Friday, November 11, 2011

Aristotle and Wittgenstein

I’m afraid that some readers will be growing impatient as they read the foregoing discussion of a kind of Platonic resolution to the problem of rationality. Hadn’t I just, in the first half of this same chapter, argued for an operational theory of intentional predicates? Not only that, but when one suggests that the non-physical property of “meaning” can be washed out of the ontology of mind and language (replaced with an externalist account of intentional predicates as describing relations between persons and environments), that would be about as nominalist as one could go, surely? Maybe not. The Platonism that I am offering has only one element of basic ontology besides matter. Form is indivisible, not divisible; a unity, not a multiplicity. There is only one form really: only one (perhaps inexplicable) ontological fact beyond the fact of the existence of something rather than nothing. Putting the question of Plato and Aristotle’s own views of species as “fixed natural kinds” to the side in favor of a view of species informed by evolutionary biology, it can be seen that putative “forms” such as the property of “cowness” or “lyrehood” are not genuine examples of form. Some categories (types of species, types of artifacts) come-to-be and pass away. In its Aristotelian version Platonic form-matter dualism becomes a kind of non-reductive materialism: primary being is substance, the unity of form and matter. From the doctrine of the unity of form, though, it appears that this must be a kind of “non-reductive formalism” as well, as every particular with a formal property has that property, not by virtue only of the formally-organized parts of that particular, but by virtue of the entire formal organization of the material world: all geometric shapes (for example) are tokens of the one thing. We know that by this point Wittgenstein would be fuming, but as usual with him we might not be certain exactly why. Of course Wittgenstein would have none of this Platonic talk. “The idea that in order to get clear about the meaning of a general term one had to find the common element in all its applications has shackled philosophical investigation; for it has not only led to no result, but also made the philosopher dismiss as irrelevant the concrete cases, which alone could have helped him to understand the usage of the general term. When Socrates asks the question, ‘what is knowledge?’ he does not even regard it as a preliminary answer to enumerate cases of knowledge.” (Blue and Brown Books; italics in original). Wittgenstein’s operationalist account of functional-role semantics is an arch-nominalist position: there is human behavior, a highly-adaptive and plastic process that changes over time, whose constants are determined by the biological (probably the best choice) nature of the human body and the human “mode of life.” “Property” names (like all names) really pick out parts of language, and the criteria for the proper application of language are essentially operational. Insofar as this line is developed as a strategy to naturalize meaning I think it’s a good one. But I have never thought that philosophy of mathematics was a particularly strong point for empiricists, and that is troubling considering that Wittgenstein devoted a considerable portion of his writings to the development of an operationalist theory of mathematics. In any event I am unpersuaded by Wittgenstein’s view that extending the known proofs of mathematics is nothing more than an elaboration of a kind of “language game,” specific to humans by virtue of our particular “form of life,” such that there was no such system of entailments until some human (for example) elaborated it. It’s counterintuitive: isn’t the fact, that we can work our way from one part of mathematics to another, evidence that mathematical reasoning is coherent? Doesn’t Wittgenstein’s ultra-nominalist view of mathematics overstate the possibility space: the different ways “mathematics” could go? However, it may be that the two treatments of the two different parts of intentionality - an eliminativist, operationalist argument to the effect that mental representation/content is not part of the reference of intentional predicates, on the one hand, and an Aristotelean argument to the effect that rationality is nothing more nor less than a formal property and that formal properties, if they exist at all, are ubiquitous – are compatible. According to Wittgenstein there are no abstract entities, of course, but it is important to appreciate how far Wittgenstein went in his naturalization of meaning, and how central to this were his ideas about mathematics. Wittgenstein saw mathematical behavior as a “technique,” a technique for living. “Living” is the operational verb that replaces the Cartesian verb “knowing”: a case of knowing how rather than knowing that. Wittgenstein rejected the passivity of the representational theory and insisted on viewing language as a physical behavior that aimed at getting on with the business of life. Granted that the Aristotelean world is one where every concrete particular is a union of matter and form, the “form of life,” understood as the vital activities of a being of that kind, would exhibit formal properties. In fact “behavioral ecology” develops an entire narrative, largely mathematical, about the ratio of nutrients per square meter to species population per square meter, showing the correlations between these functions and genetic transmission and so forth. The human “form of life,” if it is anything at all, is a product of the same natural history as that of the human organism; the rationality of humans, like the harmony of musical instruments, is an expression of form. Within this form of life, that stress made no more emphatically by anyone than Wittgenstein himself, the criteria for use of psychological predicates can be understood operationally such that no mental content is implied. In fact Wittgenstein and Aristotle come together in a sense around “form of life” or what Aristotle would call the telos of an organism. They both suspected that explanations about what sort of thing a thing was and what sort of life a thing led were more informative than explanations about what sort of things a thing thought. Wittgenstein thought that the notion of mental content made no sense. I take the argument from the form-matter distinction to show that “computation” need not necessarily entail mental representations; organizational complexity equivalent to the syntactical complexity of language is found throughout nature. Finally, Wittgenstein’s functional-role semantics and Aristotle’s teleological account of biological explanation are very similarly motivated. They come together in the area where functionalism replaces reductive materialism as a response to the supervenient nature of the functional property.

Sunday, October 9, 2011

The Mereological Fallacy and representational Theories of Mind

Stomachs don’t eat lunch. Eating lunch is something that a whole, embodied person does. We understand the role that stomachs play in the lunch-eating process; we appreciate that people can’t eat lunch without them. Brains don’t think. They don’t learn, imagine, solve problems, calculate, dream, remember, hallucinate or perceive. To think that they do is to commit the same fallacy as someone who thought that people can eat lunch because they have little people inside them (stomachs) that eat lunch. This is the mereological fallacy: the fallacy of confusing the part with the whole (or of confusing the function of the part with the telos, or aim, of the whole, as Aristotle, who as usual beat us to the crux of the problem, would say). Nor is the homunculus a useful explanatory device in either case. When I am asked how we might explain the workings of the mind without recourse to mental representations (students often ask this), the reply is that we fail to explain anything at all about the workings of the mind with them. “Remembering my mother’s face is achieved by inspecting a representation of her face in my mind.” This is explanatorily vacuous. And if reference to representations does nothing to explain dreaming, imagining and remembering, it is particularly egregious when mental content is appealed to for an explanation of perception itself, the original “Cartesian” mistake from which all of the other problems derive.

    A person is constantly developing and revising an idea of his or her world; you can call it a “picture” if you like (a “worldview”), but that is figurative language. A person does not have a picture inside his or her body. Brains don’t form ideas about the world. That’s the kind of thing people do. This original Cartesian error continues to infest contemporary cognitive science. When the brain areas in the left hemisphere correlated with understanding speech light up and one says, “This is where speech comprehension is occurring,” the mereological fallacy is alive and well. Speech comprehension is not something that occurs inside the body. Persons comprehend speech, and they do it out in the “external” world (the only world there is).

   Positing representations that exist inside the body is an instance of the mereological fallacy, and it is so necessarily, by virtue of the communicative element that is part of the definition of “representation,” “symbol” etc. Neither any part of the brain nor the brain or nervous system considered as a whole interprets anything. The key to developing a natural semantic of intentional predicates is to realize that they are predicated of persons, whole embodied beings functioning in relation to a larger environment. Brain/body dualism can be presented as non-dualist (isn’t the brain a physical organ of the body?), but it is an insidiously Cartesian view that gets us no farther in naturalizing intentional predicates.

   Suppose that you are driving down the freeway searching for your exit, and you’re worried you might have passed it. You remember that there are some fast-food restaurants at the exit, and you think that one always feels that they have gone too far in these situations, so you press on, keeping an eye out for the restaurants. However you manage to do this, it is no explanation to say that you have done it because your brain remembered the fast-food restaurants, and has beliefs about the phenomenology of being lost on the freeway, and decided to keep going and so forth. That’s like saying that the way you had lunch was that your stomach had lunch.

  This realization may also be momentous for brain science. Go to the medical school bookstore, find the neurophysiology textbooks and spend a few minutes perusing them. Within the first minutes you will find references to the “movement of information” (for example by the spinal column), “maps” (for example on the surface of the cortex), “information processing” (for example by the retina and in the visual cortex) and so on. (Actually my impression is that brain scientists are relatively sophisticated in their understanding of the figurative nature of this kind of language compared to workers in other areas of cognitive science; the point is just that representational talk does indeed saturate the professional literature through and through.) But if brain function does not involve representations then we don’t know what brains actually do, and furthermore the representational paradigm is an obstacle to finding out: think of all those experimentalists developing protocols to try to “locate the symbolic architecture.” They might be looking for something that isn’t there. If there is any possibility that this is true these arguments need to be thoroughly explored at the very least.

  Taking the argument from the mereological fallacy seriously also draws our attention to the nature of persons. It follows from what has been said that the definition of “person” will be operational. Operational definitions have an inevitably circular character: a person is any being that takes intentional predicates. In fact there is not a “machine-language” explanation of personhood. Kant, writing in the late 1700s, is fastidious about referring to “all rational beings,” he never says “human beings”; he understands that when we are discussing the property of personhood we are discussing (what I would call) a supervenient functional property (Kant would call personhood “transcendental”), not a contingent physical property. However Kant is programmatically intent on limiting the scope of materialism as such and thus fails to develop non-reductive materialism. Instead he imports the mental (“reason”) from the noumenal world and ignores the problem of the relationship between transcendental reason and the human body (this is not to say that he does not acknowledge the role of our particular, contingent sense organs in shaping our representations of the world to the extent that those representations are themselves contingent and particular to us).

  With Kant we remain in our bodies but not of them. Once one recognizes that intentional predicates are predicated of whole persons – once one sees that positing mental representations necessarily commits the mereological fallacy – the question of representation is settled. It is I, and not some “brain state,” that is remembering my mother’s face. However there is a tight network of arguments and assumptions, centered on a model of intentional states as “propositional attitudes,” that will have to be disentangled to the satisfaction of readers who are disposed to defend representations. After that unpacking is done the reader will also reasonably expect some account of a non-representational analysis of intentional predicates, something that is not achieved by simply pointing out the mereological fallacy.

Sunday, June 19, 2011

A Short Introduction to the Philosophy of Artificial Intelligence

This is a rough draft, I participated in an interdisciplinary class and I'm thinking of submitting to maybe "Teaching Philosophy."

I. The historical background

AI is not only a rich source of new technology produced by interdisciplinary syntheses. It also, in its theoretical component, is an extension and elaboration of some of the central, canonical debates about “intelligence,” “mind” and “rationality” that have defined philosophy and psychology for hundreds of years. Specifically we find ourselves participating in the conversation that dates back to the “Early Modern” period of philosophy, roughly the 17th and 18th centuries, between so-called “Rationalists” (Descartes, Spinoza, Kant) and so-called “Empiricists” (Locke, Berkeley, Hume). The Rationalists, impressed by humans’ apparently unique ability to formalize mathematics and logic, held that the human mind was endowed with innate abilities and knowledge, and that these abilities could not be understood using the methods of natural science (these views were anticipated by Plato). The Empiricists of the 18th century Enlightenment, eager to develop a naturalistic account that integrated humans into nature, proposed a simplified psychology that essentially saw the mind as a learning machine and concentrated on perceptual psychology and learning theory. (Nowadays historians of philosophy tend to see the Rationalist/Empiricist distinction as a bit overstated, as we can see in perspective that they were all discussing the same set of issues with many of the same premises.)

An important product of this Early Modern discussion, introduced by Descartes in the first half of the 1600s (Descartes 1637) but crystallized by Kant at the end of the 1700s (Kant 1789) was the representational theory of mind. According to this view the mind works by constructing a representation of the world; Kant developed the idea of a

“conceptual framework” such that our “picture” of the world was as much a product of our own innate mental structure as it was of our perceptual experiences. Thus the issue of mental representation is an essential issue in the elaboration of the nativist/learning theorist divide as it plays out across the 19th and 20th centuries. For example, the behaviorists of the early 20th century are nothing neither more nor less than Humean empiricists: they applied “operationalist” ideas from the philosophy of science to try to develop a psychology that was cleansed of any reference to unobservable, “internal” mental “states,” including representations (mental content). On the other side the phenomenologists of the same period advanced the thoroughly Kantian argument that the study of the structure of experience would always necessarily stand apart from physical science. (Here we can stop and notice an even deeper root: the medieval question of the duality of the body and the soul.) In the middle of the 20th century the “nature/nurture” debate, as this same set of issues was then called, was of central importance in debates about the social sciences in general, a central battleground of the “culture wars” of the 1960s and 1970s. The nativist/learning theory divide also shaped the 20th century ethological literature about the mental lives of non-human animals.

II. Computation and representation

The issue of representation is central to contemporary debates about models of computation. In fact the theory of computation is yet another version of the same argument that constitutes the theory of the social sciences and the theory of ethology. Alan Turing in 1936 introduced his “Turing machine,” a thought-experiment that showed that a simple machine could instantiate any algorithm of mathematics and logic. This was a seminal moment not only in the development of computers but also in the course of artificial intelligence research. For the next fifty years many in the cognitive science community and the public at large saw “artificial intelligence” as just synonymous with computer science. Two crucial points here: first, to understand what is happening in artificial intelligence research today it is necessary to understand the computationalist era, because what we are currently living through is a departure from that era. Second, computationalism, as conceived by Turing and others, required representation: classical computation is rule-governed symbol-manipulation.

At this point we can consider some basic premises of linguistics. The classical computationalist view reached its apotheosis in 1975 with the publication of Jerry Fodor’s The Language of Thought. Noam Chomsky had launched what seemed for a time a devastating attack on behaviorism with his critique of B. F. Skinner’s 1957 book Verbal Behavior and Chomsky’s subsequent Aspects of a Theory of Syntax (1965). Chomsky argued that a syntactical structure (a grammar, or set of rules for constructing sentences and statements) was generative (it could generate novel linguistic representations and therefore novel thoughts), and was thus necessary for higher-order thought (this argument led to the sign-language research with chimpanzees of the 1960s-80s). This was, as Chomsky himself stressed, Cartesianism in a new bottle.

Fodor applied these ideas to cognitive science in general. Any representational theory of mind requires a symbolic architecture: this is simply the material instantiation of the symbols: the pixels in the computer screen, the ink marks on the page, the sound-compression waves caused by vibrating vocal chords, the chalk marks on the board. If the nervous system is a symbol-manipulating system then there must be a material instantiation of the symbols as part of the physical structure of the system. Fodor proposed that syntactical structure (the program, if you will, of the brain) could account for the causal role of the seemingly semantic mental content. This arch-computationalist view took it as axiomatic that the mind/brain necessarily involved representations.

III. Computers and the brain

Computers are our own creations, so their workings are not mysterious to us. The same thing cannot be said of the brain. Each age draws on the current technology as a metaphor/theory about how the brain works: the 17th century physicalist Thomas Hobbes, for example, drew heavily on hydraulics in his discussion of the mind. He speculated that memory might be a kind of vibration, as in a spring, that lost coherence as other vibrations passed through. In our time it is commonplace to speculate that the brain is a kind of computer and that a computer is a kind of a brain. However there are two very different approaches to developing this idea.

Classical computation is based on codes (programming languages) that contain explicit instructions for the transformation of states of the machine. The actual “machine language” is binary code (this is the meaning of “digital”). The symbolic architecture in a traditional computer is located in the “chip.” This is a series of gates that might either allow or block an electrical impulse to pass through. Thus the “1s” and “0s” of digital codes stand for actual physical states of the machine. If the human brain is also a system that functions through instantiating representation than the goal of cognitive science is to uncover the machine language of the brain: to make the connection between the psychological description of the subject and the actual physical state of the nervous system.

The brain does, in fact, possess physical features that lend themselves to a theory of symbolic architecture similar to that found in digital computers. The brain is a massive assemblage of individual neurons that interact with each other through the flow of electrical impulses (“cascades”). The impulses do not pass arbitrarily, of course; the brain shows immense organizational complexity. But essentially one neuron or group of neurons will, upon being “lit up” by a cascade of electricity, either send the event onward to the downstream neurons of fail to do so, and this can be seen as the “1/0” analog. What’s more, between neurons there is a space, the synaptic cleft, which contains a soup of neurotransmitters that buffer the electrical connection (they can be more or less conductive). So instead of an “on/off” potential, like a light switch, there is a gradient potential, like a volume control. This vastly increases the potential number of physical states of which the brain is capable. All of this constitutes a non-arbitrary reason for thinking that the brain may indeed function like a traditional computer: the synaptic pattern could be the symbolic architecture of the brain just as the disposition of the gates in the chips is the symbolic architecture of the computer.

However a new generation of computer models now challenges classical computation and its axiom that representation is necessary for computation. In this new generation of research, computers are actually modeled on brains while at the same time the new computers are contributing to new insights into how brains themselves work. This movement is sometimes referred to as “parallel distributed processing” and as the “neural net model,” but it has come to be popularly known as “connectionism.”

Classical computation has some limiting and apparently intractable problems. As anyone who has worked with computers knows, they are insufferably single-minded. This is natural, as they can only do what they are told to do by their programmers; “garbage in, garbage out.” One of the central problems for traditional computers is the “framing problem.” Consider any homonym, for example “bank.” An ordinary human has no trouble during conversation distinguishing between the two senses in sentences like “I was laying on the bank of the river” versus “I made a withdrawal from my bank.” Traditional computers are strictly limited in terms of contextualizing. This is because computers don’t actually know anything. They are devices for manipulating symbols and nothing more.

What’s more, traditional computers can’t learn anything new. They know what they are told. Now, remember the Rationalist/Empiricist debate. The Rationalists thought that there was an innate conceptual structure, incarnate in language, of essentially mathematical and logical principles, and this structure (the mind, or soul) was the source and basis of rational behavior. The Empiricists argued that a naturalistic psychology required that there be nothing more than an ability to learn from experience on the basis of trial and error, and were skeptical of non-physical states and entities. Connectionist computer models are empiricist approaches to computing in the same way that behaviorism is an empiricist approach to psychology. Connectionist machines do indeed show some primitive ability to learn on their own; they function (ideally) with no recourse to internal codes or representations; and they are solidly based on basic principles of evolutionary biology.

Connectionist machines function, as brains do, by forming patterns of activation. An input layer of nodes are electrically stimulated and this layer accordingly stimulates some number of “hidden,” internal layers which ultimately stimulate the output layer. Activation potentials can be weighted in various ways but the basic mechanism is the number of nodal connections which can constitute a threshold for downstream activation:

(Insert figure of simple connectionism: input layer, hidden layer, output layer)

This technology underlies handwriting-, voice- and facial-recognition functions that are now commonplace (an original application was for submarine sonar submarine-recognition and missile-recognition). This is achieved through trial-and-error. A trainer adjusts the activation potentials to increase correct outputs and to extinguish incorrect ones. This process does not require any internal symbolic content.

Here it is useful to note that Darwin’s model of evolution as outcomes-based selection over random variation is very much a product of empiricism. In fact Darwin was reading the Scottish Enlightenment economist Adam Smith’s 1776 Wealth of Nations, with its account of larger economic structures formed from the bottom up through iterations of economic exchanges between autonomous, self-interested individuals when he was developing his account of natural selection (Darwin 1859). An important distinction between the Rationalist program and the Empiricist one is that Rationalists tend to see complex systems as organized from the top down whereas Empiricists see complexity as emerging from the bottom up. The distinction between classical computation and connectionist computing mirrors this distinction.

However the field of AI is moving in even more radical directions. Although modern cognitive scientists will obviously disavow Cartesian dualism about the mind and the body, in a sense the Cartesian model has often been simply transposed into a brain-body distinction. On a common view it is the brain that is (now) the “cognitive theater,” the seat of representations, the CPU where thinking takes place: the same role Descartes assigned to the res cogitans (Hacker). This view underlies the assumption that AI research is simply an extension of computer science. That collective assumption is now collapsing.

IV. Robotics

On a representational model, “beliefs” and other mental states are instantiated in the form of mental content: language, images and so forth “in the head.” As I said, this is recognizably a continuation of a kind of Cartesian dualism. Indeed representational models are essentially dualistic if representations are taken to have semantic properties that are not analyzable as physical properties (this is one of a number of philosophical issues that I went into to some depth in the class). An alternative view is that psychological predicates are predicated not of brains but of whole persons.

Stomachs don’t eat lunch. People eat lunch. True enough that one needs a stomach to eat one’s lunch, but it doesn’t explain how a person eats lunch to say, “Their stomach eats lunch for them.” Brains don’t think. They don’t imagine, dream, solve problems or recognize patterns. People do those things, just as people believe, desire, hope, fear, etc. In fact, committing this mereological fallacy – the fallacy of confusing the part with the whole – obstructed our ability to learn what it is that brains actually do. We were sidetracked by the misconception that brains are little people in our heads.

“Embodied cognition” is the name given to a recent movement in cognitive science that rejects representational models of thought. The idea is that “thinking” is an activity that is distributed over the whole body. This movement has been in a particularly fertile dialectical relationship with robotics. (Not surprisingly this community has developed some excellent internet resources where students can see footage of robots in action.) It is clear enough that the future of AI lies as much with the field of robotics as with the field of computer science. What is important in an interdisciplinary context is to see the underlying, and quite old, philosophical considerations that make that clear. This also presents an opportunity to discuss the history and philosophy of science.

Sunday, May 29, 2011

A little bit of naturalist apologia

We live in a world where most natural phenomena, from the micro level of atoms, cells and molecules up to the macro level of galaxies and the universe itself, seem to be describable and explicable in physical terms. Physicalism (I mean by this term the metaphysical position that only the physical universe exists) is, as I said, by no means triumphant (and it is a reasonable point that contemporary physics itself presents us with a still-mysterious and newly-strange picture of the universe). There are ongoing popular metaphysical arguments about evolutionary biology and about cosmology, for example. But it is a striking fact about contemporary culture that psychology (and by extension the behavioral and social disciplines) are still not considered to be integrated into our otherwise generally physicalist metaphysics. Put another way, while many people today have firmly internalized physicalist intuitions about organic life, say, or about distant celestial objects, physicalist theories of mind still meet with resistance today, even among secular people who have broadly physicalist attitudes.

I’m not someone motivated mostly by ideology. I’ve always been impressed by Socrates’ description (in the Theatetus) of the search for knowledge as the activity of becoming aware of what it is that one truly believes, and then stating that belief, above all to oneself, as clearly and courageously as possible (in fact Socrates is claiming, contra his relativist antagonists, that this is the essential, unavoidable human activity). I’ve had the salutary experience of changing my mind and reversing myself several times during my relationship with philosophy of mind. Now I just want to develop the soundest view of the matter that I can, as one climbs a mountain. One of the worst faults a philosopher can have is the tendency to magical thinking: trying to make a brief for what one wants to be true.

However another couple of paragraphs of self-explanation are warranted. I know this because I have spent the past ten years or so teaching the philosophy of mind to undergraduates at two large state universities. Inevitably this involves, among other things, leading a lay audience to discuss, usually for the first time, the topic of naturalism with regard to human nature and to the mind. Of course there are people who arrive in the classroom already thoroughly naturalistic. But consistently there are people who struggle with this topic for deeper, cultural reasons. Although this discussion is rarely included in books on philosophy of mind, I have found that it is best to present some apologia at the beginning for students who may have some preconceptions that can turn them off, as well as to reassure them that I too think that these are legitimate concerns that can be discussed if they wish.

OK, so here’s an ideological argument: Humans are depressingly alienated from nature. Our relationship with the rest of the biota on this planet is not a good one. Urgent action is necessary to stem climate change, species extinction and other environmental problems that pose grave threats. However we also need longer-term cultural evolution, a change in our attitude towards our relationship with nature, and this change is effected to some extent by cultural workers such as artists, philosophers and writers.

It is my opinion that human exceptionalism, and a lot of bad metaphysics down a lot of centuries that came with it, is one root of our dysfunctional relationship with nature. I think that naturalism about psychology is the most progressive view. I think that naturalism is also the most spiritual view. And it is the healthiest view of human nature. I may be all wrong on all of that. But the reader ought to understand that I concede no quarter of the argument between physicalism about the mind and its alternatives, including discussions in terms of enlightenment, ethics, freedom, spirituality and so forth. Those discussions, however, will form little or no part of this book.

A programmatic point that will be discussed, though, is the importance of clearing up some of the logical and linguistic problems that we continue to have with our concept of “mind” in order to make progress in experimental science. Theory can have a good deal to say about the development of experimental protocols, and good theory will make these implications clear.

Sunday, May 22, 2011

Metaphysics and the philosophy of mind

This is a book about the metaphysics of the mind/body problem. Metaphysics (or “ontology”) is the study of what exists (Aristotle called it the study of “being”). To many people today metaphysics seems anachronistic. Haven’t we settled the issue of what exists, they might ask, in favor of the physical universe? And isn’t natural science the way we produce knowledge about this universe? How could more work in metaphysics possibly generate any persuasive arguments, if “metaphysics” is not simply “physics”? Arguments about the relationship between the mind and the body that aren’t grounded in empirical research of some sort can’t hope to be legitimate in a world awash in data from experimental psychology, neuroscience, computer science, evolutionary biology, linguistics and the myriad of interdisciplinary areas of research that today we call “cognitive studies.” Isn’t a metaphysician a mere poet of speculation? Diverting at best, but such a person has no hope of producing useful knowledge. That, anyway, is often the initial reaction one meets with the topic of the metaphysics of the mind/body problem. I will respond to this initial “meta”-challenge in two ways.

First, I completely embrace the spirit, and much of the letter, of this initial objection. I too take it as axiomatic that what exists is the physical universe (by “physical” I mean the universe of matter and energy, or maybe matter/energy; I don’t pretend to be sophisticated about theoretical physics). I don’t think that humans are composed of physical bodies and non-physical souls, like a traditional mind/body dualist. I think that humans are physical through and through, animals that evolved here on earth through a long process of evolution the contingencies of which were, and continue to be, bounded by the constants of biology, chemistry, and physics. I don’t expect to discover that humans are angels, or that the physical universe is an illusion and humans are non-physical spirits, or anything like that.

The universe is as magical, mysterious and mystical as it may be; I don’t know anything about the ultimate composition or nature of the universe. I have no interest in making a brief for reduction, as if natural science can address every one of our wonders, or even potentially could. I don’t even know what we’re talking about when we use that kind of language. My claim is much humbler: whatever nature in general is like, humans are like that. Humans are not miracles, if a “miracle” is defined as an exception to the laws of nature. Call me an “anti-humanist.” I hold the anti-humanist view simply because I know of no reason to think that humans are miracles; I stress it because a deeply internalized assumption of human exceptionalism continues to be a barrier to progress across the whole range of the behavioral and cognitive sciences.

Which brings me to the second response to the objection that metaphysics is anachronistic: it is certainly not true that the contemporary society of educated people embraces anti-humanism as I just defined it. A great many college students, most people walking down the street and the overwhelming majority of the world’s population today continue to think that the mind is something distinct from the body or, at least, that mental phenomena cannot be adequately described and explained in wholly physical terms. This conviction has various sources that range from traditional, usually religion-based beliefs about souls, afterlives and so forth to more modern notions, such as the view that a naturalistic view of human nature is perniciously reductive and to be resisted by the liberal-minded, or perhaps that science itself is nothing more than a socially-constructed “conceptual scheme” with no particular claim to legitimacy, and so on. For another thing, very sophisticated versions of human exceptionalism exist in the academy today (for example among some linguists), such that it is by no means established conventional wisdom that physical science subsumes psychology by metaphysical axiom.

Metaphysics is not something that is replaced by physics. Physicalism is a particular metaphysical position. Everyone has metaphysical assumptions, articulated or not, whether they want to or not, and they always will. The person who chafes at the idea that there is still a need for explicitly metaphysical discussion is claiming that our shared metaphysical assumptions are currently stable, not that “there is no such thing as metaphysics,” although they may unreflectively put it that way. It’s true that physicalism is currently the ruling metaphysical paradigm among cognitive scientists, psychologists, philosophers and so on, and I too labor within this paradigm, albeit with some important qualifications that are discussed in the second part of Chapter Two.

Sunday, May 15, 2011

Elan Mental

The claim that there is something (the quality of phenomenal experience) that cannot be explained by physical science is strictly analogous to the 19th century “vitalist” claim that the property of being alive could not be explained by physical science (the phrase “élan vital” was actually coined later, in 1907, by Henri Bergson in his book Creative Evolution). Consider all of the physical facts about physical states and processes in the body, the vitalist argued: singularly or together none of these facts entail that the body be alive.

This “hard problem” was never “solved.” It simply faded away as organic chemistry and physiology steadily explicated the physical mechanisms and processes occurring in various parts of cells, and in the various organs of the body. This took some time, well into the 20th century, but by the 1940s, anyway, it was no longer credible to claim that “life” was something that might not be present when these mechanisms and processes of organic chemistry were present, or might be present in their absence. “Life” will always be an ambiguous concept to some extent (there is ongoing debate as to whether viruses are living, for example), because it is an emergent property, but its physical nature is no longer seriously challenged. The concept of “consciousness” is now undergoing the same evolutionary process – not a similar process, the very same process.

This analogy has been prominently rehearsed by Patricia Churchland and by John Searle, among others. I will consider Searle’s version a little more closely by way of setting up the last chapter, where I will discuss the relationship between intentionality and consciousness. Searle makes an analogy between the solidity of a table and the consciousness of a brain: the table’s solidity is a macro-property that emerges from the micro-properties of the wood molecules (which are lattice-like). Consciousness, he suggests, is a macro-property that emerges from the micro-properties of neurons (although he doesn’t claim to know which micro-properties or why).

There are two problems with Searle’s analogy. First, in the case of the wood molecule and the table, they share the same property in the first place: the lattice-like structure of the wood molecule, like a folded piece of paper, just is solid (can bear weight by virtue of its structure). So solidity is not an “emergent macro-property,” solidity is already a property of the “micro” ingredients. If the question is “How can physical objects support weight?” then appeal to the weight-bearing nature of the wood molecule only pushes this question back a step. This problem with the analogy is irremediable: if the argument is that brains are conscious because neurons are conscious we have once again committed the hard-to-avoid error of including something mental in our purported recipe for the mental. If not, then the analogy does not go through: the wood molecules and the table share a property in common, so we do not have an actual example of a macro-property emerging from a micro-property (that is not to say that we couldn’t find such an example, only that this one isn’t it).

The second problem is more serious and to the point of the present discussion. Psychological predicates, as I argued at length in Chapter Two, are not predicated of brains or nervous systems but of whole persons. This goes for consciousness every bit as much as it does for intentionality. Brains no more feel or sense things than they think about or imagine things. Persons think and feel. Asserting this does not exclude me from the club of materialists in any way.

The crucial difference between intentionality and consciousness is that while intentional states are supervenient and therefore unexplainable through reductive materialism, phenomenal states are not supervenient and so a legitimate answer to the question “Why does it feel like that?” is “Because it is that specific physical body interacting with that specific physical feature of the environment (chocolate molecule, blue-reflecting surface, soft pillow etc)” – strict reductive materialism. We can say this, I think, even if we accept the argument that the question “Why does it feel like that?” is itself in a sense illegitimate since there is no way to fill in the sense of “that,” as Hume, Wittgenstein and the Buddhists argue. The basic insight is that having these conscious experiences is indistinguishable from having this physical body in this physical world.

There is only one sense in which we can coherently say that our own phenomenal experiences are in any way similar to those of other conscious beings, such that we can grasp a link between intentionality (universal among all intelligent beings) and consciousness (unique to each conscious being). In this book I have emphasized the distinction between intentionality and consciousness. The last chapter will explore the connection between them.

Sunday, May 8, 2011

Hyper-chauvinistic type-to-type reductive materialism

OK, it’s not really hyper-chauvinistic type-to-type reductive materialism, because there isn’t anything to “reduce,” if the above arguments are persuasive. But the fact is that so far as qualitative experience is concerned, yours is what it is because you have the body that you have. “Experiencing” is just identical to living in a body like that.

A crucial difference between intentional states and phenomenal states is that phenomenal states are not picked out operationally, while intentional states are, even though the criteria for use of phenomenal predicates is operational just as they are for the use of intentional predicates. This is because phenomenal experience is outside of the reach of language altogether: precisely because it is unique to oneself and thus incommunicable to another. This is a difference between bodies.

Saul Kripke, much of whose work was inspired by Wittgenstein, argued that reduction was impossible on linguistic grounds. A phenomenal word like “pain” could never be defined as, say, “C-fibers firing” because the word “pain” referred to the feeling of pain (that phenomenal experience), and, Kripke argued, one can imagine being in pain without one’s C-fibers firing (or without having C-fibers at all) and that one’s C-fibers might be firing (or what have you) without one feeling pain. Of course Kripke’s point is about all phenomenal language but, as we have seen, there is no coherent way to separate “experience of the world” from “the world.”

Kripke’s claim amounts to saying that one can imagine experiencing this world without this world (which includes one’s body), or that this world (including this body) could exist without these experiences. I take Hume’s point that these sorts of claims about what can be conceived or imagined are meaningless, because phenomenal experience of the world and the world itself cannot be metaphysically distinguished from each other.

However this does not mean that “pain” might be defined as C-fibers firing: it could not. Use of the word “pain” will be determined operationally (as David Lewis insisted) as the use of all words is determined operationally. Constructions such as “pain-for-me” have no functional role in communication, but one’s (actual) pain can be mentioned even if there is no use for a term that designates it: it is no less real for being inexpressible. Meanwhile Kripke is not entitled to the claim that one’s body (one’s C-fibers firing) is causing one to have a sensation of pain that is distinct from its cause. Such a claim is irremediably dualist: one’s body is not the cause of one’s phenomenal experience. One just is one’s C-fibers firing etc.

Better, then, to drop the “reductive.” But “type-to-type” is also a dubious phrase. Unlike in the case of intentional states, there is no distinction between types and tokens when we are referring to conscious experience: each body is to some extent unique, and consciousness (unlike intentionality) is not supervenient. And the suffix “hyper” is perhaps a bit of rhetoric on my part. Better, perhaps, to call our theory of phenomenal mind simply “chauvinistic materialism,” if the need is still felt for a “theory” to account for a pseudoproblem.

This is no problem for science; since science, understood as a cultural artifact, is limited to the intersubjective, and phenomenal experience is wholly subjective (that’s why it’s a little silly to say that we have a “theory” here at all). Thus we can contemplate the resolution of the so-called “hard” problem of consciousness.

Sunday, May 1, 2011

Phenomenal states are not supervenient

Intentional states are multiply realizable, and functionalism was motivated by this fact. The supervenient nature of intentional states constitutes a real block to reductive materialism for intentionality. However intentional states can be individuated operationally. The intuition that the psychological description “He likes chocolate” involves a reference to the subject’s qualitative experience of tasting chocolate is wrong (or, we can mention the qualitative experience but we cannot actually convey it). In the case of the Martians we might not know if they even “taste” things at all; nonetheless we might come to know that they like chocolate.

A consequence of the necessarily operational basis of intentional descriptions is that, to use an example made famous by Daniel Dennett (although I don’t know that Dennett would agree with my line here), a lowly thermostat is a kind of intentional system: we can determine when it thinks that the room is too cold, just right or too hot. The intuition that this can’t be, that a thermostat is clearly not a mind, is a consequence of internalizing the traditional homogeneous concept of “mind” (because the thermostat has no Nagelian experience), aggravated by the prevailing dogma that thinking necessarily involves representations. When we disambiguate “mind” and see that intentionality is something altogether different than consciousness there is no denying that the thermostat is, in fact, an intentional being; nor does that fact in any way compromise our philosophical use of the term “intentionality.”

The qualities of experience, on the other hand, are not supervenient. It is plainly true that humans, dolphins, probable intelligent extraterrestrials and possible intelligent artifacts, among an indefinitely large set of other beings, can all believe that the chocolate is in the box, desire the chocolate and so forth. But there is no reason whatever to think that chocolate tastes like that (the way it tastes to me, say) for all of the members of the set: there isn’t even any reason to think that all of the members of the set of chocolate-desirers taste anything at all.

Ironically what this amounts to is that intentional properties are more ontologically mysterious, not less, than phenomenal properties. Consciousness has been called the “hard problem,” but in fact the right metaphysical account of consciousness is, relative to that of intentionality, positively straightforward.

Sunday, April 24, 2011

Troubles with the zombie argument in philosophy of mind

First of all it is important to keep in mind my initial disambiguation of intentional states and phenomenal states. It is already established that intentional states, since they are supervenient on physical states, can only be categorized operationally. Phenomenal states, as the friend of qualia has characterized them, cannot be categorized operationally (since they are beyond the reach of language). Of course an android could distinguish color surfaces without recourse to qualitative experience. We don’t think that there is “something that it’s like” for welder-directing robots in automobile factories to locate the seams with lasers. So when we are told that we can conceive of zombies it must be that we are meant to be conceiving of beings that are physically identical to humans but have no conscious experience. That functional descriptions can supervene on multiple physical descriptions has no bearing on the ontological status of qualia at all. I will make two arguments that are variations on the larger set of arguments presented here. First that the zombie counterfactual assumes something that it purports to show, and second that the zombie argument is an example of skepticism and is vulnerable to some counter-skeptical arguments.

The zombie counterfactual turns out to be question-begging: the friend of qualia has included qualia-body dualism as one of the premises. The claim is that we can conceive of beings that are physically identical to humans but that do not have immaterial phenomenal mental properties, unlike, presumably, the actual humans. To see this, try the same thought experiment only with the initial assumption (the most reasonable one, on the principle of parsimony) that qualitative experience will ultimately be explicable in wholly physical terms (that is, assume materialism as the default metaphysical position). The thought experiment now looks strikingly different.

Now (assuming materialism to be axiomatic, instead of dualism) the zombie argument is not question-begging: our initial assumption is that both the “blueness” of the surface of the plastic chair and the “blueness” of the phenomenal experience of the surface of the plastic chair are qualities had by virtue of the physical properties of the perceived chair and the perceiving body respectively. Now try this assertion on for size: “I can conceive of the plastic chair as having the identical physical properties that it now has (including its light-absorbing and –reflecting properties) but not having any color at all.” This is definitely inconceivable. When I try to conceive of an object with no color I for one try to imagine that the objects are transparent - I’m not sure if that counts as an object that has no color but it’s the best I can do. I cannot conceive of a physical object that is physically identical to an object that is colored but that has no color. This is just an application of the argument that to the extent that what I know, I know through “experience,” it makes no sense to draw a distinction between the phenomenal and the physical.

A second line of argument is suggested by the way Wittgenstein’s linguistic arguments can be deployed against skepticism. “Philosophical skepticism” is any argument to the effect that you don’t know something that you’re certain that you do know. Modern skepticism is closely connected to rationalism: Descartes thought that logical proof was the paradigm of “knowledge,” and it turned out, perhaps unsurprisingly, that by that standard we know very little. “The Problem of Other Minds” is a skeptical argument: you’re certain that other people have experiences and thoughts like yours but it seems that you can’t prove it. The Zombie problem is a variant on the problem of other minds: can’t you just see that other people are conscious? What would it be to doubt this? If the Zombie argument is an example of skepticism then we should be able to extend Wittgenstein’s treatment of skepticism to the Zombie argument.

A famous back-and-forth between Wittgenstein and G. E. Moore is good for illuminating this. In his essay “Proof of an External World” (1939) and elsewhere Moore is arguing in much the same philosophical spirit as Wittgenstein (they both think that skepticism is a pseudo-problem with roots in a faulty understanding of language), but Wittgenstein’s remarks on Moore’s arguments (included in On Certainty) make a crucial difference clear. Moore, presaging the “Ordinary Language” movement, wanted to show that “common sense,” by which he meant, roughly, ordinary talking and thinking about our physical selves in our physical world, was better defended with reasons than the speculative bases of skepticism and the “idealism” (really a kind of phenomenalism) still powerful in English philosophy at the beginning of the 20th century. Moore thought that our perceptions of our own bodies in the environment were instances – paradigmatic instances - of knowledge, not belief. Holding up his hand he would say, “I know I have a hand.” From the plain fact that I know I have a hand there arises similar knowledge of, ultimately, the external world: the keyboard is a reason for (believing in) the hand, the hand is a reason for the keyboard; the desk is a reason for them both and so on.

Wittgenstein’s response to skepticism is altogether different. Drawing on his premise that language must have operational criteria for determining appropriate conditions of use, Wittgenstein challenges the skeptic’s use of the verb “to know.” Take an ordinary (non-philosophical) instance of using that verb: “Do you know where your keys are?” The question makes sense because one might know or might not know, and “knowing” can be defined operationally: one can put one’s hands on one’s keys or one cannot. The question, “Do you know that the external world exists?” has no criterion for use: nothing could count as demonstrating that the external world either does or does not exist. That is, we neither “know” nor do we “not know” that the external world exists. The word “knowledge” has no place here. Where Moore proposed a kind of normative epistemology that would, if accepted, constitute an argument for the existence of the external world, Wittgenstein denies that there could be any argument one way or the other: Moore has taken the bait and tried to play a game that can’t be played.

Here we can see Wittgenstein’s “soft” arguments about language and his “hard” argument called “solipsism” come together. It is important also to see that Wittgenstein is much closer to Hume than Moore is. Hume (like Berkeley) stresses that the assertion that the external world “exists” is just as untenable as the assertion that it might not. Both sides of the disjunction are nonsense, if either side is.

The problem of other minds is exposed to the same treatment as a pseudo-problem. PI 246:

In what sense are my sensations private? – Well, only I can know whether I am really in pain; another person can only surmise it. – In one way this is wrong, and in another nonsense. If we are using the word “to know” as it is normally used (and how else are we to use it?), then other people very often know when I am in pain. – Yes, but all the same not with the certainty with which I know it myself! – It can’t be said of me at all (except perhaps as a joke) that I know I am in pain. What is it supposed to mean – except perhaps that I am in pain?
Other people cannot be said to learn of my sensations only from my behavior, - for I cannot be said to learn of them. I have them.
The truth is: it makes sense to say about other people that they doubt whether I am in pain; but not to say it about myself.

If I say “I know he’s conscious” and you challenge me by excluding all operational criteria from counting as justification for my claim, all you will do is change the ordinary meaning of the (public) word “conscious.” But then I’ll just have to start using another word for the same purpose, because what I have to say will remain exactly the same. PI 403:

If I were to reserve the word “pain” solely for what I had hitherto called “my pain”, and others “L. W.’s pain”, I should do other people no injustice, so long as a notation were provided in which the loss of the word “pain” in other connexions were somehow supplied. Other people would still be pitied, treated by doctors and so on. It would, of course, be no objection to this mode of expression to say: “But look here, other people have just the same as you!”
But what should I gain from this new kind of account? Nothing. But after all neither does the solipsist want any practical advantage when he advances his view!

As for “consciousness,” its meaning will now be anything you care to stipulate – in other words nothing - since you’ve stipulated that one can’t hear, see, smell, taste or touch anything that could even possibly indicate its presence: much like the word “god.” Thinking about the semantics of the word “god” brings us back to the first argument about the question-begging nature of the Zombie counterfactual. If the claim is that an operational theory of mind such as functionalism, specifically, has a problem with qualia, it must at least be a tacit assumption that there could be some non-operational theory of mind (a traditional one? a popular one? a philosophical one?) that does not: one that actually incorporates information about “what it’s like” into psychological descriptions and explanations. But there is no theory of mind like that, and if the arguments rehearsed in this chapter are correct there cannot be one.

Even after all of this, though, Nagel’s point still stands: there is “something that it’s like” for me to have a qualitative experience, and mine might very well be different from a Martian’s, or a dog’s or even yours. Although language cannot express these “qualia,” much less produce coherent claims that they are “properties” distinct from physical properties, our progress here has not been wholly destructive. I am now in a position to say a few things, after all, about the metaphysics of phenomenal experience.

Sunday, April 17, 2011

Turning the Inverted Spectrum on its Head

The inverted spectrum argument is first found (remarkably full-blown) in Locke’s Essay Concerning Human Understanding (II, xxxii,15, “Of True and False Ideas”):

"Neither would it carry any Imputation of Falshood to our simple Ideas, if by the different Structure of our Organs, it were so ordered, That the same Object should produce in several Men's Minds different Ideas at the same time; v.g. if the Idea, that a Violet produced in one Man's Mind by his Eyes, were the same that a Marigold produces in another Man's, and vice versâ. For since this could never be known: because one Man's Mind could not pass into another Man's Body, to perceive, what Appearances were produced by those Organs; neither the Ideas hereby, nor the Names, would be at all confounded, or any Falshood be in either. For all Things, that had the Texture of a Violet, producing constantly the Idea, which he called Blue, and those that had the Texture of a Marigold, producing constantly the Idea, which he as constantly called Yellow, whatever those Appearances were in his Mind; he would be able as regularly to distinguish Things for his Use by those Appearances, and understand, and signify those distinctions, marked by the Names Blue and Yellow, as if the Appearances, or Ideas in his Mind, received from those two Flowers, were exactly the same, with the Ideas in other Men's Minds."

Locke composed this counterfactual as part of his effort to show that “tertiary” properties (the properties of mental “ideas”) were different from secondary (the causal properties of the object of perception to cause ideas) and primary (the physical properties of the objects themselves). This was the property dualism repudiated by Berkeley and Hume. In the 20th century the inverted spectrum has had a strong career as a demonstration of the failure of functionalism to handle qualitative properties and, more to the point, as a supposed demonstration that there are such properties (in substance this is very much the same as Locke’s original application).

Imagine someone whose color spectrum was inverted (the “invert”): where normal people saw red, the invert saw blue, where blue, red. Such a person, raised among normal, English-speaking people, would be functionally indistinguishable from anyone else: asked to go out to the car and get the blue bag, say, they would perform this task exactly as anyone else would. Neither they nor anyone else would have any way of knowing that the invert’s experience of seeing the blue surface of the bag was the same experience that everyone else had when they saw a red surface, since the invert, like everyone else, would refer to such a surface as "blue." Since the invert would be functionally identical to a normal person, a functionalist is committed to the position that there is nothing different about their mental state. But (the argument goes) of course there is something different about their mental state: the quale, or phenomenal quality of the experience, is different. Thus functionalism is false.

Wittgenstein argues that the absent qualia argument demonstrates just the opposite of what the friend of qualia claims: since it is not even in principle possible for public language (the only kind of language there is) to pick out private sensations, phenomenal properties are not a problem for operationalist approaches. No theory of mind (or science of mind or description of mind) will ever include any actual discussion of the specific quality of any specific private sensations, because they cannot be discussed. As for the alleged discussion of phenomenal experience we find in philosophy, this is an instance of confusing mention with use - just as one can mention “all sentences that have never been expressed,” but cannot cite one. Outside of (misguided) philosophical conversation there is no context for use of indexically subjective language such as “blue-for-me” as opposed to the intersubjective “blue” which, like all words, necessarily has public criteria for appropriate contexts of use. This is why even the very best of the phenomenologists (Husserl, Merleau-Ponty) never seem to get beyond a sort of “Prolegomena to Some Future Actual Practice of Phenomenology”: after the manifesto there is nothing more that can be said.

So far this is a version of what I am calling the soft argument: the argument is about language, not about the ontological furniture of the world. It has conclusions that perhaps all non-philosophers would find definitive. Qualia do not constitute any sort of obstacle to the naturalization of psychology from the point of view of the scientist because science never could be expected to go beyond the limits imposed on language by its public nature in the first place. Nor is there anything inadequate about our ordinary, colloquial speech about qualitative experience (about, that is, the flavor of the sauce or the hue of the sunset), notwithstanding our individuality, for the same reason. But what about that ontological furniture?

The metaphysical argument will be about the identification of consciousness with experience. The Kantian will say (confining ourselves to the terms of the present discussion) that consciousness is a necessary precondition for the possibility of experience, hence not identical to it. This argument might gain some traction if we concede to the Kantian the point that “experiencing” an object entails bringing the object under a concept (although see the discussion of Kant in Chapter Two), but that very distinction between sensation and perception in the case of objects itself entails that sensation (phenomenal experience) is something prior to the formation of a Kantian “representation.” (If the reader is thinking of Aristotle’s nous at this point I beg your indulgence until Chapter Four.)

I have been using the word “consciousness” as synonymous with phenomenal experience, but the word is also sometimes used in regard to intentional states. Used in its intentional sense, to be “conscious” of an object is, on the traditional view, to form a representation of it or, on the view that I advocate, to be in some sort of relationship to it. But it is incoherent to say that to be conscious of pain, say, is to form a representation of pain or to be in a relationship to pain. In its phenomenal sense the word “consciousness” just refers to the sum of phenomenal experience: pain is a constituent of consciousness, not one of its objects. I may be intentionally conscious of pain at some higher level of psychological organization (one that can be picked out with operational criteria), but it makes no sense to say that I have to form a representation of pain or be in a relationship to pain to have an experience of pain.

The same is true for color qualia or for any qualitative experience. Kripke famously pointed out that the word “pain” just refers to the sensation of pain. “Blue,” in its phenomenal sense, just refers to the sensation of blue. It is one of the components of experience, not some object of experience. Experience is qualia; qualia are experience. So (to get back to the supposed metaphysical implications of the inverted spectrum) it makes no sense to say that experience has properties. Only the objects of experience have properties.

Where this leaves us is at the point of distinction between Frank Jackson’s “What Mary Didn’t Know” essay and Thomas Nagel’s equally famous essay “What is it Like to Be a Bat?” In that article Nagel makes an argument very close to but not identical to Jackson’s. No matter how much we come to know (physiologically) about the echolocation organ of the bat, Nagel argues, we will never know “what it is like” to experience the world in the way the bat does. The difference between Nagel and Jackson is that, granting the present argument that (qualitative) experience itself has no properties, experience itself does not constitute any sort of information (experience is rather the ground of information). We can concede that different conscious beings have different experiences without conceding that this entails any ontological implications.

Some writers have argued that the invert cannot, in fact, be conceived (Douglas Hofstadter holds this view, for example). I do not take that position. Since I do not see how to refute the Wittgenstein/Buddhism non-duality version of “solipsism,” I have no motive to try to prove that we do, in fact, grasp the qualitative nature of the experiences of others. Even if we could do so we could not express this “grasp” linguistically. However there is a different application of the “absent qualia” argument, one that holds that we can conceive of beings functionally (“behaviorally” is a more appropriate word here) equivalent to humans that have no qualitative experiences: “zombies.” David Chalmers’ entire argument for qualia-matter dualism hangs on this claim. I do not believe that we can “conceive” of any such thing.

Sunday, April 3, 2011

Buddhism and Qualia

“Buddhism” is the name of an ancient tradition (with roots in the older Hindu tradition) that includes both philosophical and spiritual ideas and practices. After 2,500 years it is no surprise to find that classical Indian and Chinese philosophy, including Hindu and Buddhist philosophy, encompass a gamut of positions at least as diverse as those found in the European philosophical tradition. There are, in the philosophical senses of the terms, materialist Buddhists and idealist Buddhists, foundationalist Buddhists and relativist Buddhists. By titling this section “Buddhism” I do not mean to suggest that the arguments discussed here are characteristic of all of Buddhist thought. However these particular arguments do constitute a persistent and venerable thread and I will discuss several different sources. These arguments have roots in some of the most basic elements of Buddhist teaching. While I have chosen to concentrate on the Mahayana tradition that is not meant to suggest that other traditions and schools may not include similar arguments.

There is a complex mix of philosophical, psychological, spiritual and social motivations informing the life of Siddhartha Guatama and his ultimate focus on the nature of the self. The severe caste structure of Indian Hindu society was tightly tied to traditional Hindu concepts of karma and reincarnation. Souls rose and fell over the course of thousands of lifetimes, and humans were low on the scale of karmic life: overall the effect was fatalistic and conservative. At the same time the sixth century BC in India was a period of transformation when a vernacular Sanskrit brought cultural upheavals that included the beginnings of the Epic Hindu literature and the flowering of diverse Hindu sects, Jainism and other movements besides Buddhism.

The basic Buddha Dharma is the most well-known piece of Buddhist philosophy. These are the “Four Noble Truths”: All life is suffering (“duhkha”), says the Buddha; we suffer because we are caught in a cycle of sensual satisfaction and craving; this suffering can be alleviated and even brought to an end altogether; and the “Eightfold Path” of right attitudes and practices will lead to the cessation of suffering. From this starting point Buddhism developed, among other things, a philosophical tradition with a sustained interest in the nature of consciousness, experience and the relationship between the self and the world. The idea that the “ego”-self, the self that arises through the cycle of duhkha, is the encumbrance that the “bodhisattva,” or enlightened self, needs to lose in order to achieve nirvana is prominent in the earliest teachings attributed to Siddhartha.

It is a somewhat dangerous business bringing a discussion of Buddhism into a book focused on contemporary philosophy of mind. I want to stick closely to the line of argument that has brought us to this relatively exotic territory. The overall claim is that the metaphysical problem of the alleged existence of phenomenal properties as distinct from physical properties is a pseudoproblem. I have presented arguments of Hume and Wittgenstein that I think are persuasive versions of this claim. In the case of Wittgenstein I argue that the “solipsism” argument common to the early and late works entails that phenomenal properties do not exist (are not part of this world). The view that the self is “emptiness,” and that the overcoming of the duality between the self and the world constitutes nirvana (enlightenment) is extremely similar, if not identical, to Wittgenstein’s solipsism argument. I will describe the Mahayana version of the argument and then present some textual evidence for my interpretation from classical sources.

The earlier Abhidharma School taught that “dharmas” were individual, autonomous atoms of experience; something akin to Leibniz’s infinity of monads. This is idealist ontology: primary being was dharma which was understood as consciousness (more or less: the bulk of Abhidharma metaphysics consists of discussions of just what “dharma” is after all). Although the arising ego-self, on this view, dissolves into infinitude of discrete dharmas, these dharmas are constitutive of the world. Mahayana Buddhism attempted to go further and collapse the duality between the mental and the non-mental: a kind of ultimate erasure of the self from the world that resulted in freedom from the bonds of the karmic cycle, the traditional goal of both Hinduism and Buddhism. Their formula was


One must resist the Cartesian instinct to interpret this formula as stating that the world collapses into the mind (as it does on the idealist view). The idea is that each consciousness is a universe. That universe at the middle of which you are sitting is you: it/you came into existence when it/you became conscious. When you pass away the universe you inhabit will pass away: for that universe is you. This is a line that could be defended by Hume: he might point out, for example, that if all any talk about “the world” could possibly be referring to is experience, then it makes no sense to refer to “a world” beyond experience, and thus it is incoherent to speak of a common world for all of the “microcosms” – a pseudoproblem.

An account of nirvana as the realization of the non-duality of mind and world can be found in Mahayana Buddhism and its descendents. The Prajna-paramita, or Wisdom Sutra, was traditionally taken to be the word of the Buddha, but scholars trace its origins to the first century AD and it appears that it was composed over the next several centuries. In a chapter titled “Mara” (a malevolent deity who lays traps for spiritual seekers), we find:

Subhuti: Is it then possible to write down the perfection of wisdom?
The Lord: No, Subhuti. And why? Because the own-being of the perfection of wisdom does not exist, nor that of the other perfections, the emptinesses, the Buddhadharmas or all-knowledge. That of which the own-being does not exist, that is nonexistence; what is nonexistence cannot be written down by the nonexistent.

The spiritual goal here is to free the self from the cycle of satisfaction and craving, and that is accomplished by showing that the self is not part of the world (I am not claiming that our present question about phenomenal properties is what Mahayana Buddhism is all about!). The Wisdom Sutra emphasizes the nonexistence of “own-being,” the being of oneself in one’s world. I chose this passage because the topic is language, and the message conforms to Wittgenstein’s treatment of the alleged problem that arises when we think of phenomenal language, say the word “blue,” as referring to something internal (mental) versus external (physical). Here the predicate “exists” is understood as meaning “exists in this world,” the world of experience: but the experiencing subject is not in this world. This also conforms to Hume’s argument that it makes as little sense to speak of the physical as distinct from the mental as it does to speak of the mental as distinct from the physical. Both sides of the distinction drop away simultaneously.

In the chapter titled “The Exposition of the Nonexistence of Own-Being” the point is more explicit that karma (one’s involvement with duhkha, the cycle of satisfaction and craving) is based on “own-being” and “own-marks,” worldly characteristics of persons. “Dharma” means something in the area of “consciousness,” “self” or “view.”

Subhuti: If, however, these dharmas are empty of own-marks, how can with regard to dharmas which are empty of own-marks a difference or distinction be apprehended (to the effect that one says) “this one is a being of the hells, this one an animal,…this one a god, this one a human….” And as these persons cannot be apprehended, so likewise their karma or its karma result.
The Lord: So it is, Subhuti, so it is, as you say. In respect of dharmas which are empty of own-marks no karma or karma result can be apprehended....But when those too ignorant to cognize dharmas as empty of own-marks manufacture a karma…then, through badly done karma they are hurled into the three states of woe, through what is well done they are reborn among gods and men….Here the Bodhisattva, who courses in perfect wisdom, does not see those dharmas in such a way that, when seeing them, he apprehends any dharma whatever. Not apprehending them he sees that “all dharmas are empty.”

Hundreds of years later and thousands of miles away Eihei Dogen (1200-1253), the classic avatar of Japanese Zen Buddhism, had refined the teaching of the non-duality of mind and world into a meditation practice (“zazen”) and a literary genre (“koan”) that were more minimalist and practice-oriented than the ritual-encrusted, syncretic and generally more baroque Sanskrit and Tibetan traditions, although like Tibetan Tantric Buddhism, Zen and Chinese/Japanese Buddhism in general are descendants of the Mahayana tradition. In his Shobogenzo (“Treasury of the True Dharma Eye”) the solipsistic view has crystallized considerably. Here are quotations from the section “Actualizing the Fundamental Point” (written around 1230):

As all things are buddha-dharma, there is delusion and realization, practice, and birth and death, and there are buddhas and sentient beings.

As the myriad things are without an abiding self, there is no delusion, no realization, no buddha, no sentient being, no birth and death.

The buddha way is, basically, leaping clear of the many and the one.

To carry yourself forward and experience myriad things is delusion. That myriad things come forth and experience themselves is awakening.

When you see forms or hear sounds fully engaging body-and-mind, you grasp things directly.

To study the buddha way is to study the self. To study the self is to forget the self. To forget the self is to be actualized by myriad things, your body and mind as well as the body and mind of others drop away.

Enlightenment is like the moon reflected on the water. The moon does not get wet, nor is the water broken.

When the fundamental truth of mind=experience=world is realized things are grasped directly. The idea that there are qualitative experiences that are distinct from the objects of experience reflects the same Cartesian duality of mind and world that underlies representational theories of mind; the Buddhist aim in criticizing this duality as a misconception is essentially spiritual. The Wittgensteinian position is clearly echoed in the Zenrin kushu, a 15th century Zen text, which describes consciousness as “Like a sword that cuts, but cannot cut itself; Like an eye that sees, but cannot see itself”: what is constitutive of the experienced world cannot be considered as part of, or as in, that world.

Sunday, March 27, 2011

Non-duality as a response to the "hard problem"

The respective ideas of Hume, Wittgenstein and Mahayana Buddhists are interrelated and are characterized by empiricism and naturalism. Not surprisingly the response they invite most prominently is a Kantian one. But Kant deduced that there had to be necessary preconditions for the possibility of representational experience: it was the representation of the world that had to be “conditioned.” None of these thinkers – not even Hume on my reading – defines experience as involving the formation of representations in the first place. Furthermore remember that the subject of this chapter is consciousness, and consciousness is not obviously amenable to a representational “explanation” even for those who are inclined to address intentionality that way. In any event if one balks at, say, the “solipsism” argument I certainly agree that it is a bit breath-taking, but I wouldn’t introduce it here if I knew of an argument that refuted it.

The present set of arguments, that will now be deployed against the “absent qualia” arguments that purport to show that there is a metaphysical problem about phenomenal properties, can all be attributed to some combination of Hume, Wittgenstein and/or Buddhism. The first two arguments are “soft”: they start with two respective points about the nature of language and conclude that private qualitative experience, whatever its ultimate ontological status may be, cannot be the subject of language. This is a soft conclusion because it sets aside the question of whether there really is anything that can coherently be regarded as “inner,” private experience. The remaining arguments are “hard”: they start with the inseparability of mind, experience and world and conclude that it is incoherent to posit the existence of phenomenal properties if these are taken to be mental properties distinct from the physical properties of the experienced world. However although the first two arguments are soft they do, if persuasive, suffice for the naturalization of psychology, because they demonstrate that qualitative experience, whatever that may be, could never have been a subject for natural science, and so is not a problem for natural science.

1) If language needs inter-subjective criteria of appropriate conditions of use then the phenomenal vocabulary, like all language, cannot function in virtue of referring to anything “private” to the individual. This argument is explicit in Wittgenstein, a case can be made that it is implicit in Hume’s verificationist epistemology. It is “soft”: it leaves the ontological status, if any, of qualitative experience alone. It naturalizes psychology the way that classical, early 20th century behaviorism naturalized psychology: an operationalist protocol excludes reference to the “inner” from science, so understood.

2) If the “meaning” of language/symbols is nothing more nor less than everything that has been accomplished through the use of individual, concrete tokens, then the phenomenal vocabulary cannot function by referring to anything private. This is functional-role semantics and it is unique to Wittgenstein among this group (although other philosophers, such as Pragmatists, also develop this sort of operationalist account of language). It is soft; the ontological status of qualia is not addressed directly.

3) If the self is nothing neither more nor less than the experiences of the world by the self, then there can be no duality between the “self” and the “world.” This is Wittgenstein’s “solipsism” argument, first presented at the end of the Tractatus, and the same idea is found in the Mahayana sutras of classical Buddhism and in Zen Buddhism. This is a “hard” argument: on this view the existence of phenomenal properties is denied. It is another argument that seems implicit in Hume, but Hume does give a very explicit argument to the same effect:

4) If experience defines the limit of what can be known, it is absurd to posit a distinction between the “mental” and the “physical.” Hume’s version is quite as hard as #3 because it produces the same conclusion that it makes no sense to speak of either side of the duality of mind and matter: both concepts collapse simultaneously. It is neither materialist nor idealist, but a position of non-duality. This view is shared by Berkeley.

5) Consciousness cannot be located in the world, so consciousness cannot be said to have any properties. This argument follows from #3. It is explicit (and elaborated at great length) in its Buddhist version but also clearly implicit in both Hume and Wittgenstein. It does not follow from this that consciousness cannot be a property. Just what we mean when we predicate consciousness of a thing is what is at question, although I take it as now established that the criteria for predications of consciousness are necessarily operational.

If the most tempting line of rebuttal to all of this is a Kantian one, remember that these arguments are here deployed to show that there are no phenomenal properties. This question about consciousness has been disentangled from the problem of intentionality. The motivations of the three sources of the set of arguments are varied: Hume fills out the radical implications of empiricist epistemology as far as they go; Wittgenstein has a vision about the foundations and limits of logic; and the classical Buddhists appear to take the metaphysical implications of non-duality both literally and seriously. Whatever one makes of these sources (or of my interpretations of them), it is at least fair to say that they place the burden on those who would say that there are phenomenal properties, metaphysically distinct from physical properties, to articulate reasons why anyone should think so.

However the discussion is far from over. Having assembled this set of arguments I now need to apply them to the various “absent qualia” arguments that have been the backbone of the critique of functionalism and the emergence of consciousness as a problem for cognitive science over the past thirty years.

My aim is not to show that operationalist theories such as functionalism are adequate to address the problem of consciousness; I am not “defending” functionalism from the “absent qualia” critique. My view is that intentional predicates must be handled operationally, while phenomenal predicates cannot be – that is, they cannot be to the philosopher’s satisfaction, notwithstanding the fact that all predication, to be intelligible, must adhere to intersubjective operational criteria. Functional descriptions abstract away from hardware: they include no physical descriptions. In the same way they abstract away from consciousness: they include no phenomenal descriptions. Of course this is true because there is no such thing as “phenomenal description” if by that one means reference to “private” experience. But a further point is that there is no reason to think that phenomenal experience is multiply realizable (supervenient), while intentional states are self-evidently so.

The point of the following discussions of the inverted spectrum problem and the zombie problem (both variants of the “absent qualia” problem) is not, then, to vindicate functionalism as a theory of consciousness. Theories don’t address pseudo-problems. The significance of the “absent qualia” arguments in the present context arises when they are offered as evidence that physicalism is ontologically incomplete.

Tuesday, March 15, 2011

Wittgenstein on Qualia

Anyone who has read this far understands that Wittgenstein, for better or for worse, is the canonical philosopher who has had the most influence on the arguments that I am advancing here (even if I am merely Wittgenstein’s ape, as I rather suspect, from what I have read of him, that he would say I am). But when I started drafting this book Wittgenstein worried me. My strategy is to analyze the mind-body problem into separate problems that admit to separate solutions. But Wittgenstein seemed to be addressing both the problem of intentionality and the problem of consciousness, sometimes simultaneously. Perhaps I was mistaken to try to separate them?

Wittgenstein gives us a general treatment of language, and my method is essentially grounded in linguistic analysis as well. Metaphysics is brought down to Earth when regarded as a semantic inquiry: I don’t know, after all, what “primary being” is, or the limits of nature or anything like that. The only way to naturalize psychology is to develop a natural semantics for the psychological vocabulary. If the metaphysical theory of physicalism is right then our psychological talk has had natural, physical referents all along, and we should be able to determine what those are. Wittgenstein gives us, with his functional-role semantics, what is basically an operationalist account of meaning (“meaning is use”), and an operationalist semantic is a kind of naturalist semantic.

Now we can see the apparent problem: Wittgenstein argues that all language must have operationalist criteria of meaning, including the phenomenal vocabulary. But I have conceded that the “absent qualia” problem persuasively shows that operationalist theories of mind such as functionalism can’t handle the problem of consciousness. Isn’t there a contradiction in, on the one hand, embracing Wittgenstein’s argument that the word “blue” is meaningful (as it has intersubjectively verifiable criteria of use) while the construction “blue-for-me” is not, and on the other hand insisting that the naturalization of the phenomenal vocabulary requires a different treatment than the intentional vocabulary requires?

The tension is resolved by considering two other arguments of Wittgenstein’s, both of which are common to the Tractatus Logico-Philosophicus and the Philosophical Investigations, unlike functional-role semantics (developed in the PI) which represents the major difference between the earlier and later work. A popular misconception is that there is no continuity between Wittgenstein’s two major works; this is an effect of the strikingly radical operationalist treatment of “meaning” in the PI, and a consequently radical difference in method of composition. However much is missed when one misses the common themes.

Compare these quotations, first, the famous closing sentence of the Tractatus: “What we cannot speak about we must pass over in silence.” Then PI 296: “’Yes, but there is something there all the same accompanying my cry of pain. And it is on account of that that I utter it. And this something is what is important – and frightful.’ – Only whom are we informing of this? And on what occasion?” (Italics in original). Granting that at the end of the Tractatus he is speaking broadly about something he calls “mystical,” it is apparent that he takes ethical, aesthetic and spiritual experiences to be varieties of qualitative experience that, like pain, cannot be expressed by language. (This was the point, regarding ethical “propositions,” that W. was making when he got into that brawl with Karl Popper.)

The explicitly operationalist account of language in the PI develops from this earlier awareness of the limits of language (but note that this is not the same argument as the one tagged by his famous dictum “The limits of my language are the limits of my world.” That is the other argument, discussed below). If language is necessarily intersubjective (that is, worldly) then there must be public criteria for its use, but the insight that the quality of experience is inexpressible comes before, not after, this treatment of language. Functional-role semantics is a response to the inexpressible nature of qualitative experience.

So the great logical behaviorist turns out to acknowledge qualitative experience after all? The short answer is yes: he never denied it. At PI 296 his imaginary interlocutor is unchallenged when he says “this something is what is important.” Maybe the most important thing in life: remember that value itself is part of the inexpressible (and see Chapter Four). This does not involve him in a contradiction, although it needs some more consideration here.

One objection is that Wittgenstein is what was earlier called an “atheistic” or “philosophical” behaviorist: he denies that it makes sense to think of the mental in terms of something “inner” vs. the “outer” world. But aren’t qualitative experiences essentially “inner” in this sense? Not necessarily. The nature of qualitative experience is what is at question.

More importantly and more to the point of this discussion, Wittgenstein’s claim is not about qualitative experience, it is about language. The quality of personal experience is not expressible because of the intersubjective, public nature of language. Here is a link with the argument as deployed in Chapter Two: language (representation in general) does not exist “in the head,” either literally or figuratively. We saw in Chapter Two that the notion of the “inner” as something representational was vacuous, explaining nothing. Language (symbols, “meaning”) is something that exists only in the “outer” world. Thus whatever we make of qualitative experience, all language use has public criteria.

The two related arguments, that language must have public criteria for use and that what we cannot speak about we must pass over in silence, are sufficient to show that the problem of consciousness is not a problem for science (this point will be discussed at greater length below). But by themselves they give us only “agnostic,” methodological behaviorism, which may satisfy the empirical psychologist but will not satisfy the philosopher. The philosopher still has a question about ontology. In this discussion of Wittgenstein’s first two arguments I have been careful to use the phrase “qualitative experience,” leaving open the question of what it is of which such experience consists.

The third argument of Wittgenstein’s, one that is also common to the early and later work, goes further and demonstrates that “qualia,” understood as real properties that are non-physical properties, do not exist. It gives us the “atheistic,” philosophical behaviorism that we need to naturalize the phenomenal vocabulary. As with Hume it will turn out that there is no coherent distinction between “qualitative” experience and just plain experience. (Note also that in this section I am using the word “behaviorism” rather than the word “operationalism.” Since “behaviorism” is more the standard term in the Wittgenstein literature this makes it easier to situate the present discussion in that literature, besides being much less clunky. And anyway the arguments discussed so far are in fact about language; “behaviorism” in the sense that we can use that word to describe Wittgenstein’s view is not really a “theory of mind,” although it may be a theory of psychological talk.)

The third argument is known as the “solipsism” argument, and it is found in the Tractatus at 5.6 through 5.641. The most famous aphorism from this passage is 5.6, “The limits of my language mean the limits of my world” (italics in original), but for the present argument 5.621, “The world and life are one,” and 5.63, “I am my world. (The microcosm)” may make the point most clearly. In fact on my view 5.6 is frequently misinterpreted in a sort of obvious way, a recognizably Kantian way: if one represents the world linguistically (this interpretation goes), then the world as one represents it will be limited as a function of the limits of ones’ language. This is backwards. “The limits of my language” (italicized) is the phrase under analysis, and it can only mean (it is defined by) the limits of my world, which are, exactly as in Hume, coextensive with the limits of my experience.

5.632: “The subject does not belong to the world: rather, it is a limit of the world.”

5.64: “Here it can be seen that solipsism, when its implications are followed out strictly, coincides with pure realism. The self of solipsism shrinks to a point without extension, and there remains the reality co-ordinated with it.”

5.641: “Thus there really is a sense in which philosophy can talk about the self in a non-psychological way.
What brings the self into philosophy is the fact that ‘the world is my world.’
The philosophical self is not the human being, not the human body, or the human soul, with which psychology deals, but rather the metaphysical subject, the limit of the world – not a part of it.”

The word “qualia” is a “grammatical” (to use a word ubiquitous in the PI) reification of qualitative experience, which is constitutive of the world, “the limit of the world - not a part of it” (experience is not in the world). Naturalizing psychology does not require what cannot be done, naturalizing metaphysics. “The world and life are one.” As a living being I am constitutive of my world; my life and my world cannot be distinguished: 6.431: “So too at death the world does not alter, but comes to an end.”

The ontological idea of a Leibnizian parallelism between properties of the world and properties of experience makes no sense and it is the assumption of such a parallelism (of the coherence of such a parallelism) on which the alleged problem of consciousness rests. Kant, to be fair, is not so far from this insight himself (and Wittgenstein professed admiration for Kant): the rational mind, for Kant, is not a part of the phenomenal world. Only Kant’s followers did not heed his epistemological warnings.

This crucial Wittgensteinian appropriation of the word “solipsism” remains intact and unchanged decades later in the Philosophical Investigations. In the discussion of the multiplicity of uses of language (the fact that there are many different “language-games”) that opens the book Wittgenstein writes at #24:

“If you do not keep the multiplicity of language-games in view you will perhaps be inclined to ask questions like: ‘What is a question?’ – Is it the statement that I do not know such-and-such, or the statement that I wish the other person would tell me…? Or is it the description of my mental state of uncertainty? – And is the cry ‘Help!’ such a description?

Think how many different kinds of things are called “description”: description of a body’s position by means of its coordinates; description of a facial expression; description of a sensation of touch; of a mood.

Of course it is possible to substitute the form of statement or description for the usual form of question: ‘I want to know whether…’ or “I am in doubt whether…” – but this does not bring the different language-games any closer together.

The significance of such possibilities of transformation, for example of turning all statements into sentences beginning “I think” or “I believe” (and thus, as it were, into descriptions of my inner life) will become clearer in another place. (Solipsism.)”

All statements can be rendered “as it were, into descriptions of my inner life,” and this shows the actual vacuity of the allegedly significant distinction between “the inner life” and “the outer world.” The sense of the parenthetical “solipsism” is the same as in the Tractatus. It is important to see that Wittgenstein is not (as he admits) using the word “solipsist” in its usual metaphysical sense. In fact he inverts the ordinary sense of the word. Ordinarily the solipsist is understood to be saying that he only knows that one mind exists, his own (this is the Cartesian skeptical sense of the word). Wittgenstein is saying, with reference to certain uses of the first-person “I,” that one’s own mind is the only one that cannot be conceived as something in the world.

From the Blue Book (pp. 66-69):

“There are two different cases in the use of the word ‘I’ (or ‘my’) which I might call ‘the use as object’ and ‘the use as subject.’ Examples of the first kind of use are these: ‘My arm is broken,’ ‘I have grown six inches.’…Examples of the second kind are ‘I see so-and-so,’…’I’ have a toothache’…We feel then that in the cases in which ‘I’ is used as subject, we don’t use it because we recognize a particular person by his bodily characteristics: and this creates the illusion that we use this word to refer to something bodiless, which, however, has its seat in our body. In fact this seems to be the real ego, the one of which it was said, ‘Cogito, ergo sum’.”

In fact this use of the first-person pronoun does not “refer” to anything in the world at all. PI 404:

“’When I say “I am in pain,” I do not point to a person who is in pain, since in a certain sense I have no idea who is.’ And this can be given a justification. For the main point is: I did not say that such-and-such a person was in pain, but ‘I am….’ Now in saying this I don’t name any person. Just as I don’t name anyone when I groan with pain. Though someone else sees who is in pain from the groaning.

What does it mean to know who is in pain? It means, for example, to know which man in this room is in pain: for instance, that it is the one who is sitting over there, or the one who is standing in that corner, the tall one over there with the fair hair, and so on. – What am I getting at? At the fact that there is a great variety of criteria for personal ‘identity.’

Now which of them determines my saying that ‘I’ am in pain? None.

“Personal identity theory” is a branch of metaphysics: the study of the criteria by which we identify a particular entity in the world as the “self.” But the subject, on Wittgenstein’s version of solipsism, is not an entity in the world at all, insofar as we are thinking of the subject as having qualitative experience. The experiencing subject is metaphysically identical with the experienced world.

So far I have presented two versions of this argument, Hume’s and Wittgenstein’s. I am not piling up these various demonstrations that the problem of consciousness is a pseudoproblem in order to commit the informal fallacy of the argument from authority: I have my own reservations about Hume, Wittgenstein and empiricism in general but I am persuaded by these particular arguments that the alleged metaphysical problem of phenomenal properties is a pseudoproblem. It is striking that Wittgenstein’s “solipsism” is very close, perhaps identical, to arguments found in an ancient tradition with origins very far from those of empiricism.