I compare and contrast Tyler Burge’s account of epistemic warrant (“Perceptual Entitlement” 2003, “Entitlement: The Empirical Basis of Epistemic Warrant” 2019), focusing on perceptual entitlement, with my account (“Epistemic Entitlement” 2012, “Functions, Warrant, History” 2014, “Normal Circumstances Reliabilism” 2017).

Burge’s account relies on representational functions. He argues a priori that every belief has the representation of function of being true, and every perceptual representation has the representational function of being accurate. He argues that they do not reduce, and I believe he thinks they cannot reduce, to biological functions.

He argues that for every telos (aim, goal or function) there are four natural norms associated with the telos. So for every function, there are four natural norms. These norms are explanatory levels of fulfillment or contributions to the fulfillment of a function.

For functions, the first norm is function for fulfillment. The second norm is functioning or operating as well as possible, given natural limitations and information available, on the occasion. The third norm is fulfilling the function reliably. The fourth norm is a combination of the second and the third.

He then derives representational norms from representational functions: (1) forming a true belief, (2) operating as well as possible, (3) reliably forming true beliefs, and (4) reliably forming true beliefs through operating as well as possible.

Burge then argues that epistemic warrant consists in fulfilling the fourth norm, in an explanatorily good way. That means the perceptual belief is formed through the normal functioning of the perceptual belief forming process that is reliable in normal conditions.

I argue instead that our perceptual belief forming capacities have the etiological function of reliably inducing true beliefs. It is empirical whether our belief forming capacities have this function. This kind of function is broadly biological.

I propose three norms that follow from this function. One natural norm is reliably forming true beliefs. The second is normal functioning. The third is reliably forming true beliefs through functioning normally in normal conditions. I argue that a first grade of epistemic warrant consists in normal functioning of the belief-forming process, if and only if the process has forming true beliefs reliably as an etiological function. A second grade consists in reliably forming true beliefs through functioning normally in normal conditions.

Both Burge’s account and my account deploy a historical condition in order to distinguish between accidental and non-accidental reliable success. I shall defend this shared requirement.

I shall level a criticism against Burge’s derivation of reliability norms for functions. This norm follows, I believe, only if it is a priori necessary that functional kinds reliably fulfill their functions in normal conditions when functioning normally. I do not believe this is so. And so Burge’s attempt to derive the standard for epistemic warrant from the representational function of belief falls short.