What is a door? Suppose one person says that a door is material that is intended to be placed in an opening in a structure that is used as an entrance or exit, and which doesn’t prevent that use (definition A). And suppose that another person says that a door is a hole in a structure designed for people to pass through it (definition B). Can we say that one of these two people is right, and that the other is wrong? No, we cannot; all we can say is that one of them (probably definition A) is closer to the conventional usage of the word than the other. And certainly we cannot rule out the possibility that the person who subscribes to definition B is part of a community in which that definition is standard, and in the context of which definition A would be unconventional. “What is a door?” is thus revealed as a question without any value of its own; it has value only in the context of claims about doors. For example, the claim “a building needs at least 1 door per 10 occupants to be safe in case of a fire” can be true or false, and has the potential to be informative. And in light of it the question “what is a door?” matters, not because one answer is true, but because it changes the meaning of the claim. If we are working with definition B the claim may very well be true, and relevant to architects. But if we are working with definition A it is false, or at least badly worded (a door factory is not necessarily any safer than a brick factory with the same floor plan). The key idea to take away from this example is that questions in the form “what is X?” are largely irrelevant, and reveal, if anything, only convention, and that it is the claims we make using X that are important. It follows from this that the way we define X is important only because it determines what exactly we are claiming; although the claims can be right or wrong the definition of X can’t be.
This is basically common sense when it comes to things like doors and rocks and gold. So much so in fact that many times when we are asked questions like “what is gold?” we don’t answer by saying what we commonly designate by the term, but by saying something about the things that we usually designate, in this case that they are composed of a specific atomic element. But when it comes to philosophical matters, such as justice, the intuitions of many people seem to change. Suddenly the fact that two people have conflicting definitions of justice implies that one or both of them is wrong. Personally I think this shift in intuitions is a bad one, and leads to much wasted effort (wasted because there is no right definition to find). However I am interested here more in trying to understand why our intuitions shift, for which I see three possible reasons.
One possible reason is philosophical self-preservation. If we accepted that the definition of justice is truly a matter of convention then the philosopher becomes simply a writer of dictionaries, and could do philosophy best by measuring popular opinion. Or, if we instead choose to interpret the question as a search for common features in the things conventionally called just, then the question must be answered empirically (preferably after a survey to determine what exactly the conventional usage is). Again, this takes the job of philosophy away from philosophers as traditionally envisioned. However, this strikes me as a red herring. “What is justice?” does not exhaust the philosophical questions we can ask about it. We can also consider “should we act justly?”. Even if we accept that there are many possible definitions of justice evaluating that claim in light of each can be informative and useful, since each can provide new insight into how we should act (because each definition of justice leads us to consider a different way of acting).
Another possibility is emotional investment. When we first learn to use the word “justice” we pick up both a sense of what it is and the idea that it is important. The belief that justice is important (something that we don’t believe about doors) may prompt us to cling to the idea that there is a right definition for justice. Certainly this may prompt an unexamined attachment to the idea that there is a correct definition of justice, but it can’t survive reflection. For starters the fact that we would like there to be a correct definition doesn’t mean that there is one. And, more importantly, reflection reveals that what we really feel is important is acting in a way we personally are inclined to call just. And thus this emotional attachment rationally shouldn’t prevent us from accepting the idea that there is more than one way to define justice; just because justice can have more than one definition doesn’t mean that acting in the way we intuitively think of as just isn’t important (justice as I define it may matter even though other people are free to define the word differently).
The last, and in my opinion best, explanation of the intuition that there is a right answer to “what is justice?” has to do with the fact that justice is more abstract than the concepts expressed by “door” or “gold”. People have similar intuitions about math, that there is a right way to define arithmetic over integers. I think these intuitions have their roots in the idea that we have direct intellectual access to abstractions such as mathematics. And in one sense we do, because abstractions are constructions of the intellect. But what we don’t have is any guarantee that the abstractions we create actually describe the world, or just seem to. So while this explains the intuition it also removes its force. We might understand justice as referring to an abstraction, and thus something that we have direct intellectual access to (and thus can be “right” about it). But there is no guarantee that this abstraction can actually be usefully applied in the world (even to the things we think of as just, and on the basis of which we have constructed this abstraction), or that other people have constructed that same abstraction. Again, when it comes to claims outside of our own abstract constructions involving justice, it doesn’t really make sense to say that one person may have the correct definition of justice, at best they can have one that seems to accurately describe some feature of the world.
If there is a lesson to take away from these considerations it is this: when evaluating a definition of justice what matters is not whether it captures the essence of justice or the concept of justice, but whether it is useful in making informative, interesting, and correct claims.