It is fascinating to observe the discourse between the advocates and opponents of the geometry module. The computer as a metaphor for the brain, as Ivan suggested, combined with the innate brain as sophisticated machine may shape and constrain our models for how the mind works.
Dr. Newcombe mentioned that two scientists at UC Santa Barbara propose a kind of hybrid theory and now describe the geometric module as a functional unit, rather than a structural unit. I understood this to mean that computing geometric information could be more of a distributed and opportunistic process, rather than a genetically predetermined function of a discrete chunk of the brain. How did you all understand what she said?
This orientation research seems to be looking for what is universal to humans. I wonder about the range of variation. We all know people who have extraordinary spatial intelligence and other people who have no sense of direction. Does that variation strengthen or refute the argument for a geometry module? Or, is it irrelevant?
Since I am particularly interested in how theory relates to applications in the “real world,” I wonder about how the response to proximal and distal landmarks combined with contextual geometry may influence how we design architectural space and signage, and how in the natural world, we might conduct environmental education.

It would also be interesting to know how much of the results from the geometric literature are dependent upon the simplicity of the "geometry" and "features" in the experimental setting. It would seem that the clear distinction between "geometry" and "features" would become less exaggerated in a more natural setting.

This was a great talk! I honestly had no idea that there are a number of people out there who believe in the exsitence of modules in the brain.
One of the more interesting issues that Professor Newcombe adressed was whether language enables us to think and orient in space. Although the results of her study disagree with the "language-as-bridge" hypothesis, she suggested that language does help in reorientation. I would be interested to hear to what extent language is used (and useful) in our everyday reorientation tasks.
Apart from that, I though that Professor Newcombe provided a convincing list of arguments agents the idea of a geometric module.

I was puzzled about a topic that came up during the Q&A: what counts as “geometric” vs. non-geometric or “featural” information. It was suggested that it can be hard to distinguish these. But I wasn’t clear exactly what the worry was. It seems like both kinds of information are found in any environment. So the issue must be about what information is being represented by subjects and how that information is used. And that would depend on the nature of the cognitive mechanisms involved in spatial cognition. I could see it being hard to determine from behavioral studies what kind of information is being represented and used in a particular situation. But the worry isn’t about conceptually distinguishing geometric vs. non-geometric/featural representations, right? Is there a conceptual issue I’m missing?

[belatedly]
I enjoyed the discussion on the potential difference between featural and geometric information. There are at least two ways of approaching the questions raised there. One is to regard it as a debate akin to "how many angels dance on the head of a pin", with much being made of dubious conceptual distinctions. Another perspective, though, contextualizes the debate in the literature. The problem, to my mind, is that in the literature much is being made of distinctions that have not really been thought through. For instance, in [1] we see a purported difference in how humans process "landmarks" and "boundaries", yet perhaps, in fact, this isn't the key difference in the study -- the difference, instead, might well be that what were called "landmarks" are visual entities which a person might perceive as possibly impermanent, while what were called "boundaries" in fact had a much stronger sense of permanence. So a discussion of the difference in "what counts as 'geometric' or 'featural' ", ends up being potentially important and useful because it reveals latent assumptions which may be resulting in problematic experimental paradigms.


[1] Doeller, C. F. and Burgess, N. (2008). Distinct error-correcting and incidental learning of location relative to landmarks and boundaries. Proceedings of the National Academy of Sciences, 105(15):5909-5914