This is not an easy question to pose today. For fifty years, Anglo-American philosophers have argued that it is an illegitimate or meaningless question to ask what an idea is. One is supposed to be blocked from asking such questions as "What is a proposition, or a concept?" as ontological questions. Such questions lead to the multiplied ontologies of Brentano, Meinong, and other champions of the inexistent. The way to avoid such excesses is to concentrate on epistemological questions about ideas and propositions, content and meaning. "What is an idea?" and "How does an idea represent its object?" then lead to discussions about uses of words and types of speech acts, and even to the denial that ideas in one seventeenth and eighteenth century meaning--sensible ideas, sense data--even exist. To ask what an idea is, or a proposition, or a meaning, or information, is to make the mistake of thinking that an idea or proposition or meaning or information is something that has ontological being.
For purposes of discussion, let's make the mistake. Then if an idea is really, as the Cartesians say, a property of a mind, then that idea is a mental property of an unextended mental substance. Things out there in the external world are material substances and have such extended material properties as size, shape, position, and motion or rest. So you can ask how a mental idea could make a material thing and its material properties known. The Cartesians could not give an answer. They themselves insisted that the most obvious answer--that an idea resembles its object--is ruled out because mental ideas are entirely unlike material things. All they could say was that mental ideas do represent material things, and this is because God makes it so. Nobody accepted this as a philosophical explanation.
Here is the kind of answer most seventeenth century philosophers thought had to be given to the questions, "What is an idea?" and "How does an idea represent its object?". They took these questions to imply that one can describe an ontological model of our minds and of the material world in such a way that you can understand from looking at the model how an idea is of its object. To explain how ideas work is to give an ontological model showing how they work.
Nicolas Malebranche, for example, provided a model in which ideas are not, in fact, modifications of the mind. Rather, they are in God, not as God's modifications, just in God, and they apparently make things known by being some sort of Platonic archetype, so that they resemble things. This is a reconstruction of Malebranche. What he really meant is obscure and controversial, but Malebranchean ideas do seem to be something like Platonic archetypal ideas. But how does that help us know things? Malebranche says that God shares his ideas with us on the proper occasions. For example, when we trip over a stool, God provides us with a full set of ideas that represent what is happening. It is, of course, necessary for us to have an act of mind each time God supplies us with an idea. And these acts of mind (unlike Malebranchean ideas) are mental properties of our substantial minds. Every such mental act is the same; they just have different contents, these contents being the different Malebranchean ideas appropriate on various occasions.
Few philosophers found this to be a satisfactory model. There was an empirically motivated move away from Platonic universals, and in a dualistic world of mind and matter, Malebranchean ideas are ontologically suspect: they are things of a third kind. Antoine Arnauld, for example, said Malebranchean ideas are totally unnecessary. Arnauld proposed instead that, rather than all mental acts of comprehension being the same, each act is differently signated according to the object it represents. So for Arnauld, the substantial mind has mental properties that are ideas, each idea being modified in such a way that it represents one external object rather than another. An idea is a signated property of the mind.(1)
How does an idea get signated? Descartes, Malebranche, and Arnauld agree on a causal theory of perceiving and knowing the external world. It is the same ontological model taught in physiology courses today. Here is a general picture of the physiological situation.
Some material object or event out there in the world acts, through a medium or directly, on one of your sense organs. Various transformations of a mechanical, chemical, electrical, etc.--but always physical--nature pass through your nerves to your brain. Cartesians thought that the pineal gland vibrated in different ways according to different external stimuli. Today we talk of distinctive neuron firings. In either case, the model is quite clear. The endpoint effect in the brain is correlated with its external cause in such a way that you can follow the transformations through the nerves and see that the brain effect is specifically correlated with that cause. In other words, something in the brain is earmarked in such a way as to be representative of the external object that caused that brain state. Now the question arises: How does the brain trace represent the external thing? We see how the brain trace gets to be representative of its cause. Now we must look at the brain trace itself and explain how, just from looking at it, we can tell that it is representative of its cause. How can we tell from the brain trace itself which object it represents? We know we could tell this by tracing through the transformations in the nerves and the physical media out to the external thing, but, in fact, that is all over now, it no longer is there to observe. Furthermore, that causal sequence is not meant (not meant by God or Mother Nature) to be a way of telling what the idea is about. We are not meant to trace the way through all those transformations to tell what the brain state is of; those transformations are just to get the brain state in such condition that the brain state itself is of its cause.
But why do we have to be able to tell from the brain state in itself what its object is? That is because (for the Cartesians) the mind is there waiting to examine the brain state to derive from it enough information to cause in the mind an idea, which idea represents the external object. The mind does not see the entire physiological process. The mind sees only the brain state, and must be able to tell from this brain state alone what its object is.
How does the brain state represent the external object? On the causal theory presented here, some sort of isomorphism of pattern is preserved through each transformation, so that you can trace the pattern of the material thing out there in the world through the nerves to the resulting pattern of the brain state. You can thus say that the item in the brain represents the external object by resembling it, the resemblance being exhibited through the pattern the brain item exhibits, which is the same pattern that the material thing out there in the world exhibits. The answer to the question of what it is in the brain state that makes this particular brain state represent that particular object is this: Each of them exhibits the same pattern, and the brain state is caused to exhibit the external thing's pattern by the external thing acting through the nerves on the brain. So the brain state represents the external object by having some sort of isomorphic pattern-resemblance to it. Note that the crucial thing about this physiological model is that both the brain state and its external cause are material things, so each of them can exhibit a pattern.
On this model, the mind observes the brain state or the brain state acts on the mind, with the result that an idea is caused or arises in the mind, this idea being an idea not of its proximal cause, which is the brain state, but of its distal cause, that material thing out there in the world that started the causal chain. The mental idea represents the material object. How? The Arnauldian scheme seems best. Each idea is modified or signated in such a way as to be distinctively of one rather than of any other object out there in the world.
But how, exactly, does a mental idea represent its material object? The patterns exhibited by external things and brains are physical, spatial patterns. The mental modifications of unextended mind surely cannot exhibit such spatial patterns. This sets up a crude refutation of the Cartesian way of ideas as follows: Mental substance and material substance are so different from one another that ideas that are modifications of mind could never resemble the bodies. Thus if resemblance is necessary for representation (as it seems to be, because it is the only intelligible explanation anyone ever gave for how an idea represents its object), then a mental idea cannot represent a material object because a mental idea cannot exhibit the pattern of a material object.(2)
But we can be more subtle than that. A rough pictorial or iconic resemblance is not what we have in mind, but rather isomorphic patterns sustained through many transformations. So consider this question: Through a sequence of transformations, can a mental idea exhibit the same pattern of relations as a material object? Can a pattern exemplified by a material thing be so transformed that it can be exhibited by a mental idea?
The Cartesians seem to have thought that mental ideas can exhibit the same pattern as do material objects. The question came up over the issue of intelligible extension. Material extension in this tradition is three dimensional space. That is, the Cartesian plenum just is three dimensional space, which itself just is matter. There is no empty space; space is matter. And this matter can exhibit any of the infinite patterns possible in three-dimensional solid Euclidean geometry.
Now consider that, when the Cartesians asked how we can understand the plenum, have ideas of it, comprehend it, they were taking intelligible extension to be something different from material extension. The examples are very familiar, the thousand-sided figure, the two ideas of the sun, and so on. We understand a thousand-sided figure and the sun and their properties perfectly well by thinking of intelligible extension, but not just by looking at images, figures, or real material things.
But what is intelligible extension? If this idea of material extension is not itself materially extended, what is it? And how does it represent material extension? None of the Cartesians answered these questions, but a general answer seems easily derivable from Descartes's great invention, analytic geometry. He showed how geometric figures can be represented with algebraic formulae, and how, to solve mathematical problems, one can move back and forth from the algebraic representations to the figural representations. On this basis, one could say that intelligible extension just is the set of definitions, axioms, rules, and inferences of solid analytic geometry.(3) The way we understand a thousand-sided figure is not by seeing a figure, but by thinking algebraically. If you like, this Cartesian model is the first computational model of the mind.
To pursue intelligible extension further, we can benefit from first considering Locke's position on ideas. Locke is the classic case of a philosopher who believes that real ideas really represent only if they resemble their objects. With a bit of reconstruction, Locke can be presented as a philosopher who conceives of the mind as having two kinds of ideas. Primary ideas exactly resemble the primary properties--size, shape, position, motion or rest--of material objects out there in the world, which primary properties cause those primary ideas. Secondary ideas do not resemble the collections of primary properties that--collectively--cause the secondary ideas. These secondary ideas are sensations that may indicate the presence of material bodies and give us clues as to whether these material bodies are good or bad for us, but these sensations do not really represent material bodies by showing us what they are like. Primary ideas, on the other hand, are in effect pictures of material bodies, which show us the way those bodies really are.(4)
The big problem for Locke is abstract ideas. How are we to understand what, say, a thousand-sided figure is? No picture-idea in the mind is going to provide that understanding. We just do not have accurate mental images of thousand-sided figures. Locke, and Berkeley after him, give the classic British empiricist reply to the question of how we can think of, say, the general triangle, which can be right, obtuse, and scalene all at once. We just think of a particular triangle, and have it stand in general for all triangles or for the generic triangle. This is inadequate because no explanation whatsoever is offered for how one could go from gazing at one particular, determinate triangle to having an understanding of the triangle or triangularity or the universal triangle or triangles in general.
Consider the situation again. There is a mind in the head having, experiencing, or looking at ideas to see what the external world is like. On this view, it makes some sense to say that patterns of material bodies (for example, size, shape, position, motion or rest) are transmitted into the brain, where somehow these patterns are further transmitted on to mental images, where they are exhibited for the mind to look at them. But even if all this were crystal clear, the mind would still be looking only at particular figures, not at, say, the universal triangle. So where does knowledge of the universal triangle come from? No matter how many particular triangles one looks at ... and so on ...
Of course, this model is not crystal clear. And Locke does not offer any ontological explanation of how we can know universals. Locke, does, however, focus on the problem of how a mental substance can have a property (an idea, sensation, or image) that resembles a material body or material properties. The need for this resemblance for representation is what leads him to suggest that maybe matter can think. Perhaps, he suggests, there is a more basic substance that can have both mental and material properties. Then, since mental properties and material properties are properties of the same substance, there would seem to be no reason why mental ideas could not exhibit the same patterns as material bodies and thus represent bodies by resembling them.
It does not work. Neutral monism, double-aspect theories, property dualism, all lead back to the problems of real substance dualism. Mental properties are supposed to be unextended, and bodies are extended. Note that the issue here concerns things, not words. It is not a question of whether or not we can talk about ideas in terms of material properties, but of how real mental and material things can be ontologically alike in the sense of sharing or exhibiting the same pattern. A version of this problem now plagues naturalists, neurophilosophers, and new materialists. To appreciate this contemporary problem, consider the question: What is the mind?
There are two traditional ways of characterizing the mind as a mental substance. One is to describe it as duplicating the entities, patterns, structures, functions, and operations of the brain. The other is to represent it as a black box. What happened in history is that black-box people such as Hume kept peeking inside and finding nothing, and the brain-duplicate people finally decided that the mind is redundant, that the mind is not a separate substance at all, but that the mind just is the brain. Thus today many serious philosophers of mind are materialists. This simplifies the problem of what an idea is and how it represents its object. Materialists do not have to worry about what ghostly, unextended mental ideas are, or about how they represent extended bodies.
How does a materialist answer the questions "What is an idea?" and "How does an idea represent its object?" Or, in Daniel Dennett's words: "How can a particular state in the brain represent one feature of the world rather than another? And whatever it is that makes some feature of the brain represent what it represents, how does it come to represent what it represents?"(5) Descartes and Dennett answer with a causal theory of perception. But let us look at Patricia Churchland's version. If you keep these questions in mind while reading her Neurophilosophy, you will see that ideas are brain states, or, if you like, neuron events--something in the brain-mind, something going on in the brain-mind. And if you follow through the causal theory of perception, you will see that the transformations are made from the outer initiating material object to the inner brain state, so that one can say that the inner brain state exhibits some sort of pattern isomorphic with a pattern in the outer material object. So, it makes some sense to say this brain state represents that object because the two of them exhibit the same pattern. This brain state was caused to exhibit the pattern it does by that object, and because of this we can say that this brain state represents that object by resembling it. Of course this procedure is described entirely from the outside. We do not learn about physical and physiological transformations by introspection. Furthermore, one must remember that there is no independent mind inside the brain observing this brain state to determine from it what object it represents. We have moved to the position that the brain is the mind and the brain state is the idea. But whereas it made some sort of sense to speak of the independent mind observing a brain state or even of a mind observing one of its own property-ideas, it does not seem to make much sense to speak of the brain as observing one of its own brain states, even if you call the brain the brain-mind. The brain just has the brain state.
Here is the crucial question of this paper: what is wrong with saying that the brain's knowing something just is the brain's having a brain state? The problem is this. A traditional attack on the notion that there is an independent mental substance different from a material brain is that the standard description of the mind is just a duplicate of our description of the brain's mechanisms and functions in mental rather than physical terms. But everyone knows that just having a property is not at all like--or sufficient for--knowing an idea. So it does not work to say that a mental mind modeled after the material brain knows ideas just by having them.
The mind in fact cannot be a total duplicate of the brain, because the mind knows its properties (ideas), but bodies do not know their own properties. How do minds know their own properties? When one bears down on these descriptions of mental processes and asks why it is that the mind knows its own properties, whereas a chair does not know its own properties, the answer often is that for a mind, the knowing of a property just is the having of that property. Having an idea, that idea being a property of your mind, just is the knowing of that idea. But having a physical shape is not, for a body, knowing that shape. So the mind model may seem to duplicate the brain model, but having a property for the mind is knowing that property (or at least those properties we call ideas), which is not the case for a body. Bodies neither have ideas nor know their own properties.
Philosophers such as Patricia Churchland, Paul Churchland, and Daniel Dennett suggest or imply that just having a brain state is the brain's knowing that brain state, or, rather, knowing what that brain state represents namely, the external object that initiated the causal sequence that caused the brain state to exhibit the same pattern as the external object. The brain state is a property of the brain. So here we have a material thing knowing one of its own properties (states) by having it. This certainly is the burden of Dennett's Consciousness Explained.(6)
Nobody liked this move for minds. Are we going to get away with it for brains?
Note again that the notion that having a property is knowing the property or knowing what it represents, is an aspect of the explanation of the independent mind that does not derive from the analogy to material mechanisms of the brain. For if there is anything we know (or thought we knew) about material bodies, it is that they do not know their own properties.
It is often said that the mark of the mental is intentionality, but something more than that is at issue here. Ideas represent objects. It even makes sense to say that a brain state represents an object by resembling it. Of course there are serious problems with the model in which representation is based on resemblance. Briefly, the basic problem is that everything resembles everything else in some way. Nothing seems more tautological than that, for any two objects, you can find some pattern they exhibit isomorphously (think of Leibniz). So even if resemblance is required in some way for representation, it certainly is not sufficient. That is one reason we trace the physical and physiological chain of causation and transformations between the object out there in the world and the brain state in our brains. We have to see the whole situation to determine what limited pattern in the brain state is like a pattern in its object, and thus represents that object, or at least the limited pattern aspect of it.
But what is at issue now is not representation or the intentionality of ideas or the mental, but rather the having of brain states. What is crucial is that the notion that the brain's having of the brain state is the brain's knowing the brain state or its object--this identification of the having of a property and the knowing of something--is an importation into the physical, material world of a primary mentalist doctrine. The new materialists are taking the old mentalist notion that a mind's knowing an idea just is having that idea, and applying this notion to the material brain. Historically, nobody would buy the notion that a mind's knowing a mental property (an idea) just is the mind's having that idea or its object. Is anyone going to buy the notion that a brain's having a brain state just is the brain's knowing that brain state or its object?
Maybe brain matter is different from ordinary matter. Recently, John Searle considered something like this, and earlier May Brodbeck advocated it.(7) But few materialists believe that organic matter is different from inorganic matter in such a way that organic matter (why just the brain?) knows its own properties. Perhaps the complexity of the material stuff integrated as our brains (and not just organic matter itself) makes it so that the brain knows its own brain states, or so that the brain knows external objects by way of its brain states. Thus a complex set of relations, the way brain matter is organized, would cause us to know. But such complexity is a property of the brain-mind, so appeal to complexity may be just another way of saying that the mind knows by way of ideas, i.e., by having a complex property. So again: How can having a property be knowing?
Let us suppose, however, that knowing is having a brain state. The brain state is a particular material property (or some physical phenomenon), so Locke's problem about universals arises again. How can a particular give us knowledge of a universal? Let us go back to Cartesian intelligible extension to work toward the problem again. Suppose we say that material extension is just the existing three-dimensional world of bodies in space, i.e., real, physical stuff. Then, if what we understand of this material world--intelligible extension--is just the definitions, axioms, rules, inferences, and theorems of Euclidean solid geometry, what are they? What, for example, are the axioms? Or the theorems?
Now let us agree that at least at the extremes, there is a sharp distinction between syntactics and semantics. Suppose that "[a.sup.2] + [b.sup.2] = [c.sup.2]" is taken syntactically, as uninterpreted. But of course it can be interpreted, and this is because it is a set of marks arranged in a certain way. One way of saying this is to say that the algebraic formula represents the same set of relations that, say, the figure of a circle represents. So we have two sets of marks, an algebraic formula and a figure. Note that this is to talk as though there were here some entity or set of entities of a third kind. There is the algebraic expression and the figural expression, and they each represent the same other thing; and this other thing--this intelligible extension--is a set of relations.
What is this set of relations?
It is what is exhibited both by the algebraic formula and by the figure. But what is this set of relations exhibited by the syntactical and figural expressions, or that the syntactical and figural expressions can take as an interpretation? What is this set of relations in itself?
Think of this question for comparison: What is an Aristotelian form in itself? Remember, a thing for Aristotle consists of a composite of form and matter. Form and matter do not really exist apart. The matter is the stuff that is patterned and the form is what patterns it. In Thomistic accounts, we can somehow abstract the form and emplace it in our mind so we can observe it, and thus can understand the thing. The form in our intellect does not actually inform our mind substance to make it exhibit the pattern of the thing, but by having it in our intellect, we can know the form, that is, know the pattern of the thing. But it is not clear how the form is had in the mind, nor how having it there constitutes knowing it. So what is the form?
We have arrived at the ultimate question: What is an unexemplified set of relations? In algebra and geometry, the algebraic formula is really just an exemplification of the set of relations, and the geometric figure is really just an exemplification of the set of relations. Neither of them is the set of relations in itself. We know that the algebraic formula is not in itself the set of relations. It is just a vehicle for exhibiting the set of relations. And the same goes for the geometric figure. It is just a drawing on a piece of paper or in the sand that accommodates the set of relations. The question that arises is: What is the set of relations in itself?
Suppose we go even farther back, to Plato, to see whether he has an answer. What if we say that the set of relations exhibited by the algebraic formula for a circle and the geometric figure of a circle is the universal, archetypal, Platonic Idea of a circle? Well, what is that? A Platonic Idea is what resides in the realm of Being. We know it by being in its presence. Note that such Platonic Ideas are like Malebranchean ideas of the third kind which are neither properties nor substances but are in the mind of God, and that (for Malebranche) we know when God shares them with us. An archetypal Idea, a universal, a form, just is that idea we have when we know, say, what a triangle is. And we know a Platonic Idea by being in its presence.
This makes clear that to evoke archetypal Ideas is not to give a satisfactory explanation. Explanations in mind-talk--real ontological mind-talk--very often go like this: Ideas intend their objects because that is what ideas do; our minds can understand things in general when in the presence of their archetypal ideas because this is what happens when minds are in the presence of archetypal ideas and this is just what archeypal ideas permit the mind to understand. In other words, having an idea or being in the presence of an idea is knowing it.
These are not explanations. No explantory model of how things work is given. These are statements attributing to very special entities the power to intend things and to make us know things. The classic explanation of this type, of course, is Moliere's explanation that opium puts us to sleep because it has a dormitive power.
Maybe there are such archetypal Ideas or Forms. Mathematicians are often Platonists, and Aristotelianism, at least for virtues and natural kinds, seems to be on the rise these days. The point, of course, is that the claim that knowing is having is not emanating only from mentalists, but also from materialist neurophilosophers, from Patricia Churchland and Paul Churchland, even from Searle, Dennett, Fodor, and others.
The mentalists could not solve the problem of universals this way. What about the new materialists? The only materialist who seems to take the problem of universals truly seriously is David Armstrong.(8) And his attempt at a solution suffers from a circumstance that applies to all materialist attempts. It is the problem Locke and Berkeley raise, that of how a particular thing can really give you any sort of general or universal knowledge. Perhaps your brain sets aside one brain state to which a lot of other brain states are related through causal and pattern similarities. Armstrong considers the neurophysiological correlates only in hypothetical example.(9) He goes into no details concerning inductive neurophysiology or artificial intelligence networking. Basically, Armstrong and others just set up the problem of how the work that traditional universals are supposed to do is done in the brain; they neither solve nor dissolve it. So the question arises for materialists: What is a universal?
Do the new materialists understand the dimensions or the seriousness of the problem of universals for materialism? For example, Dennett quotes Quine approvingly as saying that "to expect a distinctive physical mechanism behind every genuinely distinct mental state is one thing: to expect a distinctive mechanism for every purported distinction that can be phrased in traditional mentalistic language is another." Dennett claims to expose "the error of those who had hoped to find something in the head to settle the cases Quine's peripheralism left indeterminate."(10)
In "Three Kinds of Intentional Psychology," Dennett says somewhat more explicitly: "Folk psychology is abstract in that the beliefs and desires it attributes are not--or need not be--presumed to be intervening distinguishable states of an internal behavior-causing system."(11) He continues: "Reichenbach distinguished between two sorts of referents for theoretical terms: illata--posited theoretical entities--and abstracta--calculation-bound entities or logical constructs. Beliefs and desires of folk psychology (but not all mental events and states) are abstracta."(12) Dennett opposes his own view to that of Fodor whom he quotes as saying that "to suppose two people share a belief is to suppose them to be ultimately in some structurally similar internal condition."(13)
The above characterization of brain physiology lends itself to Fodor's position, which is that if two people know or believe the same thing, this ought to be shown in their brains by their having brain states that exhibit the same pattern. Dennett may not in fact object to this, at least not for what he calls "good theoretical entities, good illata, or good logical constructs, good abstracta," to get which he recommends jettisoning "some of the ordinary freight of the concepts of belief and desire."(14) And so Dennett eventually comes to saying that "There must be some way in which the internal processes of the system mirror the complexities of the intentional interpretation, or its success would be a miracle,"(15) the intentional interpretation being in part at least the application of ordinary folk psychology to give explanations and make predictions about human behavior.
Dennett depends on metaphors that are open to various interpretations, but the passages quoted above do suggest that something escapes the materialist net. This ontological leakage derives from Quine's not catching every "purported" distinction that can be phrased in traditional mentalistic language. This leakage is of a large number of fleeting thoughts that all of us have, which Quine and Dennett say do not have brainstate correlates in a cleaned-up psychological language. These fleeting thoughts--remember, the burden of this paper is that if we have them, they are ontological entities--float free of any neurophysiological correlates. In particular, in Consciousness Explained, Dennett does not adequately account for how our "seeming to see" mental images is illusory(16) nor explain adequately how there are no such things as "raw feels," "sensa," "phenomenal qualities," "intrinsic properties of conscious experiences," or "the qualitative content of material states," all of which he refers to as "qualia."(17) He wants to say that because sensa and so on are illusions they do not exist; but either the illusions themselves--the "seemings to see" and "seemings to feel"--must themselves be physical manifestations in the brain or they float free. Dennett surely does not want anything to float free, but his Rylean conceit that they do not even exist (an extreme view that Dennett is almost alone in holding these days) puts him in danger of turning them into non-physical, non-material things.
Here is the problem. If the mind is the brain and if thoughts are brain states, then every thought--including "purportings," "seemings," and "illusions"--must have a brain correlate because every thought just is a brain state. So, for a materialist, if we have fleeting thoughts and folkish beliefs and intentions--and purportings, seemings, and illusions--then they certainly can be discerned in the brain, because they are brain states. They must, somehow, correlate with something material in the brain. The problem at the moment is not that of determining which thoughts given brain states are identical to. It is not the problem of indeterminacy of translation that is at issue, but the problem of what those fleeting thoughts are, if Dennett and others actually allow some of them to float free of material moorings. It even looks as though Dennett casts some of them out, for example, the purportings, seemings, illusions, and bad abstracta, which seems to imply that they are something other than material things. But what else is there for them to be, but matter?
Consider intelligible extension again. The logic of the reconstruction of the Cartesian discussion of intelligible extension results in the following situation. Intelligible extension is neither the figures of geometry nor the formulae of algebra. Instead, the concrete expressions of geometry and algebra merely exhibit that set of relations that is intelligible extension. It is the set of relations itself that we understand. Perhaps we can grasp this set of relations only when it is exhibited in figures or formulae, but the set of relations itself is something different from the figures or formulae, something different from the marks on paper or on the blackboard.
But, in the Cartesian world, sets of relations --abstracta--have to be something. Descartes thought they were mental. But Malebranche recognized that even if the finite properties of a finite mind substance could represent the patterns of particular geometric figures or particular algebraic formulae, these mental properties or mental images cannot in themselves show the triangle, or universality, or general things. So he brought in Malebranchean ideas of the third kind, ideas in God, that are probably archetypal Platonic Ideas.
Fine, if you are a dualist, or think there are universals in an accessible realm of Being--or Third World--outside mind and matter. But if you are a materialist, then, unavoidably, the problem of universals has to be solved by some sort of particular manifestation in the brain. If we can think about sets of relations that can be exemplified by geometric figures and also by algebraic formulae, or if we can think about the dog that gets exemplified by Rex, Rover, and Rolf, then that set of relations must show up somewhere in the brain. For a materialist, the answer to the question, "What is, ontologically speaking, a set of unexemplified relations?" is "It is some particular set of relations physically exemplified in the brain." What is a universal? It is a material manifestation in the brain. What is the general idea of a triangle? It is something physically present in the brain. What are purportings, seemings, illusions, and folkish beliefs? The same. And you cannot get out of dependence on material brain-states by talking of uses or functions, for uses and functions must be exhibited. Something material in the brain is being used, or something physical is functioning in the brain. Speak of brain-events, brain-dispositions, brain-whatevers. That is fine. But the answer to the question about relations for a materialist has to be: There is no unexemplified set of relations, because anything that is, is material. There are relations, but if we know them, we must know them through some actual physical something in the brain.
But is some physical presence in the brain actually enough? Probably not, in fact. The brain state shows itself as an idea or as a universal only in the total context of its causal environment. What must be materially present is some physical phenomenon involving a brain state, which physical phenomenon does what we ordinarily speak of as knowing an idea, understanding a universal, and so on.
Now suppose with Fodor that a language of thought or a deep grammar were hard-wired in the brain.(18) Then we would have a situation in which one might want to say that the brain "understands" the concepts of this language of thought just by having them. Fodor says we cannot build up concepts in a natural language unless we already know them in the language of thought.(19) Here Fodor is trading (as are lots of other people) on the notion that for some material objects (human brains, the computer Hal in the movie 2001), being hardwired in a certain way so that you have certain properties and function physically in certain ways, just is knowing and intentionally using certain concepts. Again, historically this is using a mental model of the mind to explain how a material thing can know something, for certainly no ordinary material thing knows its own properties.
The problem of free-floating entities is only one of those the new materialists must face. For Fodor, propositions, formulae, and non-physical properties seem to float above their material manifestations.(20) The jargon is that they "supervene" on them. (Just what, ontologically, does supervening consist of? This, like "What is an idea?" is a forbidden question.) But Fodor does have sentences in the brain that are physically there as material brain states or neural events. Propositional attitudes have as contents (ultimately) these sentences in the language of thought. Propositional attitudes are identical with certain brain states, and their contents are other brain states, so here we have intentionality defined as a physical relation between two physical states. Two physical things are presented as being in a cognitive relationship just by being in physical proximity.(21)
Previously, only in classical talk of the mind--such as in Plato and Malebranche--have there been naked attempts to explain how one thing knows another by saying that they are simply in contiguous relation to one another or that one is in the presence of the other. And formerly, only advocates of the Cartesian way of ideas have unashamedly suggested that the knowing of an idea or its object is identical with having an idea as a property. Today these moves are being made explicitly and implicitly by Fodor, Dennet, Patricia Churchland, and many other new materialists.
But this materialist characterization of knowing as having is not an explanation. Is it then a claim that knowing is an unanalyzable property or relation, such as yellow or good for G. E. Moore, or intentionality for Brentano? If so, then again the only move the neurophilosophers have made is to transfer this unanalyzable property or relation from the mental realm of the mind to the physical realm of the brain. One might say that, anyway, we are not worse off than we were before. But if we scorned the mentalists for saying that knowing an idea is just having an idea, what are we to make of materialists who, in their materialist language and materialist ontology, propose the same thing?
Of course, if the mind is the material brain, then we had better go from the bottom up in trying to understand it. Only by studying the entire system from the object out there in the world that starts the causal sequence that results in the stimulation of sense organs that leads to the transformations in the nerves resulting in a brain state in the brain-mind, only by seeing the whole sequence from the ground up, will we be able to see what is going on in the brain, to find out what in the brain represents things outside the brain and thus constitutes our knowing them. Knowing, on this view, is nothing but a physical process; and so all thoughts and all known sets of relations have to have exemplifications there in the brain, because that is what all thoughts and knowings are--brain-states or physical brain-somethings-or-other, interrelated with the rest of the surrounding physiological and physical environment.
But consider this: Fodor talks a lot about propositional attitudes but never explains what a proposition is.(22) Do propositions float free? This problem of free-floating propositions may plague even Fred Dretske. In his Explaining Behavior, Dretske remarks on things that are "in the heads of some animals" and says that among these things is "something that has ... a content or meaning (not itself in the head, of course) that is individuated the way we individuate thoughts and beliefs."(23) Earlier he says the content is a proposition that has a meaning. But if content or meaning (or a proposition) is not in the head, then where is it? What is it? Whoever says it is just the external object simply begs the question. Again, it cannot be the whole reflex arc event between stimulus and response, because, for reasons stated above, one must be able to tell from a brain-state idea alone what its object is.
If some thoughts float free, as one might suspect Fodor, Dennett, and others of allowing, and if you decide that you can work with propositions or sets of relations that you can study in themselves, from the top down, free of messy neurophysiological muckings, then you may be slipping back over the line into dualism. You may be treating the mind and its contents as entities independent of the brain. This is because the answer to the question "What are unexemplified sets of relations?" is that either they do not exist because there is nothing but matter and all sets of relations we know of are exemplified in matter, or, if you really do think sets of relations can float free of the brain, then maybe your mind itself is in danger of floating free. Could you be just taking "The Mental Stance"? No doubt doing so is useful, but it may also commit you to dualism.
What is really under attack here? After all, nobody really thinks Dennett and Fodor, for example, are closet dualists. What is being opposed is the notion that you can do epistemology without ontology. A major gambit in twentieth century Anglo-American philosophy has been to say that ontology (metaphysics) is nonsense, and that you can work on the problems of epistemology and philosophy of mind without considering ontology. But now we have a group of new materialists, the naturalists and neurophilosophers, who are still riding on the shirttails of the anti-ontologists. They hark to the big category mistakes, to being misled by substantitives into thinking that all nouns have existing objects, by taking ideas to be things when they are really uses, they use the word magic of adverbials (do we see greenly?), and so on. But more than that, lots of words, they say, do not have ontological reference--folk psychology words, for example.
But in this the new materialists may lose their bearings. They discuss beliefs, intentions, thoughts, meanings, ideas, propositions, concepts, uses, functions, goals, qualia, supervenience, and so on, and there is a strong tendency to treat the words referring to one or another of these items as not having reference, i.e., as not having objects, as not indicating ontological entities. Of qualia, Dennett says, "There are no such things."(24) But the belief that there are no mental images, for example, is surely a member of that class of beliefs of which none is so silly that some philosopher has not believed it.
Fifty years of scornful rhetoric has been directed at anyone who might think that a meaning is a thing or that "either/or," for example, refers to an object. But for a materialist, a material correlation of "either/or" has to be in the brain in some way or another--see Armstrong's discussions of and, or, and not neurons in his A Combinatorial Theory of Possibility.(25) Armstrong does not carry through with this discussion. For example, he says: "Impossible worlds are a conception, a conception, which, like ideal gases and frictionless planes, turns out to be useful in analysing actual phenomena. The conception is a thought-instrument."(26) But he does not distinguish between possible worlds that do not exist and the conceptions that do exist. But for a materialist, the conceptions, the "thought-instruments" do and must exist--like the and, or, and not neurons--in the brain.
Abstracta, generalities, universals, archetypes, paradigms, and so on have to be in the brain, too. The attack on these items as independent existents began with Aristotle. They get dismissed by the British Empiricists with empty rhetoric of this sort: There is no universal triangle; we just use the word "triangle" or an image of a particular triangle to refer to all triangles in general. So we do. But to say that we do is not to explain how it can take place. It is rather to make the problem so stark as to appear to be totally unsolvable. How does looking at a determinate image give you knowledge of an indeterminate idea of a triangle? Certainly just saying that we assign a meaning to the word "triangle" and use it to refer to triangles in general does not explain how it is that we can do this. Yet an enormous amount of philosophy is based on the assumption that there is nothing to explain--because uses are not things, because functions are abstract and so not inherent in or exhibited by things themselves, because meanings are not ontological entities, and other such dogmas.
Now there are the new materialists. They oppose dualism. The reason dualism has been and still is a popular ontology is that we cannot operate without mentalistic items. The logical positivists and their progeny thought they were getting rid of mentalism by classifying some items as "linguistic" or "syntactic" or "logical" and thus not ontological. But it is a non sequitur to say: If linguistic or syntactic or logical, then not ontological. This fact is easily ignored, however, when attention to it threatens to summon Brentano's and Meinong's ghosts from their graves. Or we could do a Fodor and say: everybody has the problem of universals, so the hell with it--whistling in the dark.
But the problems of ontology cannot be buried. There are philosophers now who say they are materialists, probably because they take this as an inference from their conviction that there is no independent mental substance. There is only material substance. Also, they are involved in brain science, and the brain is a material object.
Note that it does not matter if the new materialists deny that they are materialists in any classical sense so long as they say the brain is something. Whatever the brain is made of, if they claim for it that having an idea--a state, an organization, an event in the brain--is knowing, they appeal to magic.
Philosophers of mind have to work with a long tradition of mentalistic language. No adequate argument has yet been given that warrants ignoring the ontological implications of the use of this language. One cannot beg off by stipulating that epistemological objects and relations do not have ontological status. Fodor says, "The properties of the world that we are epistemologically related to aren't, usually, physical properties."(27) But for materialists, meanings and other epistemological items have to be in the brain, as material,
physical entities. These entities can be activated brain-areas, brain-states, uses, functions, and so on, but they have to be physical phenomena. A meaning has to be there in the brain. The average pattern over a large area of the brain has to be in the brain, as a particular, concrete physical phenomenon. Uses are uses of physical things; functions are functions of physical things. The truth of the British Empiricists is that there are no general things, no abstracta. This is a material world, and material things are particular. Of course, there are relations, but they are physically exemplified. As Armstrong says, "Properties and relations ... depend on individuals, and are found only in [material] states of affairs."(28)
Our ability to think about abstracta, to talk about meanings, to think in universal terms, all of that must be there as particular, physical phenomena in the brain. We understand things not because we enter a realm of universal Platonic Ideas, but because specific physical things, events, and phenomena in our brains correlate with such terms as "meaning", "use", and "either/or."
So, when Dennett hints that not all fleeting thoughts correlate with something in the brain, or Dretske says not all information is in the head, the question looms: "Where are they?" And if the reply is, "Don't you understand that these are not ontological items?" The reply in turn must be, "No. Are you sure what you say aren't ontological items aren't really mental items in disguise?"
Explicitly, until these new materialists can give (or at least state explicitly as their goal) a total reduction of mentalist talk to materialist talk, until they can give a complete physicalistic description of what a meaning or a universal is, until they can show what ideas are and how we think and know things, by pointing to actual physical mechanisms or phenomena or events or things in the material brain that are (note well) in causal interaction with the surrounding physical environment, they are piggy-backing on ontological mentalism. This is because the realm of the independent mental substance is proposed precisely to take care of such difficult items as meanings and universals that do not seem to be possible in a world of matter where everything that exists is particular (which includes particular exemplifications of relations).
This critique is not meant to promote dualism. There is no a priori reason why mechanical devices (or brain phenomena) cannot do the jobs traditionally known as having an idea, thinking, reasoning with abstracta, and so on. But to do it we have to show how everything, that is, every thing--propositions, meanings, ideas, concepts, universals, etc.--is a material thing. It means that we cannot slur over describing these material things by using mentalistic language and word magic while pretending that one can do epistemology, philosophy of language, philosophy of mind, and logic without giving an ontological foundation for every last detail. It means total reduction of the mental to the material. That is, if you want to be a materialist, if you want truly to naturalize epistemology, you have to pay attention to the working scientists who write books with programs such as the following:
My purpose is to sketch a new framework for
the understanding of animal learning and the
investigation of its cellular basis. In the framework
here elaborated, quantities computed
and stored by the nervous system represent
aspects of the animal's environment and its
relation to that environment. Thus I term this
framework a computational-representational
framework. I use the term representation in
its mathematical sense. The brain is said to represent
an aspect of the environment when
there is a functioning isomorphism between an
aspect of the environment and a brain process
that adapts the animal's behavior to it.(29)
(1.) These structural interpretations are defended with reference to the original texts in Richard A "Arnauld, Malebranche, and the Ontology of Ideas," Methodology and Science, vol. 24 (1991), pp. 161- a diverging interpretation, see Steven M. Nadler, Arnauld and the Cartesian Philosophy of Ideas (Pri Princeton University Press, 1989), and Malebranche and Ideas (New York and Oxford, Oxford University Press, 1992). (2.) The development of this objection is examined in Richard A. Watson, The Breakdown of Cartesian Metaphysics (Atlantic Highlands, Humanities Press International, 1987). (3.) John Herman Randall, Jr., "Religio Mathematici: The Geometrical World of Malebranche," Studies the History of Ideas, vol. 2, (New York: Columbia Univeristy Press, 1925), pp. 185-218, and The Care Philosophy: From the Middle Ages to the Enlightenment, vol. 1, (New York: Columbia University Press, 1962), pp. 425-433; Paul Schrecker, "La methode cartesienne et la logique," Revue philosophique, vol (1937), pp. 336-367, and "La parallelisme theologico-mathematique chez Malebranche," Revue philosoph vol. 124 (1938), pp. 215-252; Leon Brunschvicg, Les etapes de la philosophie mathematique des cartesiens, 3rd edition, (Paris: Presses Universitaires Francaises), 1947; Henri Gouhier, La philoso Malebranche et son experience religieuse, 2nd edition, (Paris: Libraire Philosophique J. Vrin), 1948 Gustav Bergmann, "Some Remarks on the Philosophy of Malebranche," Review of Metaphysics, vol. 10 (1956), pp. 207-225; Harry Bracken, "Some Problems of Substance Among the Cartesians," American Philosophical Quarterly, vol. 1 (1964), pp. 129-137; Richard A. Watson, "Foucher's Mistake and Maleb Break: Ideas, Intelligible Extension, and the End of Ontology," in Stuart Brown, editor, Nicolas Malebranche: His Philosophical Critics and Successors, (Assen/Maastricht: Van Gorcum, 1991), pp. 22- (4.) Is this Locke's true position? It is not my purpose here to argue that point. My presupposition brief characterization of Locke on primary and secondary qualities is an influential standard textbo See Richard A. Watson, "Shadow History," Journal of the History of Philosophy, vol. 31 (1993), pp. 9 (5.) Daniel C. Dennett, Consciousness Explained, (Boston: Little, Brown and Company, 1991), p. 192. (6.) Consciousness Explained, p. 192. (7.) John Searle, Minds, Brains, and Science, (Cambridge: Harvard University Press, 1984); May Brodb "Mental and Physical: Identity versus Sameness," in Paul K. Feyerabend and Grover Maxwell, eds., Min Matter, and Method: Essays in Honor of Herbert Feigl, (Minneapolis: University of Minnesota Press, 1966), pp. 40-58, and "Mind: From Within and From Without," Presidential Address, Proceedings and Addresses of The American Philosophical Association, vol. 46 (1972), pp. 42-55. (8.) David M. Armstrong, Universals: An Opinionated Introduction, (Boulder: Westview Press, 1989). (9.) David M. Armstrong, A Combinatorial Theory of Possibility, (Cambridge: Cambridge University Press, 1989), p. 72-74. (10.) Daniel C. Dennett, "Real Patterns, Deeper Facts, and Empty Questions," in The Intentional Stan (Cambridge: Bradford Books/MIT Press, 1987), p. 40; the quotation is from W. V. O. Quine, "On the Reasons for Indeterminacy of Translation," Journal of Philosophy, vol. 67 (1970), p. 180. (11.) Daniel C. Dennett, "Three Kinds of Intentional Psychology," in The Intentional Stance, p. 52. (12.) "Three Kinds of Intentional Psychology," p. 53. (13.) "Three Kinds of Intentional Psychology," p. 53; see also Consciousness Explained, pp. 218-219. (14.) "Three Kinds of Intentional Psychology," p. 57. (15.) "Three Kinds of Intentional Psychology," p. 60. (16.) Consciousness Explained, pp. 362-411. (17.) Consciousness Explained, p. 372. (18.) Jerry Fodor, The Language of Thought, (New York: Thomas Y. Crowell, 1975). (19.) The Language of Thought, pp. 85,79-97. (20.) The Language of Thought, pp. 76-79,197-205. (21.) The Language of Thought, pp. 76-79. (22.) The Language of Thought, pp. 76-79. (23.) Fred Dretske, Explaining Behavior: Reasons in a World of Causes, (Cambridge: Bradford Books/MIT Press, 1988), p. 77. (24.) Consciousness Explained, p. 383. (25.) David M. Armstrong, A Combinatorial Theory of Possibility, (Cambridge: Cambridge University Press, 1989), p. 72-74. (26.) A Combinatorial Theory of Possibility, p. 75 (27.) The Language of Thought, p. 204. (28.) A Combinatorial Theory of Possibility, p. 43. (29.) C. R. Gallistel, The Organization of Learning, (Cambridge: Bradford Books/MIT Press, 1990), p.
|Printer friendly Cite/link Email Feedback|
|Author:||Watson, Richard A.|
|Publication:||American Philosophical Quarterly|
|Date:||Jul 1, 1994|
|Previous Article:||Necessary beings.|
|Next Article:||The same and the same: two views of psychological continuity.|