My fridge is an inanimate object. It doesn’t desire anything. It doesn’t have any preferences or interests. It is entirely non-sentient. If I slowly and sadistically hammered a nine-inch nail right through the side of my fridge, or even nailed it to a cross, it wouldn’t matter morally at all. It wouldn’t feel a thing.

Mind you, if you hammered a nail into my fridge, it would matter morally, because it’s my fridge, and I don’t want you to do that. I’d prefer it if you didn’t do that, and it would harm me if you did. But the harm would be done to me as a sentient individual rather than to the fridge which feels nothing, cares for nothing, and deserves nothing.

See the difference? On the one side, something that isn’t sentient and doesn’t deserve moral respect. On the other side, something that is sentient and so does deserve moral respect.

What about my society or my community? – Groupings of individuals are composed of individuals who are sentient and so do deserve moral respect, but the groupings themselves are as non-sentient as my fridge. Like a triangular arrangement of non-triangular dots (∴), the parts have a property that the whole doesn’t have. The mistake of assuming the whole has a property that the parts have is called the “fallacy of composition”. (The converse error of assuming the parts have a property of the whole is called the “fallacy of division”.) The important thing to see is that it is a mistake.

‘Individualist’ is an obvious word for someone who thinks sentient individuals deserve moral respect but thinks inanimate, non-sentient objects like fridges or society do not deserve moral respect. Please note that an individualist so understood would not normally be someone who “lacks compassion”. Why not? – Such an individualist would normally think that since individuals count, individuals have responsibilities to look after each other and to respect each other’s interests. “People matter” – in fact only individual people matter. This sort of individualist wouldn’t normally have compassion for inanimate, non-sentient objects such as fridges or society, because why should they?

But evidently, the word ‘individualist’ is also used in a pejorative sense to mean someone who “lacks compassion” like that. Presumably, this second sort of individualist thinks society should be run along the lines of “every man for himself”, with each individual protecting his or her own interests and not caring about other individuals. Philosophers often distinguish between the two sorts of individualism by labeling the first “liberalism” and the second “rugged individualism”. I hope you can see why traditionally, liberalism was associated with the left wing of politics rather than the right wing, and why the word ‘liberal’ is sometimes used in a sloppy way to mean “left wing”.

When Margaret Thatcher’s critics berate her for “not caring about society”, what do they mean? – Usually they mean that she didn’t do enough to protect the interests of weak individuals from the selfish greed of strong individuals. That strikes me as a perfectly legitimate criticism.

But some of her critics seem to mean that she was wrong to care about individual people instead of caring about an inanimate, non-sentient object called “society”. It strikes me as inhumane to care about non-sentient “collectives” of people such as nations or races instead of sentient individual people. That way lies fascistic nonsense about “destiny” and collective culpability. So I think this second sort of criticism is illegitimate and conceptually confused.

I don’t expect non-specialists to be familiar with technical philosophical terms, but I do hope that people of average intelligence can grasp the difference just discussed, and not get carried away by a rather everyday sort of ambiguity. (Such seems to have been the fate of current Irish president Michael D Higgins.)

In the hope of bringing a little bit more “harmony” where there was “discord”, let’s use language clearly!

How much should society do to protect the unusually vulnerable? For example, how much effort should be put into making sure that nut allergy sufferers can avoid nuts? We probably agree that society should insist on food products containing nuts being clearly labelled. But few would say society should ban nuts outright in all food products. Although it’s statistically inevitable that some nut allergy sufferers will die of their condition, and although we could prevent those deaths by making sure that all food production is wholly nut-free, such drastic action would surely “go too far” by placing too great a burden on everyone else.

As a preference utilitarian, I think we should decide such questions by balancing the satisfaction of preferences against the thwarting of preferences of everyone involved.

There are two components to that sort of balancing: how many and how strong the preferences involved are, and how confident we can be that this or that course of action will in fact satisfy or thwart this or that preference. Although allergy to nuts is rare, it is “normal” in the sense that it is a recognized condition forming a fairly distinct category. It is a “typical” condition which would appear as a local “bump” on a distribution graph. We know the effects of eating nuts on those who fall into this category – they are always serious and sometimes fatal. We can also be quite confident that most of these allergy sufferers have a strong preference to continue living a healthy life. So we can be confident that making sure they can avoid nuts will satisfy strong preferences. But we also know that nuts are an important part of many valuable foods, and many more people have a weaker preference for nuts to remain available. (In fact most nut allergy sufferers probably want nuts to remain available to those who do not suffer from the allergy.)

So even with as clear-cut a condition as allergy to nuts, answering the question of how far society should go to protect unusually vulnerable people is a delicate balancing act, involving some quite subjective judgement calls.

With less clear-cut conditions, this balancing act becomes even more delicate and subjective. For example, people whose bones are brittle with osteoporosis are prone to suffer quite unpredictable injuries, from the merely painful to the crippling or life-threatening. These injuries might involve other people, whom we might or might not hold morally responsible for them. If the injuries involved seem inevitable, we may deem the other agents’ involvement to be irrelevant. Suppose an osteoporosis sufferer is politely ushered off a bus, say, but breaks a hip stepping onto the hard pavement. We should say that the hip was bound to break sooner or later anyway, and that the polite fellow-traveler was in no way responsible.

The least clear-cut conditions involve mental vulnerability. Some people may be living their lives “on eggshells”, or on the brink of suicide, and they seem to be “pushed over the edge” by an insult, or personal slight, or invasion of privacy.

Mental vulnerability involves two wild cards, which correspond to the two “components” mentioned above. The first is that it’s much harder to tell how strong the preferences involved really are. The second is that it’s much harder to predict whether preferences are satisfied or thwarted by action such as passing a new law. I’ll deal with them in turn.

The first wild card is that a preference is not simply what we enjoy, but what we opt for in action. Often, we make painful choices – we literally opt for the more painful alternative. We deliberately choose to undergo experiences that we do not enjoy at all. For example, most of us want to know if our spouses are faithful. Finding out that they are not faithful is not at all pleasant. But we opt to find out, because we prefer to know the truth. Much the same applies to education in general – and indeed to any sort of personal growth. Very often, “no pain, no gain”.

However, we often protest most loudly against what we enjoy least rather than what we opt for last. That is because the very act of protesting can affect the outcome. Children who wail loudly about the horrors of eating broccoli might persuade their parents to not insist that they eat broccoli. Transgendered people who complain in the strongest terms about use of the wrong pronoun might be engaged in an act of persuasion rather than honestly expressing the strength of a preference. Bereaved parents may hope journalists will be less invasive by expressing the grief they feel for the loss of a child as if it were the anguish of a breach of privacy. And so on.

I would go so far as to say that we are systematically misled about the actual strength of preferences by the vehemence with which people tend to express them.

Now for the second wild card in balancing preferences for mental (as opposed to physical or material) well-being. It is very hard to predict whether a given act will result in the satisfaction or thwarting of such preferences. The mind is complicated, and so is the law. It is very hard to predict the effects of any law that affects minds. So a cautious conservatism is appropriate when we consider changing laws that we already know work tolerably well in practice.

For those reasons, I would argue against any law that restricts press freedom in the hope of protecting the mentally vulnerable. People who are mentally vulnerable are like people who suffer from osteoporosis, only more so. Their condition is even less “normal” in the sense that each mental vulnerability is unique, unlike more “typical” conditions such as allergy to nuts. Although it sounds callous, we might have to say “the harm was bound to happen sooner or later, and the agents who caused the harm are not morally responsible for it”.

Even if you do not agree with that conclusion, I hope it’s clear that I remain a preference utilitarian in reaching it.

Trivially, we act in ways that tend to satisfy our own preferences. Preferences are just what we prefer, and what we prefer is just what we would choose if we were able to act freely.

The satisfaction of a preference is an “objective” feature of action. For example, if I want my lover to be faithful to me, that preference is satisfied if and only if my lover actually is faithful to me. It is satisfied even if I think (mistakenly) that she is not faithful to me. It is thwarted if she is not faithful to me even if I think (mistakenly) that she is faithful.

Obviously, satisfaction of preferences has little to do with experienced pleasure. Thus preference utilitarianism is sharply distinct from traditional hedonistic utilitarianism (of the sort promoted by Jeremy Bentham).

Preference utilitarianism understands “the right” in terms of “the good”. It’s good for oneself when one’s own preferences are satisfied. It’s good for other agents when their preferences are satisfied. Morally right acts are those that promote the good of agents in general. Of course this is usually a matter of compromise between competing goods.

Since all genuine agents have preferences, and all genuine agents can act in ways that satisfy or thwart preferences of agents in general, the concepts of moral rightness or wrongness apply to the action (and inaction) of all genuine agents. This includes animals.

The meaning of the word ‘right’ in preference utilitarianism is different from the meaning of the word ‘right’ in other moral theories. This is in keeping with the broader truth that the meaning of theoretical terms depends on the theory they occur in, and on the way that theory is applied in practice. In other words, “meaning is use”.

This can cause some equivocation, as Kuhn recognized in The Structure of Scientific Revolutions. Suppose one person uses Newtonian mechanics, and another person uses Special Relativity. Then they use the word ‘mass’ in different ways. Its meaning differs between them. The first person takes mass to be an intrinsic feature of matter. The second person takes it to depend on the reference frame, and thinks it has a closer connection to energy than matter. Analogously, the word ‘right’ differs between a preference utilitarian and anyone who holds a rival moral theory.

Despite its verbose name, preference utilitarianism is probably the simplest moral theory there can be, involving the least commitment to other moral concepts such as virtue, rules, or culpability. It thus deserves to be called the minimal moral position.

Most European thinkers of earlier centuries thought that “savages” lived more primitive lives than Europeans. At the very least, they thought that “savage” culture was less highly developed than European culture, and many believed that “savages” themselves were racially “inferior”, so that they were more like children than adults. For example, JS Mill opposed paternalism in general, but made exceptions of children and “those backward states of society in which the race itself may be considered as in its nonage”.

Were these earlier thinkers racists? I don’t think so. The word ‘racism’ is often applied to the belief that one race is “inferior” to another race, and by that criterion some or all of the above-mentioned thinkers would count as racists. But I think the word ‘racism’ should be strictly reserved for a moral failing: refusing or neglecting to give due consideration to some individuals’ interests because they belong to a particular race.

Racism is motivated by malice rather than incorrect information. Our attention should not be focused on beliefs, which are “afferent” states of mind in being normally shaped by states of the world outside the head via perception. Instead, we should focus our attention on volitional states such as desires, which are “efferent” states of mind in that they normally shape states of the world outside the head via action. Racism is primary caused by bad attitudes and unjust sentiments – it’s a culpable willingness to disregard interests through action or inaction. This is different from sincerely believing something to be a fact. Having a mere belief might be an epistemic failing, especially if the belief is false, but it can’t really be a moral failing.

Why not? Let’s accept that as a matter of fact, in the real world, no human race is “inferior” to any other race. Please be absolutely clear that I do not mean to question that fact. But now let us imagine an alternative world, in which we discover a long-lost race of humans who, alas, really are “inferior” to us in some respect or other. This is a wholly imaginary world, but it should be quite easy to imagine, because we already believe something akin to this about dogs. Statistically, Chihuahuas are “inferior” to Great Danes in respect of size, and Red Setters are “inferior” to mongrels in respect of intelligence. The “inferiority” of one canine race compared to another in no way justifies ill-treatment of members of the “lesser” race, in fact if anything it helps to justify special care and concern for the less able ones to protect their own interests.

In this imaginary world, a rational, well-informed person would adopt the belief that our long-lost race are “inferior” to other human races, in whichever respect we were imagining them as being statistically “inferior”. Note that in this imaginary world the adopted belief is true. But in rationally adopting such a belief, well-informed people become “racists”, according to the criterion used above to deem Hume, Kant, Mill, Darwin et al to be “racists”.

Therefore, that criterion must be wrong. No external fact about the world could possibly turn a rational, well-informed person into a racist. The mistaken criterion is counting a mere belief that one race is “inferior” to another as the mark of racism.

As far as I know, the alternative criterion (suggested already above) was first proposed by Peter Singer in his book Practical Ethics. It does not depend on belief: instead, racism is understood as failure to give equal consideration to some individuals’ interests because they belong to a particular race. From what I know of their lives, Singer’s alternative criterion exonerates Hume, Kant, Mill and Darwin of the charge of racism. As far as I know, none of them did anything to mistreat members of other races. But it does not exonerate Heidegger of that charge: Heidegger was instrumental in getting Husserl removed from his professorship at the University of Freiberg on the grounds that he was Jewish.

It isn’t all that surprising that thinkers of earlier centuries assumed “savages” were “inferior” in some ways to Europeans. At the time, little was known of the sophistication of their cultures. We now know that there is no such thing as a “primitive” human language, and that technological advance is a poor indicator of cultural richness. But thinkers of earlier centuries had fewer opportunities to find out about other cultures.

It strikes me as especially unfair to use the mistaken criterion of racism against Hume. More than anyone, Hume helped to clarify the difference between “is” and “ought”, and the relation between belief and desire as the component mental states that explain action. To accuse him of racism on the grounds that he had a “bad belief” is not only to use a mistaken criterion of racism, but also to overlook the insights Hume himself contributed to our understanding of action and morality.

The mistaken criterion judges belief in moral terms instead of epistemic terms. That is typical of a recent wave of intolerance, which I call intolerance of creed. I have written a little more about it here.

I got into an enjoyable Twitter debate recently with another philosopher over this old nursery rhyme:

Sticks and stones will break my bones
But words will never harm me.

I think there’s a lot of wisdom in this old rhyme. (Or at least this early version of it – a later version substitutes ‘hurt’ for ‘harm’, which affects its meaning.) Construed as advice to children (or Irish politicians) it says: “if you want to be like me, don’t take offence too easily, or assume you have actually been harmed when you have merely been insulted. Try to not let your feelings get hurt too much by what people say.”

I like that, because I think people should put much more effort into not taking offence than not giving offence. What we call “offensive” is often nothing worse than a moral opinion that differs from our own moral opinions. Because it conflicts with our morality, we think it’s immoral, and that’s often a cue for condemning it and ostracizing its utterer. But it’s good for us to hear opinions that differ from our own, especially different moral opinions, because we tend to avoid listening to them. We may not smile sweetly as we hear these opposed opinions, but the widening of our horizons is salutary.

Even if what is said is motivated by genuine malice, and causes genuine mental anguish, that is not the same as actually being harmed. Personally, I would prefer to know how much someone dislikes me, even – in fact especially – if it hurt my feelings to find out. It’s usually better to know “where we stand” with others than to be cocooned in cotton wool where everyone plays “nice” and no one says what they really feel. If I really want someone to like me, I want to know what they really think of me, even if they don’t like me. We are usually harmed more by dishonesty than by honesty.

Most people tend to like those who are like them, and to dislike those who differ from them. So the ones who are subject to verbal abuse are often unusual in some way – they have might unusual talents, unusual opinions, unusual appearance, or unusual mannerisms. They may belong to another race, or have a minority sexual preference. If we try too hard to avoid causing offence instead of taking offence, it can foster an atmosphere in which differences are “not to be mentioned” – they’re swept under the carpet. But the great political “pride” movements celebrate differences rather than sweeping them under a carpet. If we insist on prohibitions or even inhibitions on mentioning difference, we may unwittingly substitute “black shame” for black pride, or “gay shame” for gay pride. I think that would be a very bad thing.

“Sticks and stones” can also be understood as saying something interesting about language and agency. When we act, a complicated causal chain links our intentions with the effects of our action. For example, inside an arsonist’s mind-brain are beliefs and desires that constitute his intention to set fire to a building. These cause activity in motor neurons, which move body parts, which result in the pouring of gasoline and the striking of a match, and so on. Many conditions are together sufficient for the fire, including the presence of oxygen. Some of them are more relevant than others if we are trying to prevent arson. To focus on the presence of oxygen – or even on the availability of matches – as “the” relevant, principal or culpable part of the cause of the fire is to focus on the wrong condition.

That’s because we can do nothing at all about the presence of oxygen, and the non-availability of matches wouldn’t be much help, as an arsonist could easily fall back some other – possibly more efficient – means of lighting fires. In that respect, we might contrast matches with assault rifles. Banning matches would probably do nothing to prevent arson, whereas banning assault rifles would probably help to reduce the scale of massacres of the sort seen recently in Newtown, Connecticut.

The “sticks and stones” rhyme reminds us that words are more like matches to an arsonist than assault rifles to a crazed gunman. It is an “error of focus” to think that words themselves are “weapons”.

I’m increasingly coming to see a wide gulf between people who see values as relative to agents and those who see values as independent of agents. We might call the first sort of person a “value relativist” and the second sort a “value absolutist”, but we must be careful to avoid assuming that value relativists think there are no psychological facts about how much an agents values something. I love spinach, in other words I value it highly among foods, whereas you may hate spinach. The value of spinach is relative, as it differs between us, but it’s a truth that I value it highly, and as a truth this is “absolute”.

Maybe the distinction between value relativists and value absolutists is completely obvious to everyone already. – I confess it hasn’t been obvious to me till quite recently.

Practically everyone thinks “beauty is in the eye of the beholder” – it’s just whatever is appreciated as beautiful, so it’s relative to whoever appreciates it. Value relativists go further and say that the same applies to all values: something is only valuable because some agent or other regards it as valuable, in other words he regards it as worth pursuing in action. And an agent only regards something as worth pursuing if he or she in fact pursues it, however attenuated their pursuit may be. It’s valuable to him because he prefers it. His choices reveal his preference for it.

Value relativists – like me – tend to be opposed to paternalism. ‘Paternalism’ is the word for forcing people to do things “for their own good”. For example, paternalists don’t just oppose smoking because it harms passive smokers, but because it harms smokers themselves by damaging their health. Value relativists who oppose paternalism think damage to the smoker’s health only counts as harm inasmuch as the smoker pursues good health. And he can’t pursue good health with all that much ardour if he smokes! So if harm is involved here, it results from some sort of internal conflict of desire, or else from ignorance on the part of the smoker. In the first case he somehow both wants good health and at the same time doesn’t care much about good health, perhaps as a result of addiction. In the second case he wants good health but isn’t aware that smoking is bad for health.

So two approaches are open to the paternalist: he can openly defend one value over another – such as good health over the pleasure of smoking – by assuming that values are absolute, and therefore can be ranked independently of individual agents. Or he can invoke his supposedly greater knowledge than that of the agent. As a paternalist, he regards himself as entitled to persuade not just through the use of reasons, but through the use of force.

Paternalists tend to avoid overt value absolutism because it leaves them with the problem of saying where absolute values come from, and why those who do not share such values should adopt them. Instead, paternalists nearly always focus on the supposed ignorance of agents they would apply coercion to. That is, they tend to defend coercion by saying that an agent often doesn’t know what’s in his own best interests. In other words, they argue that the agent has false beliefs about what is good for him.

Well, it is certainly true that everyone has some false beliefs. Not so long ago cigarettes were advertised with doctors’ endorsements as being good for the health. False beliefs were involved there, because cigarettes are in fact bad for the health. Paternalists in effect claim to have knowledge that others don’t have, just as all of us nowadays have knowledge that doctors didn’t have a hundred years ago.

Is it legitimate to use force against ignorance in this way? I don’t think so, for several reasons. First, any attempt to force someone to act in a way his own internal motivation does not underwrite is morally questionable. The agent might even regard the forced behavior as morally forbidden, in which case the paternalist is like a missionary forcing his weaker charges to act “against their conscience”. This use of force undermines autonomy. Second, attempts to impart beliefs by means of force are psychologically questionable: the aim here is a sort of “brainwashing”, and in the real world any such aim is unlikely to succeed. “Brainwashing” is mostly a fantasy from spy movies. Third, attempts to impart knowledge by coercion are epistemologically questionable. Knowledge is generally understood as true belief sustained by rational reasons and/or reliable processes, and yielding to force can hardly be counted as either. Furthermore, as JS Mill eloquently argued in On Liberty, it is always possible for any of us to be mistaken, and to presume to decide matters for others is in effect to assume one’s own infallibility. Which is ridiculous, because no one is infallible, including the Pope.

This doesn’t only apply to paternalism, but to any attempt to force behaviour, beliefs or knowledge on the as-yet unconverted. So-called “skeptics” who would override parents’ judgements by forcing them to have their children vaccinated or operated upon would do well to reflect on the threefold evils just described.

As I said, despite dubious ethics, dubious psychology and dubious epistemology, paternalists are generally more comfortable treating the discussion as being about true or false belief rather than about desires and values, which are neither true nor false. Value relativists should decisively move the discussion back into the realm of values, where it belongs, because this isn’t really a factual issue. That is, it’s not an “is” question of whether I know what’s good for me, but an “ought” question of what I want. If I want to smoke, no one can gainsay that, because in effect I create my own values.

Let’s reject the hypocrisy of paternalists who claim to act with more knowledge or better beliefs than the people they coerce. They should be seen for what they are: old-fashioned absolutists who are trying to impose their supposedly absolute values on others.

The value-relativist-versus-absolutist divide doesn’t just emerge in paternalism. It also lies behind Isaiah Berlin’s famous two concepts of freedom. On the one hand, there is “negative” freedom, the ability to do what you actually want to do by virtue of an absence of external obstacles. On the other hand, there is “positive freedom” – if it can be called “freedom” at all – which is essentially being “empowered” by wanting the right things. But the only way there can be “right things” independent of agents actually pursuing them is if values are absolute. These absolute values are skeletons in the closets of those who assume a positive concept of freedom as well as paternalists.

Most of our concepts are “family resemblance” concepts – in other words, they do not have a single criterion of application but instead apply to things that resemble each other in a variety of ways, none of which need be essential or definitive. Thus two items can belong to the same kind or “family” despite sharing none of the features used to determine membership of the family in question.

One such concept might be that of “fascism”. Fascism was unquestionably one of the greatest evils of the twentieth century, and arguably still exists as a minority political movement in most countries of the world. It is understandable if the word ‘fascism’ is over-used, as to most sane, decent people it still stands for the embodiment of political evil. There is much disagreement over what to count as examples of fascism, beyond the paradigm cases of Hitler’s Germany, Mussolini’s Italy and Franco’s Spain. Yet despite this disagreement, there is surprisingly little discussion of what should be counted a feature of fascism. What are the “family resemblances” that determine membership of the fascist kind?

Here is my own tentative list, in very rough order of centrality. Many will see an inappropriate frivolity where there should be solemnity. Many will see bias on my part for including deontological ethics and excluding right-wing economic policies. I’d defend the first by appealing to the occasionally constructive aspects of ridicule.  I’d defend the second by saying an appearance of bias would be impossible to avoid in anyone’s list of this kind. My list is not intended to be definitive – it’s just meant as a starting contribution to discussion of a surprisingly neglected subject. Here it is:

1. Anti-individualism
2. Nationalism
3. Militarism
4. Opposition to freedom of thought and expression, intolerance of non-orthodox opinion, hostility to difference in general
5. Enforcement of (4) above by a spying and/or policing network, often literally of secret police
6. Belief that there is a “way things were meant to be”, a quasi-divine “order of things” or design that it is our duty or “destiny” to re-establish
7. Belief in historical entitlements and inherited culpabilities
8. Demonization of a group whose ethnicity is uniquely “tainted” by history
9. Fixation with symbols and symbolic gestures
10. Assumption that morality is constituted by rules, so that very severe consequences of action can be overlooked as irrelevant because “the rules were followed to the letter”
11. Dressing up in kinky clothes, especially leather gear

There’s a train of thought that starts off like this: “causing physical injury is very bad, and it should be illegal”. It then proceeds to the thought that “causing severe mental anguish is quasi-injurious, so it too should be illegal”.

The idea is that it is pretty much the same thing to inflict “mental wounds” as physical wounds. Mental wounds are thought to “hurt” in much the same way as physical wounds. They involve unpleasant experiences such as terror, bereavement, and if an insult is sufficiently insulting, the anguish of being offended.

This train of thought is going off the rails. What’s sending it off the rails is hedonism: the ancient Epicurean doctrine that pleasure (which includes relief from pain) is the only good. With hedonism, the focus of any deliberation over action is conscious experience. The crucial moment of derailment occurs with the idea that what’s wrong with physical injury is “it hurts” – and therefore what’s wrong with mental injury is “it hurts too”. If what’s wrong with someone breaking your leg is that having a broken leg is really painful, then what’s wrong with someone threatening you is it’s really frightening to feel threatened. It all boils down to the quality of experience.

“We already have laws that prohibit the inflicting of physical wounds”, the train of thought continues, “so we need laws that prohibit a similar infliction of mental wounds”. This can end up in a sort of semantic meltdown: “we have a word for the first sort of abuse – namely ‘violence’ – so we need a word for the second sort of abuse. We must enlarge the extension of the word ‘violence’ so that it applies to these cases of mental abuse as well.”

Here, a bad idea in philosophy of mind leads to legal, semantic and political disaster. No good can come from the Orwellian lunacy of declaring non-violence to be violence.

The bad idea in philosophy of mind is the assumption – associated with Descartes more than any other philosopher, although it is very widespread – that we are centres of consciousness. So when bad things happen to us, what’s wrong with them is they affect our experiences in an unwelcome way.

It’s easy to see why this idea is mistaken: all sorts of bad things that can happen to us that cause no pain at all, and conversely, some good things can happen that are quite painful. For example, murder is no less wicked if the murderer kills his victims painlessly in their sleep. Or again, it doesn’t detract at all from the evil of betrayal if it’s done in secret and so is not experienced as painful by its victim. Conversely, the discovery of betrayal is better than its non-discovery, and that usually is painful. A painful truth should still be welcomed as truth.

There is an alternative to the crude hedonism of understanding good and bad in terms of pain and pleasure. Instead of seeing ourselves through Descartes’ eyes as centres of consciousness, suppose instead we see ourselves as reflective agents, so that when bad things happen to us, what’s wrong with them is that they affect our agency in an unwelcome way. They thwart our preferences.

Looked at that way, the harm of physical injury primarily consists of the way it interferes with human agency. When a cook cuts a finger, the usual procedures of cooking have to be curtailed. Losing a leg is a lot worse, even if it hurts less.

Looked at that way, the very idea of a “mental wound” is rather odd. Of course we can suffer brain injuries and so on, but unpleasant experiences are often to be welcomed as an accompaniment to the achievement of more accurate understanding. We are frightened by danger, so to avoid it we have to feel fright, unpleasant though it may feel. Please note that as long as we know the danger isn’t real, the feeling on its own is often actively sought after: that’s why people go to horror movies, or enjoy reading novels about cruelty.

JS Mill argued – convincingly, in my opinion – that we should only legislate against the causing of harm to others, and even then, only when the harm is quite severe. So we might legislate against the thwarting of strong preferences, but we should not legislate against mere mental discomfort. Sometimes mental discomfort is a by-product of a welcome adjustment to the rest of an agent’s belief-system. Sometimes an external show of mental discomfort is accompanied by internal mental satisfaction, and if the truth be told is actually welcomed as an opportunity to express moral disapproval and social solidarity against a common enemy. No good can come from rewarding the “putting on of a show” of being really, really mentally anguished or offended. That is for people who are mentally ill, or children, or “chancers” (as we call them in Ireland).

In general, words don’t get their “meaning” from definitions but through use – that is, from the habitual linguistic behaviour of people who speak the language. For example, the word ‘rot’ gets its meaning in English from the way English-speakers respond to and interact with various sorts of decay, both literal and metaphorical. German-speakers use the same word (or at least the same combination of three letters) to mean something quite different, something that is again established through habitual behaviour – in this case of German-speakers responding to and interacting with things that are colored red rather than decaying.

The word ‘rot’ was in use in both German and English before anyone tried to explicitly define its various meanings in any sort of dictionary. Definitions in dictionaries can help to guide use, but only when use is already established among competent speakers. Usually, definitions “explain meaning” or “describe use” rather than “create meaning” or “prescribe use”.

This is a fairly new discovery. For centuries, most philosophers assumed there is something like an “essence” of what a word applies to, so that a single criterion can decide whether or not the word applies. In that situation, an explicit definition might stipulate or prescribe how the word should be used from now on, thus in a sense “creating its meaning”. Many also assumed that the meaning of many words is learned through ostensive definition (through pointing, experiencing first-hand, etc.). And there were some who assumed that natural language has a foundational, axiomatic structure like mathematics. As these ideas fell out of favour in recent decades, philosophers tended to lose interest in explicit definitions. The clichéd demand to “define your terms!” no longer sways anyone (if it ever did).

But while we should reject those bad ideas, we mustn’t wholly give up on definitions. Every now and again the explicit definition of words can play an important role in clarifying discussion and in guiding our understanding. Sometimes we disagree with one another, but cannot hope to resolve our conflict of opinion because we are simply “talking past one another”. Habitual behaviour with the same words can differ between us, not because we speak entirely different languages (as with ‘rot’ in English and German), but because we have different theoretical and/or moral commitments, we move in different social circles, we react differently to similar situations, and so on. A definition can help to clarify where we stand, especially if we stand on marginally different ground.

The word ‘rape’ has been in the news a lot recently, because people who use the word obviously differ – to some extent – over what it applies to. Practically everyone agrees it applies to non-consensual sex, of course, and that non-consensual sex is immoral. But no one seems inclined to be very clear about how explicitly “consent” should be given for a sex act not to count as rape. At one extreme, few would expect a man to ask an explicit verbal question as to whether his partner gives her consent to sex – and then to wait for an explicit verbal answer – on every single occasion of a sex act. That would be mad. These things are generally understood through familiarity and “body language”. At the other extreme, few would agree with George Galloway that a man can initiate penetrative sex with a woman who is asleep, as long as consensual sex has already taken place when she was awake. I think most people would count that as rape. I do.

The moronic and ill-informed Missouri Senate nominee Todd Akin is widely criticised for using the term ‘legitimate rape’. I think he meant to use this term to refer to “unambiguous, paradigm cases of rape” rather than “cases of rape which are legally acceptable”. His inept use of language combined with ignorance in thinking rape can’t easily lead to pregnancy are good reasons to neither vote for him nor anyone else from his political party. But it isn’t a good reason to deliberately and dishonestly misinterpret what he probably intended to say.

If non-philosophers differ over the word ‘rape’, philosophers differ over the word ‘sexism’. We disagree over what to count as sexism, and why it is morally wrong. I thus find it unacceptable for any philosopher to appeal to claims such as “we all know what we mean by ‘sexism’”, as someone did recently with me. Philosophers are intellectually obliged to do more than that.

Another philosopher – one I respect greatly and often agree with – defines ‘sexism’ in terms of discrimination, prejudice and stereotyping.

But I have problems with that too. Discrimination is a necessary and central part of everyday cognition, and of morally obligatory allocation of scarce resources. For example, when we drive, we discriminate between one side of the road and the other. We can only give pedestrians the right of way by discriminating footpaths and zebra crossings from other parts of the road. Similarly, when we allocate public funds for breast cancer screening, or to support maternity leave, say, we have to discriminate between those who are likely to benefit and those who aren’t, and in practice that usually means discriminating between men and women. It is cognitively unavoidable and morally perfectly right to exercise that sort of discrimination. So discrimination per se can’t be what makes sexism immoral.

The word ‘prejudice’ refers to unjust sorts of “pre-judging”. But some sorts of “pre-judging” are morally justified. To simply have expectations and to be prepared for the eventuality that these expectations might be realized is often not only morally justified, but an important aspect of practical wisdom. For example, to solve a problem such as the old puzzle of how to transport a fox, a chicken and some grain across a river carrying only two of them at a time, I have to “pre-judge” how foxes and chickens behave. If “prejudice” and “pre-judging” differ only in that the former is immoral and the latter isn’t, how are we to distinguish – or discriminate, as we might put it – one from the other? We must have some independent moral standard for deciding between them – in which case that standard is what we should be appealing to. (This familiar pattern of argument is adapted from Plato’s Euthyphro.)

Much the same applies to stereotyping as to prejudice. In cases where it is immoral, some independent standard of morality must be appealed to – in which case it is that standard rather than stereotyping per se that we should be looking at. Taking stereotyping to be the “objectifying” of prejudices in the form of caricatures and the like, they can be painful, or even harmful. But if they are genuinely harmful, the real problem is the causing of harm rather than stereotyping. We cannot condemn stereotyping as always wrong, because it is often the stuff of humour, and like satire it can even be salutary – it can help to destroy stereotypes by drawing attention to how ridiculous they are. If we call mere humorous stereotyping “sexist”, then we trivialize sexism by dragging it up to the level of the merely annoying or upsetting.

Sexism is much more serious than that. It involves injustice and genuine harm, not just irritation of the sort that each sex has generated in the other since prehistory.

Following Peter Singer, I suggest that we define ‘sexism’ in terms of interests and the thwarting of interests. According to my tentative definition, a sexist act is one that does not give due consideration to someone’s interests because of his or her sex. “Due” consideration is usually “equal” consideration. I may change my mind about this at a later date; till then, I hope this definition helps to clarify where I stand – for now.