118 Chapter 5
Do We Really Care Whether Our Beliefs Are True
will decrease our percentage of TRUE" (and/or TRUE**, and/or TRUE*** ... ) beliefs. If we really value true beliefs, presumably we won't much care about giving up TRUTH* or TRUTH**. But do we? That is the question we are now ready to tackle.
5.5 True Belief, Intrinsic Value, and Instrumental Value
As we saw in 4.6.3, discussions of epistemic value can be conveniently organized under a pair of headings: intrinsic value, the sort of value that something has for itself, and instrumental value, the sort that something has in virtue of leading to something else. Let's begin by asking whether we should view the holding of true beliefs to be intrinsically valuable.
It might be thought that on the issue of intrinsic value very little argument is possible, since if a person tells us that she accords intrinsic value to having true beliefs, she is saying that she values them for their own sake and not because they are conducive to something else. Since she does not claim that true beliefs will facilitate the achievement of other values, we can hardly argue that her expectations about the consequences of believing the truth are mistaken. However, as we saw in the previous chapter, there is another kind of consideration that might be efficacious in persuading someone that she should not, or does not really, accord intrinsic value to the having of true beliefs. Rather than dwelling on the consequence of having true beliefs, we can try to be sure that she sees clearly the real nature of what she values-that she appreciates what having true beliefs comes to. And it is just here that the conclusions of the previous section can be brought to bear. What we saw there was that if the function pairing beliefs with their truth conditions is the one outlined by the causal/ functional theory, then it is both partial and idiosyncratic. And, in rather different ways, both of these facts entail that valuing true beliefs is a profoundly conservative thing to do.
Consider first the fact that the interpretation function has a very limited domain. What this means is that there is a huge space of possible systems of mental computation and storage whose component states have no truth conditions and thus cannot be true. No doubt many of the systems in this semantics-free space are useless and chaotic. But there is certainly no reason to suppose that they all are. A much more likely possibility is that in this huge space there ar systems that would vastly increase their user's power or happiness or biological fitness, systems that would lead to substantial reduction in the amount of suffering in the universe, and systems that would significantly reduce the probability that we will bomb ourselves into oblivion along with much of the biosphere. As Paul Churchland is fond of observing, our current cognitive processes are a tiny island in a vast, unexplored computational space, a space that we may reasonably suppose to contain riches beyond imagining. But almost all of that space is beyond the reach of the causal/functional interpretation function; it is a domain in which there is neither truth nor falsity. Those who would accord intrinsic value to the holding of true beliefs may well be reluctant to explore that vast space and will resist adopting what may be found, since we know in advance that it contains no true beliefs. But theirs is a profoundly conservative normative stand. For what they value in the end products of cognition must be semantically interpretable, and what is semantically interpretable cannot depart too radically from current patterns of reasoning or from familiar ways of causally tying mental states to extramental reality. To value true belief is to resolve that in matters cognitive, one will not venture very far from where we are now.
One further image may help to make the point. The vast space of mental storage and computation systems might be thought of as a surface on which peaks are occupied by systems that excel in some virtue, like increasing their user's power or fostering happiness or increasing biological fitness, and the valleys correspond to systems that do poorly in that regard. If we insist that the system we adopt. must be in the limited domain of the causal/functional interpretation function, the best we are likely to do is find a local maximum-a modest peak lying relatively close to our current position. There may well be vastly higher peaks farther afield. But if we take truth to be precious, we will not find them.
The conservatism entailed by the idiosyncratic nature of the interpretation function is of a rather different kind. There are endlessly many functions mapping mental states to truth conditions (or propositions or possible states of the world). In this bristling infinity of functions there is one that is singled out by common sense as providing the "real" truth conditions for mental states, in contrast with the TRUTH* CONDITIONS, the TRUTH** CONDITIONS, and all the other variations on the theme. But if, as we have been assuming, it is the causal/functional interpretation function that is sanctioned by intuition, then it is not a particularly simple or natural function. Rather, it is something of a hodgepodge, built from a more or less heterogeneous family of strategies for fixing the reference of terms and another family of strategies for transmitting reference from one speaker to another. What distinguishes acceptable groundings and transmissions is not that they share some common natural property but simply that they are found to be acceptable by commonsense intuitions.