Anti-naturalism in philosophy (II)

Here are a few more notes on the debate between Paul Horwich and Timothy Williamson which I discussed briefly in my previous post on this site.

In the video, Horwich defends his view, based on his reading of Wittgenstein, that much of what goes on under the name of (analytic) philosophy is "irrational". (Williamson calls a spade a bloody shovel and interprets him as saying that it is "rubbish".)

Horwich starts talking about 39 minutes into the video, and a brief statement of his position starts at about the 42-minute mark. His key contention is that the common philosophical supposition that abductive reasoning from intuitive data can reveal fundamental regularities (more or less as science can be seen to reveal fundamental regularities which underlie empirical data) is simply false.

He summarizes the basic argument he makes for his position in five points. Here is my summary of his summary:

1) Our concepts (I assume he means linguistically-based concepts) tend to be messy. They evolved to deal with a wide range of real-world contingencies and are used in a bewildering variety of ways. The intuitive data upon which philosophical reasoning is based is inextricably bound up with such concepts. In fact, many of the key concepts with which philosophers have traditionally been concerned are particularly prone to conflicting understandings and interpretations.

2) Via conjectured theoretical entities, models, etc., the various sciences have developed effective ways of dealing with messy empirical data to arrive at (often) elegant and relatively simple theories which reveal fundamental regularities.

3) The currently dominant strand of analytic philosophy also seeks elegance and simplicity, but any simplicity obtained is generally obtained "on the cheap" by finding a simple regularity that fits most of the (intuitive) data, and counting intuitions or interpretations which do not fit the theory as mistaken or incorrect.

4) This is a distortion of the scientific method.

5) Not all philosophers make this mistake, but a beautifully simple theory will rarely be obtainable. Almost inevitably there will be a profusion of alternative theories and no prospect of convergence. Competing theories of ethics are an obvious example.


My response...

Horwich is basically correct about the problems of philosophy as a theoretical discipline. And I believe his views on this more or less accurately reflect Wittgenstein's views.

But there are a couple of aspects of Horwich's thinking with which I am uncomfortable. Interestingly, these features are also the very features of Wittgenstein's thinking that I reject.

First, the 'scientism' issue. (Horwich uses the word.) Now, it seems to me that the term can used in at least two ways, one quite focussed, the other broader and stronger.

In the narrower sense, the term is used to highlight the inappropriate application of scientific methods or approaches. Horwich, for example, sees philosophy as inappropriately imitating science. I agree with him. This sort of thing happens and it would be better if it didn't. Philosophy (however we understand it) is not science. (Similarly, social science is not physics.)

But 'scientism' also has a broader meaning. On this view, a scientistic outlook involves having a high regard for science coupled with skepticism about other ways of gaining (anything other than commonsense and practical) knowledge.

This broader sense of 'scientism' (which I would embrace) marks a divide between profoundly different views of the world; and, because there is no agreement on basic assumptions across the divide, there is no way of satisfactorily dealing with such differences via philosophical or ordinary reasoning and argument. One set of assumptions or presuppositions will lead to certain philosophical opinions and lines of argument; another set will lead elsewhere.

Closely related to this scientism issue is the question of naturalism. Horwich rejects naturalism, which he sees as the view that every property, object and fact is naturalistic.

Mathematics is often said to involve objects, etc. which exist but do not exist in time and space. If so, then naturalism (as Horwich defines it) is incorrect. This leaves the door open for the similarly real existence of moral properties, for instance.

I haven't looked yet at Horwich's views on ethics. But I am aware of Wittgenstein's (religion-based?) views on these matters.


I am not going to try to deal here with Timothy Williamson's point of view. To do so would mean addressing topics like the dreaded Barcan formula (which was mentioned by Horwich in the video, by the way). The Barcan formula is an axiom (or schema) of quantified modal logic first stated by Ruth Barcan (who became Ruth Barcan Marcus) the acceptance of which apparently enhances and simplifies the workings of formal systems of quantified modal logic. Williamson defends the Barcan formula as being in some sense 'true', and, despite the fact that there is no general agreement on its informal interpretation, seeks to draw metaphysical conclusions from it – something along the lines that everything exists, but some things as actual objects and some as possible objects. This sort of thinking strikes me as being more in line with medieval scholasticism than with modern scientific thinking (and indeed in his writings Williamson refers to Avicenna, claiming that he informally anticipated both the Barcan formula and its converse).

There have been big advances in formal logic over the last century or so, and it would not be surprising if such advances allowed us to see certain general ideas in logic and traditional metaphysics in a new light, validating some old approaches as insightful or prescient and undermining others. But what Williamson is doing strikes me as going well beyond such modest, historical commentary. Though he equivocates about naturalism in this discussion with Horwich, backing away from Horwich's claim that he, like Horwich, is opposed to a naturalistic view of the world, he does seem, in effect, to be rejecting a modern, scientific outlook and attempting to resurrect something like traditional metaphysics.

I am not saying his motivations are religious: they may be entirely intellectual. The realm of formal logic, like the realm of mathematics, has its attractions: it can appear very appealing, especially when contrasted with the messiness and general unsatisfactoriness of mundane reality.

Horwich accepts the messiness, however, and here I am very much on his side.