Hello, Science Blogs readers. Many of you may be new to Mixing Memory, so I thought that for the first post at the new site, I would introduce myself a little. By a little, I mean a very little, because in case you hadn't noticed, I blog anonymously. Maybe that will change someday, but for the moment, I feel that I have good reasons for doing so. Here's what I can tell you: I'm a cognitive psychologist who studies a wide range of higher-order stuff (i.e., I don't do vision). For the cognitive science initiates among you, I work mostly within the symbolic tradition. That doesn't mean that I can't or haven't used connectionist models; it just means that I don't like them. Being a symbolicist, I respect and adhere to Marr's three-level approach (if you're interested in cognitive science, you should follow that link and read the chapter; otherwise, you can just consult the Wikipedia entry) to describing and explaining mental phenomena. The most important implication of this is that I don't see neuroscience as the foundation of all cognitive science, as many neuroscientists do, but instead see it as an equal member in an explanatory collective that also includes disciplines like psychology, linguistics, philosophy, anthropology, and computer science (somebody's got to write the code).
For blogging, this means that I will write about just about anything in any area of cognitive science. Sure, once I step outside of psychology, I'm leaving the comfort of my formal expertise, but I know a little bit about how the brain works, I've done a little work in linguistics, and I've worked with cognitive anthropologists, philosophers, and computer scientists (like I said, somebody's got to write the code), so I know the how theories and methods work in those areas. Below the fold, I've linked a few of my past posts, so that you can get an idea of what I blog about, and how I blog about it.
Though I'm the sole author of Mixing Memory, and thus solely responsible for its content, Mixing Memory has always been a collaborative effort in a way. I've relied heavily on the advice, suggestions, and requests of the people who read Mixing Memory, and I've borrowed more than a few ideas from other bloggers. I'm especially grateful to the helpful comments and advice (along with all the stuff I've stolen) from a few people who've been reading Mixing Memory since its early days, especially Brandon, Clark, and Richard, and more recently, fellow Science Blogger Bora. The wonderful banners on this page and the About page (and, obviously, the bulk of the banner on the Contact page) were provided by Anton Oetll, and the wonderful banner on the Archives page was was provided by Todd Hartman (who plans to start a Political Psychology blog soon).
Finally, I'm always open to requests. The main purpose of my blogging, besides just giving me something to write about other than the stuff I'm working on at the moment, is to promote cognitive science. The best way to do that is to blog about topics in cognitive science that interest people. The best way to find out what interests people is to have them tell you. So, tell me. If there's something you'd like to learn a bit more about, let me know, and if I know enough to blog about it, I will.
A series of posts on the cognitive science of art:
Goals and Motivations of Neuroaesthetics
Ramachandran's 10 Principles of Art, Principles 1-3
Ramachandran's 10 Principles of Art, Principles 4-10
Beauty and the Brain.
Lakoff and Framing:
Understanding Frames With an Eye Towards Using Them Better
Political Analogies
Do People Actually Use the 'Strict Father" and "Nuturant Parent" Models?
What Framing Analysis Is
Theory, Research, Etc.
How Evolutionary Psychology Can Make You Look Like an Ass
Cultural Differences in Cognition: The Case of Fish
Moral Psychology I: Where Is Morality in the Brain?
Causal Reasoning
If We Had a Cognitive Account of Counterfactuals, This Would Be It
Concepts IV: A Second Revolution
- Log in to post comments
It looks really good!
Thanks, Brandon. I'm trying to get used to it. It's so... white.
I miss the Eliot quote, and that friendly Blogger light blue. But the purple head is cool. Anyway, congratulations!
Dave, you know what? I miss the Eliot quote too. I'm going to stick it in there somewhere.
Looks great. Tons of posts.
BTW - I didn't realize you were a symbolist. It's kind of surprising to me given your familiarity with phenomenology and especially the Heideggarian school. I wonder what you think about people like say Herbert Dreyfus and his critique of traditional rule-based approaches to mind? How does the symbolist avoid that? I ask because Dreyfus appears much more open to the connectivist approach. Do you modify the way a symbol is treated to be more organic?
Clark, that's actually a really good question. The short answer is that, Hubert Dreyfus' biased interpretations aside, I don't think phenomenologists were as hostile to reprsentation as many have claimed. I sometimes think that Dreyfus and others miss the focus of M-P specifically. His focus is on the pre-cognitive, which mostly means pre-conscious, but can equally apply to what is today called the cognitive unconscious. He spends a lot of time talking about representations, but only after the immediate, direct, pre-cognitive "experience" that Dreyfus focuses on.
I think a symbolist position is entirely consistent with late phenomenology (after Heidegger... Sartre and M-P especially). You simply have to bring the body into the equation, and recognize that representations are action-oriented. That's not to say this is how current symbolic models work. With the exception of Larry Barsalou's perceptual symbol symbols theory, and some symbolic models in robotics, the body and action are rarely considered.
But, despite many phenomenologically-oriented philosophers and psychologists preferring connectionist models, the situation isn't really any better, from a phenomenological perspective, in connectionism either. In fact, connectionist models tend to be even more isolated from the body and action. They tend to be almost completely isolated with no discussion of how they might be integrated into a larger system. Furthermore, connectionist models are as representational as symbolist models, and the form of representation they use (spatial representations -- vectors in a multidimensional space, for example) are dubious. I definitely don't see how they are anymore valid, phenomenologically, than symbolic representations.
There is a third direction, which some phenomenologically-oriented cognitive scientists consider more valid. That's the dynamic systems approach. It is, under some views, a non-representational approach, though there are arguments that the classic examples of dynamic systems (e.g., Watt's steam engine governer) do actually use representations. The reason they're considered more valid is that they're basically feedback loops (think of the governer ). However, to my mind, they look suspiciously like the behaviorist models that M-P demolishes in the beginning of The Structure of Behavior, merely substituting "action" for "behavior" (which, both from a phenomenological and psychological perspective, is a dubious concept within the old behaviorist tradition), and more complex mathematical models for the old models of conditioning.
I confess I've long found connectivist theories difficult to separate from rule based. If only because you can code programmically any neural network as a computer program which is rule based. So I've never understood how Dreyfus could argue that neural networks offer anything beyond what he criticizes.
I ought write something up from Peirce on this. As you know he's been adopted by a lot of thinkers in this. However I think his doctrine of continuity and vagueness avoids most of the problems I see in what I've read of symbolic methods. I was just curious as to whether you modify the standard symbolist approaches in your own thought.
I agree that phenomenologists aren't opposed to representation. But I do think they significantly change what it means to be a representation.
Clark, I'd be interested to read about what insights we might get from Pierce on these issues.
I do think that symbolist positions have to be altered from old-school AI, but increasingly, symbolists are taking into account things like modality that are important from a phenomenological perspective.
Hi, Chris,
I've been meaning to suggest doing something on Daniel Wegner for the new blog, since a lot of what he does is quite cool, and interesting to just about everyone; and lo and behold, it just so happens that there was a post on him recently at another ScienceBlog (Bushwell's). So maybe you could put something up on him, either new or adapted from an older post?
Brandon, that's a good idea. I'll see what I can come up with.