To those who think that an institution must be a function of its history it must seem a considerable anomaly that when universities were first set up in the Middle Ages their main aim, apart from being communities of scholars, was to produce theologians, lawyers and doctors of medicine. For arts and what then had some connection with what we now know as science, as incorporated in the traditional seven liberal arts of grammar, logic and rhetoric, followed by arithmetic, geometry, music and astronomy, were thought of as merely propaedeutic to the study of theology, law and medicine. Those seven liberal arts occupied the student during the course for the B.A., which gave the licence to teach, and some students went no further than that. But the main concern of a university lay in what happened after the B.A., in the attainment of a mastership or in what was in effect the doctorate in one of the three areas which I have mentioned before. Indeed, in some universities, particularly those in the south of Europe, the B.A. was given only scant attention and only lip service was given to the idea of liberal arts.

There is a natural view of nonsense, which owes what attraction it has to the apparent absence of alternatives. In Frege and Wittgenstein there is a view which goes against the natural one, and the purpose of this paper is to establish that it is a possible view of nonsense.

Politics has been defined in different and contradictory ways in the last century or so. If politics is to be a single subject of study then contradictory theories should be capable of being related together. In this article I argue that they can be related in terms of what I call the Aristotelian criterion. The article is in four parts. Firstly, I discuss the problem of defining politics; secondly, I introduce the criterion; thirdly, I consider five modern theories of politics (those of Arendt, Oakeshott, Collingwood, Schmitt and Rancière) in relation to the criterion; and fourthly, I use the criterion to put forward an original and capacious definition of politics.1

If killing another human being is morally wrong on at least some occasions (as it clearly is), what precisely makes it wrong on those occasions? I have framed the question thus to indicate that I shall not be considering the view that killing another human being is always and everywhere morally wrong. I take it as read that there are at least some morally justifiable killings (for instance, in self-defence where no other means of disarming one's attacker is available). Once it is clear what is wrong with killing on some occasions it should become possible to explain why it is not wrong on others. My immediate concern is with the killing of another human being, and with applying my answer to cases of voluntary and non-voluntary euthanasia, but light will be shed on whether and, if so, why it is wrong to kill oneself, to kill unborn human beings, and to kill non-human animals especially those whose life has most in common with the life of human beings.

Gilbert Harman and John Doris (among others) have maintained that experimental studies of human behaviour give good grounds for denying the very existence of moral character. This research, according to Harman and Doris, shows human behaviour to be dependent not on character but mainly on one's ‘situation.’ My paper develops a number of criticisms of this view, among them that social science experiments are ill-suited to study character, insofar as they do not estimate the role of character in continuously shaping the direction of one's life—including what situations one is apt to get into in the first place.

In the post-Gilligan debate about the differences, if any, between the ways in which people of different genders see the moral world in which they live, I detect two assumptions. These can be found in Gilligan's early work, and have infected the thought of others. The first, perhaps surprisingly, is Kohlberg's Kantian account of one moral perspective, the one more easily or more naturally operated by men and which has come to be called the justice perspective. (What I mean by calling this Kantian will emerge shortly.) This is the perspective whose claims Gilligan initially found suspect, not because she thought it a distorted account of the way in which male subjects operated, but because she disputed its claims to be the only account or the best or dominant one. Throughout the ensuing debate Kohlberg's account has been left in place, and challenged not for correctness but only for uniqueness.

In one of the Theological Tractates, Boethius wrote ‘ we have found the definition of Person, viz: “The individual substance of a rational nature”’. He justifies the definition partly by a consideration of Latin and Greek etymologies and partly by stating ‘what Person cannot be affirmed of’. Person cannot be affirmed of Universals, accidents, relations, lifeless bodies, living bodies without sense (e.g. trees), nor of ‘that which is bereft of mind and reason’.

Exchanging views in Philosophy with a two-year time-lag is getting rather like conversation with the Andromeda Nebula. I am distressed that my reply to Messrs Mackie and Dawkins, explaining what made me write so crossly about The Selfish Gene, has been so long delayed. Mr Mackie's sudden death in December 1981 adds a further dimension to this distress.

In Philosophy 51, October 1976, 471–472, Professor Tom Regan takes ud to task for our attack on Kant's theory concerning the moral status of animals. The ground of Regan's criticism is that ‘… it is clear that Kant does not suppose, as… Broadie and Pybus erroneously assume that he does, that the concept of maltreating an animal, on the one hand, and, on the other, the concept of using an animal as a means, are the same or logically equivalent concepts’ (p. 471). Regan argues that Kant does not say that we should avoid treating animals as a means. Rather, he claims, Kant's view is that we have an indirect duty not to maltreat animals, since in maltreating them we treat, or run the risk of treating, as a mere means rationality in ourselves or in others.

John McDowell has argued that for human needs to matter in practical deliberation, we must have already acquired the full range of character traits that are imparted by an ethical upbringing. Since our upbringings can diverge considerably, his argument makes trouble for any Aristotelian ethical naturalism that wants to support a single set of moral virtues. I argue here that there is a story to be told about the normal course of human life according to which it is no coincidence that there is agreement on the virtues. Because we are creatures who arrive at personhood only by learning from others in a relation of dependency, we cannot help but see ourselves as creatures for whom non-instrumental rationality is the norm. Those who train others in personhood must view the trainee's interests as having a value independent of their interests and must imbue the trainee with a sense of that value. Extending and preserving the sense of self-worth that we must acquire if we are to acquire personhood requires we see ourselves as creatures who need something like the virtues.

Nothing makes the occupation with the great minds of the past more attractive than the fact that with the change in the whole situation of the present time, with the maturing of one’s own personality, they appear in a new light and present themselves in rejuvenated shape. I had a curious experience of this kind, when it occurred to me during the investigation of some phenomenological problems, that Goethe, though ignorant of the name, had employed a definitely phenomenological method. In occupying myself now with the revealing of this fact, it will be my leading principle to understand Goethe through himself, and I shall try not to adapt his meanings to my own theories.

My contention is that virtue ethics offers an important critique of traditional philosophical conceptions of moral status as well as an alternative view of important moral issues held to depend on moral status. I argue that the scope of entities that deserve consideration depends on our conception of the demands of virtues like justice; which entities deserve consideration emerges from a moral view of a world shaped by that conception. The deepest disputes about moral status depend on conflicting conceptions of justice. I advocate a conception of the virtue of justice that can account for the cases that pose problems for the legalistic views of moral status and discuss what ideal moral debate looks like on this view.

It is commonplace to claim that it is good to believe the truth. In this paper, I reject that claim and argue that the considerations which might seem to support it in fact support a quite distinct though superficially similar claim, namely, that it is bad to believe the false. This claim is typically either ignored completely or lumped together with the previous claim, perhaps on the assumption that the two are equivalent, or at least that they stand or fall together. Such assumptions, I argue, are mistaken. While it is not always good to be right, it is always bad to be wrong. This is an interesting and overlooked asymmetry, which calls for further investigation.

Tom Phillips' painting for the dustjacket of the hardback edition of Metaphysics as a Guide to Morals depicts a faintly translucent, darkly-coloured, multi-layered lattice of letters, in which each character abuts directly upon others above, below and beside it, each overwrites or is overwritten by others of varying dimensions, but none is immediately decipherable as part of a word; and at the centre of this array is a geometrically precise, illuminated circle—perhaps emanating from a light located behind or under the layers of letters, perhaps from one directed at them from above. This image is open to many interpretations. It could represent the sun from Plato's myth of the cave shining through the dialogue in which he presents it (a myth that irradiates Iris Murdoch's text); it could also represent the light of Miss Murdoch's attention playing over the palimpsest of texts that make up the Western tradition of metaphysical thought. But for anyone encountering it upon closing the book after a first reading, it may also seem very precisely to crystallize one's initial impression of that text.

Sometimes we draw conclusions about a thing's existence or its value by considering what are thought to be its consequences. I shall say that an argument is pragmatic when it consists in estimating an action, or any event, or a rule, or whatever it may be, in terms of its favourable or unfavourable consequences; what happens in such cases is that all or part of the value of the consequences is transferred to whatever is regarded as causing or preventing them.

Does the scope of beliefs that people can form on the basis of perception remain fixed, or can it be amplified with learning? The answer to this question is important for our understanding of why and when we ought to trust experts, and also for assessing the plausibility of epistemic foundationalism. The empirical study of perceptual expertise suggests that experts can indeed enrich their perceptual experiences through learning. Yet this does not settle the epistemic status of their beliefs. One might hold that the background knowledge of experts is the cause of their enriched perceptual experience – what is known as cognitive permeation – and so their subsequent beliefs are only mediately justified because they are epistemically dependent on this background knowledge. I argue against this view. Perceptual expertise is not the result of cognitive permeation but is rather the result of perceptual learning, and perceptual learning does not involve cognition in a way that entails cognitive permeation. Perceptual expertise thus provides a means of widening the scope of the immediately justified beliefs that experts can form.

Respect, as Kant describes it, has a duality of nature that seems to embody a contradiction – i.e., it is both a moral motive and a feeling, where these are thought to be mutually exclusive. Most solutions involve eliminating one of the two natures, but unfortunately, this also destroys what is unique about respect. So instead, I question the non-cognitive theory of emotion giving rise to the contradiction. In its place, I develop the cognitive theory implicit in Kant's work, one in which emotions take the form of evaluative judgments that determine the will. I then show that, as a purely rational emotion, respect is perfectly suited to be a moral motive.

Many contemporary philosophers writing on punishment seek to show that much of the dispute between retributionists and utilitarians springs from a failure on the part of both parties to elucidate the concept of punishment. The writers are usually utilitarians who seek to show that what is true in the retributive theory is simply a point about the concept of punishment, and that for the rest, the morality of punishment is to be explained in terms of the utilitarian theory. Those who attempt to destroy the retributive theory by reducing its element of truth to a mere conceptual point about the concept of punishment, seek to argue that the notion of being guilty of an offence is part of the concept of punishment. Against this kind of approach, I wish to consider whether a general theory of punishment is possible, and if so, what are its basic concepts—punishment, deserving of punishment, deserved, or justifiable punishment—and how they are to be elucidated. I shall be concerned to argue that much contemporary writing on punishment commits the Platonic fallacy of assuming that there is a single, core, paradigm use of ‘punishment’ which is to be found by elucidating the concept of legal punishment.