Some, such as Alvin Plantinga (former Professor of Philosophy at Notre Dame University, now Professor of Philosophy at Calvin College), think that there are problems between notions of human cognitive reliability and a standard account of evolutionary development. The reasoning goes something like this:
1. Assume that a typical evolutionary account of the development of human cognitive abilities is true.
2. Evolution works by selecting things that allow for organisms to survive and reproduce successfully.
3. There seems no necessary (or even probabilistic) connection between this and cognitive abilities that tend toward truthfulness.
4. Therefore, we have no warrant for believing that the cognitive processes by which we arrive at (among other things) theories of evolution are reliable.
5. Therefore, if a typical evolutionary account is true, it undermines its own warrant for belief.
That is, the ‘universal acid’ of Darwinism eats away at its own epistemic foundations.
How should a Darwinist respond?
To see why there might be a connection between belief (map) and reality (terrain), consider a robot navigating a terrain. If the internal representation that the robot had of the terrain were random, it probably would not be able to navigate the terrain optimally. Rather, a specific kind of representation of the terrain, coupled with a certain interface between the representation and the world, are required for the robot to navigate accurately. Applied to the biological world, navigating accurately is vital to finding resources, and so on.
The way in which a representational system points is the interface between it and the world. So, a representation will have different significance (‘meaning’) if it is applied to the world through a different interface. As a maxim, ‘beliefs’ or representations do not exist in a vacuum. Plantinga seems to believe that they do – i.e., we can talk about an organism having a belief, where the ‘content’ of that belief is completely separate from how that organism interacts with the world. As far as evolution is concerned, this seems highly questionable, because the meaning of a representation, and therefore its truthfulness, is in part determined by how the organism containing the representation interacts with the world in relevant contexts.
However, the above Plantinga-esque critique of Darwinism gains some of its bite by realizations that we have (presumably evolved) cognitive systems that in some way are misleading. Consider an area where many people say that we have a false representation of the universe, the representation of the Earth as not moving. Here, it seems like what is true and what is evolutionarily useful diverge. Just so, it might be that theories of evolution themselves are based on useful but misleading cognitive systems.
I think a more accurate way to understand something like our representation of the Earth as not moving is to say that our representation of the world as stationary is true in the relevant context. I.e., to understand how representational systems are truthful is to have to understand their applications. I.e., any representational account has to be interpreted or applied. It is only when we start to apply a representational system designed for navigating on the Earth as if it were designed for navigating the solar system that it starts to mislead. (Even there, though, in this case it rather straightforwardly plugs into heliocentric models by use of a transformation algorithm.)
So, not only is a correspondence between map and terrain obviously useful for purposes of navigating that terrain, say, but science properly understood has not shown that our (presumably evolutionarily derived) representational systems tend to be unreliable, but rather has sharpened our understanding of the scope of their appropriate application. Indeed, these sorts of considerations bring into question the scope of applicability of certain cognitive mechanisms underlying typical scientific ways of representing the universe, and so we probably are warranted in emphasizing careful consideration of the limits of the epistemic mechanisms we use to build scientific representations. Here I think there is something very useful about Plantinga’s sort of critique, but this is not the conclusion he tries to draw.
So, these sorts of considerations don’t show an incoherence between evolutionary processes and reliable cognitive processes. If anything, they lead one to think there certainly should be correspondence between representations and terrains, properly interpreted.