On Philosophy

October 6, 2008

Analytic Philosophy And Phenomenology

Filed under: Metaphilosophy — Peter @ 11:30 pm

Previously I described the framework under which phenomenology works (or, more precisely, a charitable reconstruction of that framework). With that in hand it is now possible to discuss how the analytic method is similar to the phenomenological method, and whether they are faced with common problems.

The easiest way to go about making that comparison is to see where analytic philosophy fits into the larger phenomenological framework. Under that framework each discipline studies a region, or some subset of a region. Thus we can begin by asking which region analytic philosophy aims to study. If we have the origins of analytic philosophy in mind it is natural to conclude that it studies the formal region. Originally analytic philosophy was defined, indeed named, after the a priori analytic truths it was supposed to capture: namely those tautologies sentences that follow from the meaning of words. (Note that other a priori analytic truths include mathematics. Since mathematics falls under the formal region it is extremely natural to conclude that analytic philosophy, as such, does so as well.) Thus we could say that the goal of analytic philosophy is to describe formal systems, almost extensions to logic, from which all the truths about, say, justice can be deduced.

The problem with this picture is that the conception of analytic philosophy as a formal discipline doesn’t seem to fit actual analytic philosophy. First, when it comes to formal disciplines right and wrong have a different kind of meaning than in philosophy. For example, a mathematician working on non-standard analysis (calculus) does not accuse those working on standard analysis of being in error. Error in the formal arena is not something that is said of a system as a whole, rather it is something that crops up only internally, often in the form of a faulty proof. If analytic philosophy was a formal discipline we would expect that there would be no disagreement regarding whether one analysis of personhood, for example, was better than another. As with calculus, all consistent systems would be accepted as non-competing variants. If one philosopher was to correct another it would only be to point out that their conclusions aren’t entailed by their premises. But this not how analytic philosophy proceeds; analytic philosophers are quite committed to the idea that there is some best theory about personhood, and to arguing against those they see as worse.

A second problem with taking analytic philosophy to be a formal discipline is that it appears to have a subject matter. And a proper formal discipline is ideal; it has no subject matter of its own, but may be applied wherever it fits. Now this is not to say that formal approaches are never conjoined with some subject. Indeed phenomenology itself is supposed to be a formal approach to consciousness. However, we would never bill phenomenology as a formal discipline, given that it applies its formal structures to consciousness (and develops them explicitly to fit consciousness). And, inasmuch as analytic philosophy applies its formal structures to things such as personhood, it cannot be described as a purely formal discipline.

So if analytic philosophy is not purely formal then we are left with the regions characterized by objectivity, subjectivity, and intersubjectivity. Since analytic philosophy is, as was noted above, apparently committed to the idea that there is some single correct theory about its subject matter it seems natural to say that it studies some part of the objective region. (Since the existence of one universally correct answer is characteristic of objectivity.) However, we already have a discipline that studies the objective: science. And there is only one correct (or, at least, optimal) way to study each region. Thus if we claim that analytic philosophy falls under the objective region we must bite the bullet and accept that analytic philosophy is a kind of science. This would in turn mean abandoning those intuitions that analytic philosophers are so fond of appealing to, because the scientific method rejects intuition. What we would be left with might resemble Kornblith’s work: an attempt to find natural kinds that we can label with the familiar philosophical terms. But, while Kornblith (and a number of other contemporary philosophers), may be satisfied with this, it hardly is representative of the majority of contemporary analytic philosophy. It would seem that we are better off rejecting analytic philosophy’s ostensible commitment to objectivity, and looking for some other region for it to fall under.

Next up is subjectivity, i.e. consciousness. I think it is safe to gloss over this possibility without giving it much consideration. It is true that some have characterized analytic philosophy as studying our concepts (through conceptual analysis), and it is at least possible to understand that as a study of how we conceptualize our own experience. However, this seems like an extremely bad fit for analytic philosophy. If analytic philosophy is the study of concepts as we find them in consciousness it would be hard to salvage anything from its commitment to there being better or worse answers to philosophical questions. Now, admittedly, phenomenology does seem to be equally committed to the idea that it has some of the right answers, and it does study the subjective. If we must we can rescue phenomenology by construing it as a study of common forms of subjectivity, and reminding ourselves that it is not in the business of providing necessary and sufficient conditions for consciousness. But when we get into specific details of consciousness, such as how an individual is conscious of the personhood of a person it seems foolish not to expect that to vary greatly from individual to individual.

This leaves us with the intersubjective. And the most important members of the intersubjective region are senses. (Note that Husserl himself put senses as a third top level in his ontology, but I think he was wrong in doing so. Senses would not exist without thinkers and they are shared between thinkers, thus they appear to be prototypically intersubjective.) Taking analytic philosophy to be a study of senses does seem to be a good fit. Since senses often manifest in concepts or through language this approach is in general agreement with those who claimed that analytic philosophy was a study of concepts or of language. And it also manages to salvage some degree of the objectivity that analytic philosophers were after: senses are shared between people and so argument is possible about who is more accurately capturing the sense of a term. (This suggestion also agrees with one of the proposals I made in a previous paper concerning how the intuitions appealed to in analytic philosophy could be rescued from their apparent failure to capture anything objective.)

Certainly this clarifies the nature and project of analytic philosophy. But does it solve its problems? Or do its problems now apply to phenomenology as well, inasmuch as phenomenology deals with senses? In one way its problems are solved. Worries about whether intuitions are really reliable guides to things such as knowledge can be dismissed, since: a) intuitions are reliable guides to senses, and b) no one expects complete agreement about the intersubjective. However, new problems are created by this treatment of analytic philosophy, which I maintain are at least potential problems for phenomenology as well.

The first problem is that, with respect to many philosophical issues, we expect how things “really are” (how things in the objective region are) to have some bearing on philosophical matters. For example, whether we can have knowledge in a situation, one may expect, should depend on some objective facts about the world (such as whether the objective world actually exists, whether we are hallucinating or not, whether our senses are reliable, and so on). However, a subjective or intersubjective treatment of knowledge can never include such issues. It can only inform us about the conditions that we think must hold before we recognize some experience as providing knowledge. But this is not the same as asserting that those conditions must really hold. In other words, it can yield only an internalist account of knowledge. And similar issues can be raised with respect to ethics, the mind-body problem, and so on. These problems, however, can be solved. As I detailed in yet another paper, while phenomenology cannot address such issues, by its very definition, and neither can any other single-region discipline, the possibility for cross-regional disciplines exists. To summarize those conclusions: it is possible to develop cross regional theories, but only on the basis of completed theories about the regions to be bridged in this way.

In that way the traditional problems of philosophy, as conceived of traditionally (in ways that are inherently cross-regional) can be addressed, although not by philosophy alone (or at least not by analytic philosophy or phenomenology alone). Unfortunately there is still one large problem remaining that is not so easily resolved. So far all the approaches described here, both single regional and cross-regional, produce descriptive theories. With respect to the subjective and the intersubjective they report on how we in fact experience and conceive of things. And any cross-regional discipline is in the business of drawing correlations between the findings of different regions; it builds on “finished” theories about a single region, but it cannot go back and revise them. This puts us in a difficult situation, at least with respect to epistemology and ethics. In both of those fields there is the expectation that a good theory can correct what we think about ethics or epistemology. For example, a particular epistemic theory might revise the way we collect evidence. But a purely descriptive approach can never do that. If a particular subject experienced the magic 8 ball as providing evidence then a phenomenological analysis of the sense of evidence for this individual would, indeed must, sanction the magic 8 ball as providing evidence (because what it is reporting on is the forms of this individual’s experience, and would get those experiences wrong if it did not report the magic 8 ball as evidence providing). Could we fix things at the cross-regional stage? No. Certainly at the cross-regional stage it might be noted that certain natural kinds (such as reliability) are correlated with many of the individual’s experiences of events as evidence providing, except for a few oddballs, such as the magic 8 ball case. But this mismatch does not imply that there is something wrong with the magic 8 ball case; rather it simply indicates that what is experienced as providing evidence doesn’t correlate with a single natural kind, which is probably true of many senses.

Clearly solving this problem is trickier. One way to go might be to lean on the expectations that we associate with certain senses. Evidence, for example, we expect to be highly reliable; i.e. to experience something as providing evidence is to develop some specific expectations about which future experiences we will and won’t have (i.e. that we will have those that agree with the evidence and won’t have those that disagree with it). Then, building on the correlation between our experiences of particular facts and the way things objectively are, we could say that, with respect to the magic 8 ball, it is not the case that those expectations will reliably be fulfilled. Thus we go cross-regional and back again to demonstrate the experiencing the magic 8 ball as providing evidence is, in an indirect way, inconsistent with the sense of evidence (i.e. that evidence as experienced by this individual in internally inconsistent). Unfortunately while this is a step in the right direction it is not a complete solution. First of all it doesn’t motivate revision in a particular direction; in the case discussed here both ditching the magic 8 ball as providing evidence and relaxing the expectation of reliability are ways to resolve the problem, and this approach does not prefer one to the other. Secondly it is not clear that we will always be able to make this move. Even with respect to evidence I can imagine problematic pathological cases. For example, in the magic 8 ball case our subject may believe that there are malicious demons who change the way things are whenever he checks up on the information the magic 8 ball provides. Thus in the case of the magic 8 ball he does not form the expectations that we were leaning on. And I don’t think that there is any way around this. Now if we had a prescriptive approach we might point out to the individual that he might as well stop thinking about the magic 8 ball as providing evidence, given that he never acts on that information (since he believes that the demons will interfere if he does). The function of evidence is to prepare us to take action, and so it is effectively the same as taking it not to provide evidence if the evidence it provides can’t be acted on. Thus, for the sake of a simpler epistemological theory, he should revise his opinion of the magic 8 ball. But we do not have a prescriptive approach.

Create a free website or blog at WordPress.com.

%d bloggers like this: