Stalnaker against internalism
Stalnaker holds a combination of views that seem independent to me, but closely connected to him. One is a kind of reductive naturalism about intentionality. On this view, the point of attributing beliefs and desires is to give a high-level characterisation of the subject's behavioural dispositions, their functional architecture and their causal relations to the environment. Another of Stalnaker's views is externalism about mental content. This says that intentional characterisations are relational: even when two subjects are perfect intrinsic and functional duplicates, they may still differ in their beliefs and desires, depending on what objects and properties they are causally related to.
In his new book, Our Knowledge of the Internal World, Stalnaker acknowledges that one could in principle combine naturalism with internalism (fn.5, p.114), but he maintains that the former somehow 'motivates' the latter.
An example. Three microbes inhabit three separate Petri dishes. Oscar and Twoscar live on monosaccharides such as glucose and fructose. Oscar's dish contains glucose, Twoscar's fructose. The third microbe, Invoscar, is a bit more complex. Invoscar lives on disaccharides, but only if there are monosaccharides somewhere in the wider surrounding. Since Invoscar's dish contains lactose (a disaccharide) and is located right next to Oscar's, this condition is satisfied. All three are doing well.
Following Stalnaker, we can attribute proto-intentional properties to our microbes. For instance, they need certain facts to obtain in the sense that they would die if those facts didn't obtain. And their states indicate various facts in the sense that those facts are suitably correlated with their states. These notions can be made more precise, but I won't do so here, because that is what is at issue between internalists and externalists.
Externalists about need and indication hold that Oscar shares some proto-attitudes with Invoscar, but none (or no interesting ones) with Twoscar: both Oscar's and Invoscar's well-being indicates that there is glucose in Oscar's dish, and they both need it to be the case that there is glucose there. Twoscar's well-being, on the other hand, indicates that there is fructose in Twoscar's dish, and nothing at all about Oscar's dish.
Internalists hold that Oscar and Twoscar need the same thing: they both need monosaccharides in their surroundings. Likewise, their well-being indicates the satisfaction of this need. It is Invoscar who needs and indicates something else: the presence of disaccharides nearby and the presence of monosaccharides a bit further away.
The internalist classification tracks structural differences between the organisms, and thereby differences in capacities and dispositions. In these respects, Oscar is like Twoscar, and unlike Invoscar. If, for instance, Oscar were moved into Twoscar's environment, he would respond just like Twoscar, and not at all like Invoscar. The externalist classification ignores these differences and instead tracks the nature of the actual causes of the relevant states, described from a neutral, uncentered perspective.
Who gets it right? It depends on what the intentional classification is supposed to do. For the purpose of a belief-desire psychology whose aim is to explain people's actions, abilities and dispositions in terms of their internal states, the internalist classification appears better. The externalist classification seems to ignore differences in perspective -- the difference between you and me that explains our different reaction when we both see that I'm about to be attacked by a bear. And it ignores differences in how things are presented to us -- the difference between learning that Jack the Ripper murdered Annie Chapman and learning that Aaron Kosminski murdered Annie Chapman. And it introduces distinctions that are irrelevant to any behavioural dispositions, such as the distinction between Oscar and Twoscar.
Stalnaker tries to patch these problems for externalism. (More on that hopefully later.) But what are his objections to internalism? As far as I can tell, they fall into two classes.
The first line of argument is that internalism gives the wrong result about agreement, disagreement, and aboutness. The idea is that Oscar's state is about glucose whereas Twoscar's state is about fructose, while the internalist assigns them the same content. Similarly, Oscar's and Invoscar's state agree insofar as they both indicate the presence of glucose in Oscar's dish, while the internalist assigns them different content.
I'm not sure where these alleged facts about agreement etc. come from. The point of attributing intentional states, I thought, was to give a high-level characterisation of actions, abilities, dispositions etc. These fundamentally non-intentional facts are what we have to get right, not people's intuitions about agreement.
It's true that the internalist characterisation of mental states is not our everyday characterisation, nor does it directly validate our everyday judgements about agreement etc. However, I doubt that the externalist characterisation fares much better in this respect. Last year, I wrote about simple externalist accounts of agreement and offered an internalist alternative that I argued goes better with our everyday judgements. I think the same holds for aboutness intuitions and everyday attitude reports. If we want a systematic theory of belief and desire that explains people's actions and dispositions, fits the biological facts, and directly vindicates our everyday talk, we may be asking for too much.
That was the first of Stalnaker's objections. The second is somewhat more ominous. Internalism, he suggests, presupposes certain evil Cartesian doctrines: that the real content of our attitudes is somehow purely phenomenal, or that our knowledge of the world rests on direct and infallible acquaintance with our internal states, or something along the same lines. Indeed, when in earlier works, Stalnaker discusses internalism, he often discusses phenomenalism, as if the two were one and the same.
How could that be? Shouldn't it be uncontroversial that the internalist classification is at least tracking something -- the dispositions, capacities and internal states Oscar and Twoscar have in common? At what point does the internalist commit herself to any controversial doctrines?
I guess Stalnaker imagines that on closer inspection, the individualist classification runs into a problem when it tries to say what exactly Oscar's state indicates. The individualist says the state indicates the presence of monosaccharides rather than the presence of glucose because it is not sensitive to the difference between glucose and other monosaccharides. But isn't it also indifferent between monosaccharides and stuff that is disguised as monosaccharides? And what if an evil demon put Oscar directly into this state? The state appears to be insensitive between monosaccharide situations and such demon situations. So its real content is something much wider than the presence of monosaccharides. The only possibilities ruled out are possibilities where Oscar is not in this very state. On the level of perception and belief, this does resemble a kind of phenomenalism on which the real content of our perceptions is that we are in a certain phenomenal state, from which everything else is fallibly inferred.
But this Cartesian conclusion doesn't follow from internalism alone. It follows from internalism only together with the assumption that the ('real') content of Twoscar's state is something that it indicates infallibly. If we don't want Cartesian doctrines (I kinda like them), we can simply drop this assumption. We could, for instance, restrict the relevant cases to a hand-picked class of five or six common situations and classify states by their sensitivity to the differences among them. If glucose is the only monosaccharide in those cases, Oscar's state will come out as indicating glucose, but the classification is nevertheless internalist: Twoscar's state also indicates glucose, for he responds in the same way to the six cases. A more sensible characterisation may include more situations, while still excluding ones with evil demons. Like Stalnaker, we can leave it to the context of attribution where the line is to be drawn. And we can even reject the idea of a fixed space of absolutely all possible situations. Stalnaker's Anti-Cartesianism seems to combine just as well with naturalism and externalism (his preferred combination) as it does with naturalism and internalism.