ROBERT L. ARRINGTON
Wittgenstein wrote very little about ethics. There is, to be sure, his early “A Lecture on Ethics,” which consists of an application of the philosophy of language in his first masterpiece, Tractatus Logico‐Philosophicus, to ethical concepts and judgments. Then there are occasional comments, such as some of those edited as Culture and Value, which appear to have ethical implications. Wittgenstein did not write at length about ethics in his later philosophy either, to be sure. Nevertheless, the later philosophy of Philosophical Investigations and other works exerted an enormous influence on ethical thinkers, resulting in a number of treatises that speak directly to ideas central to the later philosophy. This influence is especially felt in the emotivism and prescriptivism of mid‐twentieth‐century ethical theory, although at points the influence of Wittgenstein’s earlier philosophizing and his philosophy of language is apparent as well in these “noncognitivist” theories. Furthermore, there are ideas from Philosophical Investigations that have been used by some philosophers to develop distinctively “cognitivist” accounts of moral discourse.
We shall begin here by looking at the thought expressed in “A Lecture on Ethics,” and then we shall consider the indirect applications of Wittgenstein’s later philosophy to ethics.
By invitation from C.K. Ogden, Wittgenstein presented his lecture on ethics in 1929 to the Heretics Society in Cambridge. After attributing any difficulties in communicating his ideas to the fact that English is not his native tongue, Wittgenstein proceeds to communicate, in remarkably clear fashion and beautiful English, his thoughts on a subject of “general importance.” He declines to give a “popular scientific” lecture on the subject because doing so would “make you believe that you understand a thing which actually you don’t understand” (LE 4). Such a popular lecture, he says, would gratify “what I believe to be one of the lowest desires of modern people, namely the superficial curiosity about the latest discoveries of science” (LE 4).
“What is good?” G.E. Moore had asked in Principia Ethica (1903). Wittgenstein notes that we can get an idea of what Moore is asking about by inquiring “What is valuable?” or “What is important in life?” But these verbal equivalences get us no closer to understanding what good is. Wittgenstein begins his lecture by saying that each of these expressions can be used in two different senses, viz., what he calls the “trivial or relative” sense, on the one hand, and the “ethical or absolute sense” on the other (LE 4). The trivial question is easy to answer, because it merely asks the causal question about what action will cause or lead to the fulfillment of what some person wants or has as his purpose. Thus, for instance, if we want to know if a man is a good chess player, we simply want to know if he can move the chess pieces in situations of a certain degree of difficulty and with a certain level of dexterity; or, if we ask whether or not this is the right road we are merely asking if it is the right road relative to a certain destination. Such questions have a factual answer, depending on the relation between a man’s chess moves and an agreed‐upon standard of difficulty and dexterity. And the road is the right one if following it will lead the inquirer to his desired destination. A person could always respond “But I have no interest in playing chess well” or “But I don’t want to go to Cambridge,” in which case the discussion would cease. But if someone says to me that I am behaving like a beast, even though I might reply that I don’t want to behave any better, my critic could reply “But you ought to behave better.” This of course does not imply that it is all right for me to behave as I have. The question “Are you behaving well?” is not the trivial or relative sort of question. It has the ethical sense, asking about my conduct as measured against an absolute standard. Wittgenstein says that the difference between the two types of question comes to this:
Every judgment of relative value is a mere statement of facts and can therefore be put in such a way that it loses all the appearance of a judgment of value.
(LE 5–6)
With this distinction in mind, Wittgenstein is ready to state his conclusion:
Now what I wish to contend is that, although all judgments of relative value can be shown to be mere statements of fact, no statement of fact can ever be, or imply, a judgment of absolute value.
(LE 6)
What, he asks, if we had a book containing a complete description of the world, laying out in precise detail all of the facts contained in it (in the world, that is)? “What I want to say is, that this book would contain nothing that we would call an ethical judgment or anything that would logically imply such a judgment.” He claims that all the facts contained in this book would “stand on the same level and in the same way all propositions stand on the same level.” None of them would be propositions that are value judgments, and none of them would be sublime or important. In other words, none of them would be ethical propositions, and it would no longer make sense to draw the distinction between absolute and trivial ethical judgments. What he had called trivial judgments all make reference to someone’s state of mind. But no state of mind “is in an ethical sense good or bad” (LE 6). If we describe someone as murdering another human being, this description may cause us pain or outrage. But this is just another fact – about our state of mind – and will have no ethical significance. “So far as facts and propositions are concerned, there is only relative value and relative good, right.” His conclusion is dramatic and revolutionary: “Ethics […] is supernatural” (LE 6). There is no such thing as “the absolutely right road” or the absolutely right act or mode of conduct. Such things are all “chimeras” (LE 7).
Wittgenstein thinks that when we use such expressions as “absolute value,” “absolute good,” and “absolute right” (as in fact we do), we are likely to have an experience something like the one we have when we wonder at the existence of the world. Or I may have the feeling that I am absolutely safe, which may prompt me to say something like “I am safe, nothing can injure me whatever happens.” Admitting the reality of such experiences and expressions, it remains the case that “the verbal expression which we give to these experiences is nonsense!” (LE 8) I can wonder at the existence of something only if I could conceive it not to be the case. I can wonder at the size of a dog only if I can conceive of one smaller or larger. “But it is nonsense to say that I wonder at the existence of the world, because I cannot imagine it not existing.” Hence the experience, which may be real, has no expression that is not nonsense. Similarly, the feeling I may have that some action is absolutely the right thing to do, while real enough, has no expression that makes sense.
Thus, there can be no science of ethics, just as there can be no science of miracles. It is not that science has proved that there can be no miracles, but just that the “scientific way of looking at a fact is not the way to look at it as a miracle.” Similarly, although science has not proved that there are no ethical values, the ethical way of looking at people and events is not the scientific way. Since according to the Tractatus science incorporates the only meaningful way of talking, it follows that ethical propositions are not false but meaningless. This nonsensicality is their very essence. “Ethics in so far as it springs from the desire to say something about the ultimate meaning of life, the absolute good, the absolute valuable, can be no science”; and hence ethics gives us no knowledge (ibid). Nevertheless, Wittgenstein remorsefully states, “it is a document of a tendency in the human mind which I personally cannot help respecting deeply and I would not for my life ridicule it” (ibid). Making ethical judgments is a matter of trying to go beyond the world and beyond the limits of significant speech. This we cannot do.
As mentioned earlier, in this short lecture, Wittgenstein reiterates and slightly expands on what he had said about ethics in the Tractatus. For instance, section 6.421 of this book reads thus: “it is clear that ethics cannot be put into words. Ethics is transcendental.” Ethics attempts to say what is higher, and this cannot be done. Moreover, no ethical implications can be drawn from descriptions of facts in the world: “How things are in the world is a matter of complete indifference for what is higher. God does not reveal himself in the world” (TLP §6.432). “What cannot be said can only be shown, not said.” Thus, ethics can only be shown, perhaps through actions and the course of events. “And so it is impossible for there to be propositions of ethics” (TLP §6.42).
Although ignoring his statement of remorse cited above and other things he said about the “higher” and the “mystical,” Wittgenstein’s early thought was ardently accepted by members of the Vienna Circle, a group of philosophers and scientists meeting regularly in Vienna, led by Moritz Schlick. A.J. Ayer, a British visitor who attended some of these meetings, wrote a brief treatise attempting to summarize the thinking of this group. Its title was Language, Truth and Logic (1946). In this book, Ayer also devoted a chapter to ethical discourse, in which he related many of the ideas we have been discussing in this present chapter. Moral propositions, Ayer argued, are not empirical propositions, which can be known to be true or false by observations. Nor are they a priori propositions, which can be shown to be necessarily true or false by means of truth tables. Thus, he concluded, they are meaningless propositions, nonsense utterances used simply to express the speaker’s feelings. This startling conclusion was adopted by many English‐speaking thinkers, and dismissed with contempt by many others. Neither true nor false, ethical discourse, many began to think, falls outside the realm of meaningful speech.
It was Wittgenstein himself who overturned this new, positivistic, verificationist way of thinking. Soon after returning to Cambridge and philosophy in 1929, he began to rethink the views about the nature of language and its relation to the world that he had earlier propounded in the Tractatus. No longer identifying naming as the quintessential function of language or thinking of propositions as being essentially pictures of facts in the world, he began to stress the multiplicity of linguistic uses and to conceive of propositions as having radically different forms and logic. His authorized statement of his later‐period thought has been posthumously edited as Philosophical Investigations. For later Wittgenstein, ostensive definition, as traditionally conceived, is no longer the only way in which language is tied to the world. Rather, language must be conceived as constituted by language‐games, distinctive forms of rule‐governed behavior with a variety of criteria for the appropriate or inappropriate, true or false, permissible or impermissible uses of language. These rules, in turn, are what constitute the grammar of language and specific kinds of language‐game. Moreover, these language‐games are incorporated into a multiplicity of linguistic practices, and it is only these practices that manifest the rules of grammar. The actual uses of words in specific language‐games identify the conditions of appropriate or inappropriate application of these words, and of the sentences and utterances in the context of which they function. Language is a human activity, and its rules are human artifacts that no longer reflect the logical structure of the world, a relation that would be capable of verification or falsification by comparison to the logical structure of this world. On the contrary, grammar is autonomous (viz., just as much as human beings are).
This new conception of language immediately led many philosophers, who were uncomfortable with the idea that ethical, aesthetic, and religious discourse were just nonsense, to consider the possibility that judgments in these areas might not be used to state empirical facts about the ethical, aesthetic, or religious dimensions of the world. They might not have a fact‐stating role at all, but rather a different use in distinctive human practices governed by sui generis grammatical rules. A.J. Ayer had already hinted at one such possibility when he identified ethical judgments as expressing personal feelings. Saying that lying is wrong, he wrote, might be tantamount to saying something like “Lying, boo!”. Such an expression is not true or false, but neither is it nonsense.
Ayer’s analysis was thought by some observers to be on the right track, but, as it stands, a bit primitive. The American philosopher Charles Stevenson developed a more sophisticated form of emotivism in his Ethics and Language (1944). Stevenson was intrigued by how much moral disagreement occurs in our moral discourse and why these disagreements so often go unresolved. If, as many naturalistic philosophers had maintained, moral terms refer to natural properties detectable by scientific methods, the moral judgments containing them should be open to empirical verification or falsification. But, in fact, this sort of confirmation does not often occur. Stevenson was also impressed by the fact that moral disagreements and arguments are highly dynamic affairs – the parties to the disputes are trying, it appears, not just to get their opponents to believe something different, but to feel and act differently. This suggested to Stevenson that feelings and attitudes are essentially involved in moral judgments and arguments. Feelings and attitudes, after all, are neither true nor false; and therefore, if they are being expressed and challenged in moral disagreements, there is little surprise in the fact that these disagreements are often not resolved. Stevenson shows how, to his mind, these attitudes and dynamic activities are built into the very meanings of moral terms and judgments. The latter, he maintains, have both emotive and descriptive meaning – they serve to express and evince or evoke attitudes as well as to make descriptive claims. The word “good,” for instance, is often used by a person to describe himself as approving of some object or action, but at the same time it is used dynamically to express the speaker’s “pro‐attitude” toward the object. It is also used to urge someone else to approve of the subject matter as well. Hence, according to Stevenson, “x is good” often means something like “I approve of x, do so as well.”
Stevenson’s theory expresses both the influence of Wittgenstein’s early thinking on ethics and his later idea that meaning is use and that meaning comes in a variety of forms. Moral judgments are not true or false, but meaning cannot be understood solely in terms of reference and description. The language‐games in which we employ moral terms are integrated into our lives and our culture, in which we express and argue about “noncognitive” matters such as attitudes and have disagreements about cognitive matters, including the facts about the states of mind of one another. There are disagreements in attitude as well as disagreements in belief.
One of the implications of Stevenson’s arguments was to cast doubt on the notion of validity in moral arguments. A moral argument is designed not so much to prove a moral judgment as to persuade and provoke agreement in belief and attitude in another person. By virtue of their emotive meaning, the judgments found in moral life are not such that one or more of them must necessarily be true if others are also true – thus the notion of validity does not apply to moral arguments.
This conclusion seemed implausible to the British moral philosopher R.M. Hare, who thought that the moral argumentation often found in moral discourse can be, and is frequently, evaluated and hence found to be good or bad and hence valid or invalid. Hare set out to show how this is possible in his The Language of Morals (1952).
Hare implicitly agreed with the later Wittgenstein (and Stevenson) that discourse is not necessarily descriptive. Rather, to say that something is good is to commend it, not simply to describe it. And to commend an object is to tell someone to choose it, or, in the case of actions, to do one or another of them. Moral judgments therefore serve as utterances akin to prescriptions, and prescriptions are neither true nor false. “You ought to close the door” does not describe some fact but is a way of commanding someone to close the door. Some facts are indeed implied by a moral judgment, facts such as “I want you to close the door” or “I have a pro‐attitude toward your closing the door.” It follows that these judgments have both an evaluative meaning, expressed by its imperatival nature, and a descriptive meaning that conveys some proposition about the speaker’s state of mind (his attitudes, desires, and the like).
But if moral judgments are of this hybrid evaluative/descriptive nature, how can they be true or false and sustain relations of validity among themselves? Hare showed how prescriptions can be logically related. Moral judgments imply universal standards, which are also prescriptive. For instance, “You ought to close the door” might imply the universal principle “Everyone should close the door if requested to do so,” which commands everyone to close the door under the specified condition. Hence, we are able to argue for our specific command by generating the following syllogism: (1) “Everyone ought to obey the order to close the door if requested to do so”; (2) “You have been requested to close the door”; (3) “Therefore you ought to close the door.” Hare demonstrates how, to his mind, we can justify both premises of this argument and demonstrate its validity.
In this way, Hare can be seen to exemplify Wittgenstein’s suggestion that various areas of discourse, such as the moral one, incorporate different uses of language (in Hare’s case, the prescriptive use) and how specific forms of rationality are built into these language‐games. In this sense, Hare’s ethical theorizing, like Stevenson’s, illustrates basic ideas of the later Wittgenstein’s thought.
Stevenson and Hare, however, are noncognitivists. They do not believe in moral facts, and moral knowledge, if it can be said to exist at all, is of a very distinctive kind. But other philosophers have been influenced by the later Wittgenstein in such a way that they develop cognitivist ethical theories. For instance, D.Z. Phillips and H.O. Mounce, both distinguished Wittgenstein scholars, construct a theory in their book Moral Practices (1969) that emphasizes Wittgenstein’s view that moral discourse is a kind of activity or practice, guided by rules that permit the use of moral terms under certain specifiable conditions, i.e., uses that accordingly can be considered true or false when these conditions do or do not hold. In this manner, moral knowledge and moral justification is possible within a practice. Phillips and Mounce allow that there may be alternative moral practices, differentiated by the different sets of rule that constitute them. Hence, they argue for a kind of moral relativism. What judgments are true or justified in one moral practice may be quite different from those found in another.
Still a different kind of cognitivism is found in the last chapter of the present author’s Rationalism, Realism, and Relativism (1989) in which it is argued that the rules of what is often called common morality constitute the only rules of what is properly called morality. These rules are expressions of what Wittgenstein calls grammatical statements. For example, “Lying is wrong” is a grammatical statement defining in part what it means to be moral. Such statements are necessarily true, although they are also defeasible. They can be defeated under certain conditions, which can be built into the grammatical statement, e.g., “Lying is wrong unless doing so will save a person’s life.” These rules cannot be justified or criticized, given what Wittgenstein says about the autonomy of grammar (see Chapter 15, THE AUTONOMY OF GRAMMAR). Any people who do not follow them, who, for instance, frequently engage in lying and see nothing wrong in doing so, are simply not moral people (because, due to the grammar of “morality,” “moral,” etc., it would not be correct to call them “moral”); in other words, they do not participate in the moral language‐game or practice of morality.
The ethical potential of Wittgenstein’s views can also be made apparent, for instance, by comparing and contrasting them to the intuitionism of W.D. Ross. Obviously, there are substantial differences between the two. Ross believes that we cognize the truth of the principles of prima facie duty by an act of mind that he likens to the cognition of the a priori propositions of geometry and arithmetic. Relevant epistemological aspects of Wittgenstein’s philosophy of mathematics are far removed from this conception of mathematical knowledge. However, interpreting Ross’s principles of prima facie duty as grammatical statements may allow us to draw a connection between Ross and Wittgenstein. If, for instance, “one’s prima facie duty is to tell the truth” is a grammatical statement, then it possesses a form of necessity – grammatical necessity, not logical in the sense specified by formal logic – that is not possessed by contingent, empirical moral propositions like “Mr X did something he ought not to do.” Furthermore, in saying that the duty to tell the truth is a prima facie one, one is acknowledging that the duty to tell the truth can be overridden, such as in instances in which telling the truth would result in the death of a person. In vocabulary that has aptly been used to characterize Wittgenstein’s idea of criteria, it is defeasible. As grammatical statements, the principles of prima facie duty define what it means to be moral. They tell us what it means to speak of someone being a moral person or a situation being of a moral kind. (See Arrington 2002 for further discussion.)
Finally, I shall comment briefly on the only passage in Philosophical Investigations in which Wittgenstein has something explicitly to say about ethical concepts. In PI §77, he discusses concepts that have no clear lines of distinction but rather “shade into one another.” There can be no clear answer to the question whether a definition of such a concept is right or wrong: “anything and nothing – is right.” He goes on in this section to say, “And this is the picture in which, for example, someone finds himself in ethics or aesthetics when he looks for definitions that correspond to our concepts.” G.E. Moore began contemporary linguistic ethics when he asked what is good. He denied that the term “good” could be defined, on the grounds that it refers to a simple, indefinable, nonnatural moral quality. Wittgenstein agrees that it cannot be defined, but not because of the unique, simple nature of what it supposedly refers to, but because the concept that it expresses has no sharp boundaries. Its various meanings may have a family resemblance to one another, but there is no one quality to which it refers or which would be its meaning. The concept of right has no clear meaning, but rather a multiplicity of uses that shade into one another. As he had said in his lecture on ethics, some of these uses indicate an absolute sense of “right,” while others express a relative sense. And in both categories, there may be a variety of meanings that merely shade into one another. One right may be an absolute obligation, another an action justified only because it indicates the means to achieving a purpose or desire that someone else may not have. Ethics, as its twentieth‐ and twenty‐first‐century history clearly shows, is a murky business.