Functional analysis, not to be confused with functionalism, the claim that the minds, along with certain other things, are to be identified with the relationships between the parts of objects, is a way of approaching philosophical subjects and theories. To conduct a functional analysis we ask our selves why something exists. If it exists by “accident” then we can go no further with our analysis, but many things exist because of design or an evolutionary process. In both cases we can uncover an ideal function or role that it serves, and thus understand it in those terms. (Or, even better, use that understanding as a starting place for a more precise investigation.) Functional analysis is thus in some ways like conceptual analysis, in the sense that they are both approaches that shape an entire philosophical investigation, except that conceptual analysis approaches subjects with the assumption that we already know what the objects of study are and that all we need to do is make that knowledge precise, while functional analysis approaches subjects in order to build successful theories around them, even if those theories lead us away from our intuitions.
To begin with we can distinguish between two applications of functional analysis based on whether it is used to study an object of inquiry, such as ethics, by asking questions such as “why do we act ethical?”, or whether it is applied to the theory itself, such as epistemology, by asking questions such as “why do we study epistemology?”. Let’s consider the second case first. We might apply functional analysis to our theorizing itself when we suspect that, while we may have constructed a very intricate system describing something, our investigations may have gone off track, and we may have ended up with a theory about something other than we originally intended. I used epistemology as an example earlier because I think it is one place that our philosophical investigations may have gotten a bit off-track. In many cases epistemology has been studied by trying to answer the question “what is knowledge?” And that is not necessarily a bad question to ask, as we may be interested in understanding knowledge for its own sake. But I think the answer to “why do we study epistemology?” has to do with the fact that we are interested in improving our epistemic position, we want to eliminate as many of our false beliefs as is possible, and learn how to conduct future investigations so that they result in beliefs that are more likely than not. Our interest in pinning down exactly what knowledge is stems from that, in the hopes that when we know what knowledge is we can learn to have more of it. But the definitions of knowledge developed in epistemology are often relatively useless for this purpose. Someone may go into great lengths describing how knowledge is justified true belief in the absence of defeaters, and exactly what constitutes a defeater. But unfortunately we can’t say for sure whether such defeaters are present or if the belief is true. Thus elaborating on these requirements doesn’t make for a better epistemology, because it doesn’t help us improve our beliefs (except to tell us that if our belief is discovered to be false, or that some facts are revealed which defeat our justification, that we should discard it, but we knew that already). Now I’m not claiming that everyone ignores justification (which is obviously what it is actually useful to theorize about), it just receives less attention than it properly deserves. In fact since it is only justification that we need concern ourselves with (since it is only through improving our understanding of it that we can improve the beliefs we form), and since knowledge is a term used primarily as an endorsement, we might as well simply call justified beliefs knowledge, regardless of whether we have intuitions about knowledge that insist it must be true. Of course we don’t have to, knowledge is just a term, but it does, I think, emphasize that justification is proper object of study.
But in many cases the point of our theorizing is simply to achieve a better understanding of some subject, or perhaps the purpose we wish our theory to serve requires such an understanding, such as justification in the case of epistemology. In this case we must functionally analyze the subject matter itself. Obviously the exact nature of the analysis will depend on whether the subject in question is a human invention and thus has an intelligent purpose behind it, or whether it emerged through evolutionary processes. Justification is obviously a classification developed by us for our own purposes, and not necessarily meant to reflect an existing/natural classification. And, going back to our functional analysis of epistemology, we can argue that the point of the concept of justification is to separate the best methods of reasoning, those that lead us to the most probable beliefs, from those that don’t. And, with that bit of functional analysis done, it becomes much easier to develop a theory of justification, simply consider the recommendations it makes about how to reason. Are they superior, in general, to other possible reasoning strategies? If so then the theory seems good so far, but if they don’t then clearly we have not appropriately captured justification. This leads to a very “empirical” study of justification, not in the sense that we need to dust off our microscopes and take a look at it, but in the sense that we need to actually compare strategies, an essentially objective process, and that there is always the possibility that someone will “falsify” our theory by developing better reasoning strategies.
In a sense that is the easy case, because since justification, while not independent of external reality, is essentially a concept we have developed, and so it isn’t hard for us to uncover its purpose. Functional analysis is thus less decisive when it is applied to concepts that are meant to reflect some “natural” phenomena or a “natural” division of the world. An example of such a concept is ethics. Most people believe that it is good to act ethically. Now it is possible that is just a coincidence, but I doubt it. Thus it is more likely that ethics is somehow beneficial and thus pressures exist that “select” for it, which lead to a world in which most people have the desire to be ethical. There are two obvious possibilities: ethics may be beneficial to the individual, or ethics may be beneficial to society (and thus societies that are structured to lead their members to desire to be ethical dominate). It is possible to develop a theory of ethics under either assumption, although I personally lean towards the second, in much the same way we develop a theory about justification, we consider which rules, which attitudes, which behaviors, serve that end and which don’t, and label them ethical or unethical accordingly. It may be that one of these two possible functional analysis of ethics sticks closer to our intuitive and unanalyzed conception of it than the other does, but that doesn’t make the other wrong. It is more accurate to see the theories that develop from such a functional analysis as being about ethics-A and ethics-B, two new concepts that are closely related to, but not necessarily identical with, the old, and which capture everything important about the old and are thus suitable as replacements.
But don’t be fooled by the usefulness of functional analysis into thinking that it is suitable for every occasion. The less abstract a concept is the less likely it is that we can conduct a successful functional analysis of it. For example, it is unproductive to analyze “volcano” in functional terms, volcanoes are best understood through a direct empirical study. And neither does functional analysis provide the whole theory, it is simply a jumping off point that provides standards which can guide further investigations. But applying the standards and developing candidate theories to apply them to is a task for other skills.