I hope I’ve made clear in my previous posts about Ladyman and Ross and their wonderful book All Things Must Go (here and here) that my critical remarks about them are to be understood as supportive and constructive.  I’m trying to buttress their position and make it stronger.  I’m on their side, as are certain other sympathetic critics, who have pointed out other problems with their approach, not directly related to what I’ve said in those previous posts.  Of particular importance, I think, is the critique by Kyle Stanford in a 2010 symposium on the Ladyman-Ross book in Metascience. Stanford points out that the concept of “structure” central to Ladyman-Ross’s structural realism serves several distinct purposes in their book: it is what remains continuous through theoretical transitions, for instance, and is also what explains the novel predictive successes of those theories. Stanford doubts whether a single concept of “structure” can do all these jobs. Ladyman and Ross candidly acknowledge in their response that this is a serious problem, to which they (as of then, five years ago, anyway) have no real answer. Well, I have a suggestion for them. They wouldn’t need such a single concept of structure if they weren’t attempting to cling to a minimal form of metaphysical realism. So: drop the realism — i.e. drop the metaphysics.  They claim their metaphysics is nothing but the “articulation of a unified world view derived from the details of scientific research” (p. 65; my italics), but what exactly in those details requires or entails an intuitive notion of “reality”?

Stanford makes a different suggestion; he offers the alternative of a broadly Quinean view which, he says, avoids Ladyman-Ross’s “two-step” approach of sharply distinguishing the results and methods of “science” from the results and methods of subsequent unification efforts (i.e. “metaphysics”); he offers Penelope Maddy’s Second Philosophy as a possible elaboration of such an approach. His emphasis on using all evidence, not just some restricted class of “scientific” evidence, in the unification project is certainly welcome, but why need a Quinean view be the only possible alternative to Ladyman and Ross?

Two responses come to mind. First of all, the Quinean view suffers from a problem similar to the one I diagnosed in Ladyman and Ross; in fact the problem is worse in the Quinean case as there is no acknowledgement of a “stance” at all. At least Ladyman and Ross acknowledge that their assertions are to be understood as stance-relative, but in Quine (or Maddy, or others who take this line) stances make no appearance, nor is there any room for them. Our mother tongue, once scientifically regimented, becomes the canonical conceptual scheme. Other schemes are possible, but for Quine there is no way of stepping outside the scheme we are in to make choices — let alone bring rational considerations to bear on choices — among schemes. Metalinguistic or philosophical considerations must themselves, that is, be stated within the terms of the scheme we are already in. We can rise above it by semantic ascent, to be sure, which enables us to refer to words and linguistic artefacts just as we can use those words to refer to objects in the world. But whatever we say in the meta-language must be part of the scheme canonically specified by the scientifically regimented object language. “This ascent to a linguistic plane of reference,” said Quine, “is only a momentary retreat from the world, for the utility of the truth predicate is precisely the cancellation of linguistic reference. The truth predicate is a reminder that, despite a technical ascent to talk of sentences, our eye is on the world.” (1970, p. 12)

But second, and more importantly, why should we narrow our options to this degree? Why should we follow Williamson in considering only those options that take metaphysical problems at face value? Isn’t taking philosophical problems at face value precisely the kind of “domestication” of counterintuitive ideas that Ladyman and Ross criticize so relentlessly in their first chapter? Why on earth should the “face value,” i.e. the immediate intuitive import, of philosophical problems correspond to anything out there? Why, in particular, should the very questionable intuitive idea of some sort of ultimate or bedrock “reality” be such a shibboleth?

This is especially odd in the case of Ladyman and Ross because they devote considerable ingenuity and detailed scientific argument to their case that the most intuitively appealing instances of “reality” — i.e. “individuals” or “things” in the traditional sense — are a perspectival illusion. A “thing” within one mode of discourse dissolves into something less substantial within a more general perspective. So the identity of any “thing” is discourse-relative. Moreover, they reject reductionism, so there is no reason for them to think of discourse-relativity as merely scale-relativity (which is how they tend to talk). So why, despite all this, do they insist on “reality”?

Their project of chastening metaphysics without consigning it to the flames entirely recalls Kant, and their structural realism even more closely resembles the scientific neo-Kantian project of Helmholtz and Hermann Weyl, who tried to explicate the Ding an sich in something like structural terms. Our knowledge of reality, said Weyl (as opposed to our subjective acquaintance with the phenomenal world), could only ever be up to isomorphism; Poincaré, by the way, said something quite similar.  This isn’t exactly what Ladyman and Ross have in mind, but it would have been very informative to situate their project in the context of these illustrious predecessors, and it’s a little odd that they don’t even acknowledge them.

It was left to Carnap to just say no, and to give up on any sort of Ding an sich even in that rarefied form, and to foreswear the intuitive tug of “reality” altogether. He saw no need for it, and nothing in science seemed to require it. He was neither a realist nor an instrumentalist, let alone an anti-realist; he had no “ontological position” whatever. What Ladyman and Ross (and Quine, for that matter) call metaphysics, he called language engineering, and classified as “pragmatics.”  Carnap would have agreed with Ladyman and Ross that this is not (yet) an “an activity that has a specialized science of its own.”  Of course not; it was just a program. Carnap’s graduate student Howard Stein enthusiastically embraced the task of building this specialized science, but soon realized that to begin work on pure pragmatics, you need some descriptive pragmatics as a basis; this is how he arrived at his particular brand of philosophical history of science.

Howard, meanwhile, is also pretty relaxed about the use of the word “metaphysics,” used approximately in Aristotle’s sense, which is how Ladyman and Ross also use it (as do Quine and Maddy). So I want to reiterate that it’s not their use of the word “metaphysics” I’m objecting to, it’s the idea that the task of metaphysics (or pragmatics, or whatever) is a task of inquiry (as their “scientistic stance” demands) rather than an engineering task.  I’m thereby also objecting to the idea that we have no choice about the conceptual framework we use for our knowledge, or about the “stance” we adopt toward it (in van Fraassen’s sense).  I’m with Dick Jeffrey, who said that Carnap’s “persistent, central idea was that . . . it’s high time to take charge of our mental lives, to recognize that there’s nobody out there but us to choose our purposes and concepts to serve those purposes.”