- Chomsky is wrong again: Another installment of Chomsky's Program has been once again proven to be wrong by the insight of neuroscience, cognitive linguisticshave brought linguistics and connectionism, disciplines that have brought the study of language "back to scientific principles." This rant is really quite delightful given that it is entirely unrelated to anything anyone does or has thought of doing. So here it is. Read it an weep.
- Science Infotainment: Every five years or so, the NYTs runs another AI piece explaining how this time we are just inches away from cracking the puzzle that is the brain. This one is written by one John Markoff, who, appropriately, given the name, tells us about how new sophisticated machine learning techniques are just moments away from finally doing things the way brains do (neural networks, you know). This breathless introduction is par for the course and in a few years, after lots of money has been thrown at the problem, reputations have been made and centers launched (one is already in the offing at MIT) we will discover, once again, that associationist based investigations of mental and neural processes have not delivered on their initial promise. The one feature of this piece that is slightly novel, is the hedge: maybe this won't yield scientific insight into brains but it will allow us to make neat stuff. Here's the money quote: "The new approach in both hardware and software is being driven by the explosion of scientific knowledge about the brain. Kwabane Boahnen…said that is also its imitation, as scientists are far from fully understanding who brains function. "We have no clue," he said…" Wow: not having a clue is simply not fully understanding. Having grown up in the 60s, I can recognize 'light at the end of the tunnel' speak when I see it. Another terrific NYTs gem illustrating the burgeoning art of science BS.
- A Cute Contrast: Though the secrets of how the brain works are just moments away from being revealed, it seems that bicycles are a real mystery. This piece is a nice example of how complicated even very simple things are. One reason linguists, like all other scientists, study very simple phenomena (e.g. movement, vs how language expresses culture identity) is that simple things have a hope in hell of being understood, not because they are more interesting or important. Interaction effects are a real pain and complex systems are very hard to disentangle. So, we concentrate on very simple systems. How simple? Well, it seem bicycles are already too complicated, if this article is to be believed.
- Dogs are smart!: Ok, maybe this isn't news exactly, but I love dogs so I report on this one, the William Shakespeare of the canine world. Not only does this border collie have a larger vocabulary than your average university president, but it seems that you can teach him to understand "sentences." Of course, these are not our kind of sentences (indeed, Chaser's words are not our kinds of words), but it's nice to know that you can teach a dog to bring in the NYTs and leave the Washington Post outside to rot. The only real problem with the work reported is that it sheds absolutely no light on human language or human vocabulary competence. Why? Because Chaser's words are all tightly referentially dependent (here a simple minded referential semantics seems to be just fine, as opposed to how our words work) and the sentences are effectively linearly dependent templates. It is nice to know that dogs can do this, showing that Dolphins (who demonstrated this capacity over 20 years ago) have nothing on them in the "language" domain. So, if you want to go "ahh, cute" a lot, read this, or troll for cat pix on the web.
- Evolution: One of the boundary conditions for understanding how language emerged in the species is the apparent fact that it seems to have occurred quite suddenly. It seems that this is not a problem limited to language. Darwin, apparently, considered the rapid emergence and proliferation of flowering plants a similar kind of "abominable mystery." Well, it's here reported that a key to understanding this problem is a "genome doubling event" that paved the way for all that followed. Equipped with twice as many genes, flowers explored a new range of phenotypic options. It seems that subtle genetic changes can have vast phenotypic effects. Who would have thunk it?
- Where's the data?: There is a popular view that science mainly consists in the patent gathering of loads of info and that theories mainly serve to compact this mess into digestible form. Well, if this is what science is, then we are in big trouble for it seems that much of our data has been lost, see here. I have always found this view very odd and have tended to the opinion that large chunks of what we consider our best efforts have been motivated more by artificially constructed facts and thought experiments than by careful data mining. However, it seems that as an institution, until now, scientists have agreed that protecting the data has not been a major priority. Maybe because, the standard empiricist picture is wrong? See here as well for further thoughts on how data functions in the sciences.
- There's no business like science business: Last here's another shot at big science as pursued by out leading journals. It seems that some of our thought leaders are getting miffed about how journals like Science and Nature are more interested in being trendy than in being scientific. I am still not sure that I fully agree with this, but it is worth thinking about. As Harry Frankfurt pointed out a while ago, BS is the endemic vice of our modern culture. It's easy to get and easy to spread. Sadly, even our leading institutions are more than a little prone to tossing it around
That's all folks. Happy New Year.