Is GENERICIDE a bad choice or morphemes?

[Laurence Horn]

>         But there are other words that are used in senses at odds
>with their morphologies.  Consider escapee; one would suppose that
>the person who escapes is the escaper, and the person or thing
>escaped is the escapee, but a different meaning prevails.  Another
>example is looker, which means a person who is looked at, not one
>who looks.  I was also going to trot out informant, which is used in
>lieu of the stigmatized informer, but I guess that isn't really at
>odds with its morphology.

-ee is indeed interesting, but note that the animacy restriction is
much stronger than the semantic/grammatical role restriction
(essentially = non-agent), so an escapee could never be a jail, for
example.  (There are some counterexamples, such an attested reference
to a tomato plant as a "drownee", but they always seem to involve
personification.)  Could it be the jailer, or a police officer?  It
doesn't seem as though it really could, in that -ee doesn't seem to
be that easy to get with source arguments (cf. *evadee).  I've argued
(no doubt unpersuasively) that an escapee is so-called because s/he
was involuntarily housed in whatever place s/he escaped from, in
which case "escapee" is sort of like "standee" in the
nonvolitionality dimension--a glarfee is someone who was forced to
glarf--even if the "escapee" does seem like an agent even more than a
"standee" does.  (A very comprehensive treatment of the -ee case is
given in Chris Barker's 1998 paper,  "Episodic -ee in English:  A
thematic role constraint on new word formation", Language 74:
695-727.)  The result is that "escapee" and "standee", while they may
need some explanation (or at least more sophisticated hand-waving)
don't really mean the opposite of what they would be predicted to
mean given the word-formation rule involved, and in any case if
there's only one argument of the verb, we just assign whatever role
we have around when we're interpreting the -ee noun.

"Looker", on the other hand, is a real problem.  And yes, it has
always bothered me, I admit it.  I'm not claiming that no ambiguities
of this sort are possible, even with -ee ("masturbatee", for example,
is attested for both 'someone who is masturbated' and 'someone who is
masturbated to', à la "fantasizee"), but part of the problem is that
(as Barker's article nicely demonstrates) it's really quite hard to
give the actual semantic constraints on -ee affixation.  In the case
of -cide, on the other hand, it seems entirely straightforward:  an
X(i)cide is a killing (or killer) of (an) X.  X is always (except for
"genericide", to be sure) the theme/patient, the killee, never (just)
the agent or the instrument, the killer.  (In a suicide, X is by
definition both killer and killee.)

Larry