0.1: What is the Mind?
To even begin to ask questions about the mind we must have some idea about what the mind is, not a precise definition of course, since that would leave no room for questions, but at least a general idea about what we are supposed to be looking at. One broad way to characterize the mind is simply those internal factors that govern behavior. Obviously this is a bit too broad, since under such characterization even my house’s central heating would have a mind, since its behavior is governed by a thermostat, the workings of which are not directly accessible to me. But what is the difference between my mind and the thermostat? It seems obvious that one difference between us is that my mind has a complex structure while the thermostat is simple. There is more than one way to define this difference in complexity, for example we might define the mind as only those systems that adapt to changes in their environment. But even this may admit some systems that are simply too simple, so I will thus define the mind as an internal (in the sense of not publicly accessible) mechanism governing behavior that is conscious. Of course consciousness itself is a very imprecise notion, but I shall assume that we can work with this imprecision until we form a precise definition of it, in section 7. Although this definition of mind is broad enough to encompass most views it does not encompass them all. For example some define the mind as only that which is conscious. However, this is not a real difficulty, simply a terminological dispute. Since consciousness is included in our investigations into the mind those who define the mind as only that which is conscious can simply treat only the sections on consciousness itself as really being about the mind, and simply accepting that the rest describe the framework and surroundings of this “mind”.
In what I have to say below I shall be approaching the mind from the “outside”, so to speak, meaning that I will start from what we know, objectively, about what the mind must be, and from that foundation attempt to deduce facts about consciousness, as well as other structures of the mind. Of course this is not to say that I am ignoring our inner experience of the mind, or am denying that it exists. Many of the problems that I will attempt to tackle are only problems because we recognize that inner experience exists, for example I experience my thoughts as directed at things (intentionality), and thus an explanation of this phenomena is needed (section 3). This approach is the exact opposite of phenomenology, another approach to studying the mine, which attempts to study the mind from the “inside” by describing and categorizing our experience of being conscious. But I am not doing phenomenology here, not because I think phenomenology is wrong, but because it can’t answer the questions I am primarily concerned with here. Specifically it can’t tell me why the thermostat isn’t conscious, and what kinds of beings, in general can be conscious. It can’t answer these questions because any suitable answers must be framed in terms of objective facts about the thermostat, and because phenomenology only explains human consciousness, and not consciousness in general, and these are simply not the kinds of things that phenomenology reveals to us. Of course a successful “outside” theory must meet phenomenology somewhere; in explaining which brain processes are responsible for which features of the mind it must explain why our inner experience is one way instead of another. But the possibility of such a meeting is still a long way off, because not only would it require a complete and correct “outside” theory about the mind, it would also require a highly developed neuroscience in order to provide the necessary details about the structure of the human brain in particular. So if you want to understand the structure of experience as experienced in all its detail phenomenology is still your best bet for the foreseeable future.
1.1: What is Material?
Before we can even begin to explore the thesis that the mind is material we must define what it means to be material. Obviously we don’t mean simply made of matter, energy and space-time are “material” as well. (Perhaps a better word would be physical, but I stick with material simply for the sake of tradition. And although physical might be more suggestive it still must be defined.) Here I will define as material anything that interacts with the observable world in a regular (law governed) way, allowing us to theorize about it and predict its behavior. Generally spirits and other such non-material entities are not considered accessible to such scrutiny and prediction, and hence non-material. If one could build spirit detectors, and create reliable theories of spirit particle physics, then it would make such spirits simply another part of the material world, and not really “spiritual” at all. In another way of speaking to say something is material is to say it falls within the domain of things that can be studied by rigorous science.
1.2: Three Arguments For Materialism
The thesis that the mind is material, and not something outside the domain of nature and science, is one that is attractive, since materialist theories have explained every other phenomena, but at the same time may seem impossible. The mind is such a strange thing, how can it be basically the same as the non-mental world? Perhaps not all our doubts can be put to rest until we have a complete materialist theory of the mind, but one place to start is to show that whatever the mind is it must be material. To that end I present three arguments. The first, and weakest, is the argument from explanation, which states that a materialist theory is better at explaining the mind than a non-materialist theory. Even if we accept the argument we may not be convinced to be materialists, since it does nothing to show that a materialist theory must be true. The second is the argument from rationality, which shows that given a materialist and a non-materialist theory about the mind that we should prefer the materialist theory. Again, we may not be convinced if we don’t think there are any viable materialist theories. And finally we have the argument from causal closure, which shows that the material world must be causally closed, and thus that, if the mind has an effect on the world, it must be material. In my opinion this argument is the hardest to reject.
1.3: Argument From Explanation
The argument from explanation attempts to show that materialism can explain more about the mind then its primary rival, dualism, and thus that we should devote most of our resources to developing such theories. Obviously this doesn’t show that Materialism is correct, certainly the universe has no obligation to favor the theory that explains more. Instead the argument from explanation serves more as motivation to pursue materialism, even if it turns out to be incorrect. Thus this argument is the weakest of the three.
Materialism attempts to explain, and predict, the mind by arguing that certain kinds of collections of physical properties are behind what we think of as the mental properties. This can function as an explanation of the mental properties because by knowing about the physical properties that underlie them we can know what is “really” going on. Ideally we would be able to say why some creatures are conscious and some aren’t by appealing to some physical features of their constitution. Likewise, we could explain our perceptions, thoughts, and experiences by describing how the physical facts are responsible for mental life as we know it. Essentially we would be able to answer all the “why” questions about the mind (such as “why does being angry feel different than being happy?”) by appealing to our theory and the physical facts. And of course if the mental is really the physical then we can make predictions about it too, since we can have no problem in making predictions about physical facts.
In contrast dualism, in both its property and substance forms, can neither predict nor explain mental life. According to dualism even if the mental facts aren’t fundamental and un-analyzable they might as well be, since we have no way, except through our experience, to know about them. And this makes dualism a bad explanation of the mind. If you ask the dualist why animals are conscious and plants aren’t they can’t tell you (assuming they don’t appeal to some supernatural force, like god, to explain who has “souls” and who doesn’t, and I don’t think many would accept this as a good explanation). Similarly, they can’t tell you how or why thinking about dogs is different than thinking about cats, all the dualist can say is that one experiences the thoughts as different. And, despite the fact that we can make observations about our own mental life, dualism can’t hope to make predictions about the mental either. To make predictions requires the ability to formulate general laws (as in: thoughts of type X will be followed by type Y). But, since the mental properties are basically un-analyzable, it is hard to tell how we would be able to determine if a thought fell within one category or another. It is true that upon experiencing the thought we would know, but it is hard to piece together laws from such observations. We need to know what features of type X make type Y follow it, so that we can speculate as to why those features have that causal effect.
Of course we might feel that the predictability problem could be resolved by the right kind of epiphenomenalism, which states that for any physical state there is only one possible accompanying mental state, even though the mental properties are separate from the physical properties. And thus we might reason that since we can predict future physical states we could also determine the accompanying future mental states, and thus predicting them indirectly. Unfortunately there is a flaw with this plan, which is that epiphenomenalism doesn’t give us a way to determine which mental properties will be connected with which physical state. Perhaps our version of epiphenomenalism could borrow part of a materialist theory, and where the materialist says that certain physical configurations are certain mental properties the epiphenomenalist could say that they ensure that those mental properties accompany them. But the epiphenomenalist cannot really adopt the materialist’s theory and change the wording, because the epiphenomenalist doesn’t have a reason to think that the materialist theory they are adopting should work. Why should certain physical states give rise to certain mental ones? The epiphenomenalist cannot say, because they have added a gap to the materialist theory without a way to fill it, so even if they can make predictions they can’t say why those predictions are likely to be correct.
There are those who think that the mind cannot be understood. Certainly I can’t prove that it can be, but to say that understanding it is impossible without a very good reason seems like simply giving up. And if we are attempting to understand the mind then materialism certainly looks like the best way to go.
1.4: Argument From Rationality
The second argument for materialism, the argument from rationality, attempts to show that the only theory about the mind that we can endorse rationally is a materialist one. However this argument rests on two principles about what can be rationally endorsed, which not everyone may share. One is that we should only believe to be true statements that are supported by reasons (either by evidence or deduction). And secondly, that when several mutually exclusive alternatives are supported equally the easiest to falsify should be adopted (or, in other words, the one that makes the strongest claims, or the simplest; all three criteria usually support the same choice). Obviously these principles are behind the “scientific” worldview, but not everyone may accept them, thinking that some things may be known by intuition, or “revelation”, although I would disagree.
However we can allay some of our worries about these principles examining the reasons to accept them as guides for thought. After all there are many possible ways of reasoning, labeling this one as rational doesn’t necessarily make it the best. It might be possible to construct a defense of rationalism based on some other, more universal principles, but I won’t attempt to do so here, instead I will simply point out history has shown that such “rational” methods of approaching problems have always led to the best results, the most promising theories, the most reliable technologies. Other approaches may have had more followers than the principles outlined above, for example the method of believing whatever you have been told (dogmatism), but none of them have come anywhere close to them in achieving real results. Of course this doesn’t prove that these rational principles are the universally best way of reasoning right off the bat, but given that we accept it within some domains we can ask the question: “are these domains all that these rational principles can be applied to?” Since there is no rational reason to believe that they are so limited we should be compelled to accept that it is a universally good strategy, until proven otherwise.
So, given then these rational principles, and reason to think they are universally applicable, we should be naturally led to ask: is it possible to reconcile rationalism with the idea that something non-material exists? Another way to put this idea might be to ask: is there any observation where a non-material explanation will be better than a material one? Obviously the answer to that question must be no, because generally it is impossible to falsify non-material explanations for phenomena, but it is easy (or at least easier) to falsify materialist explanations, since one expects material causes to be observable. For example, the thesis that tiny creatures cause disease is more rational than the idea that disease is caused by evil spirits; the tiny creatures hypothesis could be disproved with the correct observations. This holds true even if one currently can’t build a microscope, because could eventually hope to construct one, but never an evil-spirits detector.
Of course when considering explanations about past events considerations of what is rational become trickier. Certainly we can still use our criterion of being supported by reasons to eliminate some bad theories, for example the hypothesis that aliens built the pyramids is supported by evidence to a smaller degree than the theory that slaves built the pyramids. In some ways we might also still be able to use the criterion of falsification, since surely more information will be revealed to us by archeological finds, and this new evidence will have the possibility of falsifying some theories (for example, evidence of alien graffiti). However, let us assume that we are in a position of having all the evidence and that the evidence equally supports two theories. In this case I would have to say that neither is more rational than the other, which doesn’t invalidate our rationalist requirement, but does illustrate how much harder it is to acquire knowledge when new experiments can’t be done to test hypotheses.
The reason I bring up considerations of past events because it is possible to form a hypothesis to the effect that something non-material existed in the past, and not have this hypothesis be defeated simply on concerns of testability. However, as mentioned above, the historical data will support different hypotheses to different degrees, and thus the non-material hypothesis will be passed over in favor of material ones (Since Evidence A can only support explanation B if it is known that B is the most likely cause of evidence A. The non-material however will never be thought of as the most likely cause, because it must be, by its nature, unpredictable, as argued in 1.1, and thus we would never have reason to believe that it was the most likely cause. If the non-material could be reliably known to be the cause of certain events that would make it simply another part of the material world, as mentioned above.). And since the non-material will never be supported by rationalism, either as an explanation of phenomena in general or past events in particular, we can conclude that rationalism, or at least this form of rationalism, leads avoidably to materialism.
1.5: Argument From Causal Closure
The arguments from explanation and rationality, however convincing, rely on some premises that may not be universally accepted. To address this problem we have the third argument, the argument from causal closure. This argument attempts to prove that the universe is causally closed, and thus, if the mind has a real effect on our actions (rejecting epiphenomenalism, see section 1.5.1 below), that it must be material. Of course few arguments proceed without any assumptions, and this argument rests on three. However, unlike the previous arguments for materialism I think that these assumptions would be universally accepted.
The first assumption is that what is material can be equated with what is observable. Of course I am not using observable in its customary loose sense here; I am using observability to indicate something that can be studied on the basis of its effects (because you could build a machine to detect its effects and thus “observe” it via the machine). Hopefully this definition of what is material doesn’t seem too far-fetched, since it is essentially the definition I introduced in 1.1. [*] The second is that we only have reason to say that something, X, is the cause of some Q (either a necessary or sufficient cause), only if we have reason to believe that the absence of X alone can prevent Q from occurring in some situations, or if X by itself can bring about Q. Finally, the third assumption is that any possible unobservables still cause events in a law-like fashion (i.e. if some particular unobservable is the cause of some effect that same unobservable will always cause that same effect (or will cause that effect with some fixed probability), assuming it is in the same internal state and in the same circumstances). Perhaps that may not be the case for all unobservables, but if the mind is one of them it certainly is law-like, since our will has a very regular effect on our actions (if it wasn’t law-like the consequences of willing a certain result would be effectively random).
Now, if it is indeed the case that the observable world is causally closed it must be the case the all the causes of something observable are themselves observable (this is simply what it means to be causally closed). This should seem intuitively plausible, since if we observe effects it seems reasonable to suppose that we can deduce information about the causes, making the causes observable (in the sense of observable introduced above). This is, of course, not to say that the causes are necessarily observable with our current level of technology, simply that they are observable in principle. For example, we can observe atoms through a scanning tunneling microscope by observing their effects on a stream of electrons; atoms did not suddenly become observable when the scanning tunneling microscope was invented, rather it was the case that they were observable but we simply lacked the tools to directly investigate them.
We can prove this intuition by examining all the possible cases of how some observable phenomena may be caused, and showing that in all cases the causes must all be observable. Let us say then that some observable phenomena, X, can be caused by two factors, A and B, either independently or in combination. If we observe X then we can only deduce that A or B (or both) are the cause, and this is not enough to consider A and B by themselves to be observable (since as things stand we wouldn’t have the ability to determine which of the two was present). However, if A and B are observably different they must differ in some observable property. And if they differ in some observable property that must mean that they can cause different observable effects, and thus we could set up additional tests to determine which of the factors is a cause of X, making them observable. Given this there are only two possibilities if the observable world is not to be completely causally closed. One is that X is caused by two different factors, A and B, which differ only unobservably. This claim, however, is essentially saying that only the observable properties that A and B have in common are responsible for X, and thus the observable phenomena are still causally closed. (A possible rejoinder to this case is to suppose that there is some third factor, C, that shares the same observable properties as A and B, but differs in an unobservable property, and isn’t a cause of X. However, this is a contradiction, because by observing X we would know that A and B didn’t have the unobservable property that differentiate them from C, and thus the property isn’t unobservable after all.) The other possibility is that there are two factors that can cause X, A and B, and that while A is observable B is unobservable. However, to avoid the pitfalls outlined above, we further stipulate that whenever B is the cause of X A is as well (perhaps because of some other factor). If this was a possible situation then X would indeed have an unobservable cause, but this situation too hides a contradiction, based on our assumption of what it means to say that one thing can cause another. B can neither bring about X by itself, nor can its absence alone be known to prevent X, since we stipulated that it is necessarily accompanied by A. Thus B isn’t really a cause at all (or at least we don’t have grounds to say that B is a cause). Again, the observable phenomena are completely causally closed.
Since this exhausts all the possibilities (as our considerations of two causes can be generalized to any number of causes) we can conclude, by our definitions of observable and cause, that the observable world must be completely causally closed. And so the material world is completely causally closed, given our identification of observable with material. Some may object, saying that there are no material causes for observable conscious states, and hence that we can’t claim the observable is identical with material. This would be a problem if we allowed all conscious experience to count as observable, however, as presented here, we can take as observable to mean only those phenomena observed with our outward-directed senses, and the causal closure holds just as well over that domain. And from that we could deduce by observing other people that our conscious states really do have a basis in the material / observable, or that they are unobservable by our outward-directed senses but have no causal powers (they are epiphenomenal). And if we reject epiphenomenalism (see below) then we can conclude that the mind really is identical with some material process.
1.5.1 Argument Against Epiphenomenalism
But since the strength of the causal closure argument rests partly on the rejection of epiphenomenalism I must provide reasons to make that rejection, in order to complete the argument.
At first it seems that we have an obvious rebuttal: if our consciousness doesn’t have a real causal effect then how can we explain our experience of being conscious? Surely it seems like our will or desire to do something has an effect on the world. It would be astronomically improbable to suppose that our minds and the physical world simply happened to align by coincidence, especially if you aren’t willing to claim that there is some super-natural causality-violating force such as god to ensure the proper alignment.
Epiphenomenalism does have a possible response to this objection. If we adopt the hypothesis that consciousness is a causally ineffective residue of the brain then it is possible that our desire to act in a certain way is created by the brain’s physical processes, the very same processes that caused the action to occur. Our previous argument against epiphenomenalism assumed that consciousness and the brain were running on basically their own courses, so let us now assume instead that the conscious state of a person at a given moment is a by-product solely of their physical state at that moment, thus eliminating the possibility of mismatches. (It is also a corollary of epiphenomenalism and the above statement that consciousness has no real causal powers at all, not even on subsequent conscious states.)
But if this is the case we should still wonder why we are conscious at all. Certainly there is no evolutionary reason that creatures whose minds generated this conscious residue would be favored over those who had minds that did not, since this residue has no causal effect. Thus, since creatures that don’t generate that residue are presumably simpler we would conclude that it is more likely that we wouldn’t have evolved to generate such a reside. And this means that even if we were sure of our own consciousness the most likely possibility is that we are a fluke, and that no one else is conscious in this way. This is worse than being unable to resolve the problem of other minds (unable to provide reasons to believe that other people are conscious); it gives us a reason to believe that other people aren’t conscious. And thus I think we should reject epiphenomenalism, since our belief that the mind is non-material should surely be weaker than our belief that other people have minds.
1.6 Arguments Against Materialism
But, just as there are several arguments for materialism, there are also arguments against it. Before I proceed further I think it is necessary to refute, or at least defuse these arguments, in order to show that there aren’t any hidden pitfalls that will prevent us from developing a successful materialist theory.
1.6.1: The Argument From Gaps
The argument from gaps is not really a single argument, but rather a collection of arguments, all of which attempt to show that there is some feature of the mind that materialist theories cannot explain. The most common candidates are intentionality, long held to be the defining feature of the mental, and qualia, which seem to be a central part of our subjective experience. I think the best way to refute this objection is simply to show how a materialist theory can explain those features, which I shall do in sections 3 and 4. So let me set aside this objection until then.
1.6.2: The Modal Argument
A more fundamental objection to materialism is the modal argument, developed by Kripke, which attempts to show that the mind and the material world are necessarily different things. His argument rests on the existence of rigid designators, such that when x and y are rigid designators x=y -> N(x=y), where N stands for the necessity operator. (In other words if x and y are identical then they are necessarily identical; i.e. identical in all possible worlds). Based on this Kripke then argues that since the mind and the brain are not necessarily identical, because we can conceive of minds existing without brains, the mind is not identical to the brain in this world (~N(x=y) -> x≠y). There are many possible responses to this claim. One is to argue that the conceivability of minds without brains doesn’t show that it is actually possible. There is also another, stronger, argument against the conclusion, which is to maintain that the mind is not a rigid designator. And both of these arguments can be seen to be rooted in Kripke’s own analysis of rigid designators.
Let me first tackle the argument that the conceivability of x≠y doesn’t imply ~N(x=y). The best way to see this is to consider two different rigid designators that refer to the same person, let us say Samuel Clemmons and Mark Twain. Certainly it is conceivable that they were different people, and I am sure that some people really don’t know that they are the same person. However they are rigid designators, and Samuel Clemmons = Mark Twain, so N(Samuel Clemmons = Mark Twain). Really all that conceivability shows us is that we simply don’t know enough about the rigid designators to say for sure that they are identical, or are simply thinking about them in the wrong way (if we mistakenly say that Samuel Clemmons was possibly someone other than Mark Twain). It is easy to see where the mistaken belief that conceivability of non-identity implies that necessary identity is impossible comes from. Specifically it comes from the counterpart function (which I introduced here). When we say that N(x=y) we really mean that for all w member of W(C(w,x) = C(w,y)). Therefore if we can think of a world in which the counterparts of two rigid designators aren’t identical then the rigid designators aren’t necessarily identical. The problem with this lies in the fact that when we consider counterparts for some rigid designator, about which we don’t know all the facts, we are really only considering objects in other worlds that meet some limited description that we have associated with the rigid designator. Thus if we don’t know much about Samuel Clemmons and Mark Twain we might think that surely there could be counterparts of Samuel Clemmons that weren’t the same as Mark Twain (one man has one name and a different man has the other). What we are missing is that the counterpart function operates on the objects themselves, not on the rigid designators, and thus even if we are considering two different rigid designators if they refer to the same object then they will have the same counterparts. But this means that in order not to make mistaken claims about the necessity of identity we must first determine if two rigid designators refer to the same object in this world. And this means that to settle the question of whether the mind is necessarily identical to the brain we must settle it in this world first. And appeals to conceivability and necessity can’t hold any weight without the analysis of our world first, as the example of Samuel Clemmons and Mark Twain shows.
But perhaps we place more weight on our ability to conceive of minds without brains. Perhaps we think that anything with the right kind of first person point of view is conscious (and hence a mind), and surely there are things that can meet that requirement which aren’t brains. In fact it is possible that it might meet all the requirements to be our consciousness and not be a brain. And I find this argument compelling. But if we are to accept it then we must accept that consciousness is not a rigid designator, but is instead a definite description. Why? Well because I can certainly also conceive of worlds in which brains are conscious, and in those worlds they would be identical to minds. But if they are identical to the mind in some worlds, but not in all worlds, then they aren’t necessarily identical to minds. And, as Kripke himself points out, the only way this is possible is if “the mind” is some kind of definite description, and not a rigid designator. And if “mind” is really a definite description then Kripke’s argument falls apart, since the necessity of minds being identical with brains has nothing to do with whether they are identical to brains in this world.
So it seems that either way we go Kripke’s argument against mind-brain identity is flawed. If the mind and brain are rigid designators then conceivability can’t tell us whether they are necessarily identical or not. And if we do think that conceivability can tell us whether they are necessarily identical or not then we must accept that the mind isn’t a rigid designator, and thus the argument can’t proceed.
2: Internalism / Externalism
Already then we have narrowed our focus, from every possible theory about the mind to materialist ones. But we must narrow our focus still further if we are to address the ability of materialist theories about the mind to bridge certain explanatory gaps. Specifically we must come down on one side of the internalism / externalism debate. When I refer to internalism here I mean specifically the class of theories that hold that the mind / mental properties depend only on physical facts about the brain (and possibly some other parts of the nervous system as well). In contrast, externalism is the class of theories which hold that the mind / mental properties depend on both physical facts about the brain as well as physical facts about the external world as well.
2.1 The Argument For Externalism
The arguments for externalism are usually derived from theories about content and reference. In the philosophy of language there is a strong movement to define reference, and thus content, partly in terms of the external. For example, Hillary Putnam’s famous twin Earth thought experiment is supposed to show that the content of “water” depends at least partly on what water really is. And if content is defined on the basis of external facts then it seems natural to extend this idea to the mind as well. After all the content of our thoughts is seemingly part of our consciousness, as when I am thinking about a tree I am aware that my thoughts are about trees. And thus if the content of a thought depends on external facts it seems only natural to say that the mind, which the thought is part of, depends on those external facts as well.
2.2: Is Externalism Compatible With Materialism?
It is possible to argue against externalism in two ways. One is to argue against its foundations in the theory of language. If it could be established that reference depended only on internal facts then externalism would be left foundationless. This is not, however, how I will argue against externalism here, since to thoroughly address the problem of reference would require an sequence of papers. Instead I will peruse the other avenue of argument against externalism, which is to show that a mind that depended on external facts would be epiphenomenal (have no influence on behavior). But if the mind is material it cannot be epiphenomenal (since material things always have causal effects), which contradicts the premise that we are developing a materialist theory of the mind, and thus demonstrating that no externalist theory can be a materialist theory.
One way to show that an externalist theory implies that the mind is epiphenomenal is simply to invoke the principle of causal locality. The principle of causal locality states, quite simply, that given an effect, E, the only things that could have been immediate causes of E were in the immediate region. (More precisely, only things contained within a sphere of radius t * c can have had a causal effect on the outcome, where t is the amount of time before that outcome and c is the speed of light.) Obviously the principle of causal locality can be violated in quantum physics, but that only happens when very small particles become entangled, and neither the brain nor the objects it is thinking about are small enough to make this a possibility. And, in addition, these non-local quantum effects happen only once, at which point the particles are unengaged, not over and over again as would be required to explain our ability to think of the same object repeatedly. From the principle of causal locality it is obvious that only the physical properties local to the signals sent out to cause behavior can have an effect on that behavior. This means that distant facts are epiphenomenal to this behavior, and thus that a mind that depended on those facts would be epiphenomenal as well.
There is, however, one way to avoid this problem in an externalist theory, proposed by Dretske, which is to argue that the external facts that the mind depends on are not co-temporal with us, but rather are historical facts, such as the fact that we interacted with a certain object in the past, or the fact that our ancestors evolved in a certain kind of environment. But this too is a problematic solution. Consider a perfect duplicate of some person formed spontaneously (by some very unlikely coincidence). Since this person is a perfect duplicate they will behave in the same way as the “original”, but will share none of the historical facts that the mind as supposed to depend on. An externalist might be tempted to then assert that behavior is over-determined, both by the external facts and local state. However, the assertion that behavior is over-determined in this way would require that if the historical facts were left constant and the internal state was disrupted that behavior would remain unchanged (since the historical facts are also causing that particular behavior). But this is contrary to fact; we know that brain damage, which leaves historical facts unchanged, has an effect on behavior. And thus the historical facts cannot be said to cause behavior (as they are neither necessary nor sufficient causes), at best they can be said to be the cause of the internal state, and that it is really the internal state that is responsible for behavior. Again, this version of externalism is only consistent if we accept that the mind is epiphenomenal.
2.3 Externalism About Reference
But if externalism isn’t a viable theory about the mind what does this mean for the theories of reference that it was based upon? It might seem that a rejection of externalism would imply that those theories too should be discarded. And if we insist that the facts about reference must play some role in the mind then this is the conclusion we must accept. However, it is perfectly consistent to believe that the facts about reference, which are determined by the contents of the world to some extent, are not part of the mind. All this would mean is that when I am thinking of a tree whether my though really makes reference to the tree is not knowledge that is immediately and infallibly accessible to me. Surely this isn’t too drastic a move; it isn’t hard to accept the similar theses that there are mathematical facts such as “2+2=4” that don’t have a causal effect on the world, or that there are ethical facts that are similarly causally ineffective. It would simply mean viewing the facts about reference as a description of the world rather than a principle by which it must operate.