What is conceptual analysis and how do we perform it? An example of the kind of conceptual analysis I'm talking about is the analysis of the property of being a bachelor into the property of being unmarried and being male. Is such analysis just taking one word and defining it in terms of multiple other words? What about impredicative definitions, then?

  • 1
    See sec. 4 of "Analysis" in the SEP, where they cover the "decompositional" conception of analysis. The practice goes back to Plato's dialogues, though in the Meno, the utility of such analysis is significantly questioned. Jul 22 at 11:20
  • It's not that easy as term splitting and substitution as discussed in one of today's post, you also need to master modal and contextual analysis and what level of (meta)-languages you are using for a particular case... Jul 23 at 20:18

1 Answer 1


To illustrate the case of the bachelor-concept: suppose you are at a party, and you see a man you don't know across the room. He's attractive, so you turn to your friend and whisper, "Is he a bachelor?" Your friend answers yes or no, or maybe maybe. At any rate, your motive for asking was recognition that if a man is a bachelor, then you might "have a chance" with him (at least, a chance that would not be adulterous, or polyamorous (so we have to suppose that you yourself aren't polyamorous)).

Suppose instead you've had too much to drink already, and you see a cardboard cutout of a famous celebrity through the door to the host's room. You stumble over to your friend and say you saw a handsome fellow in the host's room. "Ish he a batch...lllurrr...?" you ask. The friend, who is more sober than you, chuckles and says, "No, that's a cardboard cutout of the host's favorite movie star."

So when you ask, more broadly, "Are bachelors unmarried men?" and you say, "Yes," are you "analyzing" the concept of bachelors? I daresay not quite. Immanuel Kant dwells on the fact that we can easily define concepts that we self-consciously construct, e.g. various mathematical definitions ("A triangle has three sides": why, that's why I used the prefix tri- and then angle, isn't it?); but obscure philosophical notions are harder to pin down like that. How do you "break down" the concept of substance? You don't want to start from the bare term substance, but you want to start from things you apply the term to, and then explain why you're applying the term to them.

So, for substance, let's say you start with a chunk of iron. "Iron is a substance," you say. Your friend asks, "What does that mean?" "Iron," you explain, "is a primary subject of objective predication, it is matter that has properties. You couldn't turn around and say that iron is a property of its own properties, and only in an extended sense can iron be a property of something else. We'd more likely say that this chunk of iron could be part of something else, but parts aren't properties."

Your friend, who is still more sober than you, might object, "Are you sure that parthood and being-a-property are so different? What if the boundary between them is vague?" So now you have to go and break down the concepts of parts and properties, and so on down the line, until you hopefully reach terms that cannot be broken down further.

Per Plato's "Meno," and also many criticisms since, this talk of "breaking down concepts" can seem empty, pointless, and misguided. A lot of the time, a lot of communicative effort can be saved by pure stipulation, coupled with the appropriate ostensive definitions by examples. How much is gained by trying to define our terms in other terms, that are then also "redefined," etc.?

Actually, that's the only value conceptual analysis can have: if we can "prove" (or strongly indicate) which terms we use have irreducibly ostensive definitions at best. It is hard to parse the notion of syntax, for example: we might say, "Syntax is the structure of speech," but eventually you'll run into arguments about structures being the syntax of reality. So perhaps conceptual analysis leads us to the conclusion that syntax/structure is an irreducible concept. A taxonomy of such irreducibility is what we're really looking for in decompositional analysis, not quite just the meaning of individual words that we can solve for by stipulation. (For example, if knowledge were "reducible," then we would be free to rearrange its conceptual "parts" at will, and it would no longer be as significant to claim that we do or do not know something (it would matter, but not as much).)

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .