What I will call the Dictionary Fallacy is the fallacy of thinking that certain philosophical questions can be answered by consulting dictionaries. The philosophical questions I have in mind are those of the form What is X? or What is the nature of X? High on the list: What is justice? Knowledge? Existence? Goodness? But also: What is hypocrisy? Lying? Self-deception? Envy? Jealousy? Schadenfreude? Socialism? Taxation? Motherhood? For example, if a woman is pregnant, and if the pregnancy is well-advanced, in the eight month, say, is the woman the mother of the fetus growing within her, or is she a mother only when she gives birth? Can this question be resolved by consultation of a dictionary? The dictionaries I am referring to are ordinary dictionaries, not philosophical dictionaries.
1. Suppose I want to know what tolerance in the social or political sense is. I pull a dictionary from the shelf and read, "a sympathy or indulgence for beliefs or practices differing from or conflicting with one's own." (Webster's New Collegiate, 1977, pp. 1227-1228) This is not a bad definition but it seems to go wrong in one respect. Arguably, toleration entails dislike of or dissent from that which is tolerated. If I tolerate your socialist views, then I dissent from them. Otherwise I wouldn't be tolerating them. This entailment seems built into the very concept of toleration. That with which I sympathize or agree I do not tolerate, strictly speaking. Now it is easy to imagine people disagreeing with the notion that tolerance entails dislike or disapproval or dissent. "Why can't I be said to tolerate what I approve of? I approve of my wife's frugality, and therefore I tolerate it."
At this point one must ask what the debate is about. Is it about the way 'tolerance' and cognates are actually used by English speakers? Then it may well be that the lexical definition cited is correct. This is an empirical question for lexicographers. But if the question is about the nature of tolerance, then the appeal to actual usage is not conclusive, though it is undeniably a good starting point. For actual use may be misuse. Everyone knows that change is not the same as improvement. Every improvement is a change, but not every change is an improvement. Some changes are for the worse. But suppose the Obama Crowd brings it about that the word 'change' comes to be used interchangeably with the word 'improvement.' The dictionary meaning of 'change' would then give us a very poor understanding of the nature of change. The nature of change does not depend on how people use or misuse the word 'change.' It is therefore a fallacy to think that one can answer the philosophical question as to the nature of X by recourse to the dictionary definition of the word 'X.'
2. The dictionary I just cited gives 'envious' as a synonym of 'jealous.' (p. 620) Now they are as a matter of fact often used synonymously in English. But surely the nature of envy is different from the nature of jealousy. To feel envy is to feel diminished in one’s sense of self-worth by another’s success or well-being. Thus if A feels bad because B won an award, then A envies B his winning of the award. It is a misuse of language to say that A is jealous of B in a situation like this. Jealousy requires three people, whereas envy requires only two. Suppose A and B are married, and C shows an amatory interest in B. A may well come to feel jealous of C. To use ‘envy’ and ‘jealousy’ interchangeably is to ride roughshod over a simple distinction, and that is something that clear-headed people will want to avoid. Someone who claims that there is no difference between envy and jealousy on the strength of the fact that one or more reputable dictionaries give 'envious' as a synonym of 'jealous' has committed what I am calling the Dictionary Fallacy. They are falsely assuming that the way a word 'X' is used is a sure guide to the nature of X.
3. I am of course assuming that there are certain distinctions 'out there in the world' and that responsible language-users must use words in such a way as to do justice to them. For example, there is a difference between inference and implication even if everyone comes to use 'infer' and 'imply' interchangeably. And there is a difference between raising a question and begging a question even if the corresponding phrases come to be used interchangeably.
4. I am also assuming that the philosophical enterprise is constructive and theoretical: the aim is to construct theories that illuminate our experience. To understand the moral life, for example, one must understand such phenomena as hypocrisy, idealism, cynicism, jealousy, envy, resentment, irony, Schadenfreude, lying, dissembling, self-deception, bad faith, and others besides. Presumably, the analyses of these individual phenomena must fit together to from a coherent whole. An adequate theory cannot consist of a disconnected assortment of piecemeal analyses. This implies that the definitions of key terms such as 'hypocrite' cannot be purely lexical. As already pointed out, it is not the word but the thing we wish to understand, and the word as actually used is no sure guide, and in some cases no sort of guide, to the nature of the thing. As theorists of the moral life we must therefore craft precisifying definitions. These involve a substantial stipulative component, though they cannot of course be merely stipulative. One cannot, for example, stipulate that a hypocrite is a person who perceives a gap within himself between what he is and what he ought to be. That deviates too far from the lexical meaning of the term. But one may have to adjust the meaning of 'hypocrite' away from its lexical definition somewhat to make it cohere properly with other terms in the vicinity, terms that figure in our theory of moral phenomena.