A Narrative in Which your Humble Blogger Finds the Twitterverse Wanting in Intellectual Depth and Calls Out a Respected Scientist as an Exercise in Futility
Obviously, Twitter is no place for nuance but I just couldn’t really figure out what Steven Pinker, linguist/cognitive scientist/writer, was trying to say in this recent tweet:
Steven Pinker @sapinker
Neurosci is too atheoretical but €1B brain simulation won’t help-no computational/evo theory of what the brain is for- http://bit.ly/SJr5WT
The link points to a year-old feature in Nature about Henry Markram and his new Human Brain Project, which is apparently version 2.0 of his Blue Brain Project. Considering the few papers that have come out of that multi-million Big Science project, I’m hardly going to defend Markram as embodying the ideal for computational neuroscience, but it certainly is one approach. He does have a tendency for hype, but I imagine that is required for the job. Anyway, Pinker has written at length about his problems with early connectionist models of language acquisition (Cognition 1988), where he took a position that rule-based symbol manipulation was more predictive of language learning data than the distributed processing models proposed by Rumelhart and McClelland. In this tweet Pinker seems to be taking on all of neuroscience by providing a relatively narrow example of a modeling project. Of course I especially take issue with the accusation that we neuroscientists are atheoretical or that we have no general computational/evolutionary theory. No, we certainly don’t have any grand-theory of biological computation, but we have a great many deeply insightful approaches to understanding interactions between neural/sensory adaptive systems and the environments in which they evolve. If he needs to brush up on the recent work, I’d suggest anything in neuroethology (evolutionary neuroscience of animal behavior), or perhaps this recent paper in Nature that lays out mechanisms for the evolution of complex cognitive skills through gene duplication.
Well, I called him out:
Michael S Carroll @nucAmbiguous
@sapinker Are you just trolling comp neuro? Perhaps you could make a less facile claim in a forum allowing more detailed response.
No response yet.