Copyright © 1996, Allan Randall


The Rejection of Abstract Ideas in the Metaphysics of George Berkeley

Allan F. Randall
Toronto, Ontario, Canada,


An overview of Berkeley's subjective idealism, and its relation to his rejection of abstract ideas. I argue that Berkeley's idealism goes wrong largely due to the assumption that ideas must be passive, while spirit is active. If one accepts that abstract ideas can be active entities--that is, the idea itself as the thing the thought is about, not just the mind of the thinker--then there is no reason to reject abstraction. Berkeley's subjective idealism is, in addition, not even a consistent system, since his entire argument is based on the incoherency of matter, but he requires something equally coherent called spirit to support his ideas. His argument can only be consistent if taken all the way, and spirit is rejected along with matter, i.e., if his idealism is absolute.

I. Introduction

George Berkeley's main philosophical thesis is, on the face of it,quite remarkable: there is no such thing as matter [Berkeley 1710].This seems startling, even bizarre, to some. But Berkeley saw his thesis as aninevitable consequence of the philosophical tradition he inherited, largelyfrom Descartes and Locke. Berkeley's anti-materialism can be seen as a directconsequence of Locke's position on substance. Substance, for Locke, was someunseen, unfelt stuff in which the qualities we perceive around usinhere. But we can never know pure substance directly, without thequalities. All we know of a substance are the ideas of qualities. So how, askedLocke, can we have any idea at all of what pure substance is, when we can nevereven in principle see it, feel it, or otherwise experience it? All we can knowdirectly in our minds are ideas. As Descartes himself pointed out, there isnothing intrinsic about our ideas that tells us they must be caused bycorresponding physical, material objects external to us. An evil daemon couldbe inducing hallucinations in us, and the entire physical world might literallynot exist [Descartes 1637, pp. 22-23]. Pure substance is something we cannot bydefinition know directly, since it is fundamentally a different sort of thingthan an idea. So we can never really know anything about it at all [Locke 1689,I.4.18, II.23.1-3].

 So why even suggest that it exists? Locke thought that it had tosomehow exist, but he was extremely uncomfortable with this position[Locke 1689, II.12.4-6, II.23.15-29]. He recognized that there is a realproblem with insisting on an entity that is completely unknowable andindescribable; yet he insisted that qualities surely could not exist out therein the world all by themselves. They needed something to support them.Otherwise, there could be no structure to things. All there would be is a bigsoup of qualities. If the world is to be ordered, some qualities need to begrouped together and other left out. An apple somehow has its qualitiesof redness, roundness, et cetera. The table it is resting on has itsqualities of hardness, flatness, et cetera. Without this grouping of somequalities together, while leaving others out, there can be no order in theworld--no things at all. No substances. Yet, this grouping seems torequire a substratum to support the grouping. Something must be there tohave some of the qualities and not the others. The qualities cannot justbe grouped together. An idea in the mind of such a collection of ideasof qualities is an abstraction. Such abstract ideas are central to thephilosophies of both Descartes and Locke.

 Berkeley, recognizing Locke's pain in being unable to find a truly consistentconception of substance, decided that he could dispense with it--or at leastwith the idea of material substance. One might think that the obviousway to go about this would be to reject Locke's assumption that qualitiescannot exist on their own. But Berkeley did not question this assumption, andagreeing with Locke on this point [Berkeley 1710, sect. 91], he calledinto question instead another Lockian/Cartesian assumption: the very notion ofan abstract idea [Berkeley 1710, Intro.]. According to Berkeley, Descartes andLocke were wrong--there is no such thing as an abstract idea after all.Ideas are passive objects that we observe and perform mental operations on.Such a thing cannot possibly be abstract, but is always a particular. ThusBerkeley reasoned that the mind had to be fundamentally different than idea,and he split reality into two kinds of things: ideas and spiritual (notmaterial) substance. Only spirits can operate and be active. Spirit isprocess, ideas are objects. Spirit acts on ideas; ideas cannotact.

 Since pure substance out there in the world cannot be material (since this wecan have no idea of), it must be spirit. But it cannot be our spirit,for we do not cause our own ideas--those of sensation, at least, seem to bethrust upon us. Therefore, our ideas of sensation must be caused by a greaterspirit [Berkeley 1710, sect. 6, 72]. It is a short step from this to thenotion that God is causing all our ideas of sensation, leaving no room formaterial objects. You see a chair in front of you? God put that idea in yourmind directly. It was not a material chair that did so. If yourideas of the chair seem to have structure and order, that is because God iswise and has ordered and structured your ideas in a special way.

 One of the biggest problems with this whole system, of course, is Berkeley'sinsistence on spirit as a real substance. Every single argument he mustersagainst matter can be turned back on him and used against spirit. I will arguethat Berkeley is every bit as inconsistent as the materialists he scorns. Forhe has the idea of substance in his system--as spirit rather thanmatter. He tries to get around this by saying that he can have a notionof spirit even if not a true idea [Berkeley 1710, sect. 27], but thisis the same sort of mental gymnastics that Locke tried in order to keepsubstance around. Berkeley becomes his own worst enemy, exemplifying in hisargument for spirit what he rejects in other's arguments for matter. I willargue that the idea of an abstract idea is actually quite sound. The road to aconsistent idealism, if there is one, must lie, not in rejecting abstractideas, but in rejecting the assumption that ideas, or the qualities that causethem (if they be any different), cannot exist on their own. Indeed, afterBerkeley, absolute idealists such as Hegel and Bradley did just this.

II. Ideas

Berkeley actually talks about two kind of abstraction [Berkeley 1710, Intro. sect. 10], only one of which he rejects. Abstraction consists inplucking out certain features or qualities from reality and leaving the restbehind. In some cases this leaves us with an idea of a collection ofqualities that could feasibly exist on their own (without other ideas, althoughnot, in Berkeley's view, without a mind to think of them). For instance, I seean apple on the table. In order to have an idea of that apple, I need toabstract it out of the world--taking only the apple qualities andleaving behind all the other qualities I see before me, like the table theapple sits on, the floor, et cetera. But the resulting idea, an idea of anapple, is something I could imagine existing on its own. Of course, I neveractually do experience a "Pure Apple Sensation". Try to imagine such athing. You sense an apple, but no table on which it rests, no hand that holdsit, no background colours, no floor, nothing but apple. It seemsunlikely that you will ever experience this apple idea without other qualitiesaround, but it is possible to have such an idea. One couldimagine having an idea of this particular apple without any other ideasattached. It is possible. The fact that the apple sensation idea is always inpractise accompanied by other ideas is not what is important. The apple ideacan exist without what it was abstracted from, so the abstraction is valid--itis a real idea. These valid abstractions we would often not want to even call"abstract". Since the abstracted idea can exist on its own, we just think of itas an idea of a particular thing. We abstract the apple from the realworld, but this gives us an idea which, on its own terms, is particular. Theidea itself, although obtained through abstraction, is not inherently abstract.It is actually a particular idea. In this case, it is a sort of mental image ofan apple with a particular shape, color, et cetera.

 But at other times, an abstraction will give us something that could notpossibly exist on its own. For instance, my concept "dog" is abstractedfrom all the sense ideas of particular dogs I have had in my life. Each ofthese sense ideas, although abstracted from the rest of my sensation at thetime, is a particular idea in its own right. I have grouped these ideas ofparticular dogs together (leaving out my ideas of hippopotami and cats) intothe concept "dog". But, according to Descartes, I have done even more thanthat: I have generalized this collection of particulars, so that it includesdogs I have never actually seen nor imagined. In fact, it includes alldogs and all possible dogs. This, Berkeley says, is impossible. Yes, I canabstract the idea of one dog, since the idea can stand on its own. Yes, I canuse the particular ideas of dogs in a similar fashion, thus effectivelytreating them as a collection of ideas. But I cannot possibly claim to have anidea of all possible dogs. Is my mind infinite? Does it containliterally an infinity of ideas of particular dogs, all of which are possible?That is surely untenable. So what is this idea of dog-in-general? It cannot bean image, like my idea of my pet poodle Woomsy--a picture in my mind. For topicture dog in my mind, surely I can only picture a dog, or atbest I can picture some kind of strange thing that is a conglomeration of partsof various dogs I have seen. But this is not an idea of dog-in-general, it is amixture of ideas, each of which is of a particular dog [Berkeley 1710, Intro. sect. 9-11].

 Ideas, to Berkeley's way of thinking, are static. A visual idea is a mentalimage--a picture in the mind. Usually when I say that I have an idea ofdog-in-general, I use an idea of a particular dog as a stand-in for all thedogs I have ever seen. Even when I try to think beyond that, I still only thinkabout various particular dogs. I never have a mental image of dog-in-general.That is impossible. My so-called abstract idea "dog" is only universalin that it has some relation to the rest of the particular dogs I havein my memory [Berkeley 1710, sect. 16]. So abstract ideas are nonentities.They are not actual ideas that we have, but incoherent descriptions of ideas wethink we have [Berkeley 1710, sect. 20-22]. Our system of ideas consists ofparticular ideas (abstracted from their original context, but able to existapart from that context), and relations between them.

III. Spirit

So ideas are particular and passive. They are related in various ways,and these relations make us think we have general abstract ideas. So where dothese relations come from? Is Berkeley, in allowing these relations, admittingabstraction in the side door? No, says Berkeley, because the particular ideasget related to one another by the mind, which performs mental operationson the ideas: it compares and contrasts them, combines them, takes themapart and recombines them, and, most fundamental of all, it perceivesthem. All of this process is of a completely different nature thanthe ideas themselves which are completely devoid of process, and are simplypassive objects. Spirit perceives and manipulates ideas--it is active.Ideas are perceived and manipulated--they are passive.

 Furthermore, we as spirits only perceive ideas. There is nothing else we everexperience. Therefore, the only idea we can ever have of anything existingis that of ideas. Thus, "to be is to be perceived." But what about spiritor mind; does it not also exist? Berkeley certainly thinks so, since the veryexistence of ideas implies the existence of spirit. Ideas(that-which-can-be-perceived) cannot exist without something to perceive them:a spirit (that-which-can-perceive). So we really need to say "to be is to beperceived or to perceive".

IV. Matter or God?

One of Berkeley's biggest problems is that most, if not all, of hisarguments against matter can be turned right around and levelled againstspirit. True, we have no idea whatsoever of matter. But, as Locke pointed out,we also have no idea whatsoever of spirit. Matter and spirit could turn out tobe equivalent, for all we know, since they are both equally unknowable [Locke1689, II.23.30-31].

 Berkeley, using Locke's basic argument against the idea of substance, makes aconvincing case against matter. If all we can perceive are ideas, any knowledgeof an external material substance apart from the idea would have to becompletely in terms of ideas. So all we are left with are more ideas. There isno way to know something in terms of ideas, which are the only terms we have,unless that thing itself consists of ideas. But Locke's material substance isthe substratum that supports the qualities from which we get our ideas. Somatter cannot be ideas, it is, by its very definition, something apartfrom ideas. So we can have no idea of it. It is a nonentity. Berkeley uses verystrong language in putting down the intellectual competence of anyone who woulddare defend such an inconsistent notion. When we say "matter" we have no ideaof what we are speaking.

 But Berkeley recognizes that some ideas, although they are passive asperceived, seem to be thrust upon us quite beyond our control. When I see anapple on the table, I cannot help what I see--the sensation of apple/table isjust there in my mind. I did not will it. It must have been put there bysomething outside my mind, else how did it get there? It could not have comefrom matter, which Berkeley has already shown to be incoherent. So it must havecome from another perceiver or spirit, greater than myself. God, then has putthe idea of apple in my mind. If the ideas seem complex and structured, it isonly because God is wise and puts ordered ideas in my mind for my own good[Berkeley 1710, sect. 31-33].

 One of the main difficulties here concerns why God is an appropriateexplanation for the ideas that are thrust upon us, and matter is not. Infact, could I not say that this thing Berkeley calls "God" is simply thethings physical scientists call matter along with physical law? Berkeleypresumably would say something like this: "Only spirit is active, ideas arenot. Matter is also passive, since it is supposedly a substratum. Whateverplaces sense ideas in our minds must be active, so it must be another mind orspirit. A passive thing that does nothing but sit there and be a holding tankfor qualities cannot actively place ideas in our minds. Only a spirit--aperceiver and manipulator of ideas--can place ideas in our minds."

 However, the materialist view of things is not one of just passivematter. There are also physical laws that state how the matter can act.These physical material laws of causation are as unknowable as passivematter, to be sure, but they are part and parcel of the standard materialistview of the world, and it is not clear how Berkeley's God differs in any way,other than in His personality, which is really not required by any of thearguments Berkeley actually advances. "But," Berkeley might respond, "actionimplies by its very nature, a will, which is necessarily mental. Yourphysical laws, then, require a spirit, which must be God, and that is all thatis needed to generate sense ideas in our minds. Matter becomes extraneous anduseless."

 So we have seen that there are two fundamental components to Berkeley'sthought. If we reject either one, we need not accept his version of idealism:

 (1) There is a basic distinction between action and idea.
(2) Action is necessarily mental in nature.

V. Abstract Ideas Reconsidered

Berkeley has divided the world into two spheres, in the tradition ofDescartes. Like Descartes, he maintains spirit. But, going against Descartes,he has replaced matter (or extension) with ideas alone. To be as consistent ashe would like to be, and a "pure" idealist, he really needs to get rid ofspirit as well, so that only ideas exist (or perhaps qualities, although inthis extreme brand of idealism, the distinction usually just gets erased).There would then be nothing existent that we could not in principle have anidea of. Indeed, this is the step the Absolute Idealists, such as Hegel,Schelling and Bradley [Bradley 1994], would later take. In fact, this viewpredates even Descartes. It is the view held by Parmenides of Elea, who gotmetaphysics started 2500 years ago [Parmenides c475 BC].

 But Berkeley does not take such a step, and the whole reason is that hebelieves it impossible for an idea to act. If an idea could act, then itcould be the thing that unites various ideas of particular dogs--i.e.,it would be an abstract idea. Berkeley's position against action asidea, then, leads directly to his position against abstract ideas. Such anabstract idea would have to include process or action as anintegral part of what it is, and Berkeley thinks ideas can only be passive.

 So is Berkeley actually right that ideas cannot act, and therefore cannot beabstract? His argument has a lot of force. It is true that, try as we might, wecannot get an image of dog-in-general into our heads. But, we say to ourselves,I can use these mental images of particular dogs in such a way thatshows an understanding of dog-in-general, so I must have the generalidea, too. But Berkeley says no, you don't have this ability to manipulate thedog-images as an idea, you have it as part of your spirit--your set ofmental abilities. The same goes for your so-called "idea" oftriangle-in-general.

 But Descartes, from whom Berkeley supposedly gets his idea of "idea" in thefirst place, disagrees with this. My idea of a piece of warm wax, for instance,is such that I understand that it can take any of an infinite variety of shapes[Descartes 1637, pp. 30-32]. My idea of triangle-in-general is such that Iunderstand the process of generating all possible triangles, even if I neveractually do it. I understand the process and therefore I have an idea ofit. An idea need not be a mental image the way Berkeley views it. It canbe active. I can imagine a 1000-sided polygon without actually seeing onewith 1000 sides in my mind [Descartes 1637, p. 73]. So what am I understanding, ifnot an image? I am understanding a process or procedure, an algorithm, forgenerating the 1000-sided polygon, which is also related to my idea of ageneral process for creating regular shapes of 3 sides, 4 sides or N sides--inother words, my general abstract idea of polygon. These are not images,they are procedures; in twentieth century terms, they are like computerprograms running inside my head. I have an idea, an understanding, of them inmy mind, although this idea is not an image, nor is it passive.

 If this is so, and I can understand abstract ideas directly in my mind, thenan idea can be active and Berkeley's whole system loses much of itsbite. Spirit need no longer be a separate kind of thing from idea. Now bothidea and spirit can be active. Spirit, then, can be a kind of idea (or at leastthe same kind of thing as an idea--perhaps both are kinds of "qualities"). Bothare active-stuff. There is no need to separate the two, and we are on the roadto Absolute Idealism, where everything is Idea.

 So why does Berkeley insist that ideas are passive, when Descartes himselfalready had a cogent argument against it? Berkeley's entire world view wasextremely religious. He viewed his entire philosophy as an attempt to vindicatescripture and the Christian God. He saw matter as a concept used by Atheists tomake the world into a purely mechanical place--a huge machine. So Berkeley tookmechanism--i.e., activity, process, abstraction--and removed it from bothmatter and ideas and placed it in spirit. This way, the ultimate Spirit,God, is the explanation for the universe, not matter or ideas, both of whichwould make God nonfundamental.

VI. Ideas On Their Own

Personally, I see no reason to remove process from ideas. We can accept allof Berkeley's arguments against matter, and still maintain quiteconsistently that ideas are a kind of process. In fact, we will be moreconsistent than Berkeley, since we no longer need Spirit to operate onideas--the ideas can be operated on by a mind that itself consists only ofideas (or qualities, or algorithms, or whatever). Berkeley would protest thatideas need a perceiver, but we can look all the way back to Plato to see thatnot everyone thinks so. Plato's Ideas had their own independent existence,apart from any mind that perceived them. Why does Berkeley assume that hisideas contain somehow the notion of a perceiver within them? I have an idea ofapple, and yes, I am currently perceiving this idea, but so long as I can atleast in principle completely describe the contents of this idea withoutdescribing my mind, why does it need my mind in order to exist? True, I cannotdispense with process, mechanism, abstraction, computation or some suchnotion, but I do not see why these things require mind. Is it not atleast possible for the idea to exist unperceived? Assuming so leads usto an "objective" idealism like that of Plato or Bradley, as opposed toBerkeley's "subjective idealism".

 Today, in the late twentieth century, computationalism holds sway asthe most popular theory of mind. Under this view, the mind is simply anemergent property of a computational process. While many today think of this asa materialist position, it does not need to be. If we accept a Platonic view ofideas as nonmental: i.e., "what can be thought" rather than "the thinkingitself" and we accept them as processes, then the mind and its ideas canboth consist of computational processes. Computation, in fact, isprocess, so this view would reduce everything to computation. Everythingcan be described as a computer program. The abstract idea "triangle-in-general"would be a computer program capable of generating any possible triangle. Themental version of this idea would be a triangle-program in our minds.The nonmental version would be a purely mathematical description of theprogram, or the program instantiated in physical computer hardware, or anyother instantiation that is not mental. Either way, the idea is a process, andwhen in the mind, it is the kind of active idea that Berkeley thought could notexist.

 Some absolute idealists and twentieth century computationalists [Barrow 1986]have developed idealism in just this way. Bradley [Bradley 1994] is anexcellent example. He was adamant that all ideas necessarily were activeprocesses [Randall 1996]. In fact, anything that was not a process, that wassupposedly "passive", could not really be said to exist on its own. For anypassive thing, such as a relational structure, requires process, which as wehave seen is just another word for abstraction, to unite the items together. Soqualities in the world cannot be united together by something passivelike matter. In that, Berkeley is right. They must be united by somethingactive. But if Bradley and the modern computationalists are right, ideasalready have this necessary ingredient. All ideas are abstractions, notparticulars. There is no longer any necessity to postulate spirit as an activeunifier of passive ideas--ideas can do the job all on their own.

VII. Conclusion

To be consistent, Berkeley really cannot hold both that matter is anunacceptable nonidea, and that spirit is acceptable as anundefinable "notion". In this, he is falling prey to the same inconsistency asLocke. At the same time, he is right to see that any passive structure requiresprocess to make it a something. However, I do not think he hasadequately defended his contention that this process cannot be an idea. Infact, one could argue that this process itself is what thinking is all about,and that all ideas have this feature. Even Berkeley's "passive" mental imagesare not really passive. Some mental process has extracted lines and shapes andother forms and put them together in a certain way. That is what yousee, not some static image. Idea is process. Descartes saw this basicpoint, more or less, long before Berkeley wrote his treatise.

 The thing-perceived and the thing-perceiving cannot be separated out as twofundamentally different kinds of thing, as both Descartes and Berkeley have it,without intractable problems. Descartes is left with the problem of interactingmatter and spirit. Berkeley is left with the problem of interacting idea andspirit. Only by making both the same kind of thing, and erasing the distinctionas something primary and fundamental, can we get rid of the inconsistencies. Ifwe take the view of computational absolute idealism (which has much of itsroots in Berkeley's thinking, in spite of its points of difference), we canmodify Berkeley's slogan and say instead: "To be is to be perceivable." Thisapplies to both the idea and its perceiver. They are both the kinds ofthing that can be perceived. In Berkeley's view, this is absurd. A spiritcannot, even in principle, be perceived. But in the computationalist view,spirit is just a computer program in the brain, which can, at least inprinciple, be picked apart and its component programs understood as active entities--as abstraction, process, computation or mechanism--in other words, as Ideas.


Barrow, John D. and Frank J. Tipler (1986). "Teleological Ideas inAbsolute Idealism," In: The Anthropic Cosmological Principle, pp.153-159. Oxford U. Press, Oxford, 1986.

Berkeley, George (1710). A Treatise concerning the Principles ofHuman Knowledge. In: A Treatise concerning the Principles of HumanKnowledge. Three Dialogues between Hylas and Philonous, G.J. Warnock (Ed.),pp. 43-146. Open Court, La Salle, Illinois, 1710, 1713, 1962.

Bradley, F.H. (1994). Writings on Logic and Metaphysics.Clarendon Press, Oxford, 1883, 1893, 1897, 1914, 1922, 1935, 1994.

Descartes, René (1637). Meditations on First Philosophy.In: Discourse on Method and Meditations on First Philosophy (3rd Ed.),pp. 46-105. Hackett, Indianapolis, 1637, 1641, 1993.

Locke, John (1689). An Essay concerning Human Understanding (5thEd.). Clarendon Press, Oxford, 1689, 1975.

Parmenides (c475 BC). On Nature, Allan F. Randall (Ed.)., Toronto, c. 475 BC, 1996.

Randall, Allan F. (1996). Truth, Coherence and Correspondence in theMetaphysics of F.H. Bradley., Toronto,1996.

Go Back to Allan Randall's Home Page.