistinguishable from a pain, for which (MET) is obviously true (DHP1 175-8). A so-called ‘Master Argument’ is often attributed to Berkeley (Gallois, 1974), namely that it is impossible to imagine or conceive an unperceived tree (PHK 22-3, DHP1 200). Whatever the merits of that claim about what we can and cannot conceive, it would only support an argument for (MET) with the further premise that what is inconceivable is impossible, and there is no evidence that Berkeley accepted that. What he did accept was the reverse thesis – what is conceivable is possible (e.g. PHK Intro 10, PHK 5) – and thus needed to show that unperceived trees are not conceivable in order to avoid an obvious and decisive objection to (MET) (see Stoneham, 2005, 159-62). His most original and challenging argument is his denial that there is any substantive difference between primary qualities such as shape, size and motion, and secondary qualities such as colour, taste and texture. One way he makes this point is to follow the example of the 17th century sceptic Pierre Bayle and apply the arguments from perceptual variation to the primary as well as the secondary qualities (PHK 15-5, DHP1 188-91). But he has a much more powerful and general argument to the effect that whatever reason you have for thinking (MET) is true of the secondary qualities, you must also think it is true of the primary qualities. Berkeley summarizes his argument thus (PHK 10, see also DHP1 194):1 Now if it be certain, that those [primary] qualities are inseparably united with the other [secondary] sensible qualities, and not, even in thought, capable of being abstracted from them, it plainly follows that they exist only in the mind. All Berkeley's arguments for (MET) rely on the claim that the only true objects of perception are the objects of immediate perception, though ordinary language is careless on this point. The distinction between immediate and mediate perception is not quite the same as the distinction between direct and indirect perception. Berkeley’s thought is that what is immediately perceived can always be experienced in a single perception, that our perceiving it now does not depend upon our perceiving anything else, either now or at some other time. His examples, such as our immediately perceiving a painting or words in a book but not what the painting is of or the words describe, tend to emphasize that what is immediately perceived is present. But it is important to see it is not just being present but being wholly present that matters for immediate perception. Specific instances of colours, textures, shapes, etc. clearly fall in to this category for the presence of such qualities does not depend upon anything which is not given in a single perceptual experience. However, the 1

There is another reading of Berkeley’s argument here which goes: 1. Secondary qualities exist only in virtue of being perceived. 2. So an unperceived object would have only primary qualities. 3. We cannot conceive of an object having primary qualities without secondary qualities. 4. So it is impossible for an object to have only primary qualities. 5. So it is impossible for there to be unperceived objects. While this is an interesting argument, I doubt it is Berkeley’s for the same reason I doubted he used the so-called ‘Master Argument’, namely that he does not accept the principle that what is inconceivable is impossible.

4

properties which correspond to sortal concepts, such as being an apple, do not seem to be immediately perceivable, since to be an apple something must have a complex combination of qualities, thus one can only perceive that something is an apple by perceiving it in a variety of sense modalities and over a period of time: no particular look or smell or taste or feel is sufficient for appleness. Of course, we will sometimes judge there to be an apple in the fruit-bowl on the basis of a single glance, but that is only because of past connections between what we see and the other experiences which go to make up perception of apple-ness. As Berkeley says, if we have previously had these other experiences, the look of an apple will ‘suggest’ the taste and smell, but we do not thereby immediately perceive that taste and smell and thus we do not immediately perceive appleness. Furthermore, while we do not immediately perceive properties such as being an apple, we also do not immediately perceive the object, the apple itself, which has the sensible qualities which we do immediately perceive: we do not have distinct perceptual experiences of physical objects over and above our experiences of their properties (PHK 1, DHP1 174-5). So Berkeley concludes that only a very limited range of things are immediately perceived, and thus properly considered objects of perception, namely ‘light, and colours, and figures, … sounds, … tastes, … odours’ and textures. But this list can include all primary and secondary qualities: what it excludes are objects, if those are taken to be something more than collections or bundles of properties, and sortal properties. Berkeley’s inseparability argument is meant to work against anyone who accepts the mind-dependence of secondary qualities. There has been almost universal agreement among philosophers and scientists since the beginning of the 17th century that there is something subjective about the secondary qualities; unfortunately for Berkeley, this recognition of a subjective aspect to the secondary qualities does not have to take the form of accepting (MET). In particular, some philosophers, including Locke (Essay, II.viii.15), have thought that the connection between possessing a given secondary quality and appearing a particular way can be captured by saying that the secondary qualities are dispositions to cause appearances in suitable observers: to be red, for example, just is to have a certain visual appearance, but red things need not actually have been perceived to be red – it is enough that they would look red. Consequently, (MET) is false of secondary qualities, since the dispositions to cause experiences may exist even if they are and remain unperceived. Berkeley does not have much sympathy for this dispositional account of secondary qualities for a very simple reason (DHP1 187). Dispositions, as opposed to their manifestations, are not immediate objects of perception. We do not see the disposition of the leaf to look green, rather we see a manifestation of that disposition: the leaf actually appears green to us. And if the disposition in question is a disposition to appear a certain way, all manifestations of that disposition exist when, only when and in virtue of being perceived. If we want to identify the greenness of the leaf with the disposition to look a certain way, then we have in effect denied that the secondary quality is a sensible quality, i.e. it is not an object of immediate perception. However, when we go back to the data of experience and consider a case of looking at a leaf, it is undeniable that among the things we see, that is, among the objects of perception, are both the shape of the leaf and its colour. Since (MET) is

5

concerned with objects of perception, it is those actual sensible qualities, rather than any dispositions the object might have, which Berkeley’s argument addresses.2 Some commentators claim that Berkeley only accepted the inseparability claim because he had an imagistic theory of conception, namely that to conceive of something is to form a mental image of it. True, one cannot form a mental image of, say, the shape of a coin, without also imagining the coin to have some colour, so if one thought that all conceiving was imagining, Berkeley’s argument would be effective. Whether or not Berkeley held this rather implausible view of conception, there is another Berkeleian reason to accept inseparability. This reason begins with the point that we cannot perceive primary qualities without perceiving secondary qualities – e.g. you only see the shape of the coin because you see its edges and you only see its edges because of a colour contrast between the coin and the background – and once we understand why this is the case, then we can see that no conception of an object lacking secondary qualities will be a conception of it possessing the very primary qualities which we immediately perceive. I will first present the argument in the abstract and then give a concrete example. When we perceive a property, the property we perceive, be it a shape or a colour, has a qualitative character: there is something it is like for the subject to perceive that property, which is why we call it a sensible quality. A property with no qualitative character cannot be immediately perceived, and thus cannot be a sensible quality. Nothing can have a qualitative character without having some secondary qualities. So if, per impossibile, someone conceives of an object possessing some property but lacking secondary qualities, they conceive of it lacking qualitative character and thus lacking all sensible qualities. Hence no such conception can show the object possessing those very primary qualities we perceive in the absence of secondary qualities. Take, for example, a round coin I can see and feel. Suppose, for the sake of argument, I can conceive (as opposed to imagine) the coin lacking colour but still being round: perhaps I conceive of it rolling down a gentle gradient without conceiving of anyone perceiving it doing so. The question Berkeley will then ask about that conception is whether the property called ‘roundness’, which in my conception it has in virtue of how it rolls when unobserved, is the very same property as the one I see it to have. If it is not, then the conception shows nothing about the separability of the primary from the secondary sensible qualities. In response, I might insist that it is the same property, because I conceive of its unperceived roundness as its having all the points on its edge equidistant from a single point, and that is also true of its perceived roundness. However, this only proves my point if I have conceived of it as having an edge without conceiving of its secondary qualities and to do that we would have to conceive of the boundary between coin and non-coin in terms of a difference in some primary property. But a perceived edge is always and necessarily marked by a difference in secondary qualities and hence no difference in primary properties alone is sufficient for the existence of a perceivable edge. Of course, some differences in primary properties are sufficient for there to be a 2

It is worth noting here that Berkeley’s argument at this point does not need the strong thesis that we can never perceive dispositions, though I think he would have accepted that, but merely that at least sometimes when we immediately perceive secondary qualities, we are not perceiving dispositions but their manifestations.

6

boundary between the coin and the non-coin, but such boundaries are only perceivable in virtue of some difference in secondary qualities. So if I have successfully conceived of the coin as being round without conceiving of it as having any secondary qualities, I have conceived of it as having a property distinct from any property I can perceive. Hence I have not shown that I can separate the primary sensible qualities from the secondary. Berkeley’s point is that when we try to conceive of an object possessing primary qualities and no secondary qualities, the most we could achieve is a conception of it as possessing some properties and no secondary properties, but those properties are not ‘the very figure and extension which you perceive by sense’ (DHP1 188, my emphasis). And because of this inseparability of primary and secondary sensible qualities, whatever persuades you that (MET) is true of the secondary qualities will thereby require you to think (MET) is true of the sensible primary qualities as well. All we perceive are sensible qualities and it is in the nature or essence of those sensible qualities to be perceived, hence ‘their esse is percipi’ (PHK 3). There are two ways one might object to this argument. First, one might argue that we can separate the cognitive from the sensuous elements of perception and the former does not depend upon the latter. Thus, even though whenever we do perceive something as, say, square, we also perceive it as having a certain qualitative character, that qualitative character is not constitutive of our perceiving it as square. Berkeley would say that to argue thus confuses perception, which is passive and involuntary, with thought or judgement, which is active (e.g. PC 286). Secondly, one might argue that some secondary qualities are not in fact mind-dependent. Perhaps such a view can be defended, but it will not be easy. Against (MAT) Berkeley gives a very blunt and direct argument against (MAT) at the beginning of the Principles: But with how great an assurance and acquiescence soever this principle [MAT] may be entertained in the world; yet whoever shall find in his heart to call it in question, may, if I mistake not, perceive it to involve a manifest contradiction. For what are the forementioned [ordinary physical] objects but the things we perceive by sense, and what do we perceive besides our own ideas or sensations; and is it not plainly repugnant that any one of these or any combination of them should exist unperceived? (PHK 4) Unfortunately, this argument looks like a classic equivocation (Stoneham, 2003). The premises are: 1] Ordinary physical objects are amongst the things we perceive by sense. 2] All the things we perceive by sense exist only when perceived. The problem is that, as we have seen, the reasons Berkeley gives for accepting 2] require us to restrict ‘the things we perceive by sense’ to the immediately perceived sensible qualities, but the reasons we have for 1] do not hold with this restriction in place: ordinary physical objects do not appear to be amongst the immediately 7

perceived sensible qualities, for those are only colours, shapes, textures, sounds, odours, etc. Berkeley’s solution is to say that the ordinary physical objects consist in collections of sensible qualities, thus the argument does not equivocate. But the Indirect Realist has a different approach, which Berkeley needs to rule out first. According to the Indirect Realist, ‘the things we perceive by sense’ picks out two classes of object: the sensible qualities (which exist only when perceived) and the physical objects. She then claims that we perceive the physical objects by perceiving the sensible qualities, so objects of each kind are perceived, but the relation is different in each case. If sensible qualities are a different class of objects from physical objects and we perceive the latter by perceiving the former, there must be some relation between the two types of object which makes this possible. Berkeley considers two: the sensible qualities represent the physical objects, and the sensible qualities inhere in the physical objects. Either would explain how we perceive one indirectly by perceiving the other directly. Against the first, Berkeley makes the important point that nothing can resemble an idea but an idea. Of course, there are other sorts of representation than resemblance, but the position Berkeley is objecting to here accepts that sensible qualities are minddependent objects of perception and that by perceiving them we perceive something else, something material and not otherwise perceivable. So this is a case of perceiving something indirectly by perceiving something else. The relation which makes this possible cannot be a matter of convention, nor can it be one which requires us to experience both relata to know that it holds, such as a causal relation. So it does seem that resemblance is the best candidate. But it will not do because only things which are themselves perceivable can resemble each other, and according to the Indirect Realist, the material objects are not themselves perceivable. As Berkeley puts it (DHP1 206): But how can that which is sensible be like that which is insensible? Can a real thing in itself invisible be like a colour; or a real thing which is not audible, be like a sound? In a word, can any thing be like a sensation or idea but another sensation or idea? Against the second proposed relation, Berkeley offers three arguments. One is that he simply does not understand inherence (DHP1 190, DHP2 234). The second is that, whatever the inherence relation is, it implies an existential dependence, but if we accept (MET), sensible qualities existentially depend upon being perceived, so they cannot inhere in unperceiving matter (PHK 7 and 76, DHP1 197). The third is most general: if we perceive material things by perceiving sensible qualities, we must thereby gain some sort of idea of those material things. But what sort of idea could that be? It could not be the sort of idea found in direct perceptions, so it must be a ‘relative idea’, i.e. our idea of material objects is as the thing related in such a way to the sensible qualities (PHK 16). Relative ideas are not uncommon: most readers of this chapter will only have a relative idea of me, namely as the author of this chapter. But there is a plausible necessary condition upon having a relative idea: one must have a grasp of the relation holding between things of which one has non-relative ideas. This creates two problems for the Indirect Realist. One is that we are supposed 8

to have a relative idea of matter even though we cannot have a non-relative idea. The other is that it is self-defeating to explain the relation of inherence in terms of relations we have experienced because (MET) applies to them (DHP1 198). So the Indirect Realist’s claim that ‘things we perceive by sense’ equivocates between direct and indirect objects of perception cannot be sustained (see also Hume, Treatise, I, iv, 2 paragraph 4 for a phenomenological objection). However, if OPOs are just bundles or collections of sensible qualities, then it can be true both that we only immediately perceive sensible qualities and we immediately perceive OPOs. Against other material things We now need to consider how Berkeley argues against someone who accepts his arguments for (MET) and against (MAT), but who thinks that this does not prove that there is no matter, merely that material objects are unperceivable. One form of this view is the occasionalism of Malebranche and Norris, against which Berkeley has a panoply of arguments turning on the fact that the matter they postulate is unknowable, does nothing and would have been pointless for God to create (DHP2 passim). But a more plausible form of the view is that through science we come to know of the unperceived material world: science shows (ONT) is false. We can call this Scientific Realism. One way of reaching this conclusion is from our experience with scientific instruments like microscopes and telescopes. These show us that as we improve our perceptual acuity, we discover previously unperceivable features of the world. Since there are no non-contingent upper bounds to acuity, this implies that there could always be more to the world than we can perceive, so (ONT) is implausible. Furthermore, microscopes do not just show us new things, as looking over a mountain might, but show us what appear to be the otherwise imperceptible inner workings of things we can perceive (PHK 60). Berkeley’s response is that microscopes do just show us new things and there is no prior reason to think they will have any connection with the ordinary objects of experience. Where we do find such a connection, it is entirely contingent. Thus when we look at some object through a microscope, it is not necessary that we will see anything, let alone something which will help us understand the behaviour of that object. Of course, that we do see such things makes the working of nature more lawlike and easier to predict which, if we believe in such a thing, we might count as another instance of Divine Providence (PHK 62, DHP3 245). Another way of denying (ONT) on the basis of science is via the positing of theoretical objects. According to this view, ordinary physical objects like tables and trees consist of sensible qualities which do not exist unperceived. However, there are other objects, the objects of scientific theory, which lack sensible qualities but do exist unperceived. We know about these purely theoretical objects because the theories in which they figure are successful in predicting what happens to perceptible objects. The first problem for this view comes when we ask where these imperceptible material things are located. It is tempting to answer that they are in the same place as 9

the ordinary physical objects whose behaviour they explain. Thus I see a table in front of me, what I see is just a collection of sensible qualities, but in the very place I see that table, there is a collection of unperceivable, material things, perhaps a swarm of atoms. But this will not do, for the spatial properties of the table are among the sensible qualities of the table, and thus cannot be possessed by the unperceivable matter which the Scientific Realist introduces. So if there is this matter, it is not merely unperceivable, it is also not spatially related to anything we can perceive (PHK 67): But secondly, though we should grant this unknown substance may possibly exist, yet where can it be supposed to be? That it exists not in the mind is agreed, and that it exists not in place is no less certain; since all extension exists only in the mind, as hath been already proved. It remains therefore that it exists no where at all. This point is in fact devastating for the Scientific Realist, for science postulates unperceivable objects to explain perceivable phenomena and the explanations require those theoretical objects to be spatially related to the phenomena. If nothing unperceivable can be spatially related to something perceivable, then science cannot be giving us reason to believe in unperceivable material objects. In fact, if we accept (MET) and include spatial relations in the objects of perception, the only consistent interpretation of these scientific theories which postulate unperceivable objects is instrumentalist: the theory and its postulates are just a tool we use to predict and explain phenomena and thus we are not committed to the existence of those postulates, only of the phenomena they predict. Here Berkeley’s position relies very heavily on the objection to the primary-secondary quality distinction, for if the spatial properties we perceive are such as might exist unperceived, then they could be possessed by theoretical entities. Conclusion A complete defence of Berkeleian idealism would have two parts. First, one would have to argue against the existence of matter. Then one would have to show that the denial of matter does not bring with it any insuperable philosophical problems. In this chapter I have tried to make some of Berkeley's arguments against matter as plausible as possible. Of course those arguments are not watertight, but objecting to them incurs costs elsewhere in one's metaphysics. Only in the light of a consideration of whether Berkeley's denial of matter is itself a cogent metaphysics can we properly assess the true merits of immaterialism.3

3

I would like to thank Steven Appleby for his kind permission to reproduce part of his cartoon 'The Science of Philosophy' which first appeared in the Sunday Telegraph on 17th November 2002.

10

References: Berkeley, G. 1707-8: Philosophical Commentaries. In Jessop & Luce 1948. – 1710: A Treatise Concerning the Principles of Human Kowledge. In Jessop & Luce 1949. – 1713: Three Dialogues between Hylas and Philonous. In Jessop & Luce 1949. Gallois, A. 1974: Berkeley's Master Argument. The Philosophical Review 83, 55-69. Hume, D. 1738: A Treatise of Human Nature, Book 1. Jessop, T and A. Luce 1948: The Works of George Berkeley, Volume 1. London: Thomas Nelson. Jessop, T and A. Luce 1949: The Works of George Berkeley, Volume 2. London, Thomas Nelson. Locke, J. 1690: An Essay Concerning Human Understanding. Stoneham, T. 2003: On Equivocation. Philosophy 78, 515-9. – 2005: Berkeley's Principles of Human Knowledge. Central Works of Philosophy, vol. 2, ed. J.Shand, Chesham: Acumen, pp. 137-65. Note: There are many different editions of Berkeley's main philosophical works and they differ only in the finest details. I have used the following conventions to refer to Berkeley's writings: PC 154 – Philosophical Commentaries (1707-8), entry 154 PHK 89 – Principles of Human Knowledge (1710), section 89 DHP2 216 – Three Dialogues between Hylas and Philonous (1713), 2nd dialogue, page 216 of the Jessop & Luce edition (these page references are included in many recent editions). Further Reading: (1) Selected interpretations of Berkeley: Cummins, R. 1975. Berkeley's Ideas of Sense. Noûs 9, 55-72. – Defends Thomas Reid's interpretation of Berkeley's ideas as sensations. Dancy J. 1987: Berkeley: An Introduction. Oxford: Blackwell. – Interprets Berkeley as making heavy use of the principle that what is inconceivable is impossible. Fogelin, R. 2001: Berkeley and the Principles of Human Knowledge. London: Routledge. – Interprets Berkeley as taking (MET) to be intuitively obvious Pitcher, G. 1977: Berkeley. London: Routledge & Kegan Paul. – A thorough and systematic, if slightly unsympathetic, account of all aspects of Berkeley's thought. Roberts, J. 2007: A Metaphysics for the Mob. Oxford: OUP. – An original interpretation of Berkeley as a 'spiritual realist' Stoneham, T 2002: Berkeley's World. Oxford: OUP. – Part 3 tries to show that idealism can be developed into an cogent metaphysics. Winkler, K. 1994: Berkeley: An Interpretation. Oxford: OUP. – A sympathetic account with detailed discussions of other scholars.

11

(2) Other works relevant to the arguments discussed here: Allen, K. 2007: The Mind-Independence of Colour. European Journal of Philosophy 15, 137-58. – Defends the view that colours are mind-independent, which would undermine the inseparability argument. Austin, J. 1963: Sense and Sensibilia. Oxford: OUP. – Argues against (MET) and the restriction of the immediate objects of perception to qualities. Ayer, A. 1973: The Central Questions of Philosophy. London: Weidenfeld and Nicholson. – Chapter 3 responds to Austin and chapter 4 attempts to give a non-idealist 'construction' of the world. Berkeley, G. 1709: An Essay Towards a New Theory of Vision. In Jessop & Luce 1948. – Many of Berkeley's views about the immediate objects of perception, and vision in particular, are developed in this book. Campbell, J. 1993: A Simple View of Colour. In Haldane, J. and C. Wright eds. 1993: Reality, Representation and Projection. Oxford: OUP. – Distinguishes between objectivity and mind-independence and applies the distinction to colours. Campbell, J. 2002: Berkeley's Puzzle. In Gendler T, and J. Hawthorne eds. Conceivability and Possibility, Oxford: OUP. – Presents Berkeley as offering a very serious challenge to realism. Dummett, M. 1979: Common Sense and Physics. In Macdonald G. ed. 1979: Perception and Identity, London: Macmillan. – A sophisticated discussion of Scientific Realism. Section III on what it is to be immediately perceived. McGinn, C. 1983: The Subjective View. Oxford: OUP. – Discusses the relation between the inseparability thesis and the imagistic account of conception (see pp.80ff). McDowell, J. 1994: Mind and World. Cambridge, Ma.: Harvard UP. – Tries to reconcile the passivity of perception with its being cognitive. Strawson, P. 1979: Perception and its Objects. In Macdonald G. ed. 1979: Perception and Identity, London: Macmillan. – Offers a form of empirical realism strongly influenced by Kant.

12