Auto-Sexuality: A Puzzle


Auto-Sexuality: A Puzzle


We observe a great many varieties of sexuality in the human population: heterosexuality, homosexuality, bisexuality, pedophilia, bestiality, necrophilia, and perhaps others. People can be sexually attracted to many things, and each category has non-trivial numbers of members. But there is one potential category that appears to have zero membership: sexual attraction toward oneself. There is no one who is attracted to herself and to no one else. Of course, there is no shortage of self-love, narcissism, and autoeroticism, but what we don’t see is a class of people who find themselves sexually attractive and no one else. No one reaches adolescence, looks around, and finds no one and nothing to engage their sexual interest, except the person they see reflected in the mirror. No one is aroused only by his own body, finds only auto-pornography interesting, and wishes to date himself alone. No one ever falls in sexual love with herself to the exclusion of all others. There are no auto-sexuals.

This seems odd: why does human sexual variety run out of steam at this point? We all regard ourselves as potentially sexually attractive to someone, but we don’t find ourselves sexually attractive to ourselves. We might do so de re, if we see our own reflection and mistake it for a reflection of someone else; but we don’t desire ourselves de dicto, knowingly finding ourselves uniquely enticing. No one gazes at himself and finds the exclusive object of his sexual fantasies. Sexual orientation seems essentially other-directed, unlike attitudes such as esteem, approval, aesthetic appreciation, amusement, or sexual pleasure. The last is particularly telling: if a person can give herself sexual pleasure while alone, why can’t she find only herself sexually attractive? What is to prevent auto-sexuality? Why isn’t this a recognizable human grouping?

            It might be said that it is genetically excluded because the auto-sexual will never reproduce his or her kind. But the same is true of homosexuality and that doesn’t prevent it from occurring. Also, the auto-sexual (like the homosexual) might have sex with people she is not attracted to for the sake of children, and this might be a common way of proceeding. If people who are auto-sexual keep in the closet, while living outwardly heterosexual lives, then they will reproduce their kind. But there is no reason to believe in the existence of such people; they simply do not naturally occur. Is it perhaps that auto-sexuality is so taboo that everyone born that way represses it into oblivion, not admitting it even to themselves? There is no evidence for that and it is highly implausible (wouldn’t it have emerged in psychotherapy with some patients?). In addition, no such taboo exists, in religion or elsewhere, the reason being that there is no human tendency that requires such a taboo for its suppression. Auto-sexuality doesn’t occur because no one feels its stirrings. It is the other that engages our sexual predilections. Of course, masturbation is commonplace, but that is quite compatible with many forms of sexuality and never goes with auto-sexuality. Woody Allen’s old joke, “Don’t knock masturbation: it’s having sex with somebody I love” works precisely because masturbation is not like that: self-pleasuring is not an expression of sexual love for oneself. People don’t masturbate because they find themselves sexually attractive—not in the way they find other people sexually attractive. Autoeroticism is not auto-sexuality: the former is common, the latter non-existent.

            A more promising line of explanation is that sexual preference is tied to romantic love: no one is auto-sexual because that would require the possibility of romantic self-love, which is not possible.  [1] The objects of our sexual attraction are typically the objects of our romantic love, but you can’t be in love with yourself, so you can’t be sexually attracted to yourself. That certainly sounds on the right lines, but it is not clear that sexual preference and romantic emotion are so closely intertwined. What about a person incapable of romantic love—couldn’t this person still have a sexual orientation? If sexual orientation can be detached form romantic love, why can’t it extend to the case of auto-sexuality? You don’t have to love yourself romantically in order to be auto-sexual, just as you don’t have to love others in the case of other-directed sexual attraction. What if a person has different kinds of romantic and sexual objects—say, a man who is romantic about women but sexual about men? Is this psychologically impossible? Apparently not: so sexual attraction can float free of romantic emotion (though this is certainly not common). We therefore can’t deduce the impossibility of auto-sexuality from the impossibility of romantic self-love. The latter is impossible because of the emotions associated with romantic attachment (jealousy, insecurity, and so on), but the same is not true of sexual attraction—one need not have any of these emotions in order to feel sexually attracted to someone. So we still don’t know why auto-sexuality doesn’t occur.

            One might resort to the idea that sexual attraction primitively requires recognition of otherness—we just can’t be attracted to someone who is identical with ourselves (and is known to be so). To feel attracted to X you have to judge that X is distinct from you, or else there is no thrill, no sense of discovery, no crossing of boundaries, no release for the solitary ego. But this is unsatisfactory as an explanation: why should non-identity be so crucial, especially given that autoeroticism is common? Why couldn’t someone be born with a tendency to sexually desire only himself? What psychological law would this violate? What kind of cognitive dissonance might it produce? Wouldn’t it be very convenient and easy to achieve? There would be no need to venture abroad, no need to court and seduce, no need to take risks—you just direct all your passion at yourself. Other people you regard with a shudder, but find your own self the height of erotic fascination. You move from simple self-love and natural narcissism to sexual identity: you restrict your sexual desires to a very local object, viz. yourself. And yet this seems psychologically out of the question: you could not find yourself feeling this way about yourself and you could not will yourself to feel this way. We can perhaps imagine someone with extremely catholic tastes who finds everythingsexually attractive (an “omni-sexual”), including rocks and cactuses, but it is harder to imagine someone whose sexuality does not extend beyond his own self. We can also imagine someone who never actually has sex with another person or object, but not someone whose entire sexual orientation is self-directed. Or rather, though this seems imaginable, it is not a feature of the human sexual landscape in all its variety. It is a puzzle why we don’t find instances of auto-sexuality, though it strikes us intuitively as outside of human possibility. It seems not to be a viable option, no matter how capacious our view of sexuality may be. This is presumably why it is not a moral or political issue.

            We can distinguish type from token auto-sexuality: sexual desire limited to tokens of the type exemplified by oneself and sexual desire for the token of the type that is oneself. I have been discussing the latter, but the former is an interesting case in its own right: is it possible to desire only individuals who belong to one’s own type? That is, can one have a sexual orientation limited to people similar to oneself (a “simi-sexual”)? Could someone fancy only her twins? Suppose people had many twins, not necessarily related—you can find them everywhere. A simi-sexual will be a homosexual, but will also require a much greater degree of overlap—the other has to be like oneself down to a fine level of detail. We don’t find this kind of sexuality as things stand, but if there were many twins for any given individual we might find that there are people who fall into the category. Indeed, it seems rather natural; so it is not that people are not auto-sexual because they crave qualitative dissimilarity from their objects of desire—they might even welcome the lack of dissimilarity. It is numerical identity that puts a limit on human sexuality: you can desire people who are just like you, but you can’t desire people who are you. Everyone is identical to himself, so the objects exist to ground auto-sexuality, but no one wants to take this option, and no one is born taking it. When it comes to sexual preference we are all prejudiced against ourselves: we refuse to regard ourselves as (uniquely) sexually attractive. We turn ourselves down as potential objects of sexual love—no one ever takes the unit set of herself as the group to whom she is exclusively attracted.



Colin McGinn          

  [1] I discuss this in “Is Romantic Self-Love Possible?” in Philosophical Provocations: 55 Short Essays (MIT Press, 2017).


Analysis of Matter




Analysis of Matter



What is the general nature of concepts of matter? How are such concepts to be analyzed? Is there a general nature or only a plurality of concept-types? Ryle wrote a book called The Concept of Mind (note the uniqueness implied by “the”), arguing that mental concepts are generally dispositional in form; what would a book called The Concept of Matter contain? I will begin to answer this question by considering the concept of motion (someone could write a book with this title too), a concept vital to physics. So I am concerned with the analysis of matter-in-motion: how do we conceive of matter-in-motion, and how should we conceive of it? What is the “logical structure” of this concept? This question is not to be distinguished from the question of what motion is—or what it is for a material object to move. What is the logical form (conceptual analysis) of, say, “The earth moves”?

            Immediately we are confronted by a difficulty, because there is controversy about the nature of motion. Some say it is absolute and some say it is relative (I say it is both, but we will get to that). We don’t need to settle the question for present purposes, since we can consider my question under either assumption. So suppose it is relative: motion only makes sense against the background of a plurality of objects, consisting in relative change of position. Then we can say that motion statements are relational in form: roughly, “the earth moves” means “the earth changes position relative to some object x”. The object in question may be indefinitely distant from the object said to move, so motion is not a local property of an object (the motion of the earth is usually referred to the sun). This is not apparent on the face of the statement we are analyzing, but that is not generally any objection to an analysis. We say, then, that the predicate “x moves” means “x moves relative to y”. I shall say that the concept of motion is an object-introducing concept, meaning that it refers us to an object not initially supposed essential: it is as if we can’t speak of x moving unless we are first introduced to another object y. We can’t speak of motion in isolation but only in the context of a system of objects. Motion is essentially relational. To put it differently: motion is not locally supervenient; it depends on what is going on in the environment of the object in question. It is not internal to the object that is said to move. It is not an individualistic property. We need to be externalists about motion, recognizing that motion only occurs in a certain context—it is object-dependent or object-involving. The moving object is only so in virtue of being embedded in world in which other objects exist that confer motion on it. The state of motion of one object incorporates the state of motion of other objects.

            I have put the point in these ways because I want to explore an analogy between motion and mental content. According to a dominant tradition, motion was conceived as an inherent property of an object, intrinsic and internal. An object could be in motion even though no other object existed. But a counter-movement arose that questioned this idea: motion is something that essentially involves other objects, even remote ones. Similarly, there was a tradition that located mental content within the subject, so that what you mean or think is independent of anything in your environment: it is a matter of your brain or your inner subjective state. But a counter-movement arose that questioned this idea: content is something that essentially involves other objects, even remote ones. Thus we are treated to twin earth cases and other ways of demonstrating the object-dependence of mental content. The environment fixes content—as it fixes motion. Externalism about meaning and motion became received wisdom. And these doctrines were intended to capture the actual character of our concepts, which had previously been misunderstood. As the slogan goes, “meanings are not in the head”; and neither is motion “in the object”—it’s in the relation between objects. In my terms, the concepts of meaning and motion are object-introducing concepts. That is their logical structure—what they logically imply. Thus all the characterizations that are applied to the mind can be carried over to matter-in-motion: externalism, anti-individualism, non-locality, non-supervenience, relationalism, environmental determination, object-dependence, etc. We thought that motion was internal to objects, part of their inner nature; but now we see that motion lies in the connection between one object and another, a matter of their external relations. The concept of motion is therefore a concept with an internal complexity that extends beyond its initial appearance—dyadic not monadic, two-factor not one-factor. Note, particularly, that it characterizes a fact that extends across space to possibly remote objects, and indeed brings in every object in the universe. It is not just a property of an isolated object doing its thing locally, sublimely unconcerned about everything else. It isn’t like shape or mass or atomic structure. It is more like size or length or being up or down: things have these attributes only relative to other things not intrinsically.

            One might think that the relational analysis only works if we accept relative motion not absolute motion. But that is not quite right, because so-called absolute motion is not really absolute: it is motion relative to space, conceived as eternally static and at rest. As Newton understood it, the motion of a body occurs against the background of an unmoving spatial manifold: space stays where it is while objects pass through it. So all motion is relative to something, though not to other material bodies. This something, however, is highly local, being either contiguous with the moving body or pervading its volume. So motion is relative but local on this conception: moving through an enveloping space. And what it is relative to is of a different nature from the moving body itself—space not being a kind of matter. So the concept of motion relates the moving object to a surrounding entity—viz. space—relative to which it moves; it is still a relational entity-introducing concept (I say “entity” not “object” because space is not a material object in the style of the relative theory of motion). Logically, then, the two views are not that far apart, despite the difference of ontology. A truly internalist view of motion would suppose that motion is entirely intrinsic to the object, not even relative to space. Thus this kind of absolutist would insist that even if the surrounding space did not move relative to the object the object might still be moving: for both the object and surrounding space might both be moving! Only if we suppose space to be necessarily at rest can this possibility be ruled out, but even then the following counterfactual might be true: “If space were to move along with an object, that object would still be moving”. That is, the concept of motion allows for the conceivability of motion without change of relative position with respect to space. This makes motion super-intrinsic—independent even of space (as presumably shape is: things are not triangular relative to their surrounding space).

            Here our analogy proves helpful. Consider a super-internalist who holds not merely that mental content is independent of the environment but also is independent of the subject’s brain and inner subjective state. This internalist holds that content is completely intrinsic to concepts themselves and is not dependent on anything outside of it—not the brain and not the subject’s subjective experiences. He might maintain that the mind is not the brain but an immaterial substance, and that we could vary a person’s subjective state and keep his concepts constant. For example, we could vary his senses and their phenomenology while not changing what he thinks and means: his concepts are not supervenient on his brain states or sensory states. They are supervenient on nothing but themselves (and possibly the immaterial substance). The usual kind of internalism supposes that the independence concerns only the external environment, but this extreme kind of internalism takes concepts to be independent even of states internal to the subject (not including concepts themselves). Thus we have externalism, internalism, and super-internalism (“intrinsicalism”); and similarly we have three views of motion—relativity to remote objects, relativity to space, and relativity to nothing save itself. That last view may not be plausible—it may not even be coherent—but it exists as an option that someone might adopt. After all, geometric properties are not defined relative to space: a circle is not circular only in relation to non-circular space, whatever that may mean. In any case, the two leading contenders for the nature of motion both regard it as fundamentally relational—much as mental content is regarded as fundamentally relational. Reflection in both cases has persuaded us that a superficially monadic concept is really a dyadic one. In the case of the relative theory of motion the extra object can be remote from the given object, while in the case of the absolute theory (so-called) it is as proximate as could be. The absolute theory should not be saddled with the idea of completely non-relative motion, which makes dubious sense; instead it is a question of which entity motion is relative to and where that entity is located. The absolutist might say, “Of course motion is relative, only not to remote objects but to surrounding space!”

            But which theory is true? I will not attempt to adjudicate that question; I will merely note that both could be. That there is such a thing as change of relative position there can be no doubt, and if we choose to call that motion (not unreasonably), then relative motion exists. But it doesn’t follow that no other kind of motion exists: maybe there is absolute motion as well. Objects could move relative to each other and relative to space. If I say to you, “Don’t move till I get back!” I don’t intend to blame you for your motion as the earth moves; I mean relative to the room you are in. But I can also talk about motion with respect to space and mean precisely that (rightly or wrongly—rightly in my view). Thus we have two concepts of motion that coexist in our conceptual scheme, and hence two types of conceptual analysis.  [1] They vary in their ontology but they are similar in logical form. Accordingly, we recognize two types of property when we use motion words, so we conceive of matter in two different ways: bits of matter change relative positions, but they also change their relation to space—they are capable of doing both. It follows that our concepts of motion have different analyses. This is analogous to the claim that we have two concepts of content, wide content and narrow content, which can coexist. Both are legitimate and useful, though they are differently defined and serve different purposes.

            Do all concepts of matter fit this pattern? I have already suggested that concepts of shape or configuration don’t: here there is no submerged relationality, whether remote or proximal–internalism rules. Geometric concepts are not covertly object-introducing; they are self-enclosed and just as they appear. I think the same thing is true of the concepts of mass and charge: these are not defined relative to some environmental variable—we would not be right to be externalists about these properties. They look like dispositional concepts, and as such refer to interactions with other things; but the same thing is true of all dispositional concepts, mental or physical. No one is surprised by this kind of relationality; by contrast, it comes as something of a revelation to discover that motion is relative (it is somewhat similar with size and length). Motion is a bit like color in this respect: we start off thinking color is intrinsic to objects and then are surprised to find that it depends on relations to perceivers. Whether a given object is red depends on whether other distant objects (i.e. perceivers) see it as red; color isn’t written into the object considered in itself. So it seems that we have three types of physical concept in our repertoire: intrinsic (shape), dispositional (mass and charge), and object-introducing (motion and size). There is not a single homogeneous type; physics is made up of three distinct concept-types with three different kinds of analysis. It would be pleasant to report that psychology is likewise made up of three such types, and arguably it is  [2]; in any case, conceptual heterogeneity holds in the case of the science of matter. This is a result in the conceptual science of science.


Colin McGinn

  [1] Imagine a possible world stipulated to contain both sorts of motion: it contains an absolute space with respect to which objects move, as well as the more humdrum kind of relative motion. Then inhabitants of that world would need two concepts to cover the facts. It seems to me that in our world we have two sorts of concern to which talk of motion answers—practical and theoretical, to put it briefly—so we naturally employ two concepts. The case is somewhat like the concepts of weight and mass. Put tendentiously, one kind of motion might be designated realand the other apparent.

  [2] There are mental states with content like beliefs and desires, which are object-introducing; there are mental traits like irascibility and generosity, which are dispositional; and there are occurrences like being in pain or feeling moody that are non-relational and non-dispositional.


Ambiguity As a Species Defect




Ambiguity as a Species Defect



Ambiguity in natural languages is commonly regarded as a lapse from perfection. A perfect language would not contain ambiguity. Why is this? Because language is used for communication and ambiguity impedes communication. If an utterance is ambiguous, it is harder for the hearer to figure out the intended meaning; and in many cases it is not possible to do this without further questioning. Language is then failing in its purpose (or one of them), which is to convey information quickly and effectively. Ambiguity is the enemy of understanding. If we had invented language from scratch, or were constantly reinventing it, we would be thought guilty of poor craftsmanship—creating a defective product. Ambiguity is clearly not a necessary and unavoidable feature of language, since invented languages are often designed to be free of it. We can construct languages that contain no lexical ambiguity or syntactic ambiguity, as with standard formalized languages. So the defect of ambiguity is a contingent feature of natural human languages not a necessary feature of languages as such.  [1]

            Nor is the problem local or confined; a typical human language such as English is rife with ambiguity. Often we don’t notice it because the intended reading is so salient, but the formal structure of the language generates ambiguity all the time. The classic “I shot an elephant in my pajamas” is ambiguous in a characteristic way, i.e. it is not clear whether the modifier “in my pajamas” applies to the speaker or the elephant. The sentence “Old friends and acquaintances remembered Pat’s last visit to California” is said to have 32 different readings.  [2] The Chomsky favorite “Flying planes can be dangerous” has infinitely many counterparts (e.g. “Dating women can be dangerous”). Syntactic ambiguity is pervasive and prodigal. Thus we must be constantly on our guard against it for fear of failing to express our meaning. Language is an ambiguity trap that easily lures us into error. We are always in danger of failing to communicate given the formal nature of the vehicle. It could have been worse—our every utterance could have been dogged by ambiguity—but things are bad enough as it is. And yet ambiguity is not integral to the very nature of language. Apparently we have been sold a shoddy product, one expressly constructed to get in the way of communicating.  [3]

            Why is this? Why is human language so defective? The question acquires bite when we acknowledge that language is a biological phenomenon: it is an adaptation shaped by natural selection, encoded in the genes, part of our birthright as a species. It is as if we have all been born with a defective heart or liver that does its job only fitfully and inefficiently. True, our bodily organs are not perfect—they can become diseased and break down—but they are not like our language faculty, which has the defect of ambiguity built right into its architecture. So the question must arise as to how such a defective biological trait originated and why it has not been improved upon over time. One would think there was some selection pressure against rampant ambiguity—that it would have been remedied over time. Yet there is no reason to believe that human language is moving towards less ambiguity, lexical or syntactic. It seems content to remain stuck in its current lamentably ambiguous condition. This is a puzzle: why does ambiguity exist, especially on such a large scale, and why does it persist? It looks like a design flaw of major proportions, so why is it biologically so entrenched? Why don’t we speak unambiguous languages? Why do constructions like “flying planes” exist at all? It is doubtful that comparable ambiguities afflict the languages (communication systems) of other species such as bees, birds, whales, and dolphins; and it would be bad if that were the case given that such languages are crucial to survival. So why does our species settle for anything so rickety and unreliable?

            First we must recognize that this is a genuine puzzle—it really is strange that human language is so riddled with the defect in question. Why isn’t there a simple one-one pairing between sign and meaning? Why is the connection between sound and sense so loose? Let me compare linguistic ambiguity with what are called ambiguous figures, the kind found in psychology textbooks (e.g. the Necker cube or the duck-rabbit). These are aptly described as cases in which a given physical stimulus can be interpreted in two different ways—hence “ambiguous”. So isn’t the problem of ambiguity found outside the case of language, and isn’t it really not that much of a problem? But these cases are relatively rare and confined: they are generated by psychologists drawing sketchy pictures on pieces of paper. Seldom do we find anything comparable in nature: it is not as if vision by itself is constantly generating such ambiguities.  [4] We might wonder whether a patch of shade yonder is a black cat or a shadow, but such cases are not common and don’t generally disrupt the purpose of vision. It is not that vision is biologically constructed so as to lead to such uncertainties of interpretation. But in the case of language the problem is endemic and structural: ambiguity is both common and practically consequential. If vision were as prone to ambiguity as language, we would find ourselves in trouble (imagine 32 ways to see a snake, most of them not as of a snake). Ambiguity in vision is sometimes a problem, but it is not ubiquitous enough to thwart the purpose of vision (i.e. gathering accurate information about the environment); if it were, we would expect natural selection to do its winnowing work. But ambiguity in language really is a practical problem, as well as an inherent design flaw: it cuts at the very heart of communication. The question, “What did she mean?” can be pressing and momentous. And the reason for ambiguity in vision is obvious enough: vision is an interpretative, hypothesis-generating process, proceeding from an often-exiguous basis in the stimulus environment, so it must sometimes boldly venture alternative hypotheses. But language has ambiguity built into its syntax, its rules of sentence formation. It is constitutionally ambiguous.

            Sometimes a biological trait has a defect as an inevitable side effect of an adaptive characteristic. Thus it is with the human bipedal gait and large brain, or the giraffe’s elongated neck—there is a price to pay for the benefits conferred (in fact, this is true for all traits given that they all require nutritional upkeep). We can see this principle in operation in the case of those ambiguous figures—ambiguity as the price of inference. So could it be that the ambiguity of natural language results inexorably from some super-advantageous design feature? Suppose it resulted from the property of infinite productivity: you can only have that brilliant property if you also have some concomitant ambiguity. The trouble with this suggestion is that there is no obvious move from productivity to ambiguity—why should the former entail the latter? Mere combinatorial structure also doesn’t lead to ambiguity. Artificial languages are productive and combinatorial, but they don’t contain ambiguity. Nor can hypothesis generation be the explanation: true, we have to infer what someone means from the words he utters (along with context), but it is the words themselves that bear an ambiguous relation to meaning. The question is why language permits constructions like “Flying planes can be dangerous” to begin with. Why not just have the sentences, “Flying in planes can be dangerous” and “Planes in flight can be dangerous” (though the former sentence admits the reading, “Flying around inside of planes can be dangerous”)? The fact that the hearer is engaged in an inferential task doesn’t explain why ambiguity of this kind is so rampant and inbuilt. So it is hard to see how it could be a by-product of some desirable design feature; it looks like the product itself. If we just consider quantifier scope ambiguities, we see how inherent to natural languages ambiguity is—and that it is easily removable by some device equivalent to bracketing. So why does our language faculty tolerate it? Why not just clean up the mess?  [5]

            One possible explanation is that human language is so spectacular an adaptation that it can afford many rough edges and design failures (compare early wheels). It is so good that it can afford to harbor some vices—it’s still better than having no language at all. This may be backed up by the observation that human language is a recently evolved trait still going through its awkward adolescent phase—eventually it will mature into something more streamlined and fit for unqualified celebration. There may be something to this point, but it doesn’t remove all aspects of the puzzle, because it doesn’t tell us why language evolved with this defect to begin with when better options were in principle available, and there seems no evidence of any movement away from ambiguity heretofore. It might be suggested that ambiguity is like vagueness: it’s not a good thing, to be sure, but tolerable when the alternative is no language at all. I won’t consider this kind of answer further here, because it is difficult to evaluate without further evidence; but perhaps mentioning it serves to highlight the lengths we would need to go to in order to find an answer to our question. I don’t think we would find it plausible that the reason for intermittent blindness is that it is better to have occasionally blind eyes than none at all, but that is essentially what is being proposed by the explanation suggested—communicating by language is such a marvelous gift that serious defects in it can be lazily overlooked by the evolutionary process. Ambiguity is really not like the retinal blind spot. What if our language faculty enabled us to parse and understand only half of what is said? That would be rightly regarded as a grave defect, for which there is no obvious explanation. But ambiguity is rather like that—it really does impede successful communication. And even when it doesn’t, there have to be mechanisms and strategies that enable us to avoid its snares—it’s always less effort to understand an unambiguous sentence than an ambiguous one. Processing speech is certainly not aided by ambiguity. It’s not a blessing in disguise.

            It is an interesting question where human language will be in the distant future. Will its present level of ambiguity survive or will it become more perspicuous? Are we now placed on a linguistic path that cannot be altered? What would it take to impose selective pressures on the ambiguity-producing structure of our grammar? At present we have an evolved capacity that tolerates rampant ambiguity, yet functioning well enough to get by in normal conditions; but the architecture is fundamentally unsound, allowing for forms of words that could have many meanings apart from the one intended. Language should make things easier for the speaker and the hearer than it now does.  [6]


Colin McGinn    





  [1] I have found only one paper dealing with the question addressed here (though I am by no means expert in the linguistics literature): “The Puzzle of Ambiguity”, by Wasow, Perfors, and Beaver. There is no date on the paper or place of publication (I found it on the internet), and to judge from its content the problem it discusses is not generally recognized. I welcome any information about other published work on the subject.

  [2] My personal favorite is “Right is right”. Given that “right” can mean a direction, correctness, and moral rectitude, we already have nine different readings of this simple sentence. Now add in the ambiguity of “is” as between predication, identity, and composition and we generate many more meanings.  

  [3] I won’t discuss whether ambiguity exists in the underlying innate language prior to its expression in a particular sensory-motor format. It may be that all ambiguity exists at the level of spoken speech and results from the demands imposed by this medium; there may be no ambiguity at the more abstract level of universal grammar. I remain agnostic on this question.

  [4] “Is it a bird? Is it a plane? No, it’s Superman!”

  [5] It might be thought that ambiguity is useful as a means of concealment: say something that some people will take in one way while conveying a different message to others. But this is not a good explanation for why human languages are ambiguous in just the ways they are. It is grammatical rules themselves that allow sentences to be both grammatical and ambiguous, not pragmatic considerations of the kind just mentioned. Ambiguity surely didn’t evolve as a means of selective deception.

  [6] Philosophers are apt to speak of natural language as logically imperfect, as judged by the standards of some ideal language; but ambiguity makes natural language biologically imperfect, because the biological function of speech is communication and ambiguity gets in the way of that. Imagine if a given monkey cry could mean either “Predator nearby” or “Food in the offing”! Compare “Get thee to the bank!” said by someone in the vicinity of both a river and a lending institution. 


Against Causal Epistemology




Against Causal Epistemology



It has been suggested that knowledge of the natural world enjoys a causal foundation not enjoyed elsewhere. Thus such knowledge is deemed intelligible while other kinds are not. Accordingly, we can accept a notion of robust truth in relation to propositions about the natural world while other kinds of proposition must remain under suspicion of non-truth. Empirical science emerges as both true and knowable while mathematics and morals make dubious claims to truth and knowledge.  [1] It would be mysterious how we can have knowledge in the latter two cases, given that their subject matter doesn’t causally interact with our cognitive faculties. We must therefore seek some sort of non-cognitive account of those areas, accepting that knowledge is impossible as far as they are concerned.

            I propose to question this bifurcation at its root: causal epistemology is a misguided idea, applicable nowhere; it therefore provides no standard whereby we can judge other domains of thought and discourse. It isn’t just that causal epistemology isn’t generally required for respectability and intelligibility though it delivers those virtues in the case of some domains. Rather, there is really no such thing anywhere—it is a chimerical ideal. These are subversive claims given a longstanding orthodoxy, but they serve to undermine subversion in areas that matter to us.

            The idea of causal epistemology includes a picture and a promise. No one supposes that we currently possess a complete causal account (explanation, science) of human knowledge acquisition in the case of our knowledge of the natural world, but the idea of such an account is supposed to make us feel safe about human knowledge in the areas to which it applies (though not in others)—it provides a reassuring picture of what is going on. The promise is that this picture will be converted into a full-blown causal theory of how knowledge is acquired, which will vindicate such knowledge. The world makes a causal impact on our sense organs via physical stimuli (light, sound, pressure, chemical diffusion) and as a result we form beliefs about the world outside our skin. Empirical knowledge is an effect of physical stimuli; the causal relation puts us in “contact” with the world beyond, mediating between mind and reality. For example, I know that there is a cup on my desk because the cup on my desk is causing me to believe it is there: it is operating causally on my senses and my nervous system transmits this operation to my belief centers. Thus the connection between belief and fact is rendered intelligible; thanks to causation we can know such things. Moreover, this causal connection can be investigated by the empirical sciences, thus “naturalizing” such knowledge. By contrast, nothing like this is possible with respect to putative knowledge of mathematics and morals: numbers, sets, and moral values don’t causally impinge on us, generating the kind of “contact” required for knowledge. Such knowledge would be altogether mysterious and unintelligible; and so the truth status of mathematical and moral claims is brought into question.

            But the picture and the promise are precisely that—they are programmatic. Moreover, they are deeply mistaken about how empirical knowledge works. First we must make an important distinction—between the existence of a causal relation between belief and fact, on the one hand, and our having knowledge of such a relation, on the other. There is all the difference in the world between the cup’s causing me to believe in its presence and my thinking that this is so. This bears on the question of what my reason for belief is: is it that I seem to see a cup or is that I think that a cup is causing me to seem to see it? The latter suggestion is problematic: I don’t normally have such thoughts (I may even reject them); they presuppose the very item of knowledge they are supposed to ground (that there is a cup on my desk); and it is obscure how they can be explained in the favored causal terms—for how can the existence of a causal connection between a part of the world and my mind itself be a cause of a belief of mine? How does causation produce knowledge of causation, and how does the mind register psychophysical causation? No: the reason I have for forming my belief about the cup is simply that I have an experience as of a cup—not that I have an experience as of a cup causing me to believe in it (whatever that might mean). I have various sensory experiences as of the world being a certain way, and they function as my reasons for forming beliefs about how the world is. The concept of causation does not enter my deliberations or the reasons that figure in them. If there are such causal relations, they are not part of my reasoning process. My belief is justified to the extent that I have good reasons of this sort, but these reasons don’t advert to causal relations between mind and world.

            This raises a natural question: does it matter whether such relations exist so far as knowledge is concerned? Granted they do, but what is the relevance of that to epistemology as a normative enterprise? Here is a quick way to see the problem: suppose we abrogate the normal causal relations that (we believe) obtain between natural facts and our beliefs about them, replacing them with some other sort of relation—does that abolish knowledge? Suppose we postulate a pre-established harmony between belief and fact, superintended by God or Nature, with no causal interaction between mind and world, but exhibiting reliable counterfactual-supporting connections. By hypothesis there is no causal epistemology applicable to beliefs formed in this world, yet it is hard to deny that knowledge can exist in it: people have reasons based on their experience and there is a reliable link to truth. What if our world is like this (as Leibniz thought): does that mean no empirical knowledge is possible? Hardly. The ingredients necessary and sufficient for knowledge are present; it is just that there is no causal dependence between world and mind. The existence of causal dependence is thus extraneous to knowledge. What if physics abandons the notion of causality (as some have claimed it should): does that mean knowledge also undergoes defenestration? Rational reasons for belief would still exist in the shape of sensory evidence, despite the absence of causal relations.

            We should also note that causation by itself is never sufficient for knowledge and has nothing intrinsically to do with knowledge. The world is constantly impinging on our bodies, but it doesn’t generally produce knowledge thereby. Most of the time we don’t even notice it. And certainly causal interaction between inanimate objects has no tendency to produce knowledge in them. Epistemology is normative, but causality is not. Reasons are not merely effects. Nor is it true that a belief is simply triggered by an outside stimulus; belief acquisition operates against a background of other beliefs that are normatively relevant (“holism”)–it is not just stimulus-response psychology. Then there is the old problem of deviant causal chains. Suppose there exists an opaque barrier interposed between me and the cup preventing any light from it reaching my eyes, but by some strange quirk of nature the cup releases a chemical that enters my skin and causes me to hallucinate a cup (I know nothing about this). We have here a causal dependence between my belief and the cup—but do I know there is a cup in front of me? I don’t see the cup despite the causal connection, and it is doubtful that I have perceptual knowledge in this situation (why, is an interesting question). It can’t be just any old causal connection that leads to the kind of cognitive “contact” deemed essential to knowledge. Then too there is the question, bequeathed to us by Hume, of whether causation itself is intelligible: do we really know what causation is? Causation involves necessary connection, but isn’t that epistemologically problematic? Is knowledge of causation a secure foundation for epistemology? What if we can’t really know causation as it exists in objects? What if causation is itself a mystery? The causal theorist is remarkably sanguine about the epistemology of causation. Maybe all we are entitled to is the notion of contingent correlation, so that any attempt to invoke full-blooded causation is a mistake. It is certainly not a notion free of puzzlement.

            And there are the notorious difficulties concerning inferential knowledge. Not all of empirical knowledge is of singular facts that we can observe by means of the senses; some things we must infer. But these things are not the causes of our beliefs, so the causal theorist must resort to an indirect way of extending the causal theory in their direction (consider knowledge of the future). Knowledge of generalizations is particularly problematic, since general facts don’t impinge on our senses (e.g., laws of nature). Really the causal theory applies at best to a small subclass of knowledge claims—those that concern directly perceptible particulars. The rest get consigned to hand waving. The case of introspective knowledge is instructive: we don’t sense our own mental states yet we know about them. So we can’t apply our (rudimentary) causal understanding of sense perception to introspective knowledge: my pain doesn’t transmit physical signals to a special sense organ that enables me to perceive it. Whether the causal account of knowledge can be extended to introspective knowledge is a moot question; yet it is hard to deny that the mind makes “contact” with its own inner states. Knowledge of what I am now thinking hardly fits the supposed paradigm of perceptual knowledge: isn’t it just a straight counterexample to such a theory of knowledge in general?

            The upshot of these considerations is that a causal theory of knowledge hardly sets the standard for respectable epistemology. It is really just a vague picture of the mind linking itself to reality. This is not how knowledge works generally, and even in the cases most favorable to it there are serious problems. So it can’t be a point of criticism of some putative area of knowledge that it fails to live up to the high standards demanded by causal epistemology. So what if mathematics and morals fail to conform to the causal model? That model is defective even in the areas it is designed to cover. It is just massively tendentious to demand that other kinds of knowledge imitate this supposedly perspicuous model. Knowledge comes in many forms, covering many subject matters, and each form obeys its own distinctive principles and methods—perceptual, introspective, inferential, general, innate, logical, mathematical, modal, moral, aesthetic, political, historical, psychological, etc.  [2] Reference is much the same, and it would be equally misguided to insist that all reference conform to the case of reference to a perceptible particular. Maybe it is hard to devise a theory of reference suitable for all types of reference—and similarly for all types of knowledge—but that is no reason to deny that reference and knowledge have wide (and univocal) application. Equally, it may be that these concepts give rise to genuine mysteries, but again that is no reason to deny that they apply to the real world. We should not seek to deform (or reform) our ordinary understanding of a region of thought or discourse in an effort to squeeze that region into the causal model, given the problems inherent in that model (“theory” is too grand a word). We shouldn’t try to revise the semantics of a piece of discourse just because we can’t see how to subsume it under a causal model derived from another piece of discourse. Knowledge varies with the subject matter and is as heterogeneous as it is.


Colin McGinn      

  [1] This issue is raised by Paul Benacerraf in “Mathematical Truth” (1972).

  [2] This pluralist position is advocated by T.M. Scanlon in Being Realistic About Reasons (2014).


Achievement Concepts




Achievement Concepts



Knowledge is a certain kind of achievement, but belief is not. Belief aims at an achievement, but it is not in itself an achievement. To be an achievement something has to measure up to an ideal, and belief may not succeed in doing that (it may not be true and justified). It is therefore wrong to suppose that knowledge is a special kind of belief—as it might be, one that is especially vivid or strongly held. We might say that knowledge, being an achievement concept, cannot be reduced to belief, which is not an achievement concept: for aiming to achieve something is not achieving it. A belief is an attempt at knowledge, but it is an attempt that can fail (a lot if the skeptic is right).

            This suggests a general distinction among concepts: those that specify achievements and those that do not. What should we call the latter class? There is no preexisting label for it and the usual terms of art don’t work very well: we could say “factual” or “natural” but that would imply that achievement concepts are not factual or natural—which seems wrong. Perhaps we can say “basic” on the assumption that achievement concepts presuppose the application of non-achievement concepts, and just to have a convenient label. We need to include not just belief but also perceptual experience, sensation, dispositions, states of the nervous system, shape, color, electric charge—anything that does not amount to an achievement, i.e. measuring up to an ideal. To avoid misleading suggestions, I shall simply speak of A concepts and B concepts (think “achievement” and “basic”): ones that involve conforming to an ideal and ones that don’t. The case is like that of the primary and secondary quality distinction: these are just non-descriptive labels for a distinction that is easy enough to recognize but which has no preexisting description (to speak of objective and subjective qualities straight off would beg too many questions); better to keep things abstract and noncommittal. The essential point is that some concepts specify achievements and some don’t (merely being instantiated is not an achievement); there is something good (valuable, meritorious) about knowledge, say, in contrast to mere belief (which may be unjustified and false). To know is to live up to a norm, a standard, an ideal.

            What other achievement concepts are there? I suggest the following: seeing, remembering, justifying, acting, acting rightly, understanding, meaning something, speaking grammatically, and ability (there may be others). It is easy to see the rationale for including seeing and remembering because we have an analogy to the distinction between knowledge and belief: there are episodes of sensory seeming and apparent memories, so that we can fail to see what we seem to see or to remember what we seem to remember. There are attempts at seeing (hearing etc.) and remembering that fail to achieve their goal. Thus we have B concepts applying without the corresponding A concepts applying. The case of action is less obvious but on reflection the same structure applies: for an intention is precisely something that aims at action but can fail to be successful. We often fail to do what we intend to do. Every action is therefore an achievement, not just the obvious ones like winning a race or shooting the sheriff. The achievement consists basically in moving one’s body in the way necessary (and sufficient) to carry out one’s intentions. Whenever I act I measure up to an ideal, since my acting is the achievement of my intention. That is, in normal circumstances my action is an achievement (though not if I am stricken with a motor problem). If it is true that John went to the shops, then John achieved something thereby—moving his body in such a way that he went to the shops. He didn’t merely try to go to the shops (the analogue of belief); he actually succeeded in carrying out his plan, i.e. landing up at the shops. Even if I just crook my finger, I achieve something: my intention was fulfilled in the small movement of my finger. Just as knowledge is always a success, if often a small one, so action is always a success, if insignificant in the larger scheme of things. Acting is to trying as knowing is to believing. Thus action is an achievement concept (acting rightly clearly is).     

            Understanding is also an A concept, since one can try to understand and fail (or not try and also fail). To understand something—a theory, an argument, a sentence, a word—is to have achieved something. Similarly for acts of communication or meaning: to speak to someone and communicate is to have achieved a goal, not merely to ramble pointlessly. People often fail to communicate, to speak meaningfully, or to understand what the other person is saying. Mastery of a language is a kind of achievement, like mastery of chemistry or judo. It is something positive, valuable, good. Meaning is bound up with this norm—“meaningful” is a term of approval, while “meaningless” is not. Hence our talk of linguistic ability—and ability too is an A concept. Language, like cognition and volition, is an area of aspiration, of success and failure—unlike, say, being of a certain color or shape or electric charge. We have concepts that mark the fulfillment of such aspirations, such as knowledge, seeing, doing, grasping. But many concepts do not mark achievements, merely recording facts that do not engage our evaluative faculties—there is no particular merit in being yellow or a sensation of red or a cube or believing that there are fairies at the bottom of the garden. There is no aim such that these properties are the fulfillment of that aim; they are not any kind of success. They may be beautiful or harmonious or interesting, but they aren’t types of achievement.

            The two types of concept are different yet they are related. The following seems true: for every A concept there is a B concept (or concepts) such that the A concept can be instantiated only if the B concept is (are). Whether the B concepts are sufficient for the A concept is another question, but they are necessary. Thus knowledge requires belief, action requires intention, seeing requires seeming, understanding requires conscious experience, and so on. This dependence may encourage reductionist ambitions (or fears), but part of the point of what I am saying is that there is no conceptual reduction of achievement concepts to non-achievement concepts: the former concepts belong in a different conceptual space (language game, cognitive system) from the latter concepts. We have here a basic conceptual dualism, analogous to that between the descriptive and the normative. For some reason we are interested in achievements as well as plain facts, and we have fashioned concepts to express our interest. Conceivable creatures could have no interest in the kind of evaluation proper to achievements and hence employ no A concepts, but we are interested in such evaluative conceptualization. A cognitive creature could attribute beliefs and experiences and intentions but never speak of knowledge or perception or action—this creature is just not interested in the kinds of success or failure we see in such attributes. Indeed our ordinary concepts of knowledge, perception and action don’t figure in typical scientific accounts of the mind; instead B concepts tend to be the preferred mode of description. Whether someone’s mind measures up to an ideal is not of interest to theorists concerned with how the mind works, which is a matter of internal states and causal relations among them. We speak of belief/desire psychology not knowledge/action psychology. Who cares whether the subject really knows that minnows swim or whether she actually went shopping? In our ordinary dealings with each other we are concerned with such evaluations, but that is not the concern of the scientist (as regularly conceived).

            This may fuel eliminative fantasies, the feeling being that real facts can’t have the evaluative built into them: there is nothing good or bad in nature but only in our ways to reacting to it or describing it. Can’t everything in nature be described without using achievement concepts? I don’t agree with this point of view but I see why someone might be tempted by it: for the A concepts do seem to outstrip what can be stated using only B concepts—they conjure up more than nature objectively contains. Certainly they have no place in physics and chemistry, or even biology. Can’t we just eliminate these concepts and get on with the job of describing the basic traits of nature? What is true here is that the two sets of concepts are in different sorts of business, and one sort of business cannot be reduced to the other. Thus a skeptic may insist that in the case of meaning there is nothing real except experiences and linguistic behavior; there are no facts of meaning over and above these facts. Likewise there is nothing to action except internal acts of will and bodily movements; there are no peculiar doings—things that straddle the divide between inner and outer.

            The lesson I would draw is that the traditional metaphysical notions of property or attribute are suspect because they insinuate a monism that doesn’t fit the facts (compare the notions of a natural property and a moral property). There are achievements and there are non-achievements, but is there really a general category of “property” under which both these things fall? Doesn’t talking this way suggest that the distinction is not fundamental? Is knowledge really a property of a person (a “mental state”) just as belief is? Maybe we can talk this way for certain purposes, but it mustn’t be taken to show that the world is really divided into objects and properties—as if everything that is said of something is of the same ontological type. The temptation is to think that all genuine facts are characterized by non-achievements concepts—that the world is the totality of mental and physical states of affairs—so that we can do without (irreducible) talk of achievements in our austere account of reality. But a substantial part of our “ontology” comprises things that count as achievements, i.e. the fulfillment of ideals. To put it differently, much of reality has evaluation woven into it—hence the concepts I have called achievement concepts. Knowledge, say, is both a fact and a value.


Colin McGinn        



Absolute Deontology



Absolute Deontology



Kant’s position that there cannot be a case of morally permissible lying has not been met with much enthusiasm. The idea of absolute moral rules thus seems mistaken. W.D. Ross sought to remedy the problem for deontological ethics by qualifying the force of moral rules: instead of saying that we have an absolute duty to tell the truth, he suggested that we have we have what he called a “prima facie duty” to tell the truth—a duty that can be overridden in certain circumstances. This notion has always been obscure and the terminology less than satisfactory, though the problem it is intended to solve is real; also it seems to soften duties in a way that anyone sympathetic to deontological ethics will find unappealing. Do we never have an absolute duty to tell the truth? Are our duties always merely prima facie? I want to suggest an alternative approach to the problem, one that abandons the idea of prima facie duty.

            The alternative view appeals to the notion of different categories of lie. Consider the lying allegation intended to harm the person accused: that kind of lie is surely absolutely wrong. Lying to protect an innocent person is a good counterexample to a perfectly general prohibition on lying, but lying in order to harm someone is nothing like that—it precisely aims at injustice and suffering. I suggest that this duty is absolute: under no circumstances is it right to lie in order to get someone into trouble (and thereby get them into trouble). Call this the lying accusation: then we can say that the duty not to make lying accusations is absolute—not merely prima facie. Kant would be right about the status of this duty. By contrast there is no absolute duty not to lie to children: it depends on the circumstances—and there are plenty of circumstances in which lying is the only way to protect children. Likewise there is no absolute duty to tell people the truth about their physical appearance or level of intelligence. So some categories of truth telling are unqualified duties while some are not; it is not that all types of lying are only prima facie wrong. This is a good result because we don’t want to say that all types of lying are only wrong at first sight—but maybe not at second sight or on deeper reflection. When it comes to the lying accusation we don’t want to tell our children that this is wrong only in certain circumstances or at first glance—it is always wrong, very wrong, necessarily so. If a mother catches her child lying about another child, in order to get that child into trouble, the mother might say, simplifying somewhat: “It’s wrong to lie”. She means to be speaking of the kind of situation at hand, not cases of benevolent lying; and her message is that it is always absolutely wrong to lie in that specific way. She doesn’t follow Ross and cautiously intone: “It is prima facie wrong to lie”. If later the child asks her about benevolent lying, she will be beyond criticism if she replies: “I was talking about lying allegations not all possible forms of lying”. She expected the context to make her meaning clear, and no doubt it was clear; the child is being a philosophical pedant if she responds: “But you said the words ‘lying is wrong’ as if it applied to every possible kind of lie”.

            Semantically, the case is a bit like “camels have four legs”: that is perfectly true (absolutely true) but it isn’t true that every camel has four legs (some camels have lost or leg or two). Likewise, “lying is wrong” is not a universal quantification over every instance of lying; rather, certain kinds of lying are deemed universally wrong (necessarily so). It is intended as a conjunction of certain categories of lie, not every possible type of lie (e.g. lying to protect the innocent). It is wrong to make lying accusations about people and also wrong to lie in order to make yourself look better than you are and also wrong to mislead people for no reason—but not wrong to lie in order to protect an innocent person from falling into evil hands.

            It is the same for prudential duties. We can say, “you should eat in moderation” or “you should take regular exercise””, but we don’t intend to include eating in moderation after starving or exercising when you have the flu. This should not make us regard prudential imperatives as merely prima facie; rather, a more specific imperative would be absolutely binding, e.g. “don’t have a massive lunch every day” or “don’t sit around in the house all day (unless you are not well)”. It is not that prudential duties are merely prima facie duties; they are absolute duties once they are properly formulated. Ditto for duties of etiquette or rules of the road: of course it is sometimes right to violate the usual rules of etiquette or driving (as in cases of emergency), but it is equally true that some of these specific rules are universally binding, e.g. “Don’t make loud noises in church (for no good reason)”, “Don’t go the wrong way up a one-way street (in normal traffic conditions)”. And if we are talking about such rules in particular contexts the point is even clearer: “Don’t bang into that old lady”, “Say thank you to this shop keeper”, “Don’t turn left here”, “Don’t eat another plate of spaghetti”. This absoluteness is quite compatible with there being othercircumstances in which in which one should bang into someone or not say thank you or turn left illegally or have the extra plate. What we should not say is that lying is only prima facie wrong in all circumstances: in some circumstances—indeed in nearly all—lying is absolutely wrong. The concept of prima facie wrongness is not the right way to handle the possibility of exceptions to a universal quantification.

            It is a question whether this approach generalizes to all cases of duties, such as the duty not to steal or murder or break promises. We can all come up with examples in which it would be unduly rigid to insist that one not do anything of these kinds—to save the life of a child, to defend oneself against unwarranted attack, to prevent a catastrophe. But does that mean that these duties are merely prima facie?  No, because individual categories of these duties might be universally binding: one should never steal from the poor to give to the rich or murder an innocent child or selfishly break a promise because something better came up. A Kantian attitude about theseduties is correct, even if we don’t want to say that there are no conceivable circumstances in which stealing, murder, and promise breaking are morally permissible. The simple statement “Stealing, murder, and breaking promises are wrong” is just shorthand for a conjunction of these specific statements; it doesn’t need to be qualified and weakened by the use of the prima facie operator. In fact, moral duties are always absolute, admitting of no exception, once they are properly formulated. So we really don’t need to back off from Kant’s fundamental position—while disagreeing with him about certain kinds of cases.

            Here is a difficult case to end with: it is only possible to save the innocent child from the murderous storm troopers by falsely accusing someone of something and thereby getting them into trouble. You save the child’s life but only by falsely telling the storm troopers that your mother is a thief (thus landing her in jail). In this case I would say you have chosen the lesser of two evils, but the lying accusation was still an evil. But I can imagine someone sticking to the strict Kantian line here, and not unreasonably: you should not accuse your mother of being a thief in these circumstances even if it would save the life of the child. The case is in sharp contrast to merely flattering someone about his physical appearance when he is in fact in horrible physical shape, or not telling a child the awful details of how her mother died in a car accident. In these cases what you did was morally right not merely the lesser of two evils. Lying isn’t always morally wrong even if most kinds of lying are absolutely wrong.


Colin McGinn           



A Psychology of Philosophy




A Psychology of Philosophy



Most philosophers would agree that philosophy is a very difficult subject, in their heart of hearts if not in their practice. The problems of philosophy are difficult problems. They are not easily solved (sometimes not easily stated). The difficulty might be rated differently by different philosophers—from moderately difficult to extremely difficult to impossibly difficult. In any case there is a widespread perception of pronounced difficulty. I am concerned with the psychology that goes with that perception: what kind of mind is formed by the perception that philosophy is exceptionally difficult? Better: what kind of personality is created by the perception of philosophical difficulty? If a person gives his or her life to philosophy, while fully recognizing its exceptional difficulty, what kind of psychological formation will flow from that? If you devote your life to a subject that admits of relatively easy answers, you will likely experience satisfaction and a sense of achievement (as it might be, the flora and fauna of the Hebrides); but if you devote your life to a set of questions you believe you probably (or certainly) can’t answer, what will this do to you?

            The obvious reply is that you will not meet with success and you will suffer the pangs thereof. You want to know the answer and you strive to discover it, but you accept that you won’t succeed, probably or certainly. Suppose you have been struck with the problem of skepticism and you long to find an answer to the skeptic, but you accept that the problem is extremely difficult and that you have not discovered an answer, and probably never will. You could simply accept this fact, maybe reducing your efforts at solution, given that your chances of success are miniscule; that would be rational enough, considering. But you might also react by overvaluing your less-than-perfect efforts or by downplaying the difficulty of the problem. You might decide you were wrong about the difficulty of refuting the skeptic, or you might congratulate yourself on devising a highly ingenious or insightful response that has escaped the attention of others. This would be understandable, if not completely rational: you have succumbed to a kind of intellectual dishonesty (given that you are tacitly aware that your proposed solution doesn’t really measure up to the severity of the problem). You are engaging in intellectual bad faith. You do this because your earlier response was psychologically uncomfortable: why keep trying to solve what you believe you (probably) can’t solve? Why put so much effort into something pointless? And even if you think you might solve the problem, you still have to wrestle with the strong probability that you won’t solve it—given its extreme difficulty. It is psychologically uncomfortable to strive to do what you think you (probably) can’t do, so it is natural to revise your view of things.  [1]

            What I have just described is a situation in which cognitive dissonance is apt to occur. Cognitive dissonance theory was invented by the psychologist Leon Festinger in the 1950s and has become a standard part of psychology.  [2] The outlines of the theory are as follows. People are equipped with an overall drive towards psychological consistency. This drive is not restricted to belief consistency but applies also to harmony among desires, actions, emotions, and beliefs. If there is a lack of harmony, the subject will experience mental stress or discomfort or anxiety. This will lead to attempts at dissonance reduction by a variety of stratagems in order to restore harmony. The magnitude of the stress is a function of the magnitude of the dissonance. So a person will not be happy to act in ways contrary to his desires, or make statements contrary to his beliefs, or desire what he knows he cannot possess, or feel what he knows he shouldn’t feel, or believe what he has evidence against. Cognitive dissonance leads to dissonance reduction by modifying the dissonant psychological configuration. In Festinger’s original example, a member of a cult who is confronted by evidence undermining the tenets of the cult will be apt to deny the evidence or invent some ad hoc explanation for the evidence that preserves his cherished beliefs. Or a person compelled to work in an occupation that violates her values is likely to abandon or modify those values, or to insist that the occupation really serves to further them despite appearances (as it might be, pollution is actually good in the long term). The central point is that the drive for psychological consistency leads to bad faith of one kind or another. Dissonance is intolerable, so the subject strives to minimize or deny it, often by mental contortions and self-deception. Living with cognitive dissonance is harder than conforming one’s attitudes so as to avoid it. This is the realm of motivated belief and fake emotion and fabricated desire.

            How does this apply to philosophy? Simple: the acknowledged difficulty of philosophy induces cognitive dissonance, which is then massaged in various ways to reduce the mental discomfort. Difficulty produces dissonance because the life of a philosopher is predicated on denying or underestimating it. For the philosopher is investing time and resources in a project she knows is unlikely to bear fruit. She is working on impossible problems—problems she knows (or strongly suspects) she can’t solve. Suppose she is gripped by the problem of skepticism and is working hard to provide a convincing answer to the skeptic—burning the midnight oil, neglecting her family, not having any fun—all the while believing that her efforts have close to zero chance of success (after all, no one else has come up with anything). That is not a tolerable mental state to be in, because of the dissonance between will and belief: she wants and wills what she believes is not within her reach. Various dissonance-reducing reactions are possible: she can give up working on the problem, pretend that the difficulty is not as great as has been supposed, overestimate the value of whatever ideas she can come up with, become a dogmatist. In the extreme she could always declare that philosophical problems are pseudo problems or are meaningless or reflect mental illness. That way she can deflect the dissonance, restore her mental equilibrium, and relieve the stress. She might have started her philosophical life brimming with optimism—she will get to the bottom of these problems where others have failed (through insufficient attention to ordinary language, or by relying on a primitive type of logic, or by not knowing enough science, or because of general sloppiness). Then she felt no dissonance, because her beliefs were consistent with her desires and actions (such as protracted and expensive study of philosophy). But as her philosophical life wears on and the futility of her efforts becomes more apparent, she is likely to arrive at a more pessimistic view of her philosophical prospects—she comes to believe that she will likely not solve the problems that so gripped her (and still do). Then cognitive dissonance is apt to set in: she knows that she can’t achieve what she wants desperately to achieve. She could try to learn to live with this fact, though it would no doubt modify her practical motivations, or she could adopt a variety of dissonance-reducing stratagems. The two most obvious ones are denying the difficulty and overestimating her feeble efforts (“Yes, the problem is devilishly difficult, but my theory finally lays it to rest!”).

I take it this will seem familiar. My suggestion is that cognitive dissonance lies behind some of the characteristic types of philosophical posture. Hence the psychological appeal of logical positivism, ordinary language philosophy, the latest brand of scientism, dedication to formal logic, post-modernism, Wittgensteinian quietism, neurophilosophy, descriptive phenomenology, experimental philosophy, eliminativism, and engaging in purely historical studies. These are all attempts to avoid or downplay the recognition of the difficulty of philosophical problems. Not that they might not have arguments in their favor, but the psychological attraction of such doctrines stems from their promise to free us from cognitive dissonance. It is not that any of them are prima facie all that plausible—they are commonly represented as revisionist—but we feel compelled to accept them because of the stress produced by acknowledging philosophical difficulty. Add to this purely internal source of disharmony the institutional pressures of teaching or studying at a place of higher learning: how can you justify teaching a subject to students that you admit consists of insoluble (or at least unsolved) problems? What truths are the students supposed to learn as a result of such teaching? What results are you conveying to them? How can they justify the expense to their parents? Also: what justifies tenure in a field where no one ever makes any substantial progress? How can you be paid to work on problems no one thinks you can solve? There is a lot of pressure to deny that philosophy is as difficult as it seems to be and has proven to be. We just need that nice fat grant and we will finally answer the skeptic! The latest fad (as it might be, neuroimaging) will resolve those age-old problems, so we need not accept that things are as dire as they appear. We need not accept that we are striving to do what we know we can’t do. That is the fundamental problem, psychologically speaking—the cognitive dissonance at the heart of the philosophical enterprise.

Other subjects do not writhe under this kind of pressure. In them progress is made, large and small, so that the striving to achieve results is gratified, with the hope of further achievements in the future. True, they can contain great difficulties, but history has been kind to them, and there are many areas in which substantial progress has been made. The physicist never feels under pressure to deny that his questions are meaningful, or to resort to ordinary language, or to eliminate what he finds puzzling (“There are no receding galaxies!”). Nor does the biologist have to accept that she is getting nowhere—she has many discoveries to occupy her time. But in philosophy the main questions—the questions that bring us to the subject—remain maddeningly recalcitrant: the mind-body problem, skepticism, the nature of moral truth, free will, the meaning of life, space and time. Not that noprogress is made; rather, the core problems are so difficult that meaningful progress can hardly be expected. Once this fact is acknowledged (assuming it to be a fact) cognitive dissonance is the natural response: if it’s so difficult, why even attempt it? To that question we need a dissonance-reducing answer. Other disciplines are not so afflicted: they are not defined by problems of this magnitude of difficulty. So their practitioners are not subject to the same mental torment as philosophers; they have no dissonance to reduce (maybe psychologists have an inkling of what we go through in some of their more baffled moments).

Imagine a subject of study S that expressly advertises itself as dealing only with the most difficult problems known to man. Some people find themselves going into S. These peculiar people don’t just explain what the problems are and then sit back and marvel at them—they try to solve them. Isn’t S a ripe subject for cognitive dissonance? No one has solved a problem in S for hundreds of years and there is widespread pessimism about solving any of them; yet people persist in trying to solve them and promoting their subject as a worthwhile investment of time and resources. There is thus a lack of harmony here between beliefs about S and life commitment to S—and there is much frustration, self-doubt, neurosis, self-deception, etc. It wouldn’t be surprising if people occasionally sprung up proclaiming that they have discovered a new method for solving their problems (say, studying Sanskrit) or insisting that the problems of S are really pseudo problems—and they would no doubt find their relieved followers. It’s tough devoting your life to problems you don’t think you stand a chance of solving. That way lies acute cognitive dissonance, with its strategies of avoidance. Better not to go into S at all, despite its intrinsic interest; people only go into it because they believe (unrealistically) that they alone can solve the problems of S (ego trumps realism). Doesn’t S sound a lot like philosophy, psychologically speaking? It needn’t bephilosophy, but it would feel like it—it would reproduce philosophy’s characteristic psychological contours. The difficulty of S combined with devoting one’s life to it sets up psychic tensions that lead naturally to certain kinds of reaction—mostly involving bad faith. This is the psychological landscape occupied by the philosophical mind. In particular, we philosophers are always trying to find ways to make philosophy easier than we know it to be.

I have spoken of the individual psychology of the philosopher, which may be taken to be more or less universal given the nature of the subject, but there are also some more local sociological pressures conducive to cognitive dissonance. I mean those pressures (mentioned earlier) stemming from the institutional structure of a typical university and of the profession of philosophy, as it exists today. It is necessary to publish and compete and establish oneself as defending a certain position. You have to show that you are good at philosophy, in the sense of being capable of producing it; and this leads to excessive optimism about what can be achieved in the subject. In particular, you have to show yourself superior to others in solving philosophical problems. Thus you will develop a tendency to overestimate the quality of what you do while underestimating the quality of what other people do. Your views and theories are clearly correct while theirs are clearly incorrect. In teaching the subject you will be tempted to make it seem easier to make progress than it is, so that certain views will be favored as the demonstrably true ones, as opposed to those radically misguided alternatives. This will lead to a culture of exaggeration and overconfidence—a lack of humility in the face of difficulty. How can you stand out professionally if you meekly suggest that it is all very difficult? The cognitive dissonance created by the confrontation between the intrinsic difficulty of philosophy and the institutional structures within which it is practiced will lead to extreme ways of trying to reduce the dissonance—such as declaring your own position definitively correct and everyone else’s hopelessly confused.  [3] Thus it is that factions are formed and feuds triggered. Professionally, you have to have a thesis—a position, a doctrine. But this conflicts with the recognition that it is incredibly hard to come up with anything convincing in philosophy—there are always opponents and objections. So we have cognitive dissonance built into the structures of the institution of professional philosophy, and with it those dubious and dishonest strategies of avoidance—particularly, overestimating one’s own position and underestimating the difficulty of the problem. And isn’t this exactly what we in fact find in professional philosophy—the blowhard and the minimizer, to put it crudely? Also the cult of personality, the formation of “schools”, the withering contempt for those who refuse to see the light, the ever-changing fads and fashions, the dogmatism, the willful blindness, the haggard looks, the neurosis, the swaggering and posing—all attempts to deal with the cognitive dissonance created by philosophical difficulty as it interacts with professional existence. Just consider the familiar figure of the philosopher who thinks (or purports to think) that he has it all figured out: emotivism in ethics, materialism in metaphysics, nominalism in logic, naïve realism in epistemology—everything is bathed in sunlight with not a mystery in sight. This philosopher can see, and will brook, no objection to any of these firmly held views; all alternatives he rejects as absurd and dishonest. Philosophy contains no difficulty for this jolly optimist. Mustn’t we wonder at such a person’s brash confidence? Can he really believe it is all so simple and straightforward? Isn’t his breezy conviction the result of an underlying cognitive dissonance? He knows that things are not really so easy and yet this causes him such acute mental discomfort that he has decided to act as if he has it all figured out.  [4] This is intelligible enough from a psychological point of view, but it amounts to nothing more than a strategy for avoiding cognitive dissonance. At the other extreme someone might feel the difficulty with particular force and decide to give up the study of philosophy altogether. That would also resolve the dissonance and might impress us by its intellectual integrity. But most of us are stuck between these two extremes, suffering the symptoms of cognitive dissonance: it is only partially resolved in us, if at all. We have our cherished theories, so desperately cobbled together, but deep down we realize that they may be wide of the mark or just grotesque errors. To take an example more or less at random: there was a time when people convinced themselves that Davidson’s use of Tarski’s theory of truth supplied all that could be asked of a philosophical theory of meaning; and this position was held with almost religious fervor. It is hard not to see this in hindsight as a kind of bad faith prompted by the felt difficulty of the problem of meaning combined with the need to say something positive about it. The problem isn’t so hard after all–all we need to do is throw some fancy formal logic at it and it will disappear in a flood of biconditionals! Either that or we have to admit that we are trying to solve a problem we haven’t the faintest idea how to solve (or even formulate).

There is a psychology to philosophy, generated by the peculiar character of the subject, and Festinger’s theory of cognitive dissonance seems like a good theory of what that psychology is.  [5] It explains many of the phenomena we observe and fits the way philosophy feels from the inside. It is an empirical psychological theory like any other and should be judged accordingly. It won’t solve any of our philosophical problems, but it might alert us to psychic forces in us that distort our thinking and practice.


  [1] It is often noted that you can’t intend to do what you believe it is impossible for you to do, so no one could intend to solve a philosophical problem he believed could not be solved. But that leaves room for intending to do what you think it is quite improbable for you to do. However, that attitude is inherently unstable and disagreeable, especially as the failures and sense of futility mount. At what point do you give up? (Desire, of course, is perfectly possible in the presence of a belief that the desired state of affairs is impossible.)

  [2] Leon Festinger, A Theory of Cognitive Dissonance (Stanford University Press, 1957).

  [3] I have noticed over the years that people always seem to believe what they learned in graduate school, causally dismissing what has happened since. Pure cognitive dissonance.

  [4] It is not always a “he”, but statistically speaking…

  [5] What would a Freudian explanation of the philosopher’s psychology look like? Perhaps this: The difficulty of philosophy is experienced as a form of castration anxiety (of the intellect not the body), which is naturally repressed, and which manifests itself either as a denial of the castrating power of philosophy or as an assertion of the phallic prowess of the philosopher. Thus the philosopher rejects philosophical problems as meaningless or phony (and hence incapable of castrating him) or he elevates himself to superhuman levels of problem solving (phallic invincibility). The anxiety is thus allayed (how this fits the case of women philosophers is left for future research.) Maybe there was a time at which such an explanation would be taken seriously (in fact I invented it in the shower), but I prefer the Festingerian explanation to the Freudian one, having more to do with logic than libido. 


A Problem in Hume



A Problem in Hume




Early in the Treatise Hume sets out to establish what he calls a “general proposition”, namely: “That all our simple ideas in their first appearance are deriv’d from simple impressions, which are correspondent to them, and which they exactly represent” (Book I, Section I, p.52).  [1] What kind of proposition is this? It is evidently a causal proposition, to the effect that ideas are caused by impressions, and not vice versa: the word “deriv’d” indicates causality. So Hume’s general proposition concerns a type of mental causation linking impressions and ideas; accordingly, it states a psychological causal law. It is not like a mathematical generalization that expresses mere “relations of ideas”, so it is not known a priori. As if to confirm this interpretation of his meaning, Hume goes on to say:  “The constant conjunction of our resembling perceptions [impressions and ideas], is a convincing proof, that the one are the causes of the other; and this priority of the impressions is an equal proof, that our impressions are the causes of our ideas, not our ideas of our impressions” (p. 53). Thus we observe the constant conjunction of impressions and ideas, as well as the temporal priority of impressions over ideas, and we infer that the two are causally connected, with impressions doing the causing. In Hume’s terminology, we believe his general proposition on the basis of “experience”—our experience of constant conjunction.

            But this means that Hume’s own critique of causal belief applies to his guiding principle. In brief: our causal beliefs are not based on insight into the real powers of cause and effect but on mere constant conjunctions that could easily have been otherwise, and which interact with our instincts to produce non-rational beliefs of an inductive nature. It is like our knowledge of the actions of colliding billiard balls: the real powers are hidden and our experience of objects is consistent with anything following anything; we are merely brought by custom and instinct to expect a particular type of effect when we experience a constant conjunction (and not otherwise). Thus induction is not an affair of reason but of our animal nature (animals too form expectations based on nothing more than constant conjunction). Skepticism regarding our inductive inferences is therefore indicated: induction has no rational foundation. For example, prior to our experience of constant conjunction ideas might be the cause of impressions, or ideas might have no cause, or the impression of red might cause the idea of blue, or impressions might cause heart palpitations. We observe no “necessary connexion” between cause and effect and associate the two only by experience of regularity—which might break down at any moment. Impressions have caused ideas so far but we have no reason to suppose that they will continue to do so—any more than we have reason to expect billiard balls to impart motion as they have hitherto. Hume’s general proposition is an inductive generalization and hence falls under his strictures regarding our causal knowledge (so called); in particular, it is believed on instinct not reason.

            Why is this a problem for Hume? Because his own philosophy is based on a principle that he himself is committed to regarding as irrational—mere custom, animal instinct, blind acceptance. He accepts a principle—a crucial principle–that he has no reason to accept. It might be that the idea of necessary connexion, say, is an exception to the generalization Hume has arrived at on the basis of his experience of constant conjunction between impressions and ideas—the equivalent of a black swan. Nothing in our experience can logically rule out such an exception, so we cannot exclude the idea based on anything we have observed. The missing shade of blue might also simply be an instance in which the generalization breaks down. There is no necessity in the general proposition Hume seeks to establish, by his own lights–at any rate, no necessity we can know about. Hume’s philosophy is therefore self-refuting. His fundamental empiricist principle—all ideas are derived from impressions—is unjustifiable given his skepticism about induction. Maybe we can’t help accepting his principle, but that is just a matter of our animal tendencies not a reflection of any foundation in reason. It is just that when we encounter an idea our mind suggests the existence of a corresponding impression because that is what we have experienced so far—we expect to find an impression. But that is not a rational expectation, merely the operation of brute instinct. Hume’s entire philosophy thus rests on a principle that he himself regards as embodying an invalid inference.

            It is remarkable that Hume uses the word “proof” as he does in the passage quoted above: he says there that the constant conjunction of impressions and ideas gives us “convincing proof” that there is a causal relation that can be relied on in new cases. Where else would Hume say that constant conjunction gives us “convincing proof” of a causal generalization? His entire position is that constant conjunction gives us no such “proof” but only inclines us by instinct to have certain psychological expectations. And it is noteworthy that in the Enquiry, the more mature work, he drops all such talk of constant conjunction, causality, and proof in relation to his basic empiricist principle, speaking merely of ideas as “derived” from impressions. But we are still entitled to ask what manner of relation this derivation is, and it is hard to see how it could be anything but causality given Hume’s general outlook. Did he come to see the basic incoherence of his philosophy and seek to paper over the problem? He certainly never directly confronts the question of whether his principle is an inductive causal generalization, and hence is subject to Humean scruples about such generalizations.

            It is clear from the way he writes that Hume does not regard his principle as a fallible inference from constant conjunctions with no force beyond what experience has so far provided. He seems to suppose that it is something like a conceptual or necessary truth: there could not be a simple idea that arose spontaneously without the help of an antecedent sensory impression—as (to use his own example) a blind man necessarily cannot have ideas of color. The trouble is that nothing in his official philosophy allows him to assert such a thing: there are only “relations of ideas” and “matters of fact”, with causal knowledge based on nothing but “experience”. His principle has to be a causal generalization, according to his own standards, and yet to admit that is to undermine its power to do the work Hume requires of it. Why shouldn’t the ideas of space, time, number, body, self, and necessity all be exceptions to a generalization based on a past constant conjunction of impressions and ideas? Sometimes ideas are copies of impressions but sometimes they may not be—there is no a priori necessity about the link. That is precisely what a rationalist like Descartes or Leibniz will insist: there are many simple ideas that don’t stem from impressions; it is simply a bad induction to suppose otherwise.

            According to Hume’s general theory of causation, we import the idea of necessary connexion from somewhere “extraneous and foreign”  [2] to the causal relation itself, i.e. from the mind’s instinctual tendency to project constant conjunctions. This point should apply as much to his general proposition about ideas and impressions as to any other causal statement: but then his philosophy rests upon the same fallacy–he has attributed to his principle a necessity that arises from within his own mind. He should regard the principle as recording nothing more than a constant conjunction that he has so far observed, so that his philosophy might collapse at any time. Maybe tomorrow ideas will not be caused by impressions but arise in the mind ab initio. Nowhere does Hume ever confront such a possibility, but it is what his general position commits him to.



  [1] David Hume, A Treatise of Human Nature (Penguin Books, 1969; originally published 1739).

  [2] The phrase is from Section VII, [26], p. 56 of An Enquiry Concerning Human Understanding (Oxford University Press, 2007).