Moral Knowledge and Moral Language

Moral knowledge is one thing; moral language is quite another. The two topics need to be treated separately. I am mainly concerned here with the language question, but I will make some prefatory remarks about moral knowledge. Moral knowledge comprises knowledge-that, knowledge-how, and knowledge-what: we know (say) that genocide is wrong, we know how to reason about moral questions, and we know what various moral categories are (the right, the good, the virtuous). This knowledge is a priori, largely innate, and arises spontaneously in the human mind (but not in the animal mind, except rudimentarily, if at all). We have a good grasp of various moral natural kinds—lying, stealing, promise-keeping, beneficence, etc.—as well as very general moral kinds such as good and bad acts. These are not like natural kinds whose real essence may be unknown to us. Importantly, moral knowledge is not subject to traditional skepticism: there are no brain-in-a-vat scenarios to contend with (the inhabitants of the Matrix are not morally deluded). Thus, we are justifiably certain of what we claim to know. Also, we are not tormented by the limits of our moral knowledge—it has comprehensive scope. We are not cognitively closed to certain identifiable moral questions as a matter of principle. All in all, our moral knowledge is in good shape; arguably, better than our knowledge of the external world (let alone the past and future). We get an A for moral knowledge—and we didn’t have to study for the test.[1] We might suppose, given our epistemic credentials, that our moral language would be similarly distinguished—that our linguistic moral competence would reflect our epistemic moral competence. We should be expert moral speakers.

But, alas, it isn’t true; indeed, our moral language has struggled to express our moral knowledge. We have trouble verbally articulating our moral understanding. Not to point too fine a point on it, our moral language is inadequate, dubious, slipshod, and metaphorical; it is easy to poke holes in it and lament its imperfections. Logically perfect it is not. The linguistic moral turn is a turn towards the dark, clumsy, and impoverished. First, the moral word “good” has to share its form with the non-moral “good” (and similarly for “right” and “ought”). Second, it is hard to find true moral generalizations: we can’t say lying is always wrong, or promise-breaking, or stealing, or killing. We can’t even say that maximizing happiness is always right (it can conflict with justice). Thus, we are treated to tortuous formulations like “Lying is prima facie wrong”, which even its author (W.D. Ross) feels the need to apologize for. Or we are urged to adopt the imperative mood to express our moral knowledge: “Do this, don’t do that!” But the imperative mood has nothing essentially to do with morality and is wheeled in in a vain attempt to capture the normative force of moral truth—truth, mark you, not obedience. Imperative sentences have nothing logically to do with moral principles or precepts. Moral truths need no commanding authority to give them force (even God). Moral knowledge is not commandment knowledge—that applies to the military not the moral. To act immorally is not to disobey orders (that is neither necessary not sufficient). Nor will categorical propositions do: “Generosity is good”, “Lying is bad” etc. For these are wide open to counterexample and much too strong; they breed skepticism about the force of morality. We all know that morality is not so simple, that such statements must be taken with a grain of salt; but how should they be amended so as to correspond with our moral thought? Should we insert a ceteris paribus qualifying clause? But then, we have to be able say when the generalization breaks down. We just don’t seem to be able verbally to hit the nail on the moral head, to find the moral mot just. The case is somewhat like aesthetics and laws of nature: here too the facts, and our knowledge of them, seem difficult to fit into linguistic form. Our aesthetic knowledge is extensive and sophisticated, but our aesthetic vocabulary is comparatively limited. The word “beauty” has to do a lot of heavy lifting. No doubt this is partly because aesthetics is an affair of the senses for the most part, and our sense experience is notoriously rich compared to our vocabulary. The beauty of a landscape is hard to put into words, as we all know. Likewise, laws of nature are real and known and yet their linguistic formulation is much disputed—how exactly should we frame them? Are they universal generalizations over particulars, or modalized versions of same, or statements about universals, or not statements at all but truth-less predictive devices, or summaries of observations? We find it difficult to put our knowledge of natural laws into words (some of it is knowledge-how). Morality is like this: our knowledge of morality pre-dates our language for it, and our language lags behind, or is inherently unsuitable for the task assigned to it. Morality verges on the ineffable. We can make individual judgments well enough, but we find it difficult to formulate general principles. We can’t even say what grammatical mood best expresses it. We are a bit like Tarzan in the jungle: “Jane, good, snake bad”.

I don’t see any cure for this linguistic lack, but I think I see an area of moral language that gets neglected, and which has some signal virtues. The approach is to privilege the comparative form, so I call it “comparativism”. Thus, “It is better to work than steal”, “It is better to tell the truth than lie”, “It is better to preserve life than destroy it”, etc. In this approach we mark a contrast; we situate the moral value we wish to praise and promote in the context of its opposite. It is better to do X than Y—that is the basic logical form. This doesn’t logically imply that X is good or the best; more modestly, it says that one thing is better than another. It doesn’t make an outright claim of moral goodness. We could paraphrase these formulas as “On the whole, it is better to do X than Y”, thus allowing for exceptions (stealing food in extremity, lying to save an innocent life). Such statements are calculated to ward off accusations of unreasonable absoluteness; they therefore accommodate the principled objector to more absolute formulations—they are easier to accept as true than their stronger counterparts. Note that our moral knowledge might be less qualified, because it can take in more complexity; the comparative theory applies only to moral language—to what we should say morally. In its most forthright form, the theory contends that the basic laws of morality have this comparative form: this is what we mean by saying something is good or bad; this the correct analysis of our moral talk (our thought might be different). It is a linguistic theory not an epistemological or ontological theory. Intuitively, we are saying that one thing is morally superior to another thing with which it might be in competition. It is better, for example, generally to tell the truth rather than to tell a lie. Why this should be is another question—maybe because it is more likely to generate long-term good consequences. It might also be taken to mean that truth-telling as an act of will is inherently more righteous than lying (Kant), though in certain circumstances this intrinsic superiority might be overridden (not Kant). The important point is that we use the comparative form in our verbal explications of the content of morality, i.e., our moral knowledge. They can, of course, be modified according to need and nuance, as in “It is far better to preserve life than to destroy it”. Instead of just saying, baldly, “It is (categorically) wrong to steal”, we can say “It is better on the whole to work for what you have than to steal it from someone else”. Then we allow ourselves wiggle room to append “though in certain circumstances what counts as theft legally need not be a decisive objection to stealing”. Notice there is nothing relativistic about this approach: in every society it is better to tell the truth than to lie, even if members of the society reject such a principle (say by being out-and-out ethical egoists). It is more the idea that moral concepts naturally come in pairs between which a comparison is being made. Russell once gave as an example of an a priori moral truth “Happiness is better than misery”: that seems both true and certain, and it is a comparative claim—not merely the claim that happiness is good, period. Compared to what, one wants to ask. It also allows us to say that happiness is not better than (warranted) grief on a particular occasion—it is just better on the whole. We know what is meant in the simple formulation, because our moral knowledge exceeds our linguistic resources, but the language is clumsy and potentially misleading. Our moral competence outstrips our linguistic competence, but the comparative speech act does better at conveying what morality is all about. It is better than imperatival prescriptivism or statements of prima facie duty or emotional ejaculation or universal quantification; it is truer to the heart of the matter. We might almost say it is less intellectually offensive than other formulations, especially to critically minded people looking for something they can in good conscience get behind. No one wants to put their life on the line for a false or clumsy verbal formula. Moral comparativism has the right combination of the confident and the qualified, the assertive and the modest. We don’t want to understate the strength and seriousness of morality or to overstate its strictness and universality. The language of morals has to steer a judicious line between these two extremes. Moral knowledge meanwhile can capably go its own way, resorting to verbal expression only when necessary.[2]

[1] I am quickly summarizing here, omitting many a caveat; I don’t attempt to defend these views, merely to set the stage for my discussion of moral language.

[2] Have you ever noticed how, when searching for words to express extreme moral outrage, the right moral language seems out of reach—nothing quite measures up? One resorts to animal metaphors or excremental ones. The language seems puny and inept. Our moral language is cognitively and emotionally under-powered. Attempts at general principles often fall flat. The language faculty is not properly hooked up to the moral faculty. Morality is not inherently linguistic (like laws of nature).

Art for Art’s Sake

I am fed up to the back teeth with this slogan. It may have once seemed brave and insightful, but now it reveals a complete failure of moral judgment (and intelligence). It is the opposite of the truth. First, let’s parse it, because it conceals a slyly tendentious ambiguity. It clearly means “The purpose of art is art” (art is autotelic, in the lingo). Equally clearly it is intended as a normative statement not a descriptive one: it doesn’t say that all art is actually done for itself, but rather that this is how art ought to be. There would be no point in saying it if artists already conformed to it. It is presupposing that many artists don’t proceed according to its prescription, but they should. And what is that they do so faultily? They treat art as a means to a moral end: they think the purpose of art is promoting the good—the moral good. So, the proponent of the autotelic view is saying that art is an end in itself not a means of making the world a better place. We can agree that the artist is doing art not morality—he is an artist, after all—but the slogan is urging artists to do only this: to have no moral intentions, to advance no moral cause. The artist is concerned solely and wholly with non-moral objectives—say, aesthetic experience or emotional response. Of course, the artist has artistic intentions, but we are being told that he should have no other intentions. So, the slogan becomes “The purpose of art is only art, not morality”. It is like saying “The purpose of art is only art, not brick laying”. Morality is extrinsic to the work of art. No novel, say, should concern itself with moral themes, with good and evil, on pain of not being art (or artistic). You disqualify yourself as an artist by building moral content into your work. Perhaps the reader can now see why I am fed up to the back teeth with this slogan; it is patent nonsense, pure rubbish. All art has, or should have, moral content, especially the written kind: it is concerned with good and evil, virtue and vice, right and wrong. It would have no human interest if it were not. Of course, it should not be inartistically concerned with such themes, but it is not a condition on being art that it not be concerned with them (pace Oscar Wilde and Vladimir Nabokov). Nor should it be concerned didactically to support received morality, conventional bourgeois morality; and in fact, art frequently makes a point of criticizing such morality. It would be absurd to announce that the purpose of art is only morality, but it clearly is a concern of the artist. Why would anyone want to deny this?

Compare science: “Science for science’s sake”. Suppose we paraphrase this as “The sole purpose of science is science—and certainly not the promotion of the good”. That is, we insist that science (the institution) is concerned only with scientific standards: truth, evidence, theoretical interest, and nothing else. This is obviously not so: we have pure science and applied science, and applied science is often concerned with human (and animal) welfare (medical science). Why would we want to preclude the scientist from having moral concerns, such as the desire to cure cancer—would this make him a bad scientist?  This is palpable rubbish, is it not? The scientist is quite capable of having several distinct aims at the same time: achieving scientific truth, improving the lot of mankind, earning a living, impressing his colleagues, etc. He presumably also believes that scientific truth is itself an intrinsic good, so to that extent is enmeshed in normative notions—as is also true of the value of artistic good (beauty is a good). The moral aims have to be the right ones, naturally, but assuming they are, there is no discredit to the scientist in pursuing them. He might well be criticized for not pursuing them. There is no incompatibility between science and morality, as there is no incompatibility between art and morality. Indeed, morally bad science should not be undertaken, as morally bad art should not be. We can make the same point with respect to many human activities: history, philosophy, psychology, economics, politics, mathematics, cookery, entertainment, education, etc. Try substituting these into the slogan and see if you like the result: for example, “Philosophy for philosophy’s sake, not for goodness’s sake”. Should philosophy never be concerned with morality—is ethics not part of its remit? It can be concerned with other things too, as art can also be. The right thing to say is that art is properly concerned with beauty, truth, and goodness; they are all part of its plural purpose. In this it is like many other human activities. It is quite wrong to maintain that art is only about aesthetic thrills (Nabokov) or pleasure (Wilde). A novel, say, is generally about all these things, with morality surely at the center; some novels are clearly intended to make a specific moral point, and are no less artistic for that (e.g., Black Beauty). Contrary to what its author insists, Lolita is in fact a deeply moral novel and gains much of its aesthetic impact from that fact. How could a novel not be occupied with morality, since it concerns human action, into which morality is woven. Drama is about good and evil, and drama is the point of a great deal of art.

It is hard not to see the autotelic view of art as reflecting a general cynicism about morality and wanting to protect art from the taint of the moral. If morality is subjective, relative, shapeless, and conventional, we won’t want it infecting the artistic work; but that sophomoric position is hardly compulsory (in fact, it is ridiculous). Sound morality elevates the novel (as in the work of Jane Austen, among many others). It is interesting that the form of our slogan is correct only for morality: “Morality for morality’s sake, not for money’s sake or power’s sake”. The only purpose of morality is to secure morally right conduct, however that may be defined—doesn’t that sound eminently reasonable? The novel should be proud to contain a large helping of morality; we expect nothing less. Morality isn’t the opposite of art but its backbone. It may well be true of music, painting, and dance that they have little or nothing to do with morality; but they are the exceptions that prove the rule, since they are not narrative forms. The ear and eye can be aesthetically stimulated by these art forms without any moral thoughts flowing through the mind of the observer, but the same is not true of narrative art—i.e., stories. If someone asserted “Stories for stories’ sake, only”, we would recoil in disbelief. The original slogan possesses what prima facie plausibility it does by suggesting non-narrative art, but narrative art is transparently morally imbued. And it is surely a wild exaggeration to try to assimilate narrative art to music, painting, and dance—as if we are responding to what our eyes present to us in reading a book. It is true that the written word can possess internal artistic features, but the novel is not about such features—and we surely don’t want to suggest that the novel is only artistic in virtue of such features. Sometimes the moral shape of a novel is the main part of its artistic appeal.[1]

[1] I rattle on about these themes in Ethics, Evil and Fiction and Shakespeare’s Philosophy. I know for a fact that I had covert moral intentions in writing my two novels, Bad Patches and The Space Trap, as well as narrowly artistic intentions. Wilde and Nabokov were stern moralists, if very funny and artistic ones.

Cognitive Closure Generalized

Transcendental naturalism (TN) is the idea that nature naturally transcends our ability naturally to know about it. Ignorance is real and natural—a biological fact.[1] Cognitive closure (CC) is the fact of being cut off from nature cognitively; not all of it, to be sure, but some of it. TN is a metaphysical doctrine; CC is an epistemological doctrine. Nature is transcendent to the degree to which we are cognitively closed to it. Generalizing, any cognitive organism (conscious, sentient) has a cognitive horizon beyond which it cannot reach—a subjective view it cannot transcend. There are limits to what can be cognized. Nature exists whether it can be cognized or not; reality is not beholden to cognition. Thus, we combine metaphysical realism with epistemological modesty: truth and knowledge do not necessarily coincide. Knowledge is part of nature not the whole of it, and this is a natural fact. In this paper I will expand the concept of closure beyond what we might think of as intellectual closure—what we cannot theoretically understand. It will turn out that cognitive closure is the rule not the exception; indeed, it is a law of nature. This will be a taxonomy of mental limitation, or natural transcendence, depending on which way you look at it.

First, I will give a list, eloquent in itself: perceptual closure, sensorimotor closure, attentional closure, memory closure, temporal closure, knowledge closure, linguistic closure, logical closure, mathematical closure, emotional closure. Most of these will be obvious, some perhaps less so. Perceptual closure is just sensory limitation—what we can see, hear, smell, etc. The visual field is limited, the eyes have limited acuity, and they are not in the back of the head. There is only so much of nature the eyes (any eyes) can take in—this is a complete truism. Sensorimotor closure relates to motor skills: we (and animals in general) don’t have every motor skill, as a matter of principle. Skills tend to be species-specific. Don’t jump off a building hoping to fly. Attentional closure is the familiar idea that we can’t attend to everything simultaneously; indeed, attention is sharply limited, severely bounded. This fact limits our cognitive powers considerably. Attention has limitation built into it. Memory is notoriously partial, unstable, and selective; we remember hardly anything of our past. Short term memory is extremely confined and perishable. By temporal closure I mean knowledge of the past and future: there is an awful lot about both that we don’t possess knowledge of. Nor can we ever know everything about past and future facts; here reality vastly exceeds the humanly knowable. Knowledge closure is what the skeptic fastens onto: we just don’t know much, including what we think we know. Skepticism is only too easy to fall into, because it exposes a weakness in the foundations. Linguistic closure concerns sentence comprehension: sentences easily get too long or involved for us to understand, what with iterations, embeddings, and relative clauses. Logical closure pertains to our ability to construct or follow a logical argument: this becomes a strain even when the number of premises and deductions is relatively small, but we quickly become logically incompetent once complexity mounts. Mathematical closure concerns mainly the infinite and unsurveyable: unsolved conjectures, endless decimals, etc. Emotional closure (which I add for completeness) is the fact that our emotional responses are not as elastic and generous as we might wish: often we just cannot summon compassion when we should, or suppress anger, or love our neighbor. We are emotionally circumscribed beings, like other animals. Then there are individual variations of CC on a vast scale: some people are just better at some things than others, as a result of genetics or upbringing—music, mathematics, writing, observing, science, art, etc. We are all educationally closed to some degree, but some of us are more closed than others (some even find philosophy difficult). The intellect is not an infinitely malleable substance or a universal machine—any more than the body is. We all have cognitive biases and no-go areas.

The upshot of all this is that closure is a fact of nature—a natural law, in biological fact. The law is this: every organism has cognitive strengths and weaknesses. Every organism is built not to know some things and to know others. Ignorance is natural and universal. Let’s call this “the law of biological closure”; it’s like the law of natural selection. Natural selection weeds out the good from the bad; the law of biological closure enables some cognition but precludes other cognition. It is epistemically selective. We can now connect this law with three general types of closure: descriptive, analytic, and explanatory—in ascending order of magnitude.[2] By descriptive closure I mean the inability to describe experience: I cannot describe in words what it is like to feel pain or see red or hear bells ring. I can’t convey these things to people who never experience them (and if they did, I wouldn’t need to describe them). I could convey this by making them experience the things in question, but not by uttering a bunch of words. Thus, people say “I have no words”, and they are right: language is limited as a method of communicating knowledge. Language is communicatively closed with respect to subjective experience. By analytic closure I mean the inability to analyze one’s own concepts: it isn’t easy to analyze concepts, and in some cases downright impossible; we just don’t have that much insight into our own conceptual scheme. The concept of knowledge is still refusing to submit to analysis! The concept of the good is remarkably recalcitrant. The concept of beauty eludes us analytically. If we are honest, we recognize that many of our concepts resist (complete] analysis. The concept of a concept itself is remarkably difficult to articulate. The whole business of conceptual analysis is deeply mysterious. Explanatory closure just means the difficulty of explaining things—consciousness, the origin of the universe, the nature of numbers, meaning, the a priori, etc. This is the field covered by the label “mysterianism” (not my coinage). It is one form of CC among many others (the CC family is extensive and various). Cognitive closure is everywhere, with us always, a fact of nature, a biological law.

But what explains it? Suggestions have been made, but they are not very impressive. Is it that everything good has a downside, like the giraffe’s neck? But if that is so, it isn’t obvious in the cases I have listed: it isn’t written into these cases that closure should hold. Knowledge doesn’t entail its own limitations, or else the concept of omniscience would be contradictory (we can’t disprove God’s existence this way). The limitations appear to be contingent. Is it that everything has a nature and so not another nature? Is it that cognition must be bad at some things because by nature it is good at other things? Is it that the eyes are bad at seeing behind you because they are excellent at seeing in front of you? But this lacks self-evidence, to put it mildly: why should being good at knowing some things make you bad at knowing other things? Some animals do have eyes in the back of their head, after all. Or is it that biological resources are scarce, so evolution doesn’t install cognitive capacities that are energy-costly? This strikes me as on the right lines: it isn’t that omniscience is logically impossible; it’s just economically impossible—just too expensive. Animals have the traits they need to survive, physical and mental, not traits that can do absolutely anything—including things that have zero payoff. There don’t seem to be any cases of closure about things it’s vital biologically to know. The closure is all about things it might be nice to know, but evolution doesn’t care about nice. It cares about necessary or needed. Thus, closure is a fact of nature not logic (or meaning or metaphysical necessity). We are cognitively closed de facto not de jure—like other animals.

Lastly, I want to talk about Russell’s so-called Principle of Acquaintance: “Every proposition we understand is composed wholly of constituents with which we are acquainted”.[3] This is clearly a closure principle: it says that any proposition not meeting the stated condition cannot be understood by us (but may be by others). And it is not difficult to fail to meet this condition: you won’t be able to understand a proposition about red unless you are acquainted with that color. So, blind people will be semantically closed with respect to such a proposition. Indeed, you won’t be able to understand a proposition about a particular person if that proposition contains a logically proper name of that person (you will need a definite description instead). This is actually a strong empiricist principle, limiting all graspable meaning to acquaintance-based knowledge, and threatening to make a great many propositions inaccessible to us. But even if we relax the principle a bit, as Russell does, we still get a rather restrictive result: all understanding requires some kind of direct acquaintance with respect to all constituents of the proposition. The fact that this principle is so restrictive doesn’t strike us as absurd; it merely reflects the truism of cognive closure. Yet we easily forget that we are imprisoned in our own little world—the world of our senses and inherited cognitive structure. The genes that construct our brains are not miracle workers; they are laborers in a stingy biological universe that won’t finance anything exceeding its limited budget. They are not going to buy a Rolls-Royce if a VW will do. We live in a need-to-know world.[4]

[1] See my Problems in Philosophy (1993).

[2] I discuss this further in “Description, Analysis, Explanation, and Philosophy”.

[3] See his The Problems of Philosophy (1912).

[4] Of course, qualifications need to be made to this blanket statement, which do nothing to blunt its force. Clearly, we know many things not necessary to our survival, just as we see and feel things not necessary to our survival; but these are inevitable by-products of faculties that do serve our survival (the same thing is true of other animals). However, there are other areas of potential knowledge that are closed off biologically because they are not by-products of useful faculties. These have to be considered on a case-by-case basis. The general point is that knowledge exists because it is useful, like other evolved traits; it doesn’t exist because it would be nice to have it. Solving philosophical and scientific problems is not part of the biological agenda.

Universal Prescriptivism

It used to be held that there are two types of speech act: descriptive and prescriptive. Descriptive speech acts state facts; prescriptive speech acts recommend acts. The fact-act dichotomy underlies the description-prescription dichotomy. I think this is completely wrong: there is no such thing as a descriptive speech act in the intended sense, and all speech acts are prescriptive. There are prescriptive facts but no facts prescribed (as opposed to acts). The reason there are no descriptive speech acts is that there are no purely fact-stating speech acts: no one merely states a fact, as it were into thin air; an assertion of fact is an act designed to secure a certain audience-directed result. Assertions are communicative not solitary. The aim of assertion is belief creation or knowledge transmission: one states a fact in order to secure that result. It is like a request to believe or a belief recommendation: “Please believe that p”.[1] We need not accept that belief-formation is an intentional act in order to accept this, but it is a kind of quasi-voluntary transition. The illocutionary force of assertion is captured by “Please believe me when I say that p”. We could also paraphrase it as “You should know that p”. You are not just giving voice to a fact but causing belief in that fact in your audience. That is the whole point of the speech act. Thus, the act is prescriptive. Certainly, you can describe something to someone, but your purpose is to instill a belief or item of knowledge; it isn’t purely an exercise in one’s descriptive abilities. What would be the point of that? It isn’t as if facts need you to state them in order to be facts. You aren’t talking to them! If you went round simply stating facts, people would think you insane, not someone with a secure grasp of the practice of assertion.

Thus, it is not distinctive of moral speech acts that they are prescriptive. They may well prescribe bodily acts as well as mental acts, but they don’t differ from assertions in point of their prescriptiveness. The prescription not to steal is a prescription just like the prescription to believe that it’s raining outside; it merely prescribes a different kind of thing. The language-game is always prescriptive; moral recommendations are just a subclass. Questions are requests for information: the speaker is urging the audience to supply a certain piece of information, i.e., prescribing that action. All communicative speech aims for a response of some sort. One speaks to someone in order to secure a certain end concerning that person. In a sense, all communicative speech is pragmatic and social. Prescription is the common factor—not truth-stating or fact-representing. So, speech is not about expressing one’s thoughts in a solipsistic manner; and moral speech is likewise not expressive in the sense typically intended. One is not expressing emotion in a social vacuum; one is recommending something to someone. Often it takes the form of a proscription: Do not steal! That is, the negation of a recommendation. It may be said that one is expressing oneself to someone in order to secure a certain result, but the idea of solitary self-expression is alien to speech as we know it. One may express moral contempt or disgust about something to someone, but not in order to bring about personal nirvana or for artistic purposes. In a sense all speech is normative or evaluative: it would be right to do such-and-such (morally, epistemically). You ought to believe that it’s raining outside and you ought to avoid treading on people’s toes—that sort of thing. No speech is norm-neutral.

There is a kind of use-mention confusion running through these discussions: we should be able to read the linguistic off the metaphysical and vice versa. If there is a metaphysical distinction between facts and values, that should be reflected in the taxonomy of speech acts; so, the speech act can give us a clue to the metaphysics. There should be a principled semantic dichotomy between factual discourse and moral discourse. But this is not true: the world is one thing, language is another. Language has its own nature and purpose; it isn’t a faint copy of reality. Nor is human action somehow isomorphic with the world. All speech is prescriptive because of human purposes, not because its subject matter is somehow objectively prescriptive. Indeed, that is a kind of category mistake—a confusion of words and things. Prescriptivism is not a type of metaphysics but a contribution to anthropology—the anthropology of language. If you want to find out about facts and values, you need to look at the world not at language use.[2]

[1] The Beatles: “Believe me when I tell you, I’ll never do you no harm”.

[2] I don’t like the traditional distinction between facts and values, because it suggests that values are not facts (whatever they are). I prefer to speak of facts of value and facts of non-value. We don’t seem to have a term for the latter kind of facts, so we get horrors like “descriptive facts” and “genuine facts”. It is the non-value facts that seem suspiciously indefinable; the category is probably spurious.