2.08.2006

usage of the week

i finally got around to taking a syntax class, so of course i'm hearing all kinds of fun terms like 'recursion'.

i wonder if anyone out there in blogland has heard of this invented usage: Recurz.

as in: "If you define a verb phrase that way, then this rule can't recurz."

it seems to be a back-formation from 'recursion' and distinct from the more conventional usage.

'recurs,' i suppose, should be a "real" word, as in, "the dream recurs every few days."

'recurz' is listed at www.marsb.com as one of dozens of alternative spellings for 'recur,' but almost all other google hits are in foreign languages.

5 comments:

Anonymous said...

The usual back-formation from recursion is recurse; when I was an undergrad in the late 1970's, my computer-savvy friends always used the word with a grin, knowing it was a wacky coinage rather than a "real word". The etymologically correct verb is, of course, recur, but this can't be used because it already means something else.

Seb said...

Takes a lot of work to get people to case that word right (speaking from intro computer science TA experience...)

i.e., people slip and use "recurse" instead of "recur"

(Makes more sense as a back-formation of recursive--unvoiced phoneme, see?--which gets used a lot in contexts like:

A: Why the hell are you making that procedure recursive?
B: Because I'm a lambda calculus ideologue.
A: Fuck you!

----------------

I see ACW beat me to the punch on this. Too slow!

When you say "recur...already means something else," what are you refering to?

Anonymous said...

I find myself unable to explain myself coherently. Having said that, let me try.

On purely compositional-semantics grounds, we would expect recursion and recurrence to be approximately synonymous abstract nouns corresponding to the verb recur. Both recursion and recurrunce should mean "happening again".

However (ah, here comes my coherence! my explanatory skillz have returned!) the term recursion has undergone unpredictable semantic narrowing in its computer-science context. It no longer means just any old kind of recurrence: it now means the specific kind of recurrence that results from self-reference, self-invocation, self-inclusion: a kind of recurrence that results in nesting rather than mere repetition.

Because of this semantic narrowing, recur no longer corresponds to recursion in meaning. The abstract noun has flown the coop: it has left its mommy-verb, and there is a hole in the lexicon for verb meaning "to perform recursion". "Recur" is no good because it has not undergone the narrowing: it still means "re-occur" in the more general sense.

Cristi said...

well said, ACW. Second time's the charm. I was worried for a while that my original post was unclear.

also, this 'hole in the lexicon' concept intrigues me (and it's come up in other recent posts).

perhaps another post will follow...?

blahedo said...

I kept getting into fights with CS profs in grad school over the existence of the word "recurse" (they claimed it should be "recur", because that's the Latin form, as if that mattered). Then I pointed out that they used the verb "induct" all the time. Furthermore, there's a strong distinction: I can have a recurring dream. I can have a recursing dream. And they're different. (Indeed, I could talk about a recurring recursing dream, and that is a reasonably sensical concept.)

Related Posts with Thumbnails