A Puzzle About Perception

Here is a strange fact about the universe: nothing that can be perceived from the inside can be perceived from the outside, and nothing that can be perceived from the outside can be perceived from the inside. You feel an intense pain—you plainly perceive the pain—but no one around you can sense your pain. They may know you have a pain but they don’t perceive it, whereas you perceive it only too well. They may look you over carefully, even touch and smell you, but your pain will never be an object of perception to them. Your pain is perceptually closed to them despite its openness to you. It is as if the pain can only be perceived from one position in the universe—the one that you alone occupy. You perceive it easily and automatically, but no one else can do this; their senses are not up to the task. Hence we have the common belief that the mind is an invisible thing (it’s not just pain that only the subject himself can perceive). Likewise, if a thing can be perceived in the normal third-person way, it is not perceived in the first-person way: if a tree is perceived by the senses, it is not perceived by itself in the introspective mode. Even if the object in question has introspective awareness, the perceived fact will not be introspected by the object: we don’t introspect our own brain states though they are perceptible to others  (and ourselves, in principle). If a thing is perceived by the senses it is not perceived by introspection. One sort of perception precludes the other: if you can see it in one way you can’t see it the other way. Either the thing is perceived introspectively or it is perceived externally but never both. If it is suited to one kind of perception it is not suited to the other. This is odd, puzzling, because it is hard to see why things have to divide up this way. Why can’t pains be perceived both ways? Why can’t brain states be introspected as well as seen with the eyes? Why is perception limited in these ways? We can have beliefs and knowledge about things from both points of view, but not the means of cognition we are calling perception (apprehension, sensation, feeling, acquaintance).[1] I know I have a pain and so do you, and I can in principle know my brain states as well as you; but perception does not line up with this. If something can be sensed in one way, it cannot be sensed in the other way. The universe seems to be imposing arbitrary rules on the scope of perception.

            It might be wondered how deep these rules go. Is it just contingent that pain states can’t be perceived by all and sundry (“I see your pain and it looks like a nasty one”) and that physical properties are not perceived by introspection (“My occipital cortex feels like its firing nicely today”)? Are there possible worlds in which doctors can see what you are feeling pain-wise and ordinary people can introspectively report on their brain states? I don’t want to discuss the question of whether the current set-up is logically or metaphysically necessary (the question is by no means easy), but it seems clear that it is deep-seated so far as our actual universe is concerned. As things are, the rules of perception are rigid in this respect: there are never any exceptions, and it is hard to see how there could be. It isn’t as if you just have to put on some special goggles and other people’s pain will leap into view in vivid 3-D. Nor can concentrating your attention on your head area cause your brain to disclose itself to your introspective faculty. Perceptual closure of the type I am describing looks written deep into the structure of the universe as we have it. But it is metaphysically puzzling; there is a kind of conceptual arbitrariness to it. It’s like being told that certain facts can only be perceived from one spatial location—as it might be three feet away from the object—all the others being unsuitable for perceiving the object. One would like to see some sort of rationale for why the universe behaves this way. Why did God make pains perceivable only by the subject, and brains only perceivable by the outer senses? Why did he impose such tight rules on how perception of these things is possible?

            Do the facts themselves dictate these rules? Compare mathematics: numbers are neither perceptible by the senses nor by themselves (the number 3 does not experience itself as odd). Plausibly, it is in their nature not to be so perceptible: they are abstract entities and hence not even candidates for perception in the ordinary sense. Might mental and physical facts be similarly unsuitable for a certain type of perception in virtue of their intrinsic nature? Perhaps pains as such are unsuitable for outer perception and brain states inherently unsuitable for inner perception. But if so, it can’t be because they are abstract, since they are concrete, i.e. spatiotemporal items with causal powers. So what is it about these things that renders them capable of perception by certain means but not others? There is no straightforward deduction from the nature of the mental and the physical to the conclusion that they can only be perceived in one way—by inner and outer sense, respectively. They can both clearly be perceived, unlike numbers, but they are limited in the ways they can be perceived. Pains are only too happy to be perceived by introspection (one might wish they would remain imperceptible) but they bluntly decline to be perceived by the outer senses; they just won’t go there. Similarly, physical properties reveal themselves to the senses quite freely, but they firmly resist the probing of the introspective faculty. Why this selectivity, this snobbery almost? Pains refuse to pass through the doors of (outer) perception, and brains won’t yield to introspective scrutiny: neither will join the other’s club. But there seems nothing about them considered in themselves that generates this exclusivity: both are concrete empirical phenomena existing in the real world. Yet they are extremely choosy about how they reveal themselves to the perceiving subject.

            This is the way the universe has been constructed; it’s the way reality fundamentally is. We take it for granted because it’s so familiar to us: we don’t think scientists will discover tomorrow that pains can be seen or neurons introspected. It takes an intellectual effort to see this for the oddity that it is: how amazing that other people can’t see my pain! I can sense it in all its awful glory, but they are completely blind to it—for them it is just a dubitable conjecture (hence the problem of other minds). If they could see it that might change their attitudes, but they simply can’t—we mustn’t blame them. It would also be good if I could sense a tumor growing in my brain before it gets out of hand, but my powers of introspection are just not up to the task (“Why not!” I might exclaim angrily). Much philosophy has revolved around trying to avoid this puzzle (Wittgenstein being an arch-avoider), but really it is an inescapable feature of the given universe. And it adds yet another puzzle to an already long list.[2]

[1] It’s hard to find a good verb that covers all cases of what is intuitively a basic type of awareness. I have chosen to use “perceive”, which is traditional: thus we perceive our pains (and other mental phenomena) as well as tables and rabbits.

[2] It is an aspect of the mind-body problem: mind and body seem very different from the point of view of perception, so how can they be the same?

An Interactive Theory of Meaning

Theories of meaning are apt to emphasize one or other aspect of what is conceived as a polarity: either meaning is constituted by features of the mind (the subject, the brain) or it is constituted by features of mind-independent reality (objects, the environment). Thus we have internalist theories of meaning and externalist theories. Some theorists seek to conjoin these elements, holding that meaning is the combination of an internal factor and an external factor (“dual component” theories). But what about the idea that meaning might be a kind of melding or merging of the two factors—a third factor that involves the internal and external but doesn’t reduce to them? Could it be the nexus of mind and world—their manner of connection? Let’s see if we can make anything intelligible of this idea: could meaning consist in the process or procedure of mind-world conjoining?

            The suggestion I want to consider is that meaning results from the interaction of mind and world; it is an interactive property. Crudely, meaning is syntax interacting with the environment: it isn’t syntax alone or the environment alone, or the conjunction of the two, but rather the activity (process, procedure) of one interacting with the other. It isn’t sense and it isn’t reference; it’s the manner in which these two interact (interface, mesh). It’s the way language use and reality play off each other. It’s not the dancers but the dance. To make the idea concrete, let’s consider proper names: this will give us a feel for the interactive theory. Objects can elicit states of mind and accompanying linguistic actions, as when the appearance of someone elicits an act of recognition and an associated utterance of that person’s name: this is world-to-mind interaction. It happens all the time (sometimes the interaction breaks down, as when you can’t remember a person’s name). The environment is acting on your mind and producing a certain result (or failing to). Evidently this has something to do with what the name means. Equally, the name can operate to bring about a change in the condition of the environment, as when you call for someone by name: here the denoted individual acts in accordance with the speaker’s use of the name and his accompanying intentions. This is mind-to-world interaction. The process is two-way—it is an interaction. The world acts on you and you act on it. Evidently both types of action reflect the meaning of the name: maybe the name means what it does because of these interactions. The meaning of the name isn’t, according to this theory, an image or a description or a concept located in the mind of the speaker; nor is it the object spoken about: rather, it is the pattern of interactions characteristic of the name. It’s the dance not the dancers, the process not the particulars. The same can be said of other types of word: the word “table” is used in multiple interactions with tables, and its meaning is a function of these interactions. You can invite someone to sit down at table or request that someone bring you a table, and you can utter the word “table” in making statements about tables (“That’s a handsome table”). Think of the whole history of such word-world interactions along with their psychological background—the myriad interactions between tables and table talk (and table thought). The idea of the interactive theory is that this is where meaning resides—this is where it takes its rise. It isn’t what the interaction is between that constitutes meaning but the interacting nexus—the fact of one thing acting on another in a reciprocal manner.

            Here we might think of Wittgenstein. In the Tractatus Wittgenstein identified meaning with an isomorphism between propositions and facts (thought and world), thus locating it in the relation between mind and world, not in mental or worldly elements alone (or even together). In this respect his picture theory resembles the interactive theory—it’s all about relations not relata—but his chosen relation is geometric in inspiration and obtains at a specific moment of history. By contrast, the interactive theory adopts a dynamic and causal type of relation that runs in both directions (do facts also depict propositions in Wittgenstein’s scheme?). In the Investigations we find a theory that treats meaning as temporally extended use (as a “practice” or “custom”). This is rather in the spirit of the interactive theory except that that theory stresses a two-way relation of reciprocal action and brings the environment in directly (so it’s not a simple use theory). The interactive relation is what does the work in creating meaning. In Wittgenstein’s lingo we could say that our “form of life” is an ongoing interaction between self and world, organism and environment; and meaning is embedded within the broader range of interactive relations. So the theory has elements in common with both the Tractatus theory and the Investigations theory, while having its own distinctive character: meaning is both relational and dynamic, according to this theory, because interaction is an active relation.

            Can we say more about what an interactive relation is? Consider the friendship relation: it consists of a pattern of interactions spread out in time. This pattern is distinctive to friendship and involves a whole complex of interactions—meetings, sentiments, obligations, conversations, etc. It isn’t like a spatial relation or a genetic relation or a pictorial relation: it is formed and maintained by a temporally extended series of personal interactions (the same goes for the relation denoted by “lover of”). Many social relations work like this—they are essentially interactive. Or take the relation expressed by “plays” (as in “Roger plays tennis”): this is also an interactive relation, involving many types of interaction with equipment, places, and other people. Meaning (we could also say reference) is likewise constituted by interactive relations of the same general category, though obviously of a different type. The important point is that meaning arises from interaction not from the interacting elements considered in themselves. In paradigm cases the interaction is causal (though also psychologically grounded), as with words for material things in the environment; but it need not be restricted to causal interactions, as in the case of words for numbers and other abstract entities. Here we have interplay between mathematical reality and our minds (and brains), though the relation is not causal: we call numbers by names, and numbers may elicit from us an appropriate utterance (“I see that the number five is featuring a lot in this series”). The same may be said for moral values whether conceived as mere emotions or as robust platonic entities. Words can differ in all sorts of ways, and the nature of the interactions that define them will also differ.

            It might be said that the interactive theory, as so far formulated, isn’t a theory at all, because it hasn’t explained how interactions give rise to meaning, or even what meaning is. This is perfectly true: the theory says nothing about the choice between truth conditions and assertibility conditions theories, say, and it is silent on the process whereby temporally extended interactions “give rise” to meanings (is this a simple identity or something more transformative?). But the theory as outlined here isn’t intended to answer such questions: it simply offers a proposal about what kind of thing meaning is—its ontological category. It is aptly seen as a kind of process theory, as opposed to a substance theory: meaning consists of an interactive process not of some sort of substance-like item (a mental image or a physical object in the environment). Compare process theories of the self: the self isn’t some kind of substance, mental or physical, but is more like a series of events spread out in time. Indeed, it might be argued that the troubles faced by standard theories of meaning arise because they mistake the ontological category of meaning, failing to see that it is a matter of reciprocal actions over time not of locatable items existing at specific times. As to the question of how interactions coalesce into meanings, we can remain studiedly neutral, even postulating a locus of mystery if we are so inclined: that is, we don’t know how interactions become meanings, and we may never know. That question is above our pay grade; we are happy if we can at least find the general ontological category to which meanings belong. We are offering a “picture” of the kind of thing that meaning is not attempting to a give a set of necessary and sufficient conditions for meaning or something of the kind. Nothing stops us from invoking the causal-historical chain conception of names, say, and recruiting it as part of a more inclusive interactive account of how names function. What matters is that the phenomenon of meaning is subject to an interactive mode of description, even though many questions and puzzles are thereby raised. We start to see in general terms how mind and world together cooperate in producing meaning.

            It is a consequence of this picture that meaning is not possible without mind-world interactions, so that brain-in-vat meaning looks impossible. But we must remember that minds can interact with themselves thereby producing meaning; what is precluded is the kind of meaning our words have when we are not brains in a vat, specifically words for material objects in the environment. This is notoriously disputed territory, but there is certainly a respectable tradition that requires that without an environment with which a speaker interacts there can be no meaning concerning such an environment. And it is surely reasonable to think that our vocabulary for the environment requires the existence of such an environment: we mean what we do about the external world precisely because we interact with it on a daily basis—whatever may be said about hypothetical brains in vats. The external world plays a role equivalent to the internal world in making human language as we have it possible. A normal child’s language has the meaning it does partly in virtue of living in a world with which it interacts, verbally and otherwise. This is the kernel of truth in externalist theories of meaning.

            Obviously a lot more needs to be said to make the interactive theory into a full-blooded theory of meaning, but the general shape of it looks clear enough. It is surprising that the theory has so little visibility in the philosophical landscape. We mean what we do in virtue of the fact that we interact mentally with the world around us. Doesn’t this have the sound of truism?

Colin McGinn

There is another interview with me on YouTube, this time with Thom Jump. It’s about the mind of God.

Best At Doing Philosophy

I found myself wondering who is the best at actually doing philosophy—the activity, the skill. I mean as judged by such criteria as cleverness, ingenuity, argumentative power, intellectual penetration, insight, polemical punch, sheer philosophical IQ. This is independent of correctness or quantity of output. Here is my answer: Descartes, Hume, Berkeley, and Russell.  These are the guys who really stand out for philosophical intelligence. I imagine my readers will nod in assent, though they may wish to add someone who has impressed them particularly. They are some pretty smart cookies all right. But it may surprise you to learn that I regard Berkeley as the clear champ: he is just so sharp, so intellectually resourceful, so outright brilliant (outrageously so). Not that I agree with his conclusions, but his cleverness is second to none. But what about more recent practitioners? Yes, there have been some impressively gifted philosophers in more recent times: Frege, Husserl, Kripke, Lewis, Strawson, Fodor, and many others. But none of these strikes me as preternaturally brilliant, inhumanly so. And where do I stand in this? Actually I think it takes one to know one, so I place myself next to the idealist bishop. I feel a certain kinship with our misguided theist; I feel we speak the same language. I’m not claiming to prove this here, but it is my considered opinion. I even think I need to curb my cleverness sometimes, as if it is leading me down the wrong path. No doubt others will vehemently disagree. It’s a question worth contemplating.