Burge on Perceptual Systems and Veridicality

Often when a philosopher says that something is obviously true or true apriori, it is a good idea to pause and pay careful attention. Frequently, such claims conceal many weighty assumptions; assumptions one would do well to make explicit. In his essay “Perceptual Entitlement”, Tyler Burge takes as one of his fundamental tenets the claim that there is an apriori connection between the representational function of an organism’s perceptual system and verdicality:

I take it as obvious that it is known apriori that the central representational function of a perceptual system is to perceive. This function is apriori associated with a representational function (to represent veridically).

This, of course, does not amount to the implausible claim that an organism’s perceptual system is always successful in carrying out its representational function—i.e., that perceptual experiences are always veridical. Rather, Burge is making the highly plausible claim that it is a conceptual truth that an organism’s representational system aims at veridicality. However, as plausible as this claim is, I believe more needs to be said in its defence. In this post I will summarise Burge's views and in my next post on this topic I will present one objection to Burge's account.



The claim that representational systems aim at veridicality is central to Burge’s account, which is unapologetically teleological:

I assume that there are certain functions, ends, and commitments, which bring with them goods for animals and their subsystems….Many ends, goals, and functions can be established as such on biological grounds. I take it that survival, at least long enough to have offspring or to fulfil some other biologically basic functions, is an end for all animals—an end that can be established biologically.

I will refer to the practical and/or biological functions of a particular system, such as promoting survival and the ability to pass on one’s genes, as biological fitness. It is widely agreed that animals and their subsystems were all “designed” by natural selection. I will refer to this claim as the Darwinian assumption.

Burge points out that biological fitness is not the only type of end an organism or subsystem may have. Specifically, when it comes to the perceptual subsystems of higher animals, Burge distinguishes between the practical function of the system and its epistemic function. The former has to do with the perceptual system’s contribution to the overall goal of the organism—namely, biological fitness. However, the latter has to do with the perceptual system’s representational function—namely, to represent veridically—and the role this function plays in achieving the supreme epistemic end of forming true beliefs.

Burge impugns attempts to reduce the epistemic and representational function of the perceptual system to the practical or biological: There are those who ignore or attempt to explain away the representational functions of perceptual and belief forming systems. They see biological or practical functions as the only relevant ones. The good of belief is judged purely relative to such functions. I think such views are clearly mistaken. Burge acknowledges that there may be a “non-accidental” connection between a system’s practical and representational functions.” Typically, a system that represents the world veridically (i.e., the epistemic function) also promotes biological fitness (i.e., the practical function). However, Burge points out that “being true is not in general being useful.” Moreover, it is at least conceivable that a system that regularly gave rise to false beliefs may nevertheless promote biological fitness. Burge cites the example of rabbits, whose representation of “danger” he describes as highly unreliable “because of a dominance of false positives.” In the rabbit case, it prima facie seems as though natural selection has favoured representational systems that are unreliable.

If Burge’s assessment of the representation of danger in rabbits is correct, then we have good empirical evidence to suggest that a creature with an unreliable perceptual system may nevertheless be biologically fit. While I agree with the empirical claim, about which I will have more to say later, for the time being I wish to make a weaker claim. I wish to suggest that it is at least possible that a system may promote biological fitness even though it is not reliable (in the sense of representing veridically). In other words, even if it were to turn out that no creatures on earth were biologically fit despite being unreliable, there is certainly some possible world in which this is the case. Thus, at the very least we should be able to endorse the claim that there is no necessary connection between a system promoting biological fitness and aiming after veridicality. I will refer to this claim as the conceptual independence thesis (or CIT). Since all actualities are necessarily possibilities, but not all possibilities are necessarily actualities, CIT is significantly weaker than the position Burge himself seems to defend, the latter relying as it does on actual real-life examples. Consequently, CIT is a fortiori a claim Burge would endorse.

In my next post, I will attempt to show that CIT, when combined other claims Burge makes, presents a challenge to Burge's claim that there is an apriori connection between the representational function of an organism's perceptual system and veridicality.