Plato, Aristotle and Dangerous Things: what categories reveal about the classroom

[robinturner]

In class this week, I handled the transition from Plato to Aristotle. I hope it was interesting for the students, because it was certainly interesting for me.

First I covered the difference in text types between The Republic and The Politics (how you read an imaginary dialogue as opposed to how you read something that is supposed to be someone's lecture notes but almost certainly isn't). Then I got onto what really interests me, which is the difference between Platonic and Aristotelian categorisation. If this subject doesn't interest you, click your Back button now; otherwise ... I used a girl in the class as an example, asking "How do I know Zeynep is a girl and not a table?"

From an Aristotelian point of view, I know she is a girl because she has necessary and sufficient conditions for girlhood: she is a rational animal, not male and not adult (we got into a discussion about why Aristotle would regard femaleness as an absence of maleness later).

From a Platonic point of view, I know she is a girl because she reminds of of the ideal girl which I had in my mind before I was even born.

What interests me (and probably doesn't interest my students much) is the parellel with linguistics. The Aristotelean view is that of classical semantics (a point which other poeple have mentioned). In other words, "girl" is equivalent to +HUMAN -MALE -ADULT.

What I haven't seen in the literature (though it's so obvious, I'm sure someone must have mentioned it) is that the Platonic view is amazingly similar to Eleanor Rosch's theory of prototypes (later made famous by George Lakoff, hence the title of this post). Plato's metaphysics maybe a little counter-intuitive, but the theory of ideal forms makes psycholinguistic sense.

Neither theory on it's own is adequate. It is pretty obvious that when classifying someone as a girl, we do not laboriously go through a checklist of necessary and sufficient conditions; besides, there is the problem raised by Wittgenstein that there are categories for which it is impossible to specify such conditions, "game" being the most famous example. On the other hand, prototype theory fails to account for how we classify things at the boundaries of categories (e.g. how we decide that an ostrich is still a bird, but a bat is not).

In both sections, we did not circle this square, but got into amusing discussions. In one class we had the question of whether wearing a skirt is an essential or accidental attribute; in the other we got into why Aristotle regarded femaleness as incomplete maleness (which I wasn't planning to do until next week). I liked one student's comment: "I would have expected better from a philosopher."

Comments

More such entries please! :-)

[cibetky.livejournal.com]

(I'd come back more often.)
Your example with the girl reminds me of a similar discussion in a class I attended whether "girl" or "boy" is necessary -ADULT and that we don't have an exact borderline between + and - ADULT (18 years of age? It depends on culture and time period.), which makes the componential analysis somewhat flawed.

I wonder whether anyone would be inclined to make an online quiz which would assess where the participants stand on an imagined scale between nominalism and realism. That would be the kind of quiz I'd love to take! ;-)

Re: More such entries please! :-)

[solri.livejournal.com]

I actually did my MA dissertation on the girl/woman distinction in Turkish (and by comparison, English). The general criteria are different: in English a girl becomes a woman on reaching adulthood, whereas in Turkish the criterion is loss of virginity (which is why here you wouldn't normally call an unmarried woman a "woman"). However, it's more complicated: in both English and Turkish an adult woman can be called a "girl" for a number of reasons (e.g. establishing peer bonds or, conversely, to be condescending); similarly even in Turkish you can call a non-adult a woman if the context is "women as opposed to men". My aim was to test a theory of categorisation which was feature-based (as in classical semantics) but also took into account prototype theory by distinguishing between "defining features" (e.g. being female) and "typical features" (e.g. having breasts); these are further subdivided into strong and weak features. For example, for "girl", +FEMALE is a strong defining feature, in that you never call males "girls", whereas -ADULT (or in Turkish +VIRGIN) is a weak defining feature, in that while it would normally be used to establish the boundaries of the category, it can be over-ridden by context, as in the examples above. I also proposed a phenomenon I called "category stress", but I won't bore you with that!