Phenomenology as a Holistic Critique of Positivism

Phenomenology began as a part of positivism‹Wittgenstein used the word to refer to his search for the atomic facts and basic objects positivism claimed grounded reality. The problem with positivism, as Wittgenstein came to realize, is that it tried to create a theory of our experience that was effectively detached from the pragmatics by which we interact with our experience. The idea here is that when we encounter a chair, we do not encounter it as a set of materials put together, or a set of human comforts put together, we encounter it as a chair. As Wittgenstein put it: "But what are the simple constituent parts of which reality is composed?...It makes no sense at all to speak absolutely of the 'simple parts of a chair'" (Dreyfus and Dreyfus 27).

The hope of AI, especially in the late 1960 and early 1970's, was to specifically figure out what the facts and rules of the world to be represented were. The idea was to deal systematically with small, isolated domains of knowledge in "microworlds," as Marvin Minsky and Seymour Papert put it. Gradually, these microworlds would become more and more realistic and more broad until they approached real-world understanding. The problem is, in essence, that these microworlds represented what Heidegger called a "universe" instead of a "world." A "universe" is a set of interrelated facts, but this does not a world make; a world, on the other hand, is a domain in which human actions have meaning and make sense. This is a difficult distinction, but think about it this way: the set of rules and facts for how, say, the theater works do not by themselves have any meaning. Although they may contain information about people pretending to be other people, stages, sets, and audiences, they cannot address why there is theater. I don't mean this in the sense of "why we have art," but rather "that we have art"--that for whatever reason, the theater has meaning for humans. A similar analogy that avoids the "art" question is that of physics. The physical universe, by itself, does not contain meaning; photons and the force of gravity are not by themselves meaningful. However, the discipline of physics has meaning‹humans are driven by a desire to understand the physical world, humans use their discoveries to create technology, and so forth. Microworlds are isolated domains without meaning: knowing all the laws of physics in the universe would not enable a computer to understand human interaction with physics.

The general answer to this by the AI community is that the problem of discerning which knowledge is meaningful in the real world must be solvable, since humans have solved it. One answer to this that Heidegger proposes is that what humans possess is not necessarily a set of general rules about meaning but a set of very specific responses to very specific special cases.

The advantage of the connectionist scheme is that it does not make any claims as to the workings of the mind; it instead claims that since we know that somehow the physical brain instantiates the mind, we can try to create a mind by creating a brain, or something like one. This holistic approach is starting to prove very powerful (although it has not been around long enough to evaluate), but also very scary: if connectionism works, it seriously questions the basic notion of philosophy that there must be a theory for every aspect of reality. As a result, some have referred to the connectionist approach as "non-philosophical." This raises a question that we cannot yet answer. Western thought has always been dependent on philosophy for its ethics; although most people no longer believe that a complete ethics can be deduced and written as a set of rules, the way we make decisions in our lives has always been based on a philosophical premise. If connectionism calls Western philosophy itself into question, what will become of ethics? Can we concrete an ethics based on a practice that has no immediate philosophy behind it?

back