In the beginning was A.J. Ayer

The creation of moral philosophy as we know it: in the beginning was A.J. Ayer, and moral assertions were without form, and void. More precisely, they were of a grammatically misleading form and lacking in meaning. In Language, Truth and Logic (published in 1936), Ayer maintained that what appears to be a moral assertion (e.g. ‘Stealing money is wrong’) ‘expresses no proposition which can be either true or false’. Such sentences are unverifiable and hence meaningless, for the central tenet of his logical positivism was that the meaning of a statement is equivalent to the observations that would make it true.

What is really happening, then, when someone makes what appears to be a moral assertion? Ayer proposed that such an utterance should be analyzed as a kind of exclamation – an expression of (rather than a statement about the presence of) a feeling, like ‘Phew!’ (as against ‘I feel hot’) – or as a command, like ‘Don’t do that!’ These analyses are consistent with the claim that what appear to be moral assertions have ‘no objective validity whatsoever’. For as ‘pure expressions of feeling’, they ‘do not come under the category of truth and falsehood. They are unverifiable for the same reason as a cry of pain or a word of command is unverifiable – because they do not express genuine propositions.’ This account was refined in various ways in the following decades, but there still seemed little room for moral deliberation or for rational argument about moral issues. It was hardly surprising, then, that philosophers (at any rate in their professional capacity) tended to discuss what people were doing in saying something was right or wrong rather than talking themselves about what was right or wrong.

Logical positivism was never actually refuted but it was abandoned as what has come to be called a ‘degenerate research programme’. In moral philosophy this abandonment took the form, by and large, of turning away not only from the logical positivists’ answers but from their questions too. Instead of proposing alternative accounts of what moral assertions were, the generation that started writing in the Sixties and Seventies addressed live moral problems, in the conviction that, whatever moral judgments might ultimately turn out to be, they had enough complexity of structure to challenge the best philosophical intellects, and that whatever philosophers could contribute to their elucidation would be worthwhile.

The past few years have seen developments in two opposite directions. On one side, the movement towards ‘applied ethics’ has taken an ever more concrete form. Books have appeared on the ethics of social work, nursing and engineering; there are now internships for young philosophers in a variety of institutions (including the US Congress); and a few resourceful PhDs have set up as beeper-equipped ethical consultants. On the other side, there has been a renewal of interest in the fundamental questions about the nature and function of morality – sometimes called meta-ethics – that had been shelved by common consent. There is of course no incompatibility between the two, but ideally there should be a tension, so that applied ethics never quite becomes an extension of ‘advice to the lovelorn’ and meta-ethics is pulled away from abstract word-spinning.

Both Wong’s and Fishkin’s books belong to the new wave, in that they are primarily concerned to argue general theses about morality, rather than to argue for particular substantive moral conclusions. But the authors understand that it really makes a difference for the way in which we should think about ourselves and our relations to other people whether what they say is true or not, and there is a note of urgency in both that is most appealing irrespective of one’s agreement or disagreement with the content. Perhaps it is the fact that the authors are still relatively early on in their academic careers that accounts for their refreshing belief that philosophy matters.

It will be best to begin with Wong, because he explicitly takes up the logical positivists’ questions about the meaning and truth conditions of moral assertions but gives different answers from theirs. Wong’s book is structured around six propositions, slated on the first page, that together make up, he believes, the maximum claim for the ‘objectivity’ of morality. These are as follows:

1. Moral statements have truth values;

2. There are good and bad arguments for the moral positions people take;

3. Non-moral facts (states of affairs that obtain in the world and that can be described without use of moral terms such as ‘ought’, ‘good’ and ‘right’) are relevant to the assessment of the truth value of moral statements;

4. There are moral facts (that may or may not be claimed to be reducible in some way to non-moral facts);

5. When two moral statements conflict as recommendations to action, only one statement can be true;

6. There is a single true morality.

7. ‘When morality is called subjective,’ Wong says, ‘several and perhaps all of the above claims are denied.’ (Freddie Ayer rejected all of them in 1936 and I think that Sir Alfred would still reject them now.) Since some people call themselves objectivists on the strength of accepting only some of the propositions while others call themselves subjectivists on the strength of denying only some of them, Wong proposes a new distinction: those who reject the sixth proposition he dubs ‘relativists’ and those who accept it ‘absolutists’.

Wong himself suggests that relativism, so defined, best corresponds to our ‘moral experience’, and in the event denies not just the sixth proposition but the fifth as well, while claiming that he can accommodate the correctness of the first four. The key term in his account is ‘adequate moral system’, which he claims to be ‘a more explicit rendering of what people have in mind when they use terms such as “the right moral rules” ’. If we take as a canonical moral statement ‘A ought to do X,’ Wong suggests that we construe it as follows: ‘By not doing X under actual conditions C, A will be breaking a rule of an adequate moral system applying to him or her.’

Wong maintains that this account makes the first four of his propositions correct and the last two incorrect. ‘With respect to 4, we can say that moral facts are reducible to non-moral facts about agents, rules and empirical conditions.’ A moral assertion is true (1) if it corresponds to a moral fact – e.g. that A is in fact required by a rule in an adequate moral system to do X. There can be good or bad arguments for moral positions (2) because people may draw correct or false inferences for particular cases from the adequate moral system to which they subscribe. They may also go wrong in drawing implications as a result of false beliefs about e.g. economics or psychology (3). But at the same time one person can truly say of an act on a particular occasion that it ought to be done and another that it ought not (5). The first statement may correspond to the moral fact that the act is required in an adequate moral system and the second to the moral fact that it is prohibited in a (different) adequate moral system. This, of course, presupposes (6) that there can be more than one adequate moral system.

The joker in all this is, obviously, the notion of an adequate moral system. In the game as played by Wong it functions as a wild card, taking different values in different contexts. The initial characterisation as a ‘modest idealisation’ of what people mean by ‘the right moral rules’ suggests a collection of rather concrete guides to action accepted as authoritative by the members of a group or society. But when trying to show how moral statements trade within an adequate moral system might be false. Wong seems to equate an adequate moral system with the highest-level moral maxim to which people are prepared to subscribe, such as the Kantian idea of treating people as ends in themselves or the Christian idea of the equal worth of all human beings before God – even if these contradict many more concrete beliefs. On the basis of the latter, for example, he claims that it was simply a mistake, resting on poor inference or distortion of the (non-moral) facts about human beings, for Christians to believe that it was permissible to own slaves. There is also an intermediate interpretation that gets some play, on which the adequate moral system for a group is a sort of distillation of the moral rules and precepts of the group into a coherent set of principles designed to be consistent with as many of those rules and precepts as possible.

If I have understood Wong’s theory of reference correctly, I think his reply would be that it is up to the people involved to fix the referent of ‘adequate moral system’, so it could take any of these forms or perhaps others. But the trouble is that it is purely a deduction from Wong’s theory that people have any notion of an adequate moral system in their heads when they make moral assertions. The alternative view is that people need have no generic conception of morality when they use words such as ‘ought’ or ‘wrong’. It is for philosophers to worry about the boundaries between morality and etiquette or between moral deficiencies and other failings of character or intellect. It is not necessary to have formulated an answer to such questions to engage in what is recognisably moral discourse. One can talk morality without a concept of morality just as one can talk prose without a concept of prose. They are both skills that most of the members of any society acquire in the process of upbringing without first (or ever) having to learn the theory.

Suppose, however, that we grant Wong his analysis of ‘A ought to do X’ in terms of a rule in an adequate moral system applying to a case in hand. There is still a fatal flaw in his claim to have rendered the first four of his propositions true while making the last two false. My argument takes the form of a dilemma: if ‘adequacy’ is interpreted in a way that makes moral facts reducible to non-moral facts and thus gives them a truth value, there cannot be contradictory moral assertions about one and the same act that are both true; conversely, if there can be, then moral assertions have no truth value.

To make the application of a rule in an adequate moral system a matter of ‘non-moral facts about agents, rules and empirical conditions’, as Wong says it is, we are going to have to play down the suggestion of independence and external appraisal in the word ‘adequate’, so that it means something like ‘accepted’ or ‘operative’. This would give us a theory of morality akin to H.L.A. Hart’s theory of law, set out in The Concept of Law, according to which establishing the existence of a legal obligation on a certain person to do something entails first of all identifying a valid legal system in that person’s society and then ascertaining how the rules of the system bear on the case in hand. We can discover the existence of a valid legal system by collecting information about the actions and beliefs of the members of a society in their various capacities as legislators, judges, citizens and so on. About the application of the rules of the system to a particular case, there can be legitimate dispute. But it would be absurd to deny that somebody has a legal obligation in England on the ground that he would not have one under French law; and it would betray confusion to deny that a legal system was valid because one thought that many of the laws in it were bad.

On a parallel analysis of morality, two adherents of the same adequate moral system could disagree about its implications in a particular case, but this would be a real disagreement and it would not be possible for contradictory assertions to be true. Wong, I believe, accepts that: he certainly should do. The key to his rejection of the fifth proposition (that ‘when two moral statements conflict as recommendations to action, only one statement can be true’) is that the different moral statements may be true in different adequate moral systems. But on the analysis that makes the existence of an adequate moral system a question of fact, the only relevant moral system is the actually operative one. To say that somebody ought to do a certain thing because he would have to under some moral system operative elsewhere or under some reform of the actually operative one would simply be to make a false assertion. Thus the fifth proposition becomes true. Whether or not the sixth one does depends on the way in which we understand it. There are different moralities, but there is a universal moral proposition that people ought to do what the operative moral system of their society requires them to. Either way, it seems to me an embarrassment to any theory that it should entail the underwriting of every society’s morality – however barbarous.

To avoid this conclusion and to deny the fifth proposition, Wong needs to play up the idea that the adequacy of a moral system is contestable, and this he does by calling adequacy an ‘ideal’ or ‘standard’ of morality. But this alternative brings complementary difficulties with it. Suppose that you, on the basis of your ideal of morality, say that A ought to do X, and I, on the basis of my different ideal of morality, say that A ought not to do X. Assuming Wong’s analysis of what ‘A ought to do X’ means, then surely we are committed to disagreeing about what constitutes an adequate moral system. If we both have standards or ideals for an adequate moral system and they are different, I must either give up mine and embrace yours or insist that yours is a substandard or non-ideal version. Wong himself argues forcefully that there is no single true moral system. (The rejection of proposition 6 is at the core of the book.) And certainly it seems most implausible that any collection of non-moral facts could constitute the moral fact corresponding to the truth of a particular ideal or standard of morality. But since the analysis of ‘A ought to do X’ incorporates an essential reference to an adequate moral system, this interpretation of adequacy entails that no such proposition can have a truth value.

The only way out of this impasse would be to claim the following: when I assert that A ought to do X all I mean is that it follows from my moral ideal, and when you assert that A ought not to do X all you mean is that that follows from your moral ideal. On this account, both assertions have truth values (they do or do not follow from our respective moral ideals) and both can be true at once while making opposite prescriptions (if they both do indeed follow from our respective moral ideals), Q.E.D. But although this analysis formally satisfies Wong’s conditions, it does so in a way that undermines his claim to have steered a path between the extremes of objectivity and subjectivity.

Moral discourse becomes, on this account, a mere exchange of autobiographical information. What makes a moral assertion true is not anything about what it apparently refers to (the actions of people and their effects on human welfare, for example) but simply something about the state of the speaker. Truth is indistinguishable from sincerity. But the rankest subjectivist can happily concede that, if this is what moral assertions mean, they can have truth values. Indeed, Ayer accepted in Language, Truth and Logic that such a theory (which he called ‘orthodox subjectivism’) would be consistent with the commitments of logical positivism. He simply said he thought it was more accurate to hold that moral sentences express attitudes than to hold that they actually assert the existence of those attitudes. It is clear that Wong would not be prepared to accept either horn of the dilemma with which I have presented him, and that he does think he has found a middle ground between objectivism and subjectivism. But I do not believe that he has.

Fishkin shares with Wong the project of staking out this middle ground but he approaches it from the opposite direction. For where Wong’s Moral Relativity stresses the rejection of complete objectivism, Fishkin’s Beyond Subjective Morality emphasizes his desire to transcend complete subjectivism. In this, each is reacting to the dominant tendency in his sources. Wong’s starting-point is the work of other philosophers, among whom relativism has had a generally bad press. Fishkin’s starting point is the beliefs of people he calls ‘ordinary moral reasoners’. Among those he interviewed, he was struck by the prevalence of subjectivism, and the guiding thread of his book is the attempt to see what might be said to the various kinds of subjectivist he unearthed to lead them in a more objectivist direction.

Fishkin’s ‘ordinary moral reasoners’ are not exactly a random sample of the public: they were in fact fellow students or recent graduates of the two universities at which he did his own postgraduate work, Yale and Cambridge. (Their ordinariness consists in their not being philosophers.) But the book is not put forward as a sociological treatise, and Fishkin does not present any statistics on the frequency of different views or offer correlations between the moral views of his subjects and their other attributes. The interview material is used to illustrate the moral positions discussed, and I think that it will ring true to anyone who has taught an introductory ethics course in Britain or the USA.

Fishkin distinguishes four subjectivist positions: amoralism (‘It’s simply a question of what I want’), personalism (‘I can only say what’s right for me to do – I can’t judge for others’), relativism (‘I should do what’s right from my perspective, others should do what’s right from theirs’) and subjective universalism (‘My decision is arbitrary but I’m still prepared to say that everyone else should do the same in similar circumstances’). As Fishkin says, the last of these views is unstable, since it is hard to see how, if one’s own moral opinions cannot be supported in terms that claim the allegiance of others, it can be legitimate to impose them on others. This drives one from universalism to relativism, but relativism has the unpalatable implication (noticed by one of Fishkin’s subjects) that Hitler acted rightly so long as he believed in what he was doing. Hence, personalism and amoralism seem to be left holding the field, one denying that moral judgments can be made of other people, the other denying that they can be made at all.

The idea that shapes Fishkin’s book is that many people do not embrace subjectivism with enthusiasm but feel bound to adopt it because they cannot see any intellectually honest alternative. What he can show convincingly is that many of his interviewees made stringent demands on objective morality and defended subjectivism by claiming that their conditions could not be met. Fishkin lists six expectations one or more of which his respondents thought objective morality must satisfy. An objectively valid moral position must 1. have a rationally unquestionable basis, 2. consist in principles that hold without exception, 3. give an answer to any possible moral question, 4. abstract from all human interests and concerns, 5. be consistent with conscientious moral decisions, and 6. determine obligations with strict impartiality.

According to Fishkin, Kant is an example of a philosopher who actually claimed to have met all the demands posed by his ordinary moral reasoners. Kant, he suggests, ‘has had a great influence on our common culture’, though he also points out that Kant himself claimed that his starting-point was ‘ordinary rational knowledge of morality’. But where Kant could represent the categorical imperative as putting a philosophical foundation under the beliefs of ordinary pietistic Germans such as those he grew up among, Fishkin’s British and American contemporaries have retained the Kantian idea of what an objective moral system should look like, but have concluded from the impossibility of realising such a system that subjectivism is the only feasible position to take. Thus, Fishkin says, ‘subjectivism can be interpreted as the residue left over from a failed Kantianism.’

Now, while it is true that Kant held each of the six propositions about morality in some form or another, he did not hold most of them in the form Fishkin’s subjects exhibit. They are so rigorist as to make even Kant appear lax. Their idea of an exceptionless system of moral rules is apparently a set of copy-book maxims of the form ‘Never under any circumstances do so-and-so.’ Obviously, any system with more than one such exceptionless principle in it courts self-contradiction. But the notion of an exceptionless moral system as employed by philosophers is that the principles should not conflict – and to avoid that they will normally have to come equipped with ‘unless’ clauses. (It is true that Kant thought there was an exceptionless rule against lying, but that was a personal foible and not inherent in his system.) Again, what Fishkin’s subjects made of the requirement of impartiality was apparently that an objective morality should deny the relevance of all special obligations to family, associates, country and so on. But this, although a conceivable interpretation of impartiality, was not Kant’s and has not appealed to many other philosophers who have made impartiality the hallmark of morality.

Fishkin argues that his ordinary moral reasoners are correct in believing that the six criteria of an objective morality cannot be met. Given the dizzying expectations that his subjects packed into those criteria, he is right about this. But rationalistic moral philosophers sympathetic to Kant would, I think, claim only that some weaker version of the criteria can be defended. Fishkin himself makes a move in this direction, for his strategy is to suggest that what he calls ‘minimal objectivism’ might be constructed out of weakened versions of the first and last criteria. Instead of the kind of logically watertight demonstration from self-evident axioms that his ordinary moral reasoners demanded, we would settle for claims that are capable of rational support or defence. And impartiality would now enter, not in the strong form assumed by Fishkin’s subjects, but as defining an appropriate ‘moral point of view’. From this point of view one might endorse special obligations, as long as they were suitably generalised. Thus, one could say, ‘Everyone should defend his own country,’ but not ‘Everyone should defend my country.’

Fishkin’s claim is that the notion of ‘reasonable choice from an appropriately impartial moral perspective’ leaves us with a possibility of providing morality with a basis that is not entirely subjective. ‘If we attribute such a basis to at least some of our moral judgments, then we are claiming that they are more than arbitrary personal tastes. We are claiming that others, were they to consider the problem in a reasonable manner from what we regard as the appropriate perspective, would come to the same conclusions.’ This does on the surface sound like a claim for the objectivity of morality. But trouble lies in the phrase ‘what we regard as’, for Fishkin does not think that there is any one way of operationalising the notion of impartiality that can be argued for rationally as superior to any alternatives. Some crackpot interpretations of impartiality might be thrown out, but a number of respectable contenders would still be left in the field with no way of making a case for one over another. John Rawls’s conception in A Theory of Justice of a ‘veil of ignorance’, blotting out all information that is (according to Rawls) morally irrelevant, would be one construction, but an inherently controversial one. The upshot is that rational argument comes in at the first stage, to get us to the identification of morality with some kind of impartiality (though Fishkin does not himself undertake to show the connection), and at the third stage, to get us from a particular conception of impartiality (perhaps embodied in some hypothetical procedure) to specific moral conclusions. But the second stage, at which we move from impartiality in general to a particular conception of its requirements, is apparently outside the realm of rational argument.

If, as Fishkin says, the alternative conceptions of impartiality that are in the running will yield widely different moral implications, it is hard to see how the position he is canvassing amounts even to minimal objectivity. If I do not believe that I can offer any arguments that would give you any reason for dropping a conception of impartiality from which you derive moral conclusions diametrically opposed at various points to mine, I surely have no more than a variant on the position that Fishkin called ‘subjective universalism’. Fortunately, there is reason for thinking that the situation is not as dire as one would gather from Fishkin. The recent history of moral philosophy seems to me to show us that it is possible to score palpable hits on alternative conceptions of impartiality. There is no reason to suppose that the arguments ever need come up against a blank wall of pure assertion.

As it stands, Fishkin’s idea of minimal objectivism is more an agenda item than a thesis. But he and Wong have both, it seems to me, identified the problem correctly: to show how serious moral arguments are possible and sometimes decisive, while doing justice to our experience of the diversity of morals and the apparent interminability of fundamental moral disagreements.