Moral Scepticism and the Meaning of Moral Statements

I begin my somewhat exiguous²² consideration of the meaning of moral statements with an initial endorsement of the main views of John Mackie.²³ Mackie offers a projectivist moral theory which starts from an acceptance of the correctness of sentimentalist theories along the lines of Hume’s positive case and then offers an explication for the disparity between this sentimentalist reality and what he sees as most people’s objectivist beliefs about the status of their moral evaluations.

That moral language and ordinary thought do point towards objective moral values seems true enough to me too. Hume himself implicitly recognises this at the end of Book III i 1 when, after calling attention to the is/ought distinction, he finishes by commenting, “let us see, that the distinction of vice and virtue is not founded merely on the relations of objects, nor is perceiv’d by reason”, and this “wou’d subvert all the vulgar systems of morality”² —a clear statement that at least the great preponderance of mankind, whom Hume would classify as the vulgar,²⁶ believe, think and talk in terms of ‘real’, ‘true’, ‘objective’ moral values existing independently of the evaluator. Mackie is more explicit.

The ordinary user of moral language means to say something about whatever it is that he characterizes morally, for example a possible action, as it is in itself, or would be if it were realized, and not about, or even expressive of, his, or anyone else’s, attitude or relation to it. But the something he wants to say is not purely descriptive, certainly not inert, but something that involves a call for action or for the refraining from action, and one that is absolute, not contingent upon any desire or preference or policy or choice, his own or anyone else’s.²⁷

The question of people’s tendency to believe in and talk in terms of objective, mind-independent values (whether in fact values actually possess that status or not) is a straightforward empirical one.

The question naturally asked then is how to reconcile the claims sceptical or sentimentalist moral theories make about what is actually happening when people make moral evaluations with what people generally think, and language suggests, they are doing. Again I quote Mackie.

... we tend to project these sentiments onto the actions or characters that arouse them, or read some sort of image of these sentiments into them, so that we think of those actions and characters as possessing, objectively and intrinsically, certain distinctively moral features; but these features are fictitious.²⁸

In other words, the Humean sentiments and feelings which actually are the basis of moral distinctions are projected out on to the external world and mistaken for objective, mind-independent moral facts rather than subjective feelings. This is Mackie’s ‘Projectivist Error Theory.’

He reaches this conclusion after considering and rejecting other possibilities about meaning compatible with sentimentalism or subjectivism. This is necessary because subjectivism is indeterminate as far as the meaning to be attributed to moral statements is concerned. Accepting that reason alone cannot provide the means to make moral distinctions, that there are no mind-independent, objective moral values or primary qualities to be intuited or inferred, that the external, impartial vantage point is not a requirement of reason, and thus that the essential fact in discriminating between virtue and vice is that people have and share²⁹ certain feelings with respect to particular actions: accepting all this still leaves open a wide range of possibilities about the meaning of ethical terms.

The non-cognitivist school asserts that evaluative expressions do not make true or false statements. One branch of the school says that such moral sentences-in-use as ‘She is wicked’ express the feelings of the speaker. The other main branch says that such sentences-in- use prescribe or recommend. They are imperative and urge or dissuade action. But neither the ‘expressing’ emotivist branch nor the ‘prescribing’ prescriptivist branch of non-cognitivism says that moral sentences-in-use are at core propositional descriptions of the way someone feels. Rather moral language has at its heart, says the non- cognitivist, a non-propositional, non-descriptive meaning. It communicates a state or desired response which cannot be true or false.

Naturalist-descriptivist theories of the meaning of moral sentences-in-use, on the other hand, do take it that the phrase ‘X is wicked’ expresses a proposition that is true or false. The phrase indicates, or is a report of, a sentiment or feeling which happened to be aroused in certain natural conditions. After contemplating character X and the surrounding facts I make a report. At this point there is room for disagreement. Naturalist-descriptivist theories about the meaning of moral sentences-in-use might say the report is about my actual feelings, my usual feelings in such circumstances, what my feelings would be from an unbiased, impartial vantage, what the feelings of some group of others would be, or what the feelings of some group of others would be if they were all impartial. But despite these variations the moral sentence-in-use, in any of these cases, is a description that is, theoretically at least, verifiable. It is about the contingencies of relative sentiments in given conditions (as opposed to being about objective, real, mind-independent, external values).

To my mind neither non-cognitivism nor naturalist- descriptivism is credible as a theory of what most people mean when they utter a moral sentence. When A says, ‘B is wicked’, A does not normally mean no more than that in these circumstances Y, B aroused a feeling of disapprobation in her. Nor does A simply mean to express her reaction or command a response. People overwhelmingly tend to think of characters and actions as possessing intrinsic and objectively-valid features. B’s wickedness, it is generally believed, is not merely a subjective response in A, or some wider group of assessors, but something ‘real’ and mind-independent. Moral statements are not ordinarily meant as statements about the speaker.

Moreover, if the key to understanding morality and moral distinctions lay only in a conceptual analysis of ordinary moral thought and language, the case for an objectivist ethics would seem to me to be strong. This might partially account for its enduring appeal. But it must be insisted that what people say, even what they think, is no necessary revelation of what is actually going on. The case for moral realism or objectivism cannot rest on conceptual analysis alone. Conceptual analysis may clarify and elucidate; yet there is no reason to think what people mean reveals the factual basis of the matter ■— what is actually happening. Objective moral claims are not self­endorsing in fact, however rooted in language. When a person makes a claim X (X = ‘A is wicked’), and believes that his making claim X entails the truth of P (P = The property of wickedness exists in the world), it does not necessarily follow that P is true.

Now I believe that as regards the nature of evaluations themselves, Hume’s basic positive case is correct. What happens is that people experience sentiments and feelings in response to the actions and characters of humans. Morality, as a fact, is at core areasonable and based on emotions. The alternative to the Humean position of what is really going on must needs be some form of objectivism or realism perhaps manifesting itself in broadly intuitionist or Nagelian positions.

But one can, at the same time, note that ordinary speech and thought push one towards the objectivist camp. To remain outside that camp while holding a non-cognitivist or naturalist-descriptivist theory of the interpretation of evaluative language is not to be convincing. The meaning of the vast preponderance of moral sentences-in-use is not an expression, report or endorsement of the speaker’s, or anyone else’s, sentiments (whether as they are or as they would be if the subject were impartial). As Mackie said above, “The ordinary user of moral language means to say something about [an action or character] as it is in itself.... a call for action... that is absolute, not contingent upon any desire or preference or policy or choice, his own or anyone else’s.”³⁰

Mackie’s projectivist theory of the meaning of moral sentences-in-use accepts the prevalence of the tendency to objectify values while also accepting Hume’s moral scepticism or subjectivism about what is, in fact, the nature of evaluations. The seeming inconsistency is dissolved by accepting that ingrained assumptions of thought and language can be and in this case are mistaken. The conclusions of conceptual, linguistic analysis may be correct in asserting ‘S’ generally means that P, but nevertheless P is mistaken. What are taken for objective and absolute values are projections of subjective sentiments on to the external actions and characters that triggered the sentiments. The features of good and bad, right and wrong, virtuous and vicious, that are in thought and language ascribed to characters and actions, are in fact created by the projection of moral sentiments out on to the action or character which stimulated the sentiment.

Nor is Mackie’s projectivist theory limited to “a trick of individual psychology.”³¹ The whole system of communicated feelings and sentiments which are built up in a system of morality serves to amplify, diminish, and modify the sentiments of each person in the system. The perceived objective quality of moral distinctions furthers this process. The retort that these many feelings and sentiments of others are mind-independent and real the sceptic can readily concede. Nevertheless, admitting there be a whole system of interdependent, externally existing sentiments is not to admit that values themselves are mind-independent, unless one is prepared to equate values with the contingent sentiments and feelings that people happen to have. The moral sceptic does make that equation. In no other sense, though, are values for him mind-independent or objective. But I leave discussion of all this, and why and how this system of communicated feelings arises, for later chapters. Suffice it here to emphasise that the objectification or projectivist error theory makes a very good prima facie match with Hume’s positive substantive case. That being the case, let me look briefly at what little Hume actually hints at about the meaning of moral utterances to see if projectivism can be ascribed to him. Then I can consider two criticisms of Mackie’s error theory itself.

To be clear about Hume’s moral theory it is essential to separate the why and how, investigative questions about the actual nature of moral evaluations from all questions concerning the interpretation and meaning of moral language and ordinary moral thought.³² Hume’s concern is overwhelmingly with the former. His position, at its simplest, is that moral evaluations are sentiments, feelings or emotions in the assessor, felt on regarding a character or action from a certain perspective. The actual moral response, having an inescapable feeling³³ element, would fall under the rubric of a non- propositional, areasonable feeling.³⁴ But having made the distinction, and having stressed that Hume’s concern was with the substantive fact of moral evaluating, nevertheless efforts have been made to attribute a position to Hume regarding the meaning of moral statements. Mackie suggests³⁵ that Hume may simply have failed to distinguish descriptivist and non-cognitivist theories of meaning, largely, perhaps, because Hume was not sufficiently interested in such questions. Ardal hesitantly attributes a form of non-cognitivism³⁶ to Hume while stressing that, “what [the Treatise} contains is an account of the way the concepts of virtue and vice have their source in human emotions.”³⁷ The Treatise, “does not contain much in the way of... an analysis of moral discourse.”³⁸

I believe that attention to Hume’s concern with giving a factual account of moral evaluating and his lack of concern with analysing meaning is the proper background from which to read isolated passages³⁹ which have been taken to disclose Hume’s theory of meaning. For example:

Morality, therefore, is more properly felt than judg’d of.⁴⁰

To have the sense of virtue, is nothing but to feel a satisfaction of a particular kind from the contemplation of a character. The very feeling constitutes our praise or admiration.⁴¹

You never can find it [the vice], till you turn your reflexion into your own breast, and find a sentiment of disapprobation, which arises in you, towards this action.⁴²

So that when you pronounce any action or character to be vicious, you mean nothing, but that from the constitution of your nature you have a feeling or sentiment of blame from the contemplation of it. Vice and virtue, therefore, may be compar’d to sounds, colours, heat and cold, which, according to modern philosophy, are not qualities in objects, but perceptions in the mind.⁴³

All these passages can be read as dealing with substance not meaning — even the one in which I have emphasised the word ‘mean’. As Hume was largely blind to questions of meaning, a strict, literal reading of this passage to support a version of naturalist-descriptivism is unwarranted. Nor do I agree with Ardal’s irresolute opting for non­cognitivism. If asked straight out, “What do people mean by such utterances as ‘She is wicked’ or ‘X is virtuous’?”, Hume would not say that they mean nothing that is true or false. I think Hume was far too much the observer and student of human nature not to see that people do generally mean to assert a propositional statement about the external world.⁴⁴ Thus Ardal is also incorrect, in my view, when he concludes that, “Hume is not suggesting that ‘vice’ and ‘virtue’ are names for qualities that appear to belong to external objects (people), but that they really are in the mind of the observer.”⁴⁵ In fact, virtue does not really exist out in the world in some character or action. But ‘virtue’ does refer to some external quality in the inspiring character. The fact giving rise to an evaluation that an external character is virtuous is a sentiment in the assessor inspired by that external character. This appears to the assessor, though, as a real quality in some other person.

Mackie thinks that “there is at least circumstantial support for [the projectivist error theory] view as an interpretation of Hume.”⁴⁶ It explains Hume’s use of ‘moral sense’ in Book III i 2 and counts as an instance of his observation “that the mind has a great propensity to spread itself on external objects.”⁴⁷ Comparing the process of projecting our sentiments to the natural belief in causation which we attribute to the causal world itself, Mackie refers⁴⁸ to this passage in Appendix I to the Enquiry concerning the Principles of Morals·.

Thus the distinct boundaries and offices of reason and of taste are easily ascertained. The former conveys the knowledge of truth and falsehood: the latter gives the sentiment of beauty and deformity, vice and virtue. The one discovers objects as they really stand in nature, without addition or diminution: the other has a productive faculty, and gilding or staining all natural objects with the colours, borrowed from internal sentiment, raises in a manner a new creation.⁴⁹

It is at least plausible to urge that Hume would have accepted a projectivist error theory about the meaning of moral discourse. Such an imputation to Hume is the more attractive, naturally, the more persuasive one finds Mackie’s error theory. The foundations for that theory are, one will recall, that: a) moral evaluations in fact consist ineluctably in areasonable sentiments and feelings; b) moral statements are, by and large, treated as being objectively true and false, as having a non-subjective, mind-independent quality; c) moral statements voice a state of affairs that often impels action. Combining these three, in a way that retains the strengths of Hume’s areasonable moral scepticism or substantive subjectivism, while providing an explanation for the observably true-or-false status of ordinary moral language and thought, is the great strength of Mackie’s Projectivist Error Theory. I will now consider several critiques of that theory and then leave questions of meaning until I turn to law in Part B of this book. A fleshing out of Hume’s positive case, about what is really happening when people make moral evaluations, will then conclude this chapter.

Mackie’s Projectivist Error Theory rejects the objectivity of moral values in fact but admits that language and, for lack of a better expression, common-sense both purport to refer to objective, real values. The conclusion that language and ordinary ways of thinking about morality mislead us comes from an analysis of supposed objective values. How are they linked to natural features? What kind of faculty, call it reason or something else, can see or intuit these non-natural but purportedly real and objective qualities? Alternatively, is it plausible to imagine and believe in a purely normative realm outside the empirical causal world where values exist? Or again, what arguments can convince us of a logically necessary relation between values and humans? Or again still, how can values be tied to human nature and what-is-good-for-humans and yet somehow transcend the observed contingencies? The failure of objective ethics leads us to re-examine the beliefs which the structure of ethical thought and language foists upon us. Hume’s conclusion, that moral qualities actually result from subjective, areasonable responses to the natural features of actions and characters, although strictly unprovable, flows naturally from a rejection of objectivism. It is firmly anchored in the causal and empirical. For me, it is highly plausible. However Mackie’s further conclusion, that moral judgements, as ordinarily made, have implications which are false, is far from the only theory of interpretation which can co-exist with Hume’s substantive moral scepticism. It is attractive as a supplement to Hume because, unlike non-cognitivist and naturalist-descriptivist theories of meaning, it recognises that most people do not mean something relativist when they say ‘X is wrong and vicious.’ It is an empirical fact, I think, that most people, most of the time, do not mean by the above statement something like ‘I do not like X’, or ‘I do not approve of X’, or ‘Most people in my society do not approve of X’, or ‘I do not like X even if everyone else does’ or even that the above statement has no content beyond being a moral exclamation or exhortation. All such theories of meaning are implausible because people do mean something broadly objectivist.

Yet to accept that is not sufficient as Mackie is not alone in such acceptance. Roderick Firth suggests an analysis which he says⁵⁰ is a progeny of the theories of Adam Smith and David Hume. Firth calls his own analysis of moral statements an ‘absolutist dispositional’ theory. Moral statements are absolutist because they are in no way relative to the utterer (or his social group).⁵¹ Ethical statements do not have a meaning that can be replaced by “egocentric”⁵² statements nor are they merely emotive outbursts. Moral statements are dispositional because they assert how a certain being (real or hypothetical) is disposed to react (say, to an action or character). Firth then goes on to say that moral statements are statements about the reactions of an ‘ideal observer’, one who is omniscient of causal facts, all-feeling or omnipercipient, completely impartial or disinterested, consistent and dispassionate.

Firth’s theory has the advantage of capturing the objective or absolute tone of most moral speech and thought while retaining a Humean feeling-based core (albeit of an imaginary ideal observer). Firth himself notes of his ‘absolutist dispositional’ analysis that, “[t]his formulation may draw attention to the fact that a dispositional analysis which is absolutist may nevertheless be extensionally equivalent to one that is relativist.”⁵³ Moreover Firth is not forced to posit that our ordinary moral discourse is founded on an error. Rather, it is founded on the feelings and sentiments of an ideal observer.

1 do not find Firth’s ideal observer analysis of moral statements convincing. Moreover, 1 do not agree that it can be seen as a working out of a Humean moral theory (or at any rate of a theory of meaning which Hume, had he turned his mind to the issue, could have and would have defended). What Firth does, I believe, is to conflate the two issues, which 1 have urged be kept separate, of what is meant and of what in fact is happening. Hume concerns himself with the latter and argues that moral evaluations are sentiments in the assessor. To the extent that Firth is suggesting that the fact of the matter is a sentiment in an ideal observer⁵⁴ he is clearly diverging from Hume’s position. Of course it is evident why Firth should want to attribute ‘moral’ sentiments to an ideal observer — because ex hypothesi there is only one way that an ideal observer could feel in a given moral situation so that single ‘right’ or ‘correct’, in fact ‘absolute’, answers and responses would follow. But the substantive explanation of how and why humans actually make moral distinctions cannot be in terms of the reactions of a merely “possible being”,⁵⁵ a being essentially “irrelevant to the truth of [moral] statement^]”.⁵⁶ Despite Firth’s deliberate grounding of the ideal observer in the empirical⁵⁷ world of experienced psychological traits, a grounding which gives a Humean flavour to his analysis, that analysis still seems quite distinct in substance from one which attributes moral evaluations to actual sentiments in the actual evaluator.

But Firth says that his concern is with the “analysis of ethical statements”,⁵⁸ with “the logical or epistemological relationships between moral data and ethical statements.”⁵⁹ On this level, the level of meaning, one could theoretically accept Hume’s substantive position and yet argue that people do mean to set up the standard of an ideal observer when they say ‘X is wrong’ or make some other moral statement. This is theoretically possible but, as it happens, untrue. People do not generally mean this in my opinion. They mean that X is objectively or mind-independently wrong. They mean, for instance, that the institution of slavery is evil, not that someone — real or ideal — thinks so. What Firth does is to attempt a conceptual analysis of regular moral thinking and speech which can be meshed with Hume’s substantive case. As a possible reform of moral thought and language this may be admirable. It fails though, to my mind, as an explanation of actual, existing moral thought. Mackie’s merit is to recognise that substance and actual meaning do not mesh. But it is this meshing, or conflating, that makes the ideal observer theory so attractive. Given Hume’s position, would it not be better if language and thought communicated what an ideal observer would feel? I would say yes (and even hazard that Hume would too). I would even concede that such a meshing does not require any new moral vocabulary or language, just a change in the meaning or sense people ascribe to it. But that is a question of reform, one to which I and Mackie and Hume could assent while saying that the reality today is best described by Mackie’s Projectivist Error Theory.

The ability of Mackie’s theory of the meaning of moral statements to withstand alternatives such as Firth’s is buttressed by considerations I take from Nagel.⁶⁰ Now it is true that Nagel is concerned to argue for a real, objectivist ethics — not just of meaning but of fact and substance. It is also true that I have rejected that substantive conclusion as being implausible.⁶¹ But as a conceptual analysis of moral thought I think that Nagel is most instructive. Moral thought and moral language are objective in form (as opposed to the reality of what is actually happening which is not) because of a tendency humans have to see themselves in a particular way. As Nagel puts it, albeit in arguing for substantive objectivism, the objectifying tendency results from “... a certain feature of the agent’s metaphysical conception of himself.”⁶² For Nagel that conception is of “oneself as merely a person among others equally real”⁶³ (together with a view of one’s own identify over time). Such a conception of oneself does not, I have argued, force us to accept a substantive objectivism. But it does tend to make us speak and think as if there were objective, mind-independent values. It contributes to the mind’s “propensity to spread itself on external objects”⁶⁴ when we can take up the external standpoint that Nagel remarks.

So ideal observer analyses, in not recognising the full force of this tendency, are unconvincing as analyses of meaning. Mackie does recognise the objectifying tendency and is the more credible as a result.

Finally, 1 consider Simon Blackbum’s critique of the Projectivist Error Theory⁶⁵ and his quasi-realist alternative. Blackbum accepts wholeheartedly the projectivist half of the theory (i.e. the Humean/Mackie tenet that what is really happening is that individual sentiments are projected on to the world and ‘stain or gild’ that same world with features corresponding to these sentiments). About projectivism Blackbum comes to these conclusions:

[Projectivism can] accommodate the rich phenomena of moral life.⁶⁶

The nub of the matter, then, is that the projectivist provides explanation, making moralizing an intelligible human activity with its own propriety, and the opposition [objectivists] provides none, but gestures at an evidently lame analogy.⁶⁷

I myself think that there is precious little surprising left about morality: its meta-theory [projectivism] seems to me to be pretty well exhaustively understood. The difficulty is enabling people to appreciate it.⁶⁸

As a metaphysical view, projectivism explains what we are doing when we moralize.⁶⁹

I quote at length to show how fully Blackburn accepts the substantive case for moral scepticism (z.e. Hume’s positive moral theory and the projectivist half of Mackie’s theory). This may come to seem surprising because Blackburn completely rejects the second half of Mackie’s theory that the ordinary claims of moral statements are in error and mistaken. The crux of Blackburn’s criticism is that if the normal sense of moral discourse is false, why do we (and especially Mackie himself who is aware of the error) continue to use such a flawed vocabulary riddled with reifications? Surely it would be better if our moral vocabulary (or some sanitized replacement) were used differently. The fact that Mackie⁷⁰ goes on to use the ‘infected’ vocabulary himself in expressing all of his own views about what content a moral system should have, together with the inability to come up with a ‘hygienic’ set of concepts, condemns, according to Blackburn, the error theory. If Mackie were right we ‘should’, says Blackburn, avoid moral (z.e. erroneous) views and opt for a purged, metaphysically correct language which expresses not moral but shmoral¹' views. The silliness of trying to achieve this feat of reform casts the original diagnosis of error into doubt.⁷² All is not lost for projectivism though, according to Blackburn, because he offers a ‘quasi-realist’ diagnosis of morality which exposes the failings of objectivist theories, buttresses projectivism and is content to moralise rather than shmoralise.

I accept all of Blackbum’s attack on objectivist-realists⁷³ (including Nagel), as well as his recognition of projectivism as the best explanation of moralising. Blackburn rightly notes the unease caused by projectivism and the importance of causal explanation over ‘everything-being-what-it-is’ phenomenology. My objection to Blackburn is not one of substance; we are both moral sceptics. Rather I object to Blackburn’s attack on Mackie’s Error Theory, and to his own quasi-realist alternative. I discuss them in reverse order.

Quasi-realism is the view that ordinary ways of speaking do not in fact make claims to objectivity. What can sound like a second- order claim to objective or mind-independent status (e.g. ‘Even if we had approved of it or enjoyed it or desired to do it, bear-baiting would still have been wrong.’⁷⁴), is in fact just a first-order commitment to reduce the suffering of bears. Asserting ‘X is wrong even if everyone felt otherwise’, then, argues quasi-realism, is consistent with simply having a feeling about other hypothetical feelings, even one’s own. The assertion about bear-baiting’s wrongness need not be understood as making an objectivist, mind-independent claim. That is to say, Blackburn argues that seemingly objectivist ways of speaking can, quite consistently, be construed in non-objectivist ways. People can and do adopt such language; people can and do continue to moralise rather than shmoralise; there is no need to conduct practical reasoning in any different way. In essence, Blackburn’s programme “... is the quasi-realist identification of shmoralizing with moralizing.”⁷⁵ He thinks that “... pro^ectivism can accommodate the propositional grammar of ethics.”

My view of Blackburn’s quasi-realist alternative is that it simply misses the point. A realist-seeming grammar may well be explainable in sentimentalist, projectivist and subjectivist terms.⁷⁷ Indeed it may well even be that that discourse cannot, sensibly,⁷⁸ continue to be utilised by the non-believing Humes and Mackies of the world. But the point is surely not, as regards an attack on the accuracy and validity of the Error Theory, whether the old moral language can be retained. The issue is -whether traditional ways of thinking about and making moral distinctions involve claims of absoluteness and objectivity. As normal moral ways of speaking are consistent with both realism and quasi-realism, linguistic analysis cannot settle the matter. So here I adopt another favourite Humean technique. “I confess [that] I place my chief confidence in experience to prove so material a principle.”⁷⁹ Having looked inside myself I find that ordinary moral language and thought do involve claims to objectivity, to a certain mind-independent quality attaching to evaluations. 1 challenge the reader to do the same because my goal is an account of morality which convinces because it matches up with individuals’ experiences. And here, at the risk of repeating myself, I say that the evidence is overwhelmingly that an objective or absolute claim is involved in ordinary moral thought.⁸⁰

At this juncture the quasi-realist like Blackbum might well demur. ‘Do not look inside yourself; look at the use or the conceptual role of ethical ways of thought’, he might demand. And his demurrer would seem to be strengthened because that role can be explained by the moral sceptic, the projectivist, without recourse to an error theory. Ordinary usage is consistent with an assessor reporting her feelings about other hypothetical feelings, as I have already conceded. Unfortunately mere consistency is not enough, at least if the realist or mind-independent meaning is also consistent with the conceptual role of ethical ways of thought. One then has somehow to choose between the two. The difficulty in proving the Error Theory in any normal fashion then becomes obvious and can be illustrated as follows: Assume that my sceptical metaphysical position about what is in fact happening be correct. Assume as well that most people think and mean that there are mind-independent values. How could the latter assumption ever be shown to be true, given that quasi-realism is also consistent with the ordinary use of moral language, save by an appeal to what the asserter means and what he thinks others therefore mean? Conceptual analysis alone will not settle the dispute. Consequently my tack is to argue that the evidence from introspection and from the observable views and comments of others, to say nothing of the close causal connection between religion and morality, indicates that traditional ways of thinking about and making moral distinctions do involve claims of absoluteness, objectivity and mind-independence.⁸¹

So in an important sense the compatibility of projectivism and our ordinary moral ways of talking is irrelevant to what we think. Blackburn, like Firth, is to my mind best seen as offering a road to travel after convincing people of the fact of projectivism. His is the path of reform. It is just that reform involves no changes whatsoever to the existing moral language (although it will involve a revised view of the metaphysics of moralising for most people). Quasi-realism itself, though, is a terribly unconvincing theory of the actual meaning of moral language.

This lack of credibility is compounded moreover when Blackbum permits his ‘quasi-realist’ theory — a theory which after all rejects the existence of mind-independent values and says that values ultimately reduce to subjective feelings, albeit these feelings may be about other feelings — to metamorphose into what, to me at least, has the air of an objectivist position. For example, Blackburn says:

[I]f everyone comes to think of it as permissible to maltreat animals, this does nothing at all to make it permissible: it just means that everyone has deteriorated.⁸²

[I]t is not my sentiments that make bear-baiting wrong; it is not because we disapprove of it that mindless violence is abominable; it is preferable that the world should be a beautiful place even after all consciousness ot it ceases.

[I]t is not our enjoyments or approvals which you should look to in discovering whether bear-baiting is wrong (it is at least mainly the effect on the bear).⁸⁴

But there is a radical difference between seeking a standard of sentiment which is disinterested and delicate on the one hand, and seeming to endorse the position that there are right answers independent of peoples’ subjective sentiments on the other. For one, the former recognises moral distinctions as ineluctably tied to people and that differing feelings and characters will lead to some unresolvable disputes with no right answer. For another, Blackbum nowhere tells us from where his non-subjective ‘shoulds’ and ‘oughts’ derive or come. They seem to arrive from nowhere. Where and how formulated are the standards which allow him to point out a “defective sensibility”?⁸⁵ Indeed, if everyone felt maltreatment to be permissible then no one could possibly have any first-order commitment, any feeling to report, and so the claim would seem necessarily to be a second-order one to objective status. In addition, as I shall argue below, there are no logical criteria for Blackburn’s assertion that Mackie ‘should’ have to change to shmoralising.⁸⁶ Hence, in the absence of such standards and how they can be arrived at, such claims as Blackbum’s can appear to be a latent form of objectivism.

Thus I do not find quasi-realism persuasive save as a possible route of reform. There is an interesting attempt to make compatible the fact of metaphysical sentiments being projected on to the world with our ordinary moral speech. But the meaning of speech is ultimately a factual question, dependent on common experience, not dependent on entailment, best consequences or suitability. Where there are different possible uses of the same ordinary speech there is no logical implication to the beliefs of ordinary people; rather, determining ordinary beliefs has to rely on the ‘experimental’ method. It has to be based on observation and fact, not analysis or logic. That said however, there still remains Blackbum’s independent criticism of the alleged inconsistency of holding the Error Theory without going on to shmoralise.

This is a troublesome criticism. My response is, firstly, to note that Blackbum’s attack on Mackie’s continued use of moral language has no logical bearing on the success of Blackbum’s own quasi-realism which I have rejected above. The two are logically independent so that if Blackburn were to allege of Mackie:

(1) You say P implies Q. (Ordinary thought and language implies realism.)

(2) You say Q is false.

(3) Therefore you are committed, perhaps even logically committed, to rejecting P.

(4) But you continue to assert P. (To use language which presupposes the propositions implying realism.)

(5) I also say Q is false and offer quasi-Q as being true.

Quite clearly (1)— (4) (the attack on Mackie) is independent of (5) (Blackbum’s quasi-realist alternative). Moreover the attack on Mackie is not, strictly speaking, an argument against the Error Theory. Perhaps the Error Theory is true and we should go on to shmoralise. This is certainly a possibility but one not taken up by Mackie. Is Blackburn correct when he alleges that by continuing to moralise Mackie is being inconsistent, perhaps even logically inconsistent?

My defence of Mackie centres on premiss (1) above. Assume

(A) that the normal way of thinking and talking about moral evaluations does involve claims to objectivity. Assume also (B) that those claims are false. Assume even more (C) that Mackie knows those claims to be false and yet continues to talk in the same way. The arguments Mackie offers in support of assumptions (A) and (B) are empirically-based arguments. Based on factual experience and observation (i.e. inductive, empirical arguments), he asserts (A) and

(B). However there is nowhere a claim (A,) by Mackie, or me, that existing language implies or involves or entails, in some necessary way, belief in objective values. It is rather a contingent, empirical claim that could have been otherwise. The existing language could be used in other ways to make claims about, say, committments to first­order positions (per Blackburn) or about the sentiments of an ideal observer (per Firth), but as it happens our language ordinarily is not so used. So premiss (1) above subtly misrepresents Mackie’s position. If language can possibly be used with different meanings then the relation between P and Q in premiss (1) is merely contingent, however uniformly found in practice. And thus even if (3) follows from (1) and (2), it does not follow from a properly re-stated first premiss.

The central question is whether our existing language is open to any other conceivable meanings or whether there is only one demonstratively possible meaning. If it be the former then Mackie can choose whether to create a new language (shmoralise) or whether to continue to use the old language (moralise) with explicit caveats that for him it cannot mean there are objective, mind-independent values.⁸⁷ That is, the possibility exists for Mackie to make it clear that the existing language of morality is being used with a different meaning, one denuded of all objectivist associations. So the choice to shmoralise is no more logical than the choice to moralise (although shmoralising may, as a factual matter, lead more people to see the falsity of claims to objectivity⁸⁸). Indeed, imagine Mackie addressing the John Mackie Fan Club of Committed Projectivist Error Theory Believers and in the course of his talk saying ‘X is wrong’. I see no impossibility in thinking that everyone there would understand that no objective, mind-independent claims about X and wrongness were being made. Mackie’s moralising would in effect be shmoralising. It is just that the ethical language would quite properly sound the same as that used by others.

One consideration Mackie must face though, in deciding whether to continue to use the language of moralising, or instead to switch to shmoralising, is the very practical one of the likelihood of changing ordinary linguistic usage. What are his actual options? He can continue to use the standard moral language (making it clear that objective values do not exist) or he can switch to some bizarre, perhaps incomprehensible shmoral vocabulary. The desire to be read and considered may well point to a continuation of moralising.⁸⁹

Another, more fundamental, consideration for Mackie is whether refraining from moralising in favour of schmoralising would be consequentially wise. Would social cohesion and civilized behaviour be weakened if morality’s ‘objectified’ overlay were stripped away? Are humans who believe certain conduct has a mind­independent rightness about it better able to constrain their conduct than those who do not so believe? Mackie hedges on this question⁹⁰ although Goldsworthy thinks that implicitly Mackie’s “argument seems almost to entail that [it would be dangerous to expose the fiction].”⁹¹ If Goldsworthy’s reading be correct one might well wonder, with Hume, at the point of exposing such a useful fiction.

And though the philosophical truth of any proposition by no means depends on its tendency to promote the interests of society; yet a man has but a bad grace, who delivers a theory, however true, which, he must confess, leads to a practice dangerous and pernicious. Why rake into those comers of nature which spread a nuisance all around? Why dig up the pestilence from the pit in which it is buried? The ingenuity of your researches may be admired, but your systems will be detested; and mankind will agree, if they cannot refute them, to sink them, at least, in eternal silence and oblivion.⁹²

But I doubt that the belief in objective, mind-independent values, a fiction if Mackie and I be correct, is as useful as Goldsworthy suggests. Hume certainly saw little, if any, tendency on the part of religion to promote the interests of society. Dispensing with the belief in objective values could well prove no more traumatic. Nor does Blackburn think that enabling people to appreciate the truth of projectivism is a cause for worry. ³ Certainly in regard to promoting such values as tolerance the ‘objectified’ overlay appears to be more hindrance than help, as I argue in the concluding remarks to this book.

As for Mackie’s substantive argument, surely it is aimed at altering awareness, at disabusing the widespread — but false — faith in objective, mind-independent values. Any hedging must be seen in that light, as an afterthought. The fact of there being no objective, mind­independent values does not mean that humans living together in groups with particular and general passions, weaknesses and proclivities require no values. Mackie, and indeed Hume, think that values are needed and some are better than others. Such values exist within a public system of constraints on action. This is the social function of values. How those particular values which happen to be defended by Mackie are expressed — be it conventionally by moralising or unconventionally by shmoralising — does not condemn Mackie’s Projectivist Error Theory.

This brings to an end what I wish to say about the meaning and interpretation of evaluative moral discourse. I am inclined to accept an error theory of meaning as an ancillary to the fact of what is actually taking place when people make moral evaluations. Such a theory bridges the main gap between a sentiment-centred moral theory and common understanding by explicitly holding that a theory like Hume’s does not capture the ordinary sense of moral statements and beliefs. But it is not the use of ethical language in itself that introduces this objectivist error because that language is compatible or consistent with various possible meanings. Hence I accept the error theory only after having considered alternatives to and criticisms of that theory. Ultimately, other theories which accept the fact of sentimentalism must be seen as reform theories of meaning, of what we ought to mean. Lastly, the error theory is a position which can be imputed, fairly 1 think, to Hume himself. At any rate, it is a theory of meaning which fits well with his substantive metaphysical theory. For my own part, 1 find the error theory attractive and plausible.

c)