Philosophy as Logical Analysis
Philosophy as Logical Analysis
The method of philosophy is sometimes described as “conceptual analysis”. This is not wrong, but it can mislead and it carries unwanted baggage. The phrase makes it sound as if the philosopher’s object of interest is a mental entity not an objective feature of the world. If you are a skeptic about mental entities, this description makes philosophy about nothing (same for “meanings”). I would say that the philosopher is interested in de re essences that can be revealed by logical investigation.  I take this to be committed neither to the analytic-synthetic distinction nor to the idea that philosophy discovers a priori truths (or can be undertaken from the “armchair”)—though in fact I subscribe to both of these. If you believe that logic is continuous with empirical science, you can still accept that philosophy is logical analysis; it will also be continuous with science. Holding that philosophy is logical analysis commits you to very little, because the notion of logical analysis is itself pretty neutral (unless you already buy into certain views of the nature of logical analysis).  The import of the phrase is just that philosophy is not empirical analysis—the kind of thing done in chemistry. We needn’t get too exercised about quite what logic is in order to accept this characterization. Similarly, linguistics involves “grammatical analysis”— also an anodyne statement before we enter into controversies about the nature of grammar (it is certainly not chemical analysis).
What is the import of “analysis”? Easy: analysis is resolving a complex entity into its constituent elements. I take this to mean that a complex attribute (or sometimes object) is broken down into its constituent attributes (or objects). For example, the complex attribute of knowledge is resolved into the constituent attributes of truth, belief, and justification (let’s simplify matters and assume that these are the only attributes composing the attribute of knowledge). So my view is that we achieve this kind of analysis by logical means (not by chemical means); and this tells us the essence of knowledge. Presumably, the concept of knowledge maps onto the attribute of knowledge in a systematic way, with a constituent-to-constituent isomorphism; but that is not our primary concern—which is the nature of the attribute itself (what the fact of knowledge consists in). We find out the constituent structure of a fact by logical methods, whatever these methods are exactly (thought experiments, knowledge of entailment, use of logical apparatus). The process is like analyzing truth-functional connectives by means of truth tables: we spell out the nature of conjunction and negation (say) by employing logical analysis.
Sometimes the project of logical analysis is formulated in terms of necessary and sufficient conditions; nothing wrong with that, but it is important to see that necessary conditions are the central thing. We want to know what is essential to a given attribute—what has to be instantiated for that attribute to be instantiated. Whether we can supply sufficient conditions is a separate question: even if no (non-circular) sufficient conditions can be specified, that does not impede the provision of necessary conditions. The viability of logical analysis does not depend on the existence of specifiable sufficient conditions. Truth, belief, and justification could all be necessary conditions of knowledge without being jointly sufficient for knowledge—there may be further necessary conditions to be added, or it may be that there are no (non-circular) sufficient conditions. It is also possible that knowledge has only these three constituent elements but that they combine in a fashion that cannot be captured by simply conjoining them. The whole might be more than the sum of its parts, because the parts have to form a synthesis that can’t be captured by a mere list. The parts have to fit together in a certain way if the whole attribute is to exist (as the parts of an engine have to fit together for the engine to exist). In any case, the project of articulating necessary conditions is not committed to the possibility of providing sufficient conditions. In addition, sufficient conditions are of little interest unless they incorporate necessary conditions: it is a sufficient condition of something x being a number that x be divisible into a pair of equal whole numbers, but this is obviously not a necessary condition of being a number (not all numbers are even). Likewise, it is sufficient for knowing that p that one has learned that p or that one remembers that p, but these are not necessary conditions of knowledge, since not all knowledge is learned or remembered (consider innate knowledge). What makes a sufficient condition interesting is that it is the summation of a set of necessary conditions: the necessary conditions are where the philosophical action is. Necessary conditions tell us what belongs to the essence of a thing—what it logically requires. Sufficiency is another matter.
You might want to object that this conception of philosophy is too narrow, because it limits philosophy to the bare examination of the essence of attributes—what about solving philosophical problems, producing philosophical arguments, inventing philosophical theories? We don’t just analyze things for the sake of it; we disagree over the nature of things and try to provide theories of them. Thus we try to solve the mind-body problem or the problem of free will or the problem of skepticism. I agree—we don’t just provide a laundry list of dry logical analyses and then gaze fondly at them. But when we try to solve a problem or argue for a position we necessarily employ logical analyses of what we are talking about. There is no future in trying to solve the problem of consciousness without an analysis of what consciousness is, even if the analysis is quite primitive; you will not make headway with the problem of free will unless you have some idea of what free will amounts to; and you will never discover whether skepticism is refutable unless you have a decent analysis of what knowledge requires. Of course, not all concepts (or attributes) will be complexes consisting of simpler concepts (or attributes)—some will be logically simple—but that is not an objection to the method of logical analysis, for in that case the concept is analyzable by means of itself. Logical analysis merely recommends that we analyze what can be analyzed, i.e. what has non-trivial necessary conditions. And it can be an interesting discovery that a concept admits of no analysis (e.g. G.E. Moore on “good”).
The point I have wanted to impress on the reader is that the method of conceptual analysis is more anodyne (less freighted) than it is commonly reputed to be: it is really a truism about the nature of philosophy. There is nothing in it to make anyone recoil into some dimly defined idea that philosophy is really of a piece with empirical science—as if there is nothing else that it could be. It is not committed to some kind of psychologism about the subject matter of philosophy, and it does not assume that necessary and sufficient conditions can be supplied (these are rarely obtainable). Nor does it have anything to do with ordinary language philosophy: it isn’t about language at all and it concerns all of reality not just the part that is spoken of by the ordinary person (so it includes science). It is simply the idea that in philosophy we characteristically proceed by the logical analysis of whatever it is that interests us: from ethics to physics, from mind to causality, from necessity to identity.  In chemistry we perform chemical analysis on the chemical substances of the world; in philosophy we perform logical analysis on any aspects of reality that engage our philosophical interest. Chemistry is not thereby limited to breaking things down (it uses chemical analyses in formulating theories and explanations), and neither is philosophy. Chemistry employs analysis because it deals with complex entities (chemical substances) made up of simpler entities; philosophy employs analysis because it too deals with complex entities (attributes, objects, concepts) made up of simpler entities. There is nothing suspect or unscientific about logical analysis as the method of philosophy: that is simply the name we give to what we distinctively do. We analyze things logically.
 I defend this position in Truth By Analysis (Oxford University Press: 2012).
 Even Quine believes in logical analysis, though he sees no sharp distinction between logic and empirical science.
 Not all of logical analysis should be viewed as the educing of constituents; it may also involve tracing necessary links between separate concepts. I have not gone into these qualifications here, but see chapter 7 of Truth By Analysis for a survey of types of conceptual analysis. Let me emphasize that the point of providing conceptual analyses, of whatever type, is often to advance arguments and solve problems; by no means are we limited to merely describing the make-up of a concept or thing.
Leave a ReplyWant to join the discussion?
Feel free to contribute!