Eleanor Rosch

[JennaW]

Anybody read anything about Eleanor Rosch's work on concepts?

[Robert Campbell]

Jenna,

I've read a number of things by Eleanor Rosch.

What aspects of her work interest you?

Robert Campbell

[JennaW]

Her work on concepts and categorization. Ever since I came across concepts, categorization, etc. I've thought that it was more realistic to have a dynamic, fuzzy-boundary view. It seems like, what I've read so far of what she did, that she found this out empirically. I'm also looking at the conceptual blending theory (can't remember the researchers off the top of my head right now though).

[Ellen Stuttle]

Ever since I came across concepts, categorization, etc. I've thought that it was more realistic to have a dynamic, fuzzy-boundary view.

Same here, if I'm getting a (dynamic, fuzzy-boundaried) sense of what you mean.

Question for Robert: Someone -- was it Bryan Register? -- wrote a JARS piece critiquing AR's theory of concepts from a Piercian perspective. Do you know off-hand what the article is I'm thinking of and which issue it's in?

Ellen

___

[Robert Campbell]

In response to Ellen's query, Bryan Register's piece

The universality and employment of concepts, Journal of Ayn Rand Studies, 1(2), 211-244

appeared in Spring 2000 and was quite critical of Rand's theory of concepts. I don't think Register's approach would qualify as Peircean, though.

An article about Rand and Peirce by Marc Champagne is due out in the Fall 2006 issue (Volume 8, number 1).

Robert Campbell

[Robert Campbell]

Jenna,

I hope your hospital stay won't stretch on too much longer.

Meanwhile, I've been remiss in not responding more quickly about Eleanor Rosch's work. Now that I've finished teaching my summer school course (and gotten my slides ready for two talks at the TOC Summer Seminar), let me see whether I can net out Rosch's main contributions to the study of categorization.

In the early to mid-1970s Rosch put forward what became known as the prototype theory of concepts. (Her earliest publications are under a married name, Heider--worth keeping in mind if you're tracking them with PsycInfo or another database.)

According to Rosch, human concepts generally lack clear definitions, and whether something qualifies as an instance of a concept is therefore not an "all or none" affair. Instead, concepts are arranged around a prototype, a "most typical" exemplar. Instancehood is "graded": some instances are better than others (i.e., more similar to the prototype), and instances that are poor enough may not belong to that concept at all, but a sharp line can't be drawn between poor instances and non-instances.

For instance, most people think that a bluebird is a "good" bird and an emu is a "not so good" bird. So the protoypical bird must be fairly close to a bluebird--small, able to fly, able to sing, etc. For Rosch, such judgments of "good" or "not so good" aren't made in addition to judgments about instancehood; they are part and parcel of judging whether something something is a bird or not. In fact, people are faster, on the average, at pressing the "Yes" button in response to "A bluebird is a bird" than in response to "An emu is a bird," even though they know that both statements are true.

I've been told that Rosch later denied intending prototype theory as a general theory of concepts. It would be hard to infer this from her initial articles on prototypes, however, as none of these stated any limits to the theory.

Here are two challenges to protoytpe theory:

Armstrong, Armstrong, and Gleitman (1983) asked participants in their study to give examples of "really good" and "not so good" odd numbers. 3 is really good; 54,267, let's say, is not so good. Yet their participants had no doubts about what qualifies as an odd number and what doesn't; there's nothing "fuzzy" about the boundaries of the odd number category.

Most people would say that a perch or a trout is a "really good" fish, a puffer is a "not so good" fish, and a seahorse is "really not very good" fish.

A killer whale looks a lot more like the most typical fish a lot more than a seahorse does. Yet most people will say that a killer whale is not a fish and a seahorse is. Isn't some principle from outside of prototype theory required to explain how people make this kind of distinction?

Robert Campbell

PS. One of my favorite Rosch articles is about categorization but not really about prototype theory:

Rosch, E., Mervis, C., Gray, W. D., Johnson, D. M., & Boyes-Braem, P. (1976). Basic objects in natural categories. Cognitive Psychology, 8, 382-439.

[JennaW]

I'm reading The Big Book of Concepts by Gregory Murphy and it goes into some detail into concept formation research. From what I'm getting, there are concepts and categories that are more hardlined with a defined boundary-- like in math, which has connetions to logic-- while in other areas there are fuzzy boundaries, as when it comes to qualitative descriptions such as "warm", or "lengthy".

So, no, just going by the classical view of concept/category formation is not realistic. And this subject in re: to cognitive science is actually old news, as 20+ years of research has taken place; not to mention neuroscience has been blossoming.