There is often discussion in interdisciplinary research about what we mean by central concepts. We might insist, for the sake of the research, on agreement about the issue of how we are to understand, say, "diaspora". Or we may decide for the sake of the research that it ought not be defined. On the other hand, it sometimes happens that, for the sake of investigation, there is insistence that a term not be defined, moreover that it not be used. This has been the case with some terms in feminist theory, such as "autonomy" and even, occasionally, "rationality". There is an ethical issue involved in decisions to define or not to define central concepts. For when we decide to define a term specifically, we decide to take a position about the specific direction of research. If we decide to define a term, we express confidence in the current theoretical resources and in our perspective upon them. We are making a judgment that in fact the perspective is not in fact in question, and that we are in a position to insist upon specific criteria from the perspective that we now occupy. Sometimes, for instance, it is not appropriate to insist upon a specific definition because it is recognised that the appropriate referent of the term is dependent upon the results of empirical investigation, which is still to be undertaken. In science, for instance, research can be carried out effectively in awareness that the referent of the term is still to be discovered. This was the case, for instance, with "black holes". There was something there to be theorized about, but it would not have been appropriate to define the object of investigation. That would have been a theoretical error because the resources were not available and could not be expected to be available.

To insist prematurely upon a definition can constitute a failure to recognise the importance of the object of investigation. If someone were to insist prematurely that the term "black holes" be precisely defined, one might think such a person had failed to recognise the complexity of the investigation and the importance of that complexity. Interestingly, Armando Hart argues that the biggest mistake made by the leaders of socialist revolutions in Eastern Europe and the Soviet Union was not to have taken up seriously the task of redefining terms like "human rights", "freedom" and "civil society". How could the failure to redefine concepts be a bigger mistake than all the other mistakes made by the Soviet Union? Philosopher of science Ian Hacking points out that social and human kinds, that describe people and behaviour, have implications for how people understand themselves and for how they behave. For instance, when the term "sexual harassment" was introduced, many women were able to identify disturbing behaviour in the workplace for what it was - wrong! This had consequences for self-confidence. Hacking suggests that social and human kinds have a "looping effect" upon ways of being. The understanding of such terms loops back upon us and affects our existence and behaviour. Hart seems to be suggesting that because terms like "human rights" and "freedom" were not redefined and claimed, they were not able to be motivating in practice and theory in the way that they might have been.

Jorge Luis Arcanda also emphasises the prime ethical importance of redefining. In his view, the tendency in Cuba, for instance, has been either to reject certain concepts - like "civil society"- altogether, and to say that, not bein g a Marxist concept, it does not apply, or to use the concept but to use it in the same way as in liberal debates, with the same meaning. If the concept is rejected, there is no debate about the issue in question and no way to treat changes as part of a process toward greater democracy or freedom. And if the term is used in the same way, it also works against progress because it just turns out that there isn't any such thing as civil society, democracy or freedom, and so not much hope in pursuing or improving these phenomena. The failure to redefine terms, according to Luis Arcanda, has resulted in the promotion and strengthening of the liberal ideology, and a weakening of resistance to it.

Because of the "essentialist" problem mentioned above, it is often assumed we cannot engage critically with social or human categories. It is recognized that social or human kinds, unlike or at least differently from natural kinds, cannot be defined precisely, and that it is a mistake to look for precise sets of categories. It is also recognized that social or human kinds involve us in the way that Hacking describes. It is then concluded that social or human kinds can be used in whatever way we choose. It is often concluded from such premises that "anything “goes". This is a mistake. Sometimes it is suggested that because the term "feminist", say, has no fixed definition, it follows that as long as someone thinks she is a feminist, she is a feminist. Or people say that because "sexual harassment" has no fixed definition, as long as someone thinks she has suffered sexual harassment, she has indeed suffered sexual harassment. But although it is indeed the case that there exists no fixed set of properties defining these phenomena, it is certainly the case that I can say on the basis of reason and evidence that someone is using the term incorrectly. I don't need to think there is a fixed definition for "feminist" in order to judge that, say, it is not the case that being a feminist requires hating men.

We must distinguish between the requirement that social and human kinds be defined permanently and precisely and the expectation that there are good reasons for applying such terms in some cases and not in others. It is true that such terms do not possess fixed definitions, applic able to all situations at all times. It does not follow that there are not non-arbitrary criteria for deciding that a term is being applied incorrectly. It is in fact the case that a person can be mistaken in thinking that a certain behaviour constitutes sexual harassment, even if she believes it very strongly. We do in fact rely upon empirical evidence in arguing that what is believed to be sexual harassment is not in fact sexual harassment. And if this is the case, then there do exist non-arbitrary criteria for applying the term in some cases and not in others even though there is no fixed "essentialist" definition. In short, definitions raise ethical issues because what we do with them constitutes a commitment to a direction. Decisions about defining a term or not are also, in some cases, judgments about importance of the research, and in other cases, judgments about the adequacy of a current theoretical or practical perspective. That is, decisions about redefining a term can express a commitment or not to critical engagement. There are also, as above, issues about responsibility. The argument that "anything goes" is often irresponsible and can be used to resist the hard, sometimes uncomfortable, work of examining theoretical and empirical grounds for the use of a term.