Karthik sent me this link (here) to some very interesting work on vision. It concerns the recovery of sight in people who were blind from birth. It seems that one can be born with cataracts, which completely prevents visual perception. Pawan Sinha (PW), a neuroscientist at MIT, had the generous idea of surgically reversing these cataracts and the bright idea of seeing what would happen to the patients visual capacities once eyesight was regained. The findings illuminate what it means to claim that some capacity is innate. Let me explain.
As I’ve mentioned before (e.g. here), everyone is a nativist. The question is not whether minds/brains are structured but how they are structured. Why is everyone a nativist? Because everyone believes that minds/brains inherently generalize from experience and there is no generalization (i.e. inductive “leaps”) without natively supplied directions guiding these leaps. No “biases” or given modes of generalization (i.e. projectable predicates) no induction. No induction no thought. Unstructured minds/brains are boring blobs of neuronal protoplasm. So, everyone, and I mean everyone, is a nativist.
So, the question is not if the mind/brain has structure but what structure it has, and this latter question can only be solved empirically. Illusions provide fruitful grounds for inquiry. Why? Because, by hypothesis, illusions are not veridical. In other words, illusions do not “copy” the visual input. They may use it, but they clearly go beyond it (i.e. “add” to it, or ignore it, or distort it or…). And as we wish to know how minds/brains generalize, studying how minds/brains go beyond the information in the input (i.e. how illusions arise) is a good place to look for its biases, for what minds/brains bring to the act of induction.
This is where PW’s unfortunates come in. Because of their cataracts, these people received zero visual input from birth. A plausible way of studying what the mind/brain brings to visual perception is to see what happens when these visual impediments are removed. What do these late seers see?
Well, it seems that they are immediately susceptible to two famous illusions; the Ponzo and Muller-Lyer (see above cited paper and this more technical one). In other words, as soon as the cataracts are removed the subjects “see” the illusions (within 48 hours of the surgery). Or as PS and colleagues put it (here, R2):
…even at the very outset of their visual experience, the Prakash children already exhibit susceptibility to the Ponzo and the Muller-Lyer illusions.
And from this the paper concludes, very reasonably IMO that
This suggests that susceptibility to the Ponzo and Muller-Lyer illusions likely does not depend upon a sophisticated analysis of the scene. 
Why not? Because being able to parse scenes is something that comes along somewhat later. Thus, these illusions are likely grounded in “processing mechanisms that do not depend on visual experience.”
This is all very nice stuff. However, it strikes me that “immediate” behavioral expression of the underlying capacity sets the nativist bar very high. Let me explain. In another paper (here) PS and friends discuss a classic problem first mooted in the 17th century; Molyneux’s Problem (here).
Suppose a man born blind, and now adult, and taught by his touch to distinguish between a cube and a sphere of the same metal, and nighly of the same bigness, so as to tell, when he felt one and the other, which is the cube, which is the sphere. Suppose then the cube and the sphere placed on a table, and the blind man made to see: query, Whether by his sight, before he touched them, he could now distinguish and tell which is the globe, which the cube? To which the acute and judicious proposer answers: ‘Not. For though he has obtained the experience of how a globe, and how a cube, affects his touch; yet he has not yet attained the experience, that what affects his touch so or so, must affect his sight so or so…’
PS & Co uses the cataract subjects to test how quickly it is possible to generalize from one perceptual modality to another, in this case from touch to sight. They find that in contrast to the illusions discussed above that appear immediately, it takes time for the inter-modal transfers to become operative. As PS &Co put it (p.2):
Our results suggest that the answer to Molyneaux’s question is likely negative. The newly sighted subjects did not exhibit an immediate transfer of their tactile shape knowledge to the visual domain.
Note the word ‘immediate.’ The paper suggests that anything but “immediate” generalization implicates a non-native capacity. Why believe this? It’s not crazy to think that being able to coordinate and use two native capacities might take more time than exploiting but one. Moreover, as the paper indicates the inter-modal transfer happens very rapidly. We are talking to transfer in as little as five days. If this be “learning” it is remarkably rapid. It may not be quite one-trial (though who knows) but it’s not exactly la longue duree (here) either. Indeed, PS & Co notes (p. 2) that
Rapidity of acquisition suggests that the neuronal substrates responsible fro cross-modal interaction might already be in place before they become behaviorally manifest.
Indeed. It seems that we are inter-modal transfer “ready” and that it takes very little to get the whole thing into gear. I don’t know about you, but this seems very like saying that we are built for this, rather than that we, in any interesting sense of the word, “learn” it. This suggests that there is a perfectly good sense in which Molyneaux (and Locke who agreed with him) was, in an important sense, wrong. The mind/brain is very ready to jump to inter-modal conclusions, as the smallest input seems able to trigger the cross modal inferences.
As I said at the outset, everyone is a nativist. These papers show that for some visual behaviors no visual input is required and for others very little is needed. This tells us something about what must be native. It implies that minds/brains are (largely) pre-wired for these kinds of perception.
In addition, the work also generates an interesting question: what does the difference between “immediate” vs “very rapid” onset of behavior tell us about the built-in structures of interest? Is this a useful difference? I am unconvinced, but I could be wrong. That said, these are great little papers and usefully advance our thinking about the mind/brain’s native proclivities.
Last note: this stuff is also clearly of interest to critical period hypotheses. Recall that this work is based on systems that were largely dormant until rather late. Nonetheless they quickly became operative when the opportunity arose (i.e. the channels were unclogged). This strongly suggests that not all mind/brain capacities disappear if not used. Which do and which don’t seem like worthwhile things to try to find out.
 I personally find the hedging here unhelpful. What’s with this “suggests” and “likely” stuff. It’s clear that the authors think that this is the right conclusion. The hedging is there, I believe, to sound “scientific,” (i.e. to sound like they are very open minded). But the conclusion from their data seems much stronger than this. The authors do not really believe that a learning explanation of these effects is a reasonable given their data. So they don’t really think that their results leave the question open. And if not, why the hedging? I don’t see the point and it just confuses matters.
 Note that while early onset is a good indicator of native structure, late onset is not a good argument for “learning.” Native is what’s needed to bridge the gap from input to acquired capacity. Early onset precludes the capacity of significant environmental input (a pre-requisite for learning). However, late onset does not imply the existence of relevant input, only its possibility. That learning takes time does not mean that anything that takes time is learned. This is what POS arguments have taught us.
 I should add that this paper is not arguing against a straw man. Richard Gregory, a very good visual perception person, held that we effectively learn to see the illusion in the sense of inducing it from environmental observations.