Allow me to make a bold claim: that the method of conceptual analysis, employed in much of analytic philosophy, is flawed.
But first, what is conceptual analysis? In the simplest possible terms conceptual analysis is the process of picking apart a concept (such as mind, good, ect) in order to determine all that we can about it a priori. Often this conceptual analysis is accomplished by proposing better and better definitions of the word in question (meaningful definitions of course, not like the ones you would find in a dictionary, which define a word in terms of near synonyms). How do we know if one definition better approximates the concept than another? By showing how one of the definitions leads to contradiction or to applications that are clearly not what we mean by the concept in question. For example, when analyzing “good” in this way we may point out that certain properties are intuitively part of our notion of good, and we argue against other definitions by showing how they label as “good” situations that are intuitively just the opposite (for example the case of killing one patient to save five).
The problem with conceptual analysis is two-fold. The obvious objection is simply that not everyone has exactly the same understanding of the concept in question, and hence situations that seem intuitive or unintuitive to one person may seem just the opposite to another. The more serious problem is that our concepts may not align with the real world, and thus may mislead us. For example consider the concept of gold as understood by one of our distant ancestors. That person’s concept of gold probably is just any rock that looks like gold. But what about pyrite (fool’s gold)? It seems that their conceptual analysis would lead them to conclude that was gold as well. The conceptual analysist can of course respond by arguing that gold’s density is also part of the concept. But what then about white gold, which looks more like silver or platinum than gold? Perhaps it is only the density that defines the concept of gold, but, again, in that case an alloy of lead and some lighter metal would be considered gold. No matter how they analyzed their concept of gold it wouldn’t come to the right conclusion about what gold really is, since they have no understanding of the atomic nature of matter, or even how the metals are properly distinguished from one another. Yes, the concept of gold that this person has may be coherent, and it may be a priori knowledge, but it doesn’t match up with anything actually in the world.
There are two ways to defend conceptual analysis against such an attack. One is to argue that the concepts that are being investigated in this way are not supposed to align with the real world (they are not natural kinds). If that is true it seems hard to understand why we should care about them then. The concept of “ethics”, for example, surely cannot be divorced from reality so easily. After all what is it that can be good or bad besides real people and real actions? It seems obvious to me at least that ethics is simply a way of diving acts into two categories, just as gold and not-gold divide rocks into two categories. I hope that the connection between other concepts and reality is likewise obvious (at least if we expect them to be relevant to life in any way). The other possible response is to argue that the concept held initially and the concept that results after an investigation are really two different concepts. This might be a consistent way of looking at things, but if this is really the case I am forced to ask why the initial concept shouldn’t simply be discarded. After all there is nothing it corresponds with in reality, and we have a new concept that does which can take its place, and perform the role in our discourse and thoughts that it used to.
Even if we accept that conceptual analysis is flawed the papers published as part of such an investigation have not been in vain, as they can almost all be useful to a different kind of investigation. Let me then describe the kind of investigation that should replace conceptual analysis. The first part of the investigation is to establish the relevant features of the subject of your investigation, and this is where our tools of conceptual analysis are useful. We want to achieve a rough idea of what role the concept under investigation plays in life and in discourse. We also want to make sure that we don’t accidentally label irrelevant properties as essential, omit essential ones, or claim that it must have properties that come into conflict with each other; and this is where tools such as examples and counter examples, and of course intuition pumps, become useful. (For example our minds control our actions, a role in life, and have a first person ontology, an essential property. Note that this is not a complete first step, just an example.) The next step is to go beyond this rough sketch and through investigation discover what causes those properties, what explains them. Of course I make it sound easy here, but this second stage is at least as difficult as the first, and we may have to revise the rough outline we are working from as we go along. It may turn out that our initial concept aligned very closely with the real subject of our investigation, or it may be that the concept really includes much more or much less than we initially thought (or possibly even that there are really two or more real things that we have both been describing with a single concept, just as jade is really composed of two distinct types of stone).
Again, we can turn to our simpler example of gold in order to understand how this process could work. First our inquisitive ancestor accumulates many rocks that he or she thinks of as gold. They then examine these rocks closely for common properties. It is likely that most of them do share a few common features, and those that do not share those features are discarded as being not gold, but something that is simply very similar to gold. Eventually he or she (possibly several revolutions in physics later) comes up with an explanation of why they all have those properties; because they are made up of the same element. Of course for something as complex as “mind” the investigation is unlikely to be so simple. And, even though the investigation of gold was a matter purely for scientists, philosophers still can have a role to play in the investigation of “good” and “mind” (and other concepts). Specifically there still remains a lot to be done in developing a rough outline of the essential features of these concepts, as well as determining how the facts science has uncovered (neuroscience, sociology, and psychology) explain these features.
Finally I would like to point out that I did employ some of the tools of conceptual analysis here, specifically the investigation of gold used as an intuition pump, which just goes to show that the tools may still be useful. If I’m not basing my claims on conceptual analysis then where do they come from? Well the first step, deciding what a successful investigation into a concept is fairly easy, it is one that reveals truths. Fortunately history has already done the second step for me. In the course of other investigations methods like conceptual analysis have rarely led to the truth (or even close approximation to it, consider Descartes’ physics for example). On the other hand investigations of the type I have proposed as a replacement often have been, thus my confidence in endorsing them for philosophy as well.