know from

Laurence Horn laurence.horn at YALE.EDU
Fri Dec 27 20:35:12 UTC 2002


At 3:03 PM -0500 12/27/02, Philip Trauring wrote:
>I think there are two different constructions here. One is 'know
>from' and the  other is 'know X from Y'.

Precisely.

>
>Examples such as 'he doesn't know his ass from his elbow' are also
>idioms, whose constructions I would argue are not easily tracked and
>frequently flaunt normal grammatical rules. Does anyone have an
>example of this 'know X from Y' construction that isn't an insult?

Well, the positive counterparts would represent such cases (albeit
ones occurring only in a context in which there's some reason to
suspect that the subject did NOT have the powers of discernment
referred to).  A related example involving not "doesn't know Y from
Z" but the related "can't tell Y from Z" is (if I'm not mistaken,
courtesy of Hamlet) "I can tell a hawk from a handsaw."  I would
submit that it's even possible to get "I know my ass from a hole in
the ground", and there are several such examples available on google.
These are not insults, but defenses against (actual or implied)
insults.

Note that the possibility of getting "X can't tell Y from Z"
alongside "X doesn't know Y from Z" makes it less likely that we can
simply call the latter (or the former) an unanalyzable idiom.
There's also "X can't tell Y and Z apart", again more likely in
negative contexts, and related constructions like "How do/can you
tell Y and Z apart?" or "How can you tell Y from Z?"  No, not "How
do/can you tell your ass from a hole in the ground?" but "How do you
tell a finch from a cardinal?" or "How do you tell a restrictive
relative clause from a non-restrictive one?"  Now, I agree that it's
harder to get these with "know from" than with "tell from", but the
claim that the former is an idiom and thus non-compositional seems a
bit too strong.

>
>I'm not convinced the two constructions ["know Y from Z", "know from
>Z"] have the same origin.

As argued earlier, they pretty clearly don't.

Larry



More information about the Ads-l mailing list