Moral Minimalism

I shall explore the prospects for a minimalist theory of normative ethics. By “minimalist” I mean a theory (analogous to minimalism in linguistics) that seeks to base normative ethics on the most exiguous of foundations, viz. a single moral principle, with other aspects of the ethical life consigned to something extraneous to morality strictly conceived. The moral principle in question is exceedingly familiar: DO NO HARM. That is all that morality contains, according to the minimalism I envisage, neither more nor less. The onlymoral principle is the injunction not to do harm. Usually this principle is included in a total utilitarian package: Do no harm andmaximize wellbeing (welfare, the good, happiness, pleasure). I propose to drop the second conjunct so that morality only prescribes the avoidance of harm. Clearly the two conjuncts are logically independent, though the second is generally taken to include the first: if our aim is to maximize wellbeing, it should surely include minimizing harm. But we may live in a possible world in which there isno harm to be undone or produced, yet still we are subject to an injunction to maximize wellbeing—we must increase the level of wellbeing even if there is no suffering to be eliminated and none that can be produced (this is a world of harm-proof people). More obviously, one could accept the injunction not to harm while rejecting the injunction to promote wellbeing: I mustn’t harm anyone, but I have no duty positively to improve anyone’s lot. For example, I must not strike an innocent man for no reason, but I am under no obligation to make him happier than he already is. So I propose dropping the second injunction while insisting on the first. I call this position “disutilitarianism” because it emphasizes the avoidance of disutility not the production of utility. It is a negative prohibition: it says what we must notdo not what we must do. We must not cause harm, though we have no duty to cause its opposite (if it has a real opposite)—we have no duty to maximize the general good, or even to produce it in a particular case. There is a duty against maleficence, but no duty of beneficence.

Let me immediately address a natural objection, namely that it is clearly morally praiseworthy to promote the good. I don’t disagree, though there are notorious cases in which promoting the good is not the morally right thing to do (the bane of utilitarianism); but I would distinguish between what morality requires and what it is admirable to do. It certainly shows the virtue of generosity to help the poor and needy, but that is not the same as saying that this is a moral duty. It may just be supererogatory. We have a dutynot to harm, but we have no comparable duty to make people happier—though it might be virtuous so to do. I will come back this point, noting now only that moral minimalism does not preclude acting virtuously in promoting wellbeing; it claims only that this is not part of morality in the strict sense. We might even say that not causing harm isn’t a virtueat all, being merely our most basic moral obligation—there is nothing virtuousin declining to strike an innocent man for no reason. Duty and virtue are separate domains.

A main reason for advocating moral minimalism as against full-blown utilitarianism is that the stronger doctrine runs into well-known problems. I won’t rehearse these problems, but they concern considerations of justice and the problem of moral inflation, whereby we turn out to be the moral equivalent of murderers by not helping starving people in distant lands to the point of self-impoverishment. What is crucial, I think, is that there is a deep asymmetry between harming and benefitting: we have an absolute duty not to do the former, but the latter is optional. Partly this is because of the difference between pain and suffering, on the one hand, and happiness and wellbeing, on the other: the former are clearly defined and obviously bad, while the latter are amorphous and not invariably good (e.g. the pleasure-loving happy sadist). The dentist must do his best to avoid hurting you, but he is under no obligation to make you feel happier when you leave his office than you were when you came in—and what exactly would that be? He knows how to avoid harming you, but he may have no idea what would make you happier (a joke, a donation, a pat on the back?). So the harm principle has a different deontic status from the benefit principle. This is of course exactly how we operate in daily life: you avoid stepping on people’s toes as you walk down the street, but you don’t try to cheer everyone up as you pass them by. They will blame you for hurting them, but not for failing to improve their mood. They may think that that is none of your concern, while avoiding crushing their toes indubitably is. So we can say that the harm principle has a greaterhold on us than the benefit principle; I propose accordingly that we restrict morality to the harm principle.[1]

It is a significant fact that all the standard rules favored by the deontologist can be seen to stem from the rule against causing harm. Breaking promises, lying, stealing, assaulting, murdering, acting unjustly—all involve causing harm to others. These rules are prohibitions designed to minimize suffering, ranging from disappointment to physical agony. None of them reflects the utilitarian’s insistence that we should maximize wellbeing—as if by sitting at home doing nothing we have committed grave evils. Of course, it is possible to harm by omission—and that is equally proscribed by the harm principle. You can fail to save someone from being hit by a car, so that your omission harms him or her. But doing nothing to make people happier is not ipso facto a form of indirect harm. We can’t somehow squeeze beneficence in under non-maleficence. The usual rules of morality concern things we are not to do (“Thou shall’t not…) and they all concern the harms that result from doing these things. Bringing each of these specific rules under the harm principle effects a major simplification, making moral thinking easier to manage and sharper in focus. All we really need to remember—all we need to know—is that it is wrong to cause harm. Whenever you are faced by a difficult moral choice you need only ask yourself what action will cause the least harm and then do that. For instance, you should not break a promise to meet A because meeting B instead will increase the total level of happiness in the world; you should avoid harming A by leaving him hanging (maybe suggesting to B that she finds something else to do). It is no small advantage to morality that it should be codified in a single easily remembered slogan. Children need to be instructed in it, and many adults have no aptitude for moral complexity, so keep it simple.

Can you harm someone in order to benefit him later? If so, there is no absolute ban on causing harm. Here we need to distinguish two cases: causing harm now to prevent greater harm later, and causing harm now in order to increase happiness later. The dentist drills the tooth now in order to prevent the pain of later toothache, so she is minimizing pain in the long run: that is morally acceptable and in accordance with the harm principle. But it is another thing entirely to try to justify causing harm now by citing future benefits that don’tinvolve harm minimization—as it might be, applying the rod to the child in the expectation that she will grow to be happier than she would be otherwise. This is far from obviously acceptable and it gains no support from the harm principle, which speaks only of minimizing harm not maximizing happiness. Omitting to do something harmful today can cause greater harm tomorrow, and is therefore morally proscribed; but omitting to do something harmful today that will result in less overall happiness in the future is not to be morally condemned (except by the rigid utilitarian). Even if beating children is known to make them happier in later life, that is no ground for beating them—though if it will prevent them from excruciating suffering later, then it should be done (however reluctantly).  We must always seek to minimize harm, even if harm is necessary to bring that about; but harm can’t be justified by considerations of overall utility, as if pain now is made up for by elation later (as opposed to mere contentment).

It is important to minimalism to distinguish between what it is good for a person to do from what it is morally obligatory for a person to do. Minimalism is only a theory of the latter; it is neutral on the broader question of virtuous or admirable conduct. Living a good life includes acting generously and kindly, even if no harm is reduced thereby. That may seem to leave a lot of moral life outside the scope of the minimalist theory, but in fact it covers more than might be supposed. For much generosity and kindness involve the avoidance of suffering not merely the production of utility. You can harm someone by not being concerned about his or her welfare, as when you callously decline to give food to a starving person. But not all generosity is like that, as with the generous host: she is not avoiding harming her guests by laying on a great feast, but rather adding to her guests’ enjoyment. That is what is not morally required—increasing other people’s happiness. By not voting for tax increases to help the poor you may be harming them indirectly and by omission, so this falls under moral criticism; what does not invite moral criticism is declining to share your resources with people already amply resourced. So quite a lot falls under the prohibition against causing harm, not merely refraining from attacking people directly (animals too). Someone might be exceptionally generous with his friends, by always treating them to fancy dinners and the like; that may be commendable, but it is not morally obligatory. This is a distinction well worth preserving, and it is a virtue of minimalism that it makes the distinction firmly (unlike classical utilitarianism). Much virtuous behavior is discretionary, but moral behavior never is—it is strictly obligatory. Being a miser may not be admirable, but it is in a different category from being a sadist. The paradigm of the immoral act is maiming someone, not providing a thrifty meal instead of a lavish one.

Is the anti-harm theory deontological or consequentialist? You can take it either way, either as a moral rule or as a statement about consequences. That is, you can say that an action is right if and only if it actually minimizes harm, or you can say that the agent must always intend to minimize harm and that this is what makes it right not the actual consequences. I prefer to think of it as an absolute general rule with a number of sub-rules as special cases (such as “Don’t break promises”), but clearly the consequences are crucial in justifying the rule—pain and suffering being bad things in themselves.

I would emphasize the formal merits of the minimalist theory. It is simple, clear, manageable, and practicable. It is intuitively compelling and scarcely controversial in its recommendations (unlike utilitarianism). Its only questionable claim is that there is nothing more to morality than what it includes; but this is mitigated by the distinction between morality proper and what counts as virtuous conduct. It combines the best of deontology and consequentialism. It is what you would expect of a moral system that is designed to help people live together in close proximity. It is non-paternalist. It doesn’t seek to meddle in other people’s lives, as the prescription to make everyone as happy as possible does.  It has a pleasing homogeneity. It is readily universalized. It does not attempt to combine disparate ideas (as in W.D. Ross’s mixed theory). It is easily teachable. It does not call for extremes of altruism and intolerable guilt over never doing enough. It takes what is good in utilitarianism and discards what is bad. The disutilitarian is a realistic, clear-eyed, compassionate, commonsense type of fellow, mainly concerned to prevent pain and suffering. Everything else is icing on the cake. If he can prevent us from harming each other (animals included), he thinks he has done his moral duty. What we choose positively to do, as a matter of personal virtue, is our own affair and of no concern to morality as such.[2]

Colin McGinn

[1]A further asymmetry is this: the harm principle applies impartially to intimates and strangers, but the benefit principle applies differentially according to personal distance (at least according to common morality). You must not harm anyone equally, but it is morally permissible to benefit members of your own family over others. This suggests that the harm principle is part of non-negotiable moral law, while the benefit principle operates according to personal discretion.

[2]The disutilitarian might well contend (echoing Nietzsche) that morality since the advent of Christianity has indulged in a kind of duty-creep whereby virtuous behavior has been converted into a species of strict moral duty. Thus Jesus urges us to give to the poor and needy (defined relatively) and his followers have interpreted this as an extension of our moral duties. But that is not necessarily the right way to interpret the words of Jesus: he is not assimilating charity to the deontic level of non-violence, merely suggesting that we cultivate the virtue of generosity and not content ourselves with the mere observance of our strict moral duty. Perhaps under the influence of Christian ethics, as it came to develop, utilitarian ethics made a virtue of blurring the line between moral duty and personal virtue, thus assimilating the demerit of not being charitable with the demerit of violently assaulting people. That was a conceptual error and one the minimalist is anxious to remedy.

The Alleged Limits of Moral Philosophy

Bernard Williams wrote a book entitled Ethics and the Limits of Philosophy.[1]This title invites interrogation. What kind of limitation might be meant? We can all agree that philosophy is limited in someway: it cannot do what science does, for example, or history or geography or literature or painting. In that sense everything is limited: there is no point in using one’s philosophical faculties in order to answer non-philosophical problems. Someone could write a book called Ethics and the Limits of Scienceand we could be persuaded that science is not the answer to ethical questions, since it is not the answer to many questions, especially normative ones. But isn’t ethics precisely moral philosophy—so how could philosophy be limited in doing the philosophy of right and wrong? What if Williams had called his book Moral Philosophy and the Limits of Philosophy? Of course, real ethical questions involve factual matters, and hence are not properly part of philosophy, but what could be meant by saying that philosophy is limited in dealing with the philosophical aspects of ethics? And is philosophy limited in other areas traditionally designated philosophical too? As it turns out Williams doesn’t really mean that philosophy is limited with respect to ethics (or moral philosophy): he means a certain kindof philosophy is so limited. He doesn’t mean that a more historically rooted and humanistic philosophy is limited when it comes to ethics; he means the kind of philosophy exemplified by Kant and Bentham along with their successors. He means something theoretical, abstract, systematic, monistic, context-independent, non-psychological, ahistorical, absolute, and scientific-sounding. So his title is misleading: he thinks that a certain dominant strandof Western philosophy is limited when it comes to ethics. Not that this strand might not contain important truths and be valuable in its way, but that it has limits—it doesn’t cover the full territory of ethics. This is a less resounding thesis than that suggested by the title of his book. He might more accurately have called it Ethics and the Limits of a Certain Kind of Philosophy. The book would then have gone on to argue that the kind of philosophy in question omits certain important considerations, to be remedied by adopting a different kind of philosophical approach or style or method.

The question I want to raise is whether Williams would wish to extend his thesis to other parts of philosophy. Is it just ethics in which a certain kind of philosophy has inherent limits? Let us call this kind theoreticalphilosophy, meaning thereby to sum up the list of features I cited in the last paragraph. Would he complain that epistemology, philosophy of mind, aesthetics, philosophy of language, philosophy of science, and so on, are not sufficiently historical or humanistic or contextualized? Is his critique of theoretical philosophy as too limited itself limited to ethics? Is it that the other areas traditionally covered by philosophy are perfectly well suited to the theoretical style, but that right and wrong are not? If so, what is it about this domain that makes it stand out so? It can’t be merely that it is a normative domain, because so are aesthetics and epistemology (which concerns what we oughtto believe and is shot through with normative notions), not to mention logic. And why exactly would the normative preclude theoretical treatment while everything else invites it? I don’t recall Williams ever addressing this question—though he certainly contrasted the “absolute conception” of science with philosophical investigations. My question is whether he would be prepared to extend his critique to all of philosophy or whether he intended it as restricted to the case of ethics.

It seems to me this is an uncomfortable dilemma for him. For it is hard to see on what grounds he could restrict it, and yet extending it surely proves too much. It proves too much because clearly theoretical philosophy is not limited in any non-trivial way when it comes to these other areas. How could it be argued that logic and philosophy of language are objectionably limited in their methods and results? Of course, they can be supplemented by other disciplines, but in what way are they just the wrong way to approach the subject? Similarly for epistemology and philosophy of mind: why do they fail to provide an adequate way to approach the questions that constitute their domain of interest? Would Williams be prepared to write a book entitled Knowledge and the Limits of Philosophyor The Mind-Body Problem and the Limits of Philosophy? What other approach to these questions would he favor over the one traditionally practiced by philosophers? Does he think logic should be more historically situated and psychologically realistic? What about the analysis of knowledge or the nature of intention? I myself see no reason to distinguish ethics from other branches of philosophy methodologically, and I also believe that there is no real alternative to the usual way of doing things. So I would see no point in a book paradoxically entitled Philosophy and the Limits of Philosophy, even when that last phrase is understood to mean “limits of a certain kind of philosophy”.

In fact, Williams’ chief targets were Kantian ethics and utilitarianism. He found them too abstract and oversimplified as well as psychologically unrealistic. I can see a point to that critique, but it is an unwarranted leap to suppose that ethics in general has been blighted by the same failings. What about the work of W.D. Ross? What about Aristotle? These are theoretical thinkers in the sense intended—they purport to offer a systematic treatment of ethics valid for all times and places—but they are more pluralistic and realistic than the abstract monistic formulae of Kant or Bentham. True, philosophers are prone to defend oversimplified monistic theories, but it is no abnegation of theory as such to move in a more complex pluralistic direction. Is that all Williams is asking for? Evidently not, but I fail to see why ethics should be held to a different standard than other philosophical topics. In epistemology we can distinguish a rule-based from a consequentialist view of justification: either you follow the rules of induction, deduction and abduction, or justification is defined as simply what makes the best predictions (or has the best results for humans if you are a pragmatist). This is analogous to the distinction between deontology and consequentialism in ethics. We can certainly oppose either view as being partial or limited, but combining them is hardly a move away from the theoretical to something more historically grounded or humanistic. Similarly, we can oppose the monolithic systems of Kant and Bentham without thereby abandoning a broadly theoretical approach to ethics. Pluralism is not inherently anti-philosophical or an indication that philosophy has reached its limits. To reject bad theories, or theories that oversimplify, is not to reject theory altogether.

And is it that Williams finds nothing of value in the theories he criticizes? No: for they crystalize important aspects of morality—moral rules and good consequences, respectively. They are idealizations intended to bring out what matters, much as other philosophical theories are idealizations. There is nothing wrong with that so long as we realize what we are doing. Maybe they aretooidealized, but again that is not a point against theoretical philosophy as such. Nor do I see any real alternative to theoretical philosophy if we are going to keep on doing philosophy at all. Certainly, merely describing the moral attitudes and practices of societies present and past is not a kind of moral philosophy worthy of the name. So I don’t really see what Williams is getting at by accusing moral philosophy of failing to recognize its limits.[2]

Colin McGinn

[1]Harvard University Press: 1986.

[2]I considered Bernard Williams a friend. I admired him as a philosopher. I enjoyed talking to him. We once appeared together on television discussing animals and ethics. I taught a seminar with Malcolm Budd on Ethics and the Limits of Philosophywhen it came out. But I never felt I really understood his position in ethics—either what he objected to or what he favored. I got the flavor of it, if course, but the actual content of his views eluded me.

Remarks on Metaphysics

What kind of statement expresses the results of metaphysical inquiry? Wittgenstein famously begins the Tractatusthus: “The world is all that is the case” (1), “The world is the totality of facts, not of things” (1.1.), “The world is determined by the facts, and by their being allthe facts” (1.11). The use of the phrase “the world” is conspicuous, intended to announce a thesis of metaphysical proportions, but what does it refer to? What does Wittgenstein mean by “the world”? Presumably he means the actual world, though he could certainly be taken to include other possible worlds—they too are constituted by facts (in that world). But what in the actual world is he referring to? Not ethics, because he denies that value is inthe world, and not philosophy since its results can only be shown. Not the self either: “The philosophical self is not the human being, not the human body, or the human soul, with which psychology deals, but rather the metaphysical subject, the limit of the world—not a part of it” (5.641). The facts are what can be stated by means of true propositions, but Wittgenstein doesn’t think that everything (real?) can be stated. He means to exclude some things (if the word “thing” may be permitted). We could take him to be distinguishing the world from our attitudes to the world, including our ethical attitudes. Thus he might say that while the world is the totality of facts the mind is the totality of attitudesto facts. This would be to oppose mind and world (as in the title of a well-known book: Mind and World); the mind is not intended to fall within the denotation of “the world”. This is the narrow interpretation of  “the world” to be contrasted with the wide interpretation that includes the mind within the world.

Two pieces of evidence may be cited for the narrow interpretation. The first is that in a later section Wittgenstein says the following: “Similarly the possibility of describing the world by means of Newtonian mechanics tells us nothing about the world: but what does tell us something about it is the precise wayit is possible to describe it by these means. We are also told something about the world by the fact that it can be described more simply with one system of mechanics than with another” (6.342). Assuming that he refers to the same thing in both places by “the world”, he must be referring to what might be called “the physical world”, since he is not supposing that the mind can be described by mechanics. This certainly fits the general tenor of the book. Second, he construes facts as “combinations of objects”, and there is no reason to believe that he understands the mind that way: how is being in pain or feeling angry a combination of objects? There is no developed philosophy of mind in the Tractatusand it would be merely speculative to suggest that he understands the mind as a totality of facts constituted by combinations of objects. It is true that at one point he speaks of a speck in the visual field, musical notes, and “objects of the sense of touch” (2.0131), but these are not mental phenomena; they are the objects ofmental phenomena (not sensing but thing sensed). It is also true that he may be committed by the picture theory to regarding thoughts as combinations of (symbolic) objects, since they have to be isomorphic to external facts; but there is no reason to suppose that he regards everything about the mind in this way. In any case, it seems clear that he intends the narrow interpretation in the passages cited, so that it includes neither ethics nor the mind nor the self nor philosophy: the world is contrasted with these other domains, not taken to include them. Perhaps we could paraphrase him by saying “the objective world”. That would make sense of his remark that “the world is independent of my will” (6.373), which would make no sense if the will were partof the world. He is quite happy to assert, “The sense of the world must lie outside the world. In the world everything is as it is, and everything happens as it does happen: init no value exists—and if it did exist, it would have no value” (6.41). So he has no trouble excluding things from the world that don’t belong there, in the realm of reportable fact. He is speaking always of what may be mentally represented, not the representing itself. And his point is that the world in this narrow sense consists of facts not things, states of affairs not merely the objects that occur in them. The structure of the world is the structure of facts not objects (particulars, universals).

But now, having settled on the denotation of “the world”, we have the question of the logical form of Wittgenstein’s pronouncements. We know what he is talking about, more or less, but what is he saying about it—and how is he saying it? On the face of it the sentence “The world is the totality of facts” has the form of an identity statement combining two definite descriptions: “The F= the G”. It is not an identity statement joining two proper names, as in “a= b”, though we might substitute a name for one or both of the descriptions, calling the world (say) “Winston”, so that we have “Winston is the totality of facts”. The question then would be how to analyze these descriptions: would Russell’s theory do the job? That gives the decidedly peculiar, “There is uniquely something xsuch that xis a world and xis identical to the totality of facts”, which might also yield its second description to Russell’s analysis. The truth is that the alleged description “the world” is by no means a term of ordinary language but a philosopher’s invention; semantically, it is hard to know what to make of it. In any case, the statement in question purports to identify one thing with another—the world with the totality of facts. Elsewhere we read: “The totality of existing states of affairs is the world” (2.04) and “The sum-total of reality is the world” (2.063)—again, apparently, identity propositions employing definite descriptions. Metaphysics thus characteristically issues in statements of the form, “The world is (identical to) X”, where “X” is to be replaced by some description that purports to tell us the general nature of things.

What is notable is that Wittgenstein’s own statement falls short of what most metaphysicians aim to supply, since he is neutral as to the kindof fact that constitutes the world. All he tells us is that facts make up the world, not what these facts might be (similarly for his talk of “objects” and “states of affairs”). So far as his statement is concerned, these facts might be physical or mental or abstract or divine or unknowable. His theory is merely structural (logical), not substantive: it gives the form of the world not its substance (as he no doubt intends). Still, his statement provides a canonical formulation of a metaphysical thesis—a thesis about the general nature of reality. If we add to it the claim that all facts are physical facts, then we get metaphysical materialism. If we say that all facts are mental facts, we get metaphysical idealism. If we say the world consists of two types of substance, material and immaterial, we get metaphysical dualism. Schopenhauer wrote a book entitled The World as Will and Representation, clearly aiming to make a metaphysical statement (the book was known by Wittgenstein). Plato’s metaphysics can be expressed as, “The world is the instantiation of universals by particulars”. Hegel maintains, “The world is spirit”. David Lewis might say, “The world is the totality of all worlds”. Quine could opine, “The world is what science tells us it is”. The positivists might assert, “The world is what is verifiable”. All these views make use of the general notion of “the world”, and all could agree with Wittgenstein’s structural thesis. The metaphysician is telling us what the world is—its nature, its manner of being. Hesperus is Phosphorous, and water is H2O, and the world is spirit or matter or both or neither. We are offered a very general identity statement purporting to enlighten us about something called “the world”.

It is reasonable to be suspicious about such metaphysical statements. This is not because they are unverifiable or that ordinary language has gone on holiday but because the conditions of reference may not be met. Does the term “the world” really refer to anything determinate as used by the would-be metaphysician? It isn’t much like a regular definite description with uniquely identifying descriptive content, or an embedded demonstrative; and “world” is hardly a regular sortal noun that carries criteria of identity and counting. What kind of entity is the world? What predicates does it satisfy? How is it to be picked from among other things? How can we speak about it as concrete particular with a specific nature? Is it an object? Can it be named? To what end? Sentences containing this pseudo-description, such as “The world is the totality of facts”, are semantically anomalous, though perfectly grammatical; certainly, we can’t just assume they are meaningful, possessed of determinate truth conditions and reference. They seem parasitic on other types of sentence in which the word “world” appears doing more humdrum things (“I’ve searched the world for her”), and thus derive apparent sense from their humble origins. But metaphysical sentences sonorously beginning, “The world is…” are up to something beyond the normal routines of the words they contain: for they purport to refer to the whole of reality—whatever that might mean. Hence the lack of clarity about Wittgenstein’s use of the phrase: does he include value in the world, or logical form, or the mind, or the fact that the world is the totality of facts? (Is this fact also a combination of objects, the world being one of them?) In fact Wittgenstein excludes various things from the reference of “the world”, so the phrase can’t just be a variant of “everything”: but then we need to be told exactly what he does intend to refer to. The phrase trips easily off the tongue, to be sure, but it may still fall flat—it may fail to single out a specific entity. Similarly for “the totality of facts” or “the sum-total of reality”: do we really know what these phrases mean? Presumably they are not intended to include the non-existent or merely fictional (but what about Meinong?), but there are true propositions about them too—isn’t it true that Sherlock Holmes is a detective or that unicorns don’t exist (hence all the problems about whether the world contains negative facts)? It is just not clear that we have hold of a well-defined concept here. What if a common sense type of chap were to protest, “I have no idea what you mean by ‘the world’, though I’m perfectly happy with phrases like ‘the cat in the corner’ or ‘the queen of England’—what is it exactly that you have in mind?” Grammatically, it looks as if we have an identity statement flanked by definite descriptions that pick out entities in good standing, but appearances can be misleading—in which case the standard products of metaphysical inquiry are lacking in sense. At the least we are owed some kind of account of how such sentences work. To put it bluntly, isn’t “the world” a meaningless abstraction, however sublime it may sound—just the kind of thing on which the later Wittgenstein would pour scorn? Isn’t it suspiciously like “the holy spirit” or “the ether” or “the force”—in fact, worse than these because they at least contain relatively well-defined words? Just because I can say, “You mean the world to me” doesn’t imply that I can talk meaningfully about what kind of the thing the worldmight be. Certainly we cannot begin a sentence with, “The world is…” and expect automatic semantic propriety; we need to say more about what precisely we have in mind.

Because the sense of such sentences is unclear, we are apt to interpret them by whatever means comes to hand. And here I think semantics gives way to mental imagery: we form various picturesof what might be meant. These pictures may vary from individual to individual, but they are introduced in order to pin down the import of the proposition we are struggling to grasp. Metaphysics thrives on emergency imagery, particularly spatial imagery. Thus when I hear the sentence, “The world is the totality of facts”? I picture a heapof facts—a mountain of them, what with the world being so large. Wittgenstein tells us at one point, “The world divides into facts” (1.2), and we duly picture a divisible something—something with spatial parts. The world is an assemblage of smaller entities (“facts”) that combine into a larger whole, as rocks may form a mountain. Wittgenstein’s use of “totality” is interesting: not “set”, which might prove not concrete enough, but the more tangible idea of a spatial grouping of some sort—a pile, a stack, a pyramid maybe. The world is an agglomeration of lesser things, where these things are themselves conceived as spatial particulars (like atoms or molecules—atomic and molecular facts).  Such imagery courses through our mind as we study Wittgenstein’s enigmatic text and gives us an illusion of understanding—I know what a heap is! I conjecture that metaphysical discourse is unusually prone to this kind of imagery, as a kind of substitute or crutch. It would be interesting to do some empirical work on such imagery: how frequent is it, are there any universals, what happens when it is absent? Wittgenstein had an engineer’s mind and was fond of the notion of picturing, so it is possible that he had unusually strong imagery when composing the Tractatus: this will have encouraged him to think he was talking sense. And partly he was—but was it complete sense?  Language can carry us away, as he recognized in the Investigations, but so can the mental imagery it provokes: it can provide dubious abstractions with concrete credentials. Isn’t the Tractatusa very visual work, reliant on the reader’s complicity in visualization?

The same is true of other metaphysical visions (!): they are apt to come with pictures attached. What do you think of when you think of dualism? I imagine two entities side by side, one extended and concrete, the other wispy and amorphous (compare the image of consciousness as steam emanating from a steam engine). When I think of materialism I imagine an accumulation of geological strata: at the base we have atoms in the void, with chemistry and biology and psychology laid on this base, like bricks laid on a foundation (and just think of the imagery associated with that word!). I don’t think of the facts of the world as separated in space, like islands, but as built one upon another—vertically not horizontally. Idealism puzzles my imagination because the mind is not so readily imagined spatially, but my imagery is something like a cloud of feathers or a ghostly gathering—a weightless assembly of formless nothings. Plato tried to give imaginative expression to his theory of forms by the parable of the cave, which is full of spatial imagery, but the theory taken neat suggests (to me) nothing so much as a colony of splendid birds of paradise. Frege likened his theory of sense and reference to the optical image in a telescope, in order to make the metaphysics palatable (intelligible), with space explicitly invoked; without this analogy we struggle with mental pictures of free-floating simulacra of things (those elusive “modes of presentation”). Much of the charm of metaphysics derives from these flights of imagination: we contrive to render elusive abstractions mentally manageable. Without this we might flounder in incomprehension, with only words to play with (“the world”, “totality”, “substance”, “immaterial”, “hierarchy”, “supervenience”, etc.). When Wittgenstein remarks, “Objects make up the substance of the world” (2.021) we reach for familiar ideas of substance and think we know what he means, as in “Flour makes up the substance of the cake”. Imagery abets metaphysics—maybe makes it humanly possible. What makes metaphysics meaningful to us is the imagery we bring to its pronouncements: but this is a suspicious gift, intoxicating though not necessarily illuminating. It may simply provide spurious protection from the verbal haze (or blaze). Or it may bias us in favor of views that interact better with our imagination—that provide us with more appealing pictures. Wittgenstein spoke of being held captive by a picture—well, in metaphysics there may be no alternative. In normal discourse we can rely on words to carry us along, but when discourse turns metaphysical words struggle to keep up, and then imagination takes up the slack, or tries to. We find ourselves dependent on pictures of many kinds: of heaps, webs, steam, railway tracks, shadows, lenses, ghosts, exotic animals, shimmering mirages, tools, chess games, light, magic tricks, building blocks, cement, blank slates, sentences—all the tricks of the philosophical trade. In this way we try to give sense to what we are inclined to say. When you read the words, “The world is…” your imagination is activated: you start to form pictures of what might be meant. You would be lost otherwise, or perhaps just not interested.[1]

Colin McGinn

[1]I don’t want to give the impression that I am against all metaphysics (on the contrary), but I think certain ways of proceeding are fraught with linguistic peril, particularly pronouncements of the general form “The world is X”.