Barry Stroud’s Critique of Naturalised Epistemology

It may be worth putting Barry Stroud’s (1935–2019) critique of naturalised epistemology in some kind of context. That’s partly because this is something that Stroud himself did when he tackled philosophical problems and issues.

So here are a couple of passages from his obituary — as published in Berkeley News (a publication of the University of California):

“While best known for his work in epistemology and philosophical skepticism… Stroud’s overarching legacy, his colleagues say, was his ability to see the big picture and get to the heart of philosophy…
“As a philosopher, Stroud came of age during a time when the prevailing Western attitude was that philosophical questions could be answered by the natural or social sciences, and he challenged those ideas...
“‘One might say that, while everyone else was philosophizing about consciousness, reality and knowledge, he was philosophizing about philosophizing itself,” [Kolodny] added.
“‘Barry single-handedly brought philosophical skepticism — which gives reasons to doubt whether we can know even the most ordinary things about the world around us — back to the center of philosophical discussion,’ Bridges said.”


In his ‘The Significance of Naturalised Epistemology’ (1981), the Canadian philosopher Barry Stroud offers a realist point against W.V.0. Quine’s naturalist position on epistemology. He makes the simple distinction between

1) Beliefs which are “constructions or projections from [sensory] stimulations” [Quine’s position].


2) Beliefs about the world.

Stroud’s question is: How do we get from 1) to 2)? Or:

How do we get from sensory stimulations — which are (it is assumed) caused by the world — to truths about the world or representations which are taken to be true of the world?

This is the ancient sceptical question. Indeed, at least since Descartes, this was once the central question of epistemology.

Another way of putting all this is in terms of what and how we know.

On Quine’s account, all we know about is are sensory stimulations and what we assert in response to them. This means that we don’t know anything about the world itself. (A locution which Quine would, no doubt, have rejected.) Or, as Stroud puts it, from this

“we would not in addition have independent access to the world they are about on the basis of which we could determine whether they are true”.

If what Stroud is saying is correct, then we can’t even say that sensory stimulations — and the causally resultant assertions — are “about the world” if they aren’t true of the world (or give us knowledge of the world) in the first place.

So are beliefs and assertions about our own and other people’s beliefs and assertions not directly (or even indirectly) about the world? These beliefs and assertions (or Quine’s “projections”)

“could not be seen as a source of independent information about the world against which their own truth or the truth of the earlier beliefs could be checked”.

This is as strong a statement of the possibility of scepticism in epistemology as you could hear from a sceptic himself. That is, we can compare beliefs against beliefs; though we can’t compare beliefs (or their contents) with the world (or its parts) itself. The American philosopher Donald Davidson (1917–2003) would even argue that we can’t compare beliefs, etc. with sensory stimulations because there are no belief-free sensory stimulations “which could count as evidence” in the first place.

All this can be (partly) boiled down — or is analogous — to Bishop Berkeley’s well-known statement that an “idea” can only be like another idea and not like what it’s an idea of in the world. Stroud, then, is referring to the traditional epistemological problem of “bridging the gap between sense data and bodies” (Quine, quoted in Stroud). However, whereas traditional epistemologists saw this as a problem, the logical positivists saw it is a “pseudo-problem”. Or, as Quine put it, this move between sense-data (or Quine’s “sensory stimulations”) and bodies is “‘real but wrongly viewed’” (Quine, quoted in Stroud).

In terms of this (logical) gap between “data and bodies”, Quine neither saw it as a pseudo-problem nor saw it, as Kant did, as the greatest failure of philosophy. Quine’s

“positive account does not try to show how we rule out the possibility that the world is completely different in general from the way our sensory impacts and our internal makeup lead us to think of it”.

If Quine did believe this, then we can argue that he might as well have believed that the problem of the external world (or even its existence) is a pseudo-problem if the sceptic or realist doesn’t “try to show how we rule out the possibility” that the world may be different to what we think it is. Perhaps, then, there’s a logical gap between data (or evidence) and the world. If there is, then why didn’t Quine think that such problems are pseudo-problems as the logical positivists did? Or as Stroud puts the logical positivist (or verificationist) position:

“The traditional epistemological question of the reality of the external world and our knowledge of it was for Carnap and Schlick and other verificationists a meaningless pseudo-question; no answer to it was empirically confirmable or disconfirmable.”

We can’t get between our evidence (or data) to get directly to — or at — the world in order to match the evidence (or data) with the world. Hence the logical gap. Yet physicists - or at least many of them - had always accepted that they must rely on evidence (or “phenomena” in the case of Kantian scientists like Einstein and Mach in the early 20th century). And because physics had the last (as well as first?) word on the world or nature, and if physicists happily accepted the importance of sense-data (or Carnap’s “cross-sections of experience”) when it came to world-talk, then the logical positivists did so too. Thus, as a consequence of this “deference” (Quine’s term) to physics, Stroud comments that

“[f]or Carnap we must distinguish a philosophical (pseudo) employment of a form of words from an ordinary or scientific employment of the same words”.

We mustn’t talk about the world or reality (its nature or existence) in the way the sceptics or epistemologists do. Instead we must speak as physicists (or even laypersons) speak. However, if we take the former option, then we’ll basically be talking rubbish.

In terms of Quine’s position again.

What’s the point of accepting the possibility that the “world [could be] completely different” if Quine didn’t offer us a way out of this problem? Again, if the gap is logical, then he must surely have seen the problem as a pseudo-problem. However, if all we have are “sensory impacts” and a largely given “internal makeup”, then we must surely see why Quine took the (pragmatic?) position which he did take.

Despite Donald Davidson’s criticisms of Quine’s emphasis (or very position) on sensory stimulations (hinted at earlier), Stroud argued that Quine’s position isn’t like that of the sense-data theorists (or British Empiricists/ phenomenalists). What Quine doesn’t do, even with his sensory stimulations, is try

“to isolate a domain of pure sensory data evidentially or epistemically prior to the knowledge of nature that is to be explained”.

So perhaps Quine’s position was midway between the “atomism” of sense-data theorists and Davidson’s holism. Thus:

(1) The sense-data theorists’ position: Sense-data are untouched by belief and theory.
(2) Davidson’s position: Beliefs are (clearly) touched by (other) beliefs and theory.

And , perhaps, the happy medium:

(3) Quine’s position: Sensory stimulations are touched by belief and theory but which nevertheless must ground (future) beliefs and theories.

Stroud’s perspective seems to depend on accepting a very controversial theory of metaphysically-realist truth. (Putnam once argued that, in terms of scientific truth, Quine was himself a metaphysical realist who, for example, accepted the principle of bivalence for the statements of physics— see here.)

Stroud puts this realist position by — once again — questioning Quine’s exclusive reliance on sensory stimulations and the resultant “projections” (or “posits”) we make “about the world” because of such sensory stimulations. He asked Quine this simple question:

“[How do the] subject’s ‘projections’ or ‘posits’ turn out to be correct, and not just a question about how he comes to make them [?]”

If the (to use Richard Rorty’s phrase) “world is well lost” (or if we don’t have direct access to the world), then how do we know which projections (or posits) are correct and which are incorrect without the (as it were) world telling us so? (Rorty and Davidson would say that the world can’t tell us anything — not even metaphorically or indirectly.) How does the Quinian decide which posits (or projections) are correct and which ones are incorrect? Are these decisions made exclusively on pragmatic or instrumentalist lines?

Of course Quinians can’t only be concerned with “how he comes to make” these projections (or posits) because some people come to make such projections about goblins or the influence of ley lines. So there’s more to the Quinian story than (mere) projections or posits. And, according to Stroud. that something extra is causality (or causation),

Stroud puts this rather simple causal approach to knowledge this way. He says that we

“would see that the world around [the investigator or epistemologist] is generally speaking exactly the way he says it is and that its being that way is partly responsible for his saying and believing what he does about it”.

This is certainly largely Davidson’s position and also the reason why he argued that “most of the beliefs in a coherent set of beliefs are true”. That is, we wouldn’t say (or believe) what we do about the world if the world wasn’t (as it were) responsible for what we say (or believe) about it. That relation (or connection) between belief and the world is largely accounted for in terms of causation. (According to Davidson, “causation does not come under a description” and it’s not in itself “explanatory”.) This causal (for want of a better word) position may seem simple and even a little naïve. Stroud writes that

“[m]any philosophers nowadays would hold that that is enough for knowledge: the subject believes that p, he is right, and it is no accident that he is right”.

So say that Jeff believes that P because the world is as he says it is. That is, the world causes him to believe (or say) that P in a causal-kinda-way. Indeed this almost has the appearance of being some kind of isomorphic relation between the world and what Jeff says (or believes) about it. It’s no surprise, then, that Stroud concludes by saying that the “adequacy of any such ‘causal’ account of knowledge is still questionable at best”.

And because of everything that’s just been said about Stroud’s account (i.e., that we essentially loose the world on Quine’s alternative), then we must also accept “that countless ‘theories’ could be ‘projected’ from the sensory impacts we receive”. Yet that’s no surprise at all because Quine himself admitted that. Indeed Quine was well known for stating the following: All theory is underdetermined by the sensory evidence. (All this is also part of the story of ontological relativity, the indeterminacy of meaning and the inscrutability of reference.) And that’s precisely why Quine also believed that we must employ “pragmatic” requirements and judgements when it comes to theory choice.

Stroud puts all the above less positively. He argued that because of this theory-pluralism (or theory-liberalism),

“if we do happen to accept one such ‘theory’ it could not be because of any objectively discoverable superiority it enjoys over it competitors”.

The theory we choose won’t be a truer account of the world. It won’t give us (so to speak) more of the world. It will only be pragmatically (or instrumentally) superior (to us) than the other theories. It won’t be truer or even more correct. It won’t have been chosen because something has been (as Stroud puts it) “objectively discovered” which places it in a superior position — i.e., in terms of truth rather than (mere) pragmatic utility (or whatever). That said, according to Stroud, Quine did accept an objective component to his alternative position. Yet that objective component is only the “meagre [sensory] data” which isn’t itself the world (as well as not really data or evidence on Davidson’s position). Even this pseudo-objective component doesn’t amount to much because (as already stated) this same objective data (or evidence) can be used to construct many competing, complementary or — sometimes - even contradictory theories.


Stroud, Barry, ‘The Significance of Naturalised Epistemology’ (1981)

[I can be found on Twitter here.]

13 views0 comments