Internal Language and Consciousness



Internal Language and Consciousness


Do the words of inner speech consist of mental images of outer speech? If I say to myself “It’s raining” do I entertain mental images of my saying out loud “It’s raining”? Are inward utterances strings of such images? If so, that would indicate that inner speech is the internalization of outer speech, with the process of internalization consisting in image formation. And that would imply that inner speech is dependent on outer speech for its existence and nature—that it is a derivative phenomenon. On the other hand, if inner speech were not image-like that would imply that it is independent and autonomous, perhaps more basic. It is certainly true that we can form auditory images of the sounds of outer speech, as when you recall how someone pronounces certain words; but are these what inner sentences are made of? I think not: the two things are very different. First, images fade with time, but not so the constituents of inner language—they are not like memory images of words you have heard uttered. Second, they don’t feel the same: using words inwardly is not like hearing yourself speak, or anyone else—it lacks the sensory specificity of such images. Third, it is not to be supposed that a deaf person entertains auditory imagery when engaged in internal dialogue, yet he or she may experience inward speech in much the same way as hearing people. Nor is it that the deaf have mental images of sign language while the hearing have mental images of vocal speech—though these may make an appearance in the minds of the people involved—but rather that their inner dialogue proceeds in a modality-neutral manner. Fourth, such imagery lacks the combinatorial power of words in the formation of phrases and sentences, while inner words combine in natural grammatical ways (how would images distinguish between nouns and verbs?). Finally, people may differ in their powers of mental imagery, some being notably deficient in it, but they don’t thereby differ in their ability to talk to themselves; lacking such imagery altogether is no bar to self-communion. So it would be quite wrong to identify the lexical constituents of inner speech with sensory images of speech events; they are not “faint copies” of heard sounds (or possibly seen gestures). Accordingly, it would be wrong to suppose that internal language use derives from external language use by the process of image formation—by a process of internalization.[1]

What then is the phenomenology of inner speech? What kind of consciousness do we have of it? We are certainly conscious that we have such speech—it is not just a theoretical conjecture—but what is it like to talk inwardly? Here matters grow obscure: the phenomenology of inner speech seems curiously elusive and difficult to pin down. It hardly has a sensory feel at all, except around the edges. To judge from my own case, when I try to introspect its felt quality I come up with nothing definite, except that the harder I try the more I seem to force a sensory interpretation on my experience, as if desperate to find something concrete to cling onto. Most of the time the words slip by noiselessly, neutrally, blandly, a kind of wispy nothingness in the center of my thronging consciousness—a black hole in my sensory awareness. They are speech without the sound of the speech, experience without quality. Why is this? Here is a hypothesis: the internal language is largely unconscious.[2] The lexicon is hidden from us, except for fragmentary connections to heard language, so that we can’t say what word in this language corresponds to the English word “sound” (say). We have no conscious experience of the words of IL (as we may call this internal language). We are conscious that we have such a language, but its actual nature is inaccessible to consciousness—both lexicon and grammar. We are familiar with the idea of unconscious computations conducted in a symbolic medium; well, our inner speech is conducted in a medium whose character is hidden from our consciousness. It is a bit like vision: conscious percepts backed by unconscious symbolic processes. Thus I know what I am saying in my inner code, but I don’t know the nature of the code—I don’t introspect it. The words are nothing like conscious mental images of auditory impressions, but rather inscrutable elements of lived experience, unmistakable yet elusive. The picture, then, is of an unconscious cognitive system that is not derived from external speech and does not share its medium; rather, it is more plausible to suppose that outer speech is the externalization of this system clothed in materials appropriate to the senses. The process of externalization imposes a sensory quality on our awareness of language, or at least one expression of it; hitherto it was largely closed to conscious awareness, possessing zero phenomenology. Language is fundamentally an unconscious phenomenon only recently reaching consciousness in virtue of externalization.[3] It just happens that we are conscious of language in its sensory manifestation—and even here the sensory form is not to be identified with language as such. External perceptible speech is more of a symptom of language than the real thing. Consciousness and language are not natural bedfellows.

If that is true for inner speech, it is true a fortiori for what is called the language of thought. We should distinguish this idea from that of inner speech: LOT is not IL. The LOT theory, as propounded by Fodor, is motivated by considerations about the logical structure of propositional attitudes, particularly related to productivity: and any creature capable of propositional attitudes is credited with a LOT. It is not supposed that in thinking and belief formation we are saying the words of LOT to ourselves as we say the words of our inner language to ourselves. LOT is a theoretical construct not a datum of human conscious life (it is meant to apply to animal thinkers too, whether they engage in internal dialogue or not). LOT belongs at the level of computations in the visual system, i.e. not connected with consciousness at all. Our thoughts are indeed conscious but the symbolic medium in which they are encoded is supposed far removed from conscious awareness, so that we are not even aware that LOT exists. So the language in which we are alleged to think is not the same as the language in which we talk to ourselves: we might even call the latter language English (or Swahili), while that description is presumably false for LOT.[4] The idea of LOT is logically independent of the idea of IL—neither entails the other. The existence of IL is a matter of lived experience, while LOT is a daring hypothesis unconfirmed by direct awareness.[5] Accordingly, we can say that a typical human speaks three languages (individuating languages intuitively): LOT, IL, and EL. We have a language of thought, possibly universal among thinkers, human and other; a language for inner dialogue (or monologue), which is not possessed by all thinkers but which may be possessed by all normal humans; and a particular so-called natural language that is not universal to all human speakers (English or Swahili). These occupy different compartments of the mind, with two of the compartments hidden to consciousness and the other compartment open to conscious perception (at least as so manifested). The study of language therefore has three separate parts, with the third part distinctly marginal to the general phenomenon of language. Of course, it may turn out that these three compartments are at a deep level unified by a common set of linguistic principles, so that there is really only one language at work in all three areas; but prima facie we have three distinct linguistic faculties, operating in different realms and serving different purposes. Putting it chronologically, LOT is as old as thought itself, perhaps going back to the dinosaurs and beyond, while IL is of more recent vintage, date of origin unknown (let’s say two millions years ago), and EL is the new kid on the block, a mere 200,000 years old (and highly dependent on the larynx). Each resulted from suitable mutations driven by natural selection at vastly different times, but they each now sit in the human head guiding behavior: one language used as a vehicle of thought, another used to talk inwardly, and the third used as a means of communication. It would be quite wrong to think that the last adaptation is somehow more fundamental than the other two; they are simply different. It is as if evolution discovered language three times for different reasons and installed it in our nervous systems to serve three sorts of end. LOT is designed to enable thought to exist, IL is designed to improve thinking and solve problems, and EL is designed to get information across from one organism to another. Different functions, different languages—unless by some miracle they all converge on a set of linguistic universals.[6] Consciousness plays a minor role in this grand narrative, being mainly clueless except in peripheral ways. Nor need these languages obtrude themselves on conscious awareness given their job description: basically they are combinatorial systems designed to generate unlimited output, thus enabling a potential infinity of thoughts, internal conversations, and meaningful utterances. These are different sorts of output, and the means by which they are generated is of no immediate concern to the organism benefitting from them (compare vision). We happen to have three linguistic organs or organ-systems installed in our brain, a necessary duplication in the circumstances–as we have two eyes, two hands, etc. It is widely accepted that we have a number of memory systems, also designed with different goals in mind, possibly with different evolutionary origins, and so language is hardly anomalous in its variety (we also have five senses). We pay more attention to our communicative system for various reasons, but that is no indication of theoretical centrality or evolutionary priority. Language extends far beyond this domain and penetrates into every corner of the human mind (as well as the minds of many animals). Its relative unconsciousness is no impediment to this ubiquity. In particular, inner speech is not to be identified with some kind of echo of outer speech, as the image theory would suggest; it is a robust mental faculty in its own right, plausibly underlying the very possibility of outer speech.[7]


[1] We can also say, more generally, that the inner speech faculty is not the same as the imaginative faculty: imagination extends beyond language, obviously, but also internal language use is not a special case of imagination—which is not to say that the two don’t interface.

[2] Chomsky writes: “There’s good reason to believe that the inner language that we use all the time is inaccessible to consciousness, and that the fragments that reach consciousness reflect externalization—which, strictly speaking, is not part of language but of an amalgam of internal language and some sensorimotor system; it’s like the relation between a program in a computer and the particular printer that’s used to externalize it (so philosophers of language aren’t really discussing language).” (Personal communication, June 2020)

[3] I mean recent in evolutionary time: prior to the onset of vocal speech about 200,000 years ago language was silent and unperceived, not present to consciousness except obliquely; when it connected to the vocal apparatus it became an object of conscious apprehension (though not in its full nature).

[4] We speak this way because of the obvious connections to the language we grow up to speak not because IL is literally identical to spoken English (or Swahili), but LOT is not connected to our spoken language in that way, operating as it does entirely behind the scenes.

[5] LOT is regarded as constitutive of thought, part of its inner essence, while IL is best seen as an aid to thought, and strictly external to it. Thus one language operates to aid in the deployment of another language—IL helps LOT form the right beliefs. In other words, by speaking to ourselves we regulate what LOT will enter in the belief box. If we call the first language L1 and the second L2, then we can say that L1 helps organize L2 so as to serve the purposes of the belief system. That is, we get better beliefs if we use our inner speech faculty to think.

[6] A possible hypothesis is that IL is the origin of both EL and LOT instead of being quite separate from LOT. That would imply that LOT did not exist before IL came into existence, which would imply either that dinosaurs didn’t think or that they spoke to themselves. I suppose these are possible scenarios, but the more plausible theory is that thought came into existence before inner speech evolved, thus ruling out the hypothesis that IL is the sine qua nonof LOT. Still, we should keep an open mind on the question. Alternatively, it may be supposed that LOT is the basis of both IL and EL: I think this would be wrong for various reasons, but I won’t go into the question now.

[7] This paper is offered as a summation of several controversial lines of thinking not as a complete defense of the positions presupposed, for which the interested reader would have to look elsewhere.


A Theory of Language


A Theory of Language


In A Theory of Justice Rawls suggests a thought experiment designed to rid us of biases in our thinking about justice, labeled the Original Position. We find ourselves occupying a certain position in society, perhaps an unfairly privileged one, and we need a way to prescind from this contingency, so we imagine ourselves not to have knowledge of the facts of our actual position. In this way we can gain an impartial perspective on what constitutes a just distribution of resources. I want to suggest something similar in the philosophy of language: we are familiar with certain concepts and perspectives on our language that arise from the contingencies of our linguistic history, but these may give us a biased view of the real nature of language, so we need a way to prescind from such contingencies. We need to imagine ourselves not possessing our customary knowledge, stripped down to the basics of linguistic life; then we can start from scratch in producing a theory of language. That theory might be quite revisionary of our ordinary conceptions, as Rawls’s theory is quite revisionary of common assumptions about justice. He was trying to get to the very heart of justice; we are trying to get to the very heart of language—by suspending common assumptions (the method is also similar to Husserl’s epoche). So: what does our commonsense theory of language look like and where might it go wrong?[1]

Our commonsense understanding of language invokes a range of concepts familiar to philosophers of language: meaning, reference, sense, ambiguity, communication, intention, convention, speech act, assertion, command, question, grammar, logical form, use, word, sentence, speaker, hearer, implication and implicature, truth conditions, translation, and others. These make up the tacit theory we bring to bear on language, which are then converted into theoretical terms of an overall theory. They belong to a semantic meta-language. Maybe we acquire this meta-language in the course of our linguistic upbringing, maybe it has an innate component: in either case it comes to us naturally and spontaneously, without any self-consciously explicit theory construction. So let’s imagine that we do not acquire this set of concepts in the way we do but rather reach linguistic maturity without developing any such self-understanding: we speak an object language (say, English) but we don’t speak a meta-language about it—we have no thoughts at all about the nature of the language we speak. We don’t even have the concept of meaning. Perhaps animals that use symbolic systems—whales and dolphins, say—are in this cognitive situation: they speak and listen but they don’t theorize at all. They just act and react. Then the question is what account of language might we develop starting from this position of complete ignorance. Would it resemble the theories we currently espouse? Would it employ the concepts we employ now? We have no folk semantics to rely on, so what would a properly scientific semantics look like? If we could wipe the slate clean, what would we write on it starting from scratch? Would everything from our folk semantics survive, or just some, or none? In particular, are there any biases built into our existing theories, arising from the contingencies of our history?

Certainly there have been philosophers (and linguists and psychologists) who have detected such biases and recommended rejecting them. Quine, say, recommends rejecting most of folk semantics as irredeemably unscientific: meaning, reference, intention, synonymy, etc. He proposes replacing this obsolete theory with a behaviorist theory: the central concepts of scientific semantics are stimulus and response, stimulus meaning, holism, indeterminacy, and radical translation. Wittgenstein attacked the idea, supposedly part of common sense, that every word functions as a name, which he attributed to a natural human failure not to make distinctions; and he proposed replacing concepts like denotation with the concept of use. Others have rejected sense or meaning as anything over and above reference. Still others have expanded folk semantic concepts to include revisionary theories of language, as with Frege’s semantic theory (e.g. sentences denote truth-values construed as objects). The idea of criticizing our habitual thinking about language has not been off the table, but it has not been as radical as the sort of criticism we see in physics or biology (except perhaps in the case of Quine). And it has not been suggested that we are prone to systematic error because of a historical contingency regarding language, making us miss the essence of language. The contingency I have in mind is that we happen to use language for communication, but this is not the primary or original manifestation of language. We might never have used language to communicate, but we have shaped our understanding of language through consideration of this contingent expression of linguistic mastery. To put it directly: the use of language as a tool of thought is the primary manifestation of language, and it is not subsumable under the kinds of concepts we bring to language used as a vehicle of communication. If we had started by considering this kind of internal language, we would have arrived at a very different understanding of the essence of language; as it is we have been biased by the contingent circumstance that we use language to communicate. For various reasons communicative uses of language are salient to us, and certainly they require their own treatment, but they are not the only or the central uses to which language may be put. They are, in fact, derivative and parasitic, reflecting more about the human sensory-motor system than the inner workings of language as such.

I am not going to defend the cognitive view of language here—the idea that language is primarily a tool for thinking.[2] I will merely observe that a huge amount of language use goes on outside of any communicative context, being “in the head”. We talk to ourselves constantly and use language to help us solve problems (as well as being a resource for contemplation). A person can lack the ability to speak and hear and yet still employ language internally. A reasonable hypothesis is that language evolved initially as a cognitive aid and only later become adapted for communicative use (about 200,000 years ago by most estimates). It became integrated with our sensory-motor systems, thus inheriting their architectural features. Prior to that it functioned very much like our imagination—another human faculty that evolved as an aid to thought and which exists primarily as an internal system. We may suppose that language and imagination coexisted in the human mind for a considerable time quite cut off from any direct behavioral expression; they were not initially designed as communicative systems. Today imagination is still in this condition, but language was partially repurposed as a means of communication while preserving its cognitive function. Both imagination and language are essentially combinatorial systems with infinite potential, and this confers great utility on them in solving problems; but whether this structural feature becomes expressed in vocal speech is entirely contingent. We can imagine imagination achieving public expression in somewhat the way language has by coopting a sensory-motor system that reveals imaginative acts as they occur—say, by displaying visible images on the body of the organism—but as things stand this adaptation has not occurred. Likewise language could have remained internal and performed its cognitive function without the evolution of a sensory-motor system that embeds it in acts of communication. It seems to me entirely possible that many animals employ an internal symbolic system in their cognitive lives—a language of thought essentially—without ever transitioning to spoken language.[3] That is just an entirely different biological system with a different function and morphology (sounds, gestures, larynx, ears). The human species has been using internal language for its entire history, reaping the cognitive benefits, and only lately began incorporating it into their communicative repertoire. Presumably a mutation (or combination of them) led to this ability long before any thought of communication entered the picture: language has been aiding thought, solving problems, enabling contemplation, for lo these many years, well before overt speech took off. If speech were to cease tomorrow by some catastrophe, language would carry on performing its vital cognitive function. Of course, internal language use is inaudible and invisible, while external language use is noisy and visible (consider sign language), which may give salience to the latter compared to the former; but in point of biological priority, functional utility, and sheer pervasiveness internal language surely has the edge. We talk to ourselves all day but only occasionally do we open our mouths (and what an effort it is!). Internal language is woven into our entire conscious life (even when sleeping). Without it our minds would be much diminished.

The point of urging this perspective on language is to expose the parochial character of our current theorizing about language—the bias built into it, the skewed perception. Suppose we had never developed the ability to speak externally, confining ourselves to inner speech: what would the theory of language look like then? The entire apparatus of folk semantics for outer language (better, outer expression of language) would be redundant in so far as that apparatus is directed to communication. Two epistemic features of communicative speech may be noted: uncertainty about what the speaker has in mind and uncertainty about his or her truthfulness. We often don’t know who or what the speaker is referring to (has in mind) and we don’t know how reliable she is. Thus we must bring to bear the concept of reference and the concept of truthfulness. There is an epistemic gap between speaker and hearer—the hearer can’t read the speaker’s mind—and that gap must be negotiated in acts of communication. But no such gap obtains between the user of internal language and the recipient of such use. So there need be no concerns about what the speaker has in mind or whether he is being truthful. These reasons for bringing in the concepts of reference and truth do not exist for the internal use of language—so what purpose do they serve in the theory of internal language? The first thing we would light upon in theorizing about this use of language is its combinatorial structure, because that is what powers its central function—creative thought, problem solving. If we did come up with the concepts of reference and truth it would be an afterthought not a central concern, as it is for public language. Problem-solving potential would be the central concept, and this would lead us to focus on the constitutive structure of language—on syntax, basically. Grammar, recursion, infinite scope, embedding, transformations—these would be the properties to engage us. These are what drive the cognitive engine, making language apt for its function.

What about meaning? Would we make a point of assigning meaning to internal language? The word “mean” has two uses: applied to speakers and applied to words. The OED gives this for “mean”: “intend” and “intend to convey or refer to”. The primary meaning of the word is in application to agents who mean things, linguistically as well as otherwise. Words mean things in so far as people mean things by them (this connection is articulated in Grice’s account of meaning).[4] But this use of “mean” does not apply to internal language: a user of an internal sentence does not intend to convey information to anyone, still less to harbor Gricean intentions. We don’t mean things by the words we use when we engage in inner speech (that word itself is dubious—is this any kind of speaking?). But if internal language is not meant, does it mean? Is our inclination to say that it is meaningful a kind of projection from outer communicative speech? Would someone in the original linguistic position, with no spoken language to deal with, reach for the concept of meaning—our concept of meaning—in describing internal language? Maybe some notion of linguistic significance would be invoked, possibly the notion we apply to non-human languages, but would we say that our internal sentences have meaning? Who to and for what purpose? Who cares about what the sentences mean so long as they do their job—aid in problem-solving thought? I need to know what you mean in order to get information from you, but do I need to know what I mean when I am trying to solve a problem? My purpose is not to mean things by my inner sentences but to employ them in a cognitive task. So the concept of meaning seems irrelevant to internal sentences; maybe they have meaning, but who cares? What we are concerned about is their ability to facilitate cognitive functioning, which is a matter of their inner architecture and functional properties: syntax not semantics, roughly.

And so on down the line: speech act, intention, language game, convention, use, communication, assertion, command, and question—none of this is relevant to internal language. It is relevant to only one expression of language in acts of overt speech, which involves a sensory-motor apparatus extrinsic to language itself as a cognitive structure.[5] Just as vocal speech is not integral to communicative language as such (consider sign language), so overt speech in general is not integral to language as such—not part of its core. It is a relatively recent evolutionary step, not prefigured in the origins of the language faculty. Making outer speech the focus of language studies is like making those external signs of the imagination I mentioned earlier into the focus of imagination studies. What is called philosophy of language, as practiced by analytical philosophy heretofore, is really philosophy of one-expression-of-language, and not even the most central expression. It’s like focusing on writing while ignoring speech and calling that linguistics. Internal language has a far stronger claim to be basic and original. The mind consists of various faculties—mathematical, scientific, moral, imaginative, and linguistic (among others)—and the last item is no different from a locational point of view. True, we can manifest these internal faculties in external form by engaging our senses and motor capacities, but none of them is rightly seen as constituted by such exterior manifestations—including language. The theory of language must accordingly recognize that fact and shape its concepts appropriately. That will involve at least a shift of emphasis, possibly large-scale revision, in which some familiar concepts recede and others seize center stage. To put it bluntly, syntax will take over from semantics, at least as semantics is currently conceived. We have been bedazzled and distracted by the noises we make when accessing our internal language faculty and by our feeble attempts at interpersonal communication, when all along language was purring along inside us doing its vital work.[6]

[1] In physics and astronomy the analogue of position in society is position in the universe. We live on the surface of a small planet and our view of the universe is shaped by this contingency (e.g. the sun looks to rise); to do physics and astronomy correctly we need to prescind away from this and form a conception of the universe that excludes such factors. We need to replace the science we naturally assume given our contingent position with a science that gets to the heart of objective facts. We need to do this in physics and astronomy but also in theories of justice and in theories of language.

[2] Chomsky defends it in Why Only Us (2016) and What Kind of Creatures Are We? (2016).

[3] These symbolic systems might be as species-specific as human language is, so that internal discourse varies dramatically from species to species.

[4] Here is an interesting fact: we say “John means ‘unmarried male’ by ‘bachelor’” but not “John means ‘bachelor’ by ‘unmarried male’”, where we could paraphrase “means” by “intends”—despite the synonymy of the two terms. Likewise it is right to say “’bachelor’ means ‘unmarried male’” but it sounds wrong to say “’unmarried male’ means ‘bachelor’”—again despite the synonymy. This shows the affinity between “means” and “intends” and also the close connection between speaker meaning and word meaning.

[5] Ambiguity and the arbitrariness of sound-meaning pairings are properties of external speech, creating genuine problems for speakers and hearers, but they don’t trouble internal language users: we are not in doubt about which meaning we had in mind, and whatever arbitrariness there is in internal symbols poses no problem of inward interpretation. Internal language is peculiarly transparent.

[6] The conception of language I am working with will no doubt sound completely alien to the typical contemporary philosopher of language, but close students of Chomsky’s work will find it comfortably familiar. Time to get some reading done.


Academic Freedom

Yesterday I had the pleasure of conductng my first seminar in eight years. We discussed two papers of mine from this blog. We used Zoom. The students were excellent, the discussion scintillating; it was just what a philosophy seminar should be. You ask what institution organized this seminar–is it perhaps indicative of a thaw? Perhaps some small courageous American college dedicated to academic freedom? It was with philosophers at Moscow University at the Center for Consciousness Studies, not an American among them. Does anyone see the moral of this?


Is Reference Real?




Is Reference Real?


“Sitting next to her, talking with her about the simplest and most insignificant subjects, Prince Andre admired the joyful shining of her eyes and smile, which referred not to what they were saying, but to her inner happiness”: Tolstoy, War and Peace, Volume II, Part Three, chapter XVII.


The standard position is that the concept of reference applies to some expressions but not to all. Which expressions refer is controversial: doubt has been cast on predicates, definite descriptions, “I”, quantifiers, connectives, adverbs, numerals, feature-placing sentences, and fictional names. The paradigms of referential expressions are taken to be names of persons or things and demonstrative pronouns: there are entities for these words to refer to and they succeed in referring to them. Reference is taken to be a real relation that can hold between words and things, but it may not hold in some cases. Not just any old expression can stand in this relation; the expression must contain a special power of referring, not conferred on every expression. Perhaps the power is derivative on a speaker’s performing a referential act, so that word reference is not an intrinsic property of the word, but it is still supposed that reference is not granted to any and every meaningful word. It is an elite property, marking a word out for ontological distinction: the word reaches majestically out to reality, sword-like, while non-referential words are semantically limited to language itself, trapped inside it. The referential words provide the hooks linking language to the world. They are lines thrown out to reality.

But this position is not the only one: Frege’s position stands in stark contrast. According to Frege, every expression refers: not just names and demonstratives but predicates, connectives, quantifiers, sentences, and that-clauses. A menagerie of entities is wheeled in to provide the reference for these expressions: first- and second-level concepts, truth functions, truth-values, and indirect senses. If each word refers, there must be something to which it refers; the relation needs relata. Reference thus comes cheap and is thoroughly democratic, making no distinctions. It is not a privilege possessed by names and other singular terms but a property of every meaningful expression. Language has hooks to the world at every point. Many philosophers have felt that this is excessive, that it empties the concept of reference of all significance; they evidently have a sufficient grasp on the notion of reference to feel this in in their bones. Another position, seldom if ever occupied, is that no word refers–not even the supposed paradigms of reference. True, traditional empiricist theories of meaning, which equate meanings with “ideas” or mental images, make no mention of a reference relation to the non-mental world; but even these theories don’t explicitly deny that reference applies to language. The neglected position does just that: it asserts that reference is a pseudo-concept, applicable to nothing. Frege took language to be semantically uniform with respect to reference and so does this third position, but in the opposite direction; only the standard position adopts a mixed perspective—partly referential, partly not. The uniformly non-referential position holds that there is no such relation as reference, not even for names—though it is not denied that every expression has meaning (determinately so). Reference is a myth, a philosopher’s invention, a trick of the imagination. Words don’t refer, period (and neither do speakers).

It must be admitted that the two pure positions enjoy a theoretical advantage: they don’t divide words into two groups characterized by different semantic concepts. The standard position posits an awkward theoretical disunity: on the one hand, the robustly referential expressions (or speech acts); on the other, the etiolated non-referential expressions locked up in themselves. At least Frege’s universal referential realism makes no such division and accordingly affords a generalized conception of meaning—all meanings are designed to generate acts of reference. The mixed position, by contrast, supposes that some meanings have the job of generating reference and some do not—and that seems like an unattractive lack of unity in the concept.[1] Also, there is intractable controversy about which expressions are referentially endowed, which casts doubt on the clarity of the concept of reference, as philosophers are wont to understand it. Some people have no doubt that definite descriptions refer while others stoutly deny it, and intuitions seem hazy on the question. Some people insist that predicates refer to properties while others vehemently deny it, and the issue is difficult to resolve. In certain moods Frege’s system can strike one as theoretically compelling, but then one snaps out of it and reverts complacently to the mixed position. We may find ourselves accepting that reference has its varieties, but we stubbornly maintain that expressions either have it or they don’t—even as we lack clear criteria to make the distinction. A more uniform position would seem desirable, but Frege’s position is too extravagant to contemplate; and the third anti-referential position just seems manifestly untrue, since surely some words refer!

The view I am going to defend is the anti-referential view, but this will require some subtlety and alertness to philosophical illusions. The basic idea is that the philosopher’s concept of reference is a misplaced reification of a common sense notion: it invests more in that notion that it can really bear. I will begin by consulting the dictionary, always a solid starting point. The OED has this for “refer”: “to mention or allude to” and then “(of a word or phrase) describe or denote”, followed by secondary meanings such as “send someone to a medical specialist”. Under “mention” we find “refer to briefly” and “refer to (someone) as being noteworthy”. So we have a speaker’s use of “refer” that includes mentioning and alluding, and then a linguistic use of “refer” that includes describing and denoting. Clearly “mention” is not a synonym of “refer”, so we can’t trade on its familiarity to validate the philosopher’s use of “refer” (we often mention that such and such but we can’t refer that such and such). The closest we get to the philosopher’s use is “describe or denote,” but “describe” doesn’t cut it since not all referring expressions are descriptive—and don’t predicates describe? So we are forced to fall back on “denote”. Here is the entry for “denote”: “be a sign of; indicate” and “stand as a name or symbol for”. The first thing that strikes one here is the generality of the notion as so defined. The notion of being a sign of or indicating is not even restricted to linguistic contexts: it is the analogue for denotation of Grice’s notion of natural meaning. Clouds can denote rain under this definition, since they can be a sign of rain or indicate it. So this cannot be the definition we are seeking if we want to capture the philosopher’s notion. We turn then to the second definition: here the generality is at least confined to language, since it concerns symbols. But isn’t “red” a name or symbol for redness? Isn’t “and” a symbol for conjunction? Isn’t “all” a symbol for universal quantification? Isn’t every word a symbol for something? Then everything in language denotes something, according to this definition. We are in Frege territory, with the notion shorn of all exclusive implications: denoting is just the property of being a symbol for. Isn’t a symbol precisely something that is a symbol for something—that’s what it is to be a symbol. Calling a symbol denotative doesn’t go beyond calling it a symbol. Clearly Frege and other philosophers intend something stronger than that—but what? That is all that the dictionary tells us and it does nothing to vindicate the philosopher’s use of the term. The notion of reference so understood is trivial not substantive: no concept of a special relation possessed only by special words can be gleaned from the dictionary definition.

At this point the philosopher is apt to fall back on something called “the name-bearer relation”: surely this is a case of genuine substantive reference. Here is the name and there is its bearer, clearly distinct things, and the former designates the latter—denotes it, refers to it. But what does this alleged relation come to?  Isn’t it just that the former is a name of the latter? It is the word (symbol) that we use for a certain individual. But nothing in that locution implies some special relation of reference that holds uniquely between name and object. Notice that the alleged relation is not perceptible and has no causal powers: it is not like a bridge or a light ray or a line drawn in space. It is curiously impalpable, strangely elusive. We might imagine it as analogous to a kind of pointing: pointing fingers are perceptible things, and some reference is indeed aided by pointing. So we imagine the name as pointing to its bearer (or the speaker doing so, physically or mentally): that is what reference looks like! But this is so much mythology: there is no such pointing going on. The idea is just an imaginative crutch designed to confer substance on the supposed special relation of reference. Someone might protest: “Surely sentences containing names are about something—isn’t that what reference is?” I would make two points. First, sentences are also “about” other things such as colors and conjunction—this is part of what they mean—but the idea is supposed to be that these expressions don’t refer. Second, the relevant notion of “about” is connected to truth, as follows: a name contributes to truth conditions in virtue of the fact that a certain object is relevant to the truth of sentences containing it and not other objects. The name “Aristotle” is such that sentences containing it are true just if Aristotle is a certain way. The name “Aristotle” is a name for Aristotle in this sense: but on the face of it this truism says nothing about any supposed relation of reference.[2] The mistake is to think that the very thin notion of reference that is captured by the truism is actually a thick substantive relation that it makes sense to investigate like other relations in nature. It is not, for instance, like the relation of perception (with which reference is often compared). It is quite true that we mention things and allude to them and even make reference to them, but these commonsense notions don’t add up to the philosopher’s special notion of reference—that privileged restrictive notion that invites such philosophical perplexity. We have taken an ordinary notion, fine in its place, and run away with it, postulating metaphysical facts that exceed anything common sense can ground. We have, in a word, reified the notion. This is why it sounds so strange to say in this sense that a predicate denotes a property or that “and” denotes conjunction or that sentences denote truth-values. We have stretched and mangled the notion beyond recognition. We have conjured a semantic reality from thin air and based our whole account of language around it. We are guilty of philosophical hyperbole.[3]

Notice that abandoning the notion of reference, as a heavy-duty theoretical tool, does not mean we have to abandon key insights. We can still distinguish names and demonstratives from descriptions, predicates, and quantifiers; and we can still say (if we like) that the meaning of a name is its bearer. What we must not do is invoke a special relation of reference that supposedly links names to their bearers, or demonstratives to objects in the environment, or “I” to a certain self. We might wish to speak this way for convenience, expressing only the thin deflationary notion I specified; but we can’t suppose we are speaking of a primitive relation, analogous to pointing, that links words with things. Seeking a causal theory of reference, say, or a functional theory, is folly, since reference is not the kind thing that could consist in such facts; it’s just a device we use for talking about words, sentences, and truth. Or again, it is a potentially misleading way to express the idea that symbols symbolize—that symbols are symbols for something. When someone learns that “and” is the word for conjunction in English we can express that fact by saying “He now knows that ‘and’ refers to conjunction in English”, but nothing can be read into that way of talking of the kind assumed in standard philosophical parlance. There is no such thing as reference in that sense—it isn’t real. The right thing to say is that in the ordinary dictionary sense every meaningful word refers (shades of Frege, suitably deflated) but in the philosopher’s sense no word refers, not even names. And that seems like a welcome result given that a mixed position is theoretically infelicitous: better to have a nice uniform account of meaning. Wouldn’t it be odd if one part of language—the part occupied by names and demonstratives—had the remarkable property of reference while the rest of language hobbled along without it? Imagine lines drawn vertically below every allegedly referential word and a blank below every non-referential word: that would suggest totally different functioning as between the two kinds of word. Yet syntactically and pragmatically (as well as sense-wise) all expressions are alike–why the anomaly?

How did the notion of reference find its way into the philosophy of language? We all remember Frege’s argument for why we need sense in addition to reference, but the same argument can be construed as a defense of reference (and may subliminally operate in this way). Suppose we start by assuming that “Hesperus” and “Phosphorus” have different senses: we then observe that they nevertheless have something in common, i.e. the same planet is relevant to the truth of sentences containing either name. Let’s introduce a name for this common feature: we say that both names refer to the same planet. Then we announce that two names can have the same reference—an interesting discovery about language. We invert Frege’s argument so as to move from diversity of sense to unity of reference, in the process inventing a new concept, viz. reference. And now we are off and running, applying this notion all over the place. But the data that led to this conceptual profusion merely concerned the fact that the same object figures in the truth conditions of sentences with different senses—it has no more content than that. Then too, we are influenced by the pointing gesture and by the relational nature of perception. We thus end up with a thick (yet elusive) concept of reference whose correct application proceeds to perplex us. We need to develop a healthy skepticism regarding the notion of reference that has grown from these dubious roots. There is nothing wrong with saying that names name people and things, but we shouldn’t convert this into the fanciful idea that there is some peculiar relation that links names to their bearers—an invisible string, a mental pointing, a magical arrow.[4] Alternatively, we could view talk of reference and denotation purely mathematically, as in mathematical model theory, and give up the idea of real empirical concrete relations between words and things. There should be no metaphysics (or science) of reference.

The usual way this subject has been pursued is to assume that there are certain paradigm cases of reference (the name-bearer relation, demonstrative reference to sense-data) and then ask whether other cases approximate to the paradigm. Then we debate how far the concept of reference can be extended from the paradigm cases. The more radical position is that there are no paradigm cases, because nothing refers. We can speak of meaning and sense as real properties of language, but it is pointless to debate how extensively the notion of reference applies. It has no serious application, not as a matter of linguistic theory (though a doctor can still refer a patient to a specialist). Articles with titles like “On Denoting” or “On Referring” are about nothing (this one might be called “Against Denoting”). Questions like whether the reference relation is rigid or non-rigid with respect to a particular class of expressions are also misguided, since there is no reference relation to be one or the other (though we can reconstruct the underlying issues without assuming realism about the reference relation). We must either purge the philosophy of language of the myth of reference or show that it can get by without going beyond the minimalist view of “refers”. The standard division between the referential and the non-referential is an untenable dualism.


[1] If the sense of names involves a mode of presentation of a reference, what does the sense of other expressions involve if there is no reference to be presented? The mixed position gives us a radically divided theory of what senses are, thus undermining the idea of a general notion of sense.

[2] It is commonly stated that predicates denote their extensions, thus conjuring up a special semantic relation between predicates and sets; but this is no more than a fancy way to talk about the things a predicate is true of. There is no more content to the idea of extension-denotation than the idea of what a predicate is true of. The same should be said of names: talk of their denotation is just a fancy way to talk about their contribution to truth.

[3] I am conscious of echoes from Wittgenstein in these remarks.

[4] Notice that other non-theoretical uses of “refer” and cognates do not suggest any such “queer” facts (Wittgenstein’s term): letters of reference, references as citations, works of reference. Nor does the comical use of “I don’t know to what you are referring” by the socially aspiring lady in the train station tearoom in the film Brief Encounter (she actually knows quite well to what the station master is referring).


The Meta-Linguistic Turn



The Meta-Linguistic Turn



Suppose you are interested in the nature of numbers or causation or necessity or the mind or values. Your interests are traditionally metaphysical or ontological. You propose to think about these things and try to come up with answers. But someone tells you that you are going about it the wrong way: you should really focus on sentences or statements about these things—you should think about linguistic entities instead. You should analyze these sentences, give their logical form, and specify how they are used, engaging in syntax, semantics, and pragmatics. Then you will shed light on numbers, causation, necessity, mind, and value. You should take the “linguistic turn”, which will provide a methodology and a concrete focus: statements not facts, words not objects. You will be studying the world as it is represented in language.

But how should you study language in order to pursue your metaphysical interests? Not as a linguist might or a psychologist or a neurologist: you are not interested in these kinds of empirical questions qua philosopher. You are interested in the meaning of the sentences in question, the better to understand what they refer to (numbers, causal relations, etc.). So your methods are not those of the linguist or psychologist or neurologist (you don’t, for example, examine the brain of the user of these sentences). No, what you investigate is the way we ordinarily understand the sentences in question: you analyze their meaning, dissect their structure, and describe their standard use. We might say that you engage in conceptual analysis, understood broadly. You do this by talking about sentences, i.e. you employ a meta-language. This will involve you in using semantic concepts (as well as syntactic and pragmatic concepts): you will use the concepts of meaning, reference, entailment, truth, assertion, criterion of application, etc. In the case of statements of number, for example, you will note that numerals look like singular terms with reference, that some numerical expressions operate as quantifiers with domains of quantification, that “2+2=4” is true if and only if 2+2=4, and that arithmetical statements are grammatical and meaningful. This will bring you to consider the significance of such meta-linguistic sentences: what does it mean to say that a sentence is meaningful, or that it has this or that specific meaning, or that a word refers to an object, or that a sentence is true? You will unavoidably be drawn to consider meta-linguistic language, specifically the analysis of semantic expressions. You will be considering the theory of meaning and reference. For example, you will wonder how the causal theory of reference relates to statements of number, or whether it is possible to refer to a private sensation, or whether names are descriptions, or whether meaning consists in truth conditions. In other words, you can’t do philosophy of language without examining meta-linguistic sentences: you must not only use them in talking about words and sentences, you must understand them—they must become a direct focus of interest (mentioned not used). In understanding language you must understand language about language—and understand it philosophically. Most centrally, you will be concerned to understand sentences of the form, “s means that p”. Put simply, you will want to analyze the word “means”. And, of course, much philosophy of language is precisely concerned with such locutions and sentences—it is interested in meta-linguistic sentences. This is because it is interested in semantic concepts. So the philosophy of language is concerned with sentences and with sentences about sentences: with both the object language and the meta-language. How could it not be concerned with the meta-language?

Thus the linguistic turn involves a meta-linguistic turn: when we turn from the world to language we also turn from language to language about language. And this second-order language is part of our normal conceptual repertoire: everyone who speaks a human language also speaks a meta-language about that language. We all have the concept of meaning and associated concepts; these are not novel theoretical terms introduced by a groundbreaking scientist, like “quark” or “gene”. We recognize the central terms of art of the philosopher of language as our own; indeed, these terms are as natural and innate as the object language itself. Not only do we acquire a language (an object language) with great ease and speed; we also acquire a meta-language with comparable ease and speed. Language about language is part of our natural endowment as human beings. This means that our competence in such language can be used to develop theories of language: our intuitions about the concept of meaning, say, can be used to test theories of meaning. So the meta-linguistic turn is entirely natural to us: it embodies knowledge that we have as a matter of course. We understand our first-order sentences but we also understand our second-order sentences; it is not that we start struggling when language turns back on itself–as if the word “means” throws us into disarray, as “quark” might. So the meta-linguistic turn is methodologically sound so far as our linguistic knowledge is concerned, and it seems essential to the linguistic turn as a whole. The linguistic turn necessarily brings with it the meta-linguistic turn. Accordingly, when we turn from causation itself, say, and look at causal statements, we inevitably look at statements about causal statements, especially statements invoking the concept of meaning. The general theory of meaning thus becomes relevant to the theory of causation once the linguistic turn has been made. You can’t understand what a cause is without knowing what meaning is! Whether this counts as a welcome result or a reductio of the entire approach is another question; the point is that the linguistic turn leads us in that direction. Metaphysics and semantics become inextricably intermingled: facts can only be understood if meaning can be, i.e. meta-linguistic statements concerning meaning. This makes metaphysics quite different from physics, say, because no one thinks that we can’t do physics without a theory of meaning to guide us.

The case is similar with the epistemic turn: if we shift our focus from facts to knowledge of facts in the hope of getting a better handle on facts, we will inevitably be led to consider our knowledge of knowledge. We won’t bring in what a psychologist has to say about knowledge (as in the psychology of education) but will concern ourselves with our ordinary grasp of knowledge—for example, that knowledge involves truth and justification and that it is vulnerable to skepticism. This is meta-knowledge—knowledge about knowledge. We use this meta-knowledge in order to understand knowledge itself (first-order knowledge), but we will find ourselves examining meta-knowledge in its own right—we will take it as an object of investigation. Are we right to suppose knowledge involves truth and justification and that it is vulnerable to skepticism? Perhaps we don’t know enough about knowledge to be confident about such things.[1] Are we epistemically perfect with respect to our epistemic capacities? What is the nature of knowledge of knowledge? Is it caused by knowledge? Is it perceptual? How fallible is it? Thus we engage in an examination of meta-knowledge as part of the epistemic turn; we can’t just stop at first-order knowledge itself but must proceed up a step to consider our knowledge of knowledge. That is what philosophers do—investigate the questions that are raised by their methods. So the epistemic turn will involve us in questions about meta-knowledge: we can’t know what causation is, say, without understanding our knowledge of causal knowledge! We can examine how we come to know causal facts, but that will bring us to wonder about what we know of our knowledge of causal facts. Thus the epistemic turn involves a meta-epistemic turn. Again, whether this is a welcome result or a reductio is another question; the point is that the epistemic turn turns into a meta-epistemic turn. Combining the epistemic turn with the linguistic turn, we are not just interested in sentences like “A knows that x caused y” but also sentences like “A knows that B knows that x caused y”. Our knowledge of our causal knowledge might or might not be adequate, but we must assume that it has some veracity in order to invoke our understanding of such knowledge in the effort to understand causation. We end up scrutinizing our higher-order knowledge in the project of getting clearer about causation. We have some understanding of what knowledge is just by having the concept, which is why people can recognize the correctness of elementary truths about knowledge; but that is not to say that the theory of knowledge is all plain sailing. So the epistemic turn in metaphysics will find itself caught up in questions about our knowledge of knowledge. Whether that counts as philosophical progress is an interesting question; certainly it complicates the claim of the epistemic turn to constitute a royal road to metaphysical understanding. It isn’t as if everything becomes clear once we turn our attention to causal knowledge and away from causation as such (and similarly for the other topics listed).

The same is true for the conceptual turn: here too the shift from things to concepts of things will involve us in meta-questions about concepts, and it is not as if that subject is all goodness and light. Concepts about concepts, thoughts about thoughts, also require scrutiny. The conceptual turn becomes a meta-conceptual turn with all the questions that entails. A person might be forgiven for supposing that we had it easier when we were trying to focus on facts directly. Is it really easier to understand our concept of necessity (say) than necessity itself? Why turn from the frying pan to the fire when the fire is even hotter? In any case, meta-questions about concepts will accompany the conceptual turn. When we turn from things to concepts in order to understand the things better we will inevitably face the question of how to understand concepts: and the latter may be no more pellucid than the former.

Original proponents of the linguistic turn probably thought that language was a relatively problem-free zone compared to traditional metaphysics, so it could be reliably turned to in times of need. Aren’t statements about numbers more straightforward than numbers themselves? We can at least see and hear them, and write them down. Thus Frege’s advice to begin with the former and work back to the latter looks sensible enough. But one look at the theory of language espoused by Frege deflates that expectation, what with its apparatus of sense and reference, concept and object, function and argument, truth-values as objects, thoughts as objective platonic entities, saturated and unsaturated entities, infinite hierarchies of indirect sense, etc. Indeed, the ontology of the theory invites questions comparable to those raised by traditional metaphysics concerning numbers. The meta-language propositions of Frege’s theory of language are as puzzling as any in traditional metaphysics. So the meta-linguistic turn reveals the linguistic turn to be not on such solid ground as it might have hoped. In philosophy as in life, when you turn from one thing to another, be fully cognizant of what you are turning to.


[1] It may be said that second-order knowledge is more reliable than first-order knowledge and less vulnerable to skepticism: we can be more certain that first-order knowledge is vulnerable to skepticism than that we are not brains in a vat. But the determined skeptic might insist that this purported knowledge of the truth of skepticism is not as solid as we suppose: we might have made a slip in our reasoning or the concept of knowledge might not be as demanding as the skeptic supposes. At any rate, such questions would have to be posed once we examine second-order knowledge.


Women, Physics, and Disgust

I don’t usually like to comment on what is going on in the “the profession”, but the recent skirmish between Robin Dembroff and Alex Byrne over the definition of “woman” reminds me of two incidents in which I was involved: two young female professors wrote highly personal critical reviews of my books Basic Structures of Reality and The Meaning of Disgust which I thought overconfident and callow (I’m trying to be nice). I hope this is not a pattern. It is not good scholarship, to put it mildly.


Causation: Invisible Cement



Causation: Invisible Cement



I wish to make some inconclusive but perhaps suggestive remarks about a subject that has troubled me for 50 years. Hume’s imperishable contribution is well known: whatever the objective nature of causation may be, we can say with confidence that our knowledge of it is neither a priori nor observational. We don’t know how objects will causally influence each other by pure reason (causal statements are not analytic) and we don’t perceive causal relations (particularly causal necessity). When we ascribe a causal power to a thing we don’t do so because of a rational insight (as in mathematics) and we don’t do so because we can literally see the power (as we can see color and shape). So how do we know the causal powers of a thing—how, say, do we know the effects of a moving billiard ball on a collision course with another billiard ball? According to Hume, we know it by induction: we have seen other objects behave in certain ways and we can see that this object is similar to those. Of course such reasoning is inconclusive and fallible, as well as being annoyingly indirect, but it is the only method we have. I can know empirically that a billiard ball is red by looking at it and I can know a priori that it is extended because I know it is a physical thing, but I can’t know what it has the power to do by either method; I must resort to my recollection of how similar balls have behaved in the past and trust that it will follow suit, knowing that this reasoning is far from satisfactory. The causal power is curiously elusive, frustratingly hidden from view: it exists right there in the object just waiting to unleash itself, and yet I cannot get my mind around it. I can stare at the object all day from every angle and in every light, but the power will not reveal itself to me. If causation is the cement of the universe, it is invisible cement—I can get no impression of it. The primary and secondary qualities of the object display themselves without this coyness, but causal powers lurk unseen, refusing to show their face. Thus the epistemology of causation is subject to skepticism, to the point that causation itself comes into question.

The question I want to ask is why it is thus invisible: what is it about causation as it exists in objects that makes it closed to perception? What is it about the nature of causation that explains its imperceptibility? It can hardly be that it is too small or has the privacy of the mind or is a type of value or is abstract like numbers—none of these possible explanations applies in this case. It must somehow be intrinsically hidden from the senses, and thus only known (if at all) by the indirect method of induction, but what is its nature such that this is so? I suggest it is because causal powers relate essentially to the future; they determine how an object will behave. An object with a certain power is such that in certain conditions it will bring something about at a later date: the speeding billiard ball will have certain effects in the future as it collides with other balls. The power is future-oriented, future-involving. Yet it is possessed presently: it is now true that this object will do that. But I can only observe what is present—my senses don’t take in the future (as they don’t take in the past). It is as if a certain kind of “externalism” applies to the power: it is what it is now in virtue of what it will be later, i.e. something external to the present time. It has a trans-temporal quality, a foot in the present and a foot (many feet) in the future. It is temporally spread out. But then it will not be perceptible by the senses, since they take in only what is present (perceptions are themselves caused by present facts). If we possessed precognition as a type of sense, then we would be able to see the causal powers of objects and would not need to rely on extraneous (Hume’s word) induction; but as it is we are blind to causal powers, being forced to look outside the particular causal nexus before us and go by past similarities. If this is right, it is at least clear why causation is invisible: it is invisible because we can’t see the future. We can know the future by means of truths of reason—we know that in the future bachelors will always be unmarried—and we can know it by induction (putting aside skepticism)—we know that the sun will rise tomorrow based on its past performance—but we don’t know it by perceiving it directly. So causal powers elude our powers of perception, since they intrinsically include future happenings.

Powers are a bit like biological functions. The heart has the function of circulating the blood, but this involves future exterior events, so we can’t see its function by looking at it, as we can see its gross anatomy. We can’t look at it at a given time and simply observe that it has that function. It is possible for a heart to have this function and not have the power to carry it out (it is a defective heart) and vice versa, so these are not the same property; but they share the characteristic that they incorporate “external” events. We can see the heart but not see its function, as we can see a billiard ball but not see its causal powers—because in both cases the property in question “takes in” remote occurrences. In the case of powers this remoteness is temporal: the power is defined precisely as the power to change the future in a certain way. Powers are cross temporal not temporally confined—what it is to be a causal power is not constituted by what holds at a given instant (like color or shape).[1] Powers reach into the future, but perception is stuck in the present. That is why Hume is right bout the epistemology of causation. It is thus not mysterious that causal powers should be invisible; it stems from their very nature. It may be that this nature is itself mysterious, but given its reality imperceptibility can be predicted. As remarked, pre-cognition (pre-perception) would reveal it to the mind, but we don’t possess that—and it may be conceptually impossible. In any case our senses do not put us into epistemic contact with the future—only inductive reasoning can do that (or knowledge of analytic truths). Causal powers are something additional to the usual primary and secondary qualities, and different in nature: for they are not temporally confined. This implies that you can’t see causal powers, because you can’t see what an object will do.

It is arguable that all causes and effects involve motion so that causal powers are always powers to move things (strictly, accelerate things). If so, the future changes that are implicit in present powers are changes of motion, which means that what we can’t perceive are future motions when we fail to perceive powers. We have no perception of future motions; we can only perceive present motions. That is precluded by the fact that there is no backwards causation, so no present perception can be caused by a future event. Future motions can’t cause present perceptions of them because perceptions must be caused by what they are of and there is no backwards causation. Hence we cannot perceive powers to produce future motions. The temporal asymmetry of causation lies behind our inability to perceive causal powers. If backwards causation were possible, then we would be able to perceive future motions, and hence perceive powers; but we can’t do the latter because the former is not possible (not in the actual world anyway). So the unobservable character of causation has its roots in the (actual) impossibility of backwards causation. Hume’s critique springs from that fundamental fact. It is not that there is some (wholly) present fact about the cause that for some reason we can’t perceive; it is the involvement of future facts that stands in the way of such perception. In vain do we look for future motions in the perceived world. Therefore what we observe of causes does not include their being the causes they are. To put it another way, the universe unfurls towards the future as causal powers manifest themselves, but we have no perception of this future, so the powers remain invisible to us. Metaphorically speaking, the universe knows the future better than we do. If we lacked the ability to make inductions based on similarities, we would have no causal knowledge at all; we would not even have a conception of causation. For us history would be just one damn thing after another. It would not be apparent to us that causal powers exist, even though they do, since nothing in what we perceive can give us the idea of them. Perception and causation are just not cut out for each other, given what each of them intrinsically is. They relate quite differently to time. Perception is directed exclusively to the present while causation is inherently future-directed.[2]


[1] Thus causal powers cannot be reduced to their “categorical basis”, which is temporally confined and may be perceptible, even though such a basis may be necessary for the power to exist.

[2] It should be noted that this applies to mental causation as well as physical causation: we don’t introspect mental causal powers, as we don’t perceive physical causal powers. Hume, of course, was well aware that his critique applies as much to mental causation as physical causation. I don’t believe he ever asked why causation is not revealed to the senses, being content to note that it isn’t.


The Humanistic Turn



The Humanistic Turn



A popular narrative has it that analytical philosophy replaced an emphasis on knowledge with an emphasis on language. It is said that Descartes brought epistemology to the center of philosophy and that Frege shifted the center to the philosophy of language (assisted by Wittgenstein, Austin, et al). Thus we have the “linguistic turn”. Some add to this the idea that the linguistic turn gave way to the conceptual turn, in which the study of thoughtbecame the central preoccupation of philosophy. From knowledge to meaning to thought, there has been a succession of intellectual revolutions during the last three centuries. Continental philosophers may wish to add the phenomenological turn, initiated by Husserl and taken up by Heidegger and Sartre, in which philosophy takes as its central concern the study of lived experience. Yet others may prefer the idea of the logical turn, in which formal logic became foundational. And there may be various sub-turns: the linguistic turn itself exhibited a number of smaller turns, from propositions to speech acts to language games to theoretical linguistics. In any case, philosophy has undergone various shifts of emphasis, beginning with Descartes: moves towards something and away from something else. There are then debates about whether these moves were good or bad, and how exactly they should be characterized.

But what was Descartes turning from? I think there are two answers that are not unconnected: religion and traditional metaphysics. He was turning away from reliance on Scripture and church authority, which lie outside of the human subject; and he was also turning away from the legacy of Platonic metaphysics and Aristotelian scholasticism, which emphasize impersonal ontology. Philosophy must focus not on a transcendent God and not on external Nature, both construed as non-human realities, but on the human capacity to know—on human reason, human experience. How do we know and what do we know—with the accent on “we” (or “I”). We must study the human ability to know things, not an extra-human God or Nature. Human knowledge (its scope and limits) is something we can get our teeth into, since it is an aspect of us, part of our nature. Philosophy must turn from the non-human to the human—to human nature, in a word. Descartes accordingly meditates on himself, alone and unaided, as a natural human creature, particularly as a knowing creature. The Cogito is one of his first discoveries: he finds that he is a thinking thing that indubitably exists. The question then is whether this thinking thing can know the things it thinks it knows, and more besides; thus the stage is set for the analysis of knowledge and the attempt to rebut skepticism—central concerns of post-Cartesian philosophy, up to and including the analytical kind.[1] The important point is that this epistemic turn was a humanistic turn: a turn towards the human and away from the non-human. The focus on knowledge was the form this humanistic turn took for Descartes and subsequent thinkers.

The successive turns away from knowledge–which came to include perception and reasoning, common sense and science–were also humanistic turns: language is a human attribute too, an aspect of human nature; and the same is true of concepts and thoughts. We humans have these attributes as aspects of our given nature, and philosophy is now construed as an investigation of human nature under this dispensation. Philosophy becomes the philosophy of the human, specifically our powers of linguistic and conceptual representation—a kind of rarified psychology. Thus we find Strawson’s project of “descriptive metaphysics”: the attempt to map our ordinary natural conception of reality—our “conceptual scheme”. This is to be contrasted with theology and with science, which both attempt to describe the non-human world. It also stands opposed to the Platonic style of metaphysics in which we attempt to describe reality as it exists independently of humans—the world outside the cave. The world of Forms has nothing intrinsically to do with human beings, but exists independently of us, and will go on existing whether we do or not. And before Strawson’s humanistic metaphysics we had such works as Locke’s Essay Concerning Human Understanding and Hume’s Treatise of Human Nature—note the occurrence of the word “human” in both titles. These are avowedly books about the human animal, if I may put it so—about a certain type of creature with a specific nature. We know, we perceive, we think, we speak: that is what philosophy should concern itself with. If it is not about such things, we should turn away from it and towards its proper object. Thus, seen from a broader perspective, the epistemic, linguistic, and conceptual turns are all turns within a larger humanistic turn. The debate between empiricism and rationalism, for example, is a debate within the humanistic conception of philosophy, namely how human knowledge is arrived at and what its nature is. We today have become very accustomed to this turn and are scarcely aware of it, perhaps not even seeing anything turn-like about it. This is because it is a shift in intellectual history that goes back centuries and whose antecedents are barely comprehensible to us moderns—the time of religious domination and pre-modern metaphysics. At one time it was startling to maintain that philosophy should concern itself with the human being, whether as knower, thinker, or speaker. Aren’t we just too small and insignificant compared to God and the Platonic Forms to be awarded such a prominent place? What does the universe care about our ability to know, think, and talk (especially the last)? Isn’t the Form of the Good so much grander than our feeble human meanderings? Isn’t there something impious and vain about focusing philosophy on ourselves? Isn’t humanistic philosophy anthropocentric philosophy? And isn’t that culpably self-centered?

Not all modern philosophers have adopted the humanistic approach. Two giants stand out: Spinoza and Leibniz (we might add Schopenhauer and even Kant in some respects). Both developed traditional-sounding metaphysical systems without regard for human perspectives, and both are alien to modern sensibilities. Spinoza and Leibniz have struggled for curricular recognition and are often regarded as eccentric at best. That is not a false impression given that they don’t share the humanistic turn (proudly in the case of Spinoza, in view of his naturalism about the human creature). Both belong squarely in the tradition beginning with Plato and Aristotle (including the pre-Socratics) and pre-dating the Christianized doctrines of the middle ages; they are concerned to provide an intelligible general ontology without reliance on Scripture or the human viewpoint. They are certainly not interested in describing what the ordinary human animal thinks, or how he or she thinks it. We might call these non-humanistic philosophers: they resist the humanistic turn in whatever variety it presents itself. Kant is the odd case because of his duality of the phenomenal and the noumenal—the former decidedly human, the latter not human at all. Kant took a kind of half-turn, though historically he triggered a yet sharper turn towards the human: his view is that we can’t know anything about the non-human world, though there is such a thing, while the human world is open to our understanding. Berkeley is an interesting case: at first sight his idealism would appear to put him firmly in the humanist camp, but then we remember his placing of God at the center of all things and the non-human asserts itself. The infinite spirit is not a human spirit, and it is the foundation of the entire universe. Berkeley is a non-humanist wolf in humanist sheep’s clothing. Again, he is someone we moderns find it difficult to digest. We are more comfortable delving into our own nature accompanied by Locke and Hume. We like the humanistic turn, self-centered as we are. We prefer to think of ourselves as the measure of all things—whether by our knowledge or our meanings or our concepts. We think human nature is fantastic.

Where do more recent philosophers fall? Almost all twentieth century philosophers are humanists: Russell, Wittgenstein, the positivists, Husserl, Heidegger, Sartre, Davidson, Quine, Strawson, Dummett, and many others.[2]The anti-humanists are harder to spot, not surprisingly given the humanistic turn, but a few exhibit anti-humanist leanings—I might mention Kripke and Nagel. Kripke opposes the humanistic interpretation of modality, stressing the existence of metaphysical necessity and distinguishing it from the epistemic kind—it stems from objective reality not from our own minds (as with analyticity). Nagel qualifies because of his realism and emphasis on human limitation—in no way is the bat’s mind a version of the human mind, and our concepts are not guaranteed to catch hold of everything real (the View From Nowhere is not a human view). Mysterians (such as Chomsky and myself) are sharply anti-humanist because we reject the idea that reality is designed so as to conform to human modes of thinking. We humans are just tiny specks in reality, not the measure of reality; neither human knowledge nor human language nor human thought are constitutive of the real—the objective world is. These human attributes are merely part of the world. Construed as a general meta-philosophy, the humanistic turn was a mistake, according to mysterians–though there is nothing untoward about studying human nature as such. We do better to revert to Plato, Aristotle, Spinoza, Leibniz, and whoever else shares their anti-humanist predilections. In any case the issue has been joined: to be a humanist or not to be a humanist.[3]


Colin McGinn


[1] A problem with the epistemic turn, as exemplified in Descartes, not mentioned in the standard narrative, is that knowledge turned out to be not as unproblematic as we might have hoped. Descartes recognized the problem of skepticism from the start and tried unsuccessfully to solve it, so knowledge can hardly be a solid foundation on which to erect a humanistic philosophy alternative to what had gone before; but further, it became clear that even defining knowledge is difficult, so we don’t even know what knowledge is. If our aim is to find a sound starting point in human nature, knowledge seems like the wrong concept to invoke. In this light we can see why a switch to language might seem appealing; however, the concept of meaning soon revealed itself to be anything but pellucid, so turning to it hardly leads to the Promised Land. And the same can be said of concepts and thoughts. Human nature turns out not to be as transparent as we might have hoped. If obscurity is a count against non-humanistic philosophy, then it applies also to humanistic philosophy.

[2] There is also the humanistic turn with respect to ethics (and politics and aesthetics): instead of finding moral value in the supernatural realm or in the natural order we find it in the human individual; it is imminent not transcendent, an aspect of human nature. Moral humanists would include Nietzsche and Hume as well as many twentieth century ethicists (e.g. Bernard Williams); moral anti-humanists would include Platonists, Kant, and early Moore. In the case of ethics the subject matter at least has a clear connection to human life, so the humanistic turn is more intuitive; but it is far from self-evident that values themselves are part of human nature. In any case, we should include ethics under the grand opposition I am describing.

[3] The same kind of dilemma can also confront the sciences, physics in particular. With positivism physics took a humanistic turn, given that verification is a human attribute; and instrumentalism invites the question “Instrumental for whom?” Einstein’s relativity theory comes perilously close to building the human subject (the “observer”) into physics, and so does quantum theory on some interpretations; at least part of the appeal of these theories is surely their humanistic aura. Newtonian physics, by contrast, was resolutely non-humanist. And isn’t post-modernism really just the final expression of humanism? Even truth is a human construction: the world is nothing but the human world (culture, custom, power relations).