Of course in some languages (that's
certaily the case of French) practically all compounds are lexicalised noun
phrases which, unsurprisingly, preserve the phrasal order of elements. Such is
the case of Tibetan, where N-N compounds have the modifier-head structure (they
derive from genitive + noun phrases) but compound nouns involving an adjective
have the head-modifier order N-Adj (again, this order is normal in phrases). But
Germanic (and IE) N-N compounds (with exceptions like daisy or Wednesday) do not
contain an etymological genitive and don't necessarily correspond, historically
or synchronically, to Gen + N phrases: "toothbrush" is not equivalent to
tooth's (or teeth's) brush. The modifier has a very general function here,
"something to do with a tooth or teeth". The whole phenomenon is a matter of
word-internal syntax rather than the grammar of English phrases.
Modifier-head is the unmarked order in
endocentric N-N compounds in any language that adds inflections at the end
of words, since with this order the compound doesn't differ from the head
in isolation in matters of declension, grammatical gender, etc., and "split
compounds" with inflections sandwitched between the constituents can be avoided
at the same time. Needless to say, the marked order is not absolutely ruled
out, but there must be some special reason for its occurrence.
Of course in Modern English any phrase can
be converted into a compound, but such compounds are typically headless, i.e.
exocentric: forget-me-not, go-between, might-have-been, etc. Problems of
inflection (or whatever passes for noun inflection in English) are solved in a
radically simple way: [forget-me-not]s, etc.
V-N exocentric compounds like spoilsport,
lackland, cutthroat, sawbones, scarecrow or turnkey are most likely
modelled on similar French formations; at any rate they were unknown before
Middle English. Here again English speakers don't worry about fine points of
logic and semantics as much as the French do and form plurals as if for
simplex words: spoilsports = [[spoil][sport]]s, etc. I even think sawboneses =
[[saw][[bone]s]]s is acceptable beside haplological sawbones (pl.). This is
because the English plural ending behaves more like an enclitic than a real
desinence.
Piotr
----- Original Message -----
Sent: Sunday, January 07, 2001 5:09 PM
Subject: Re: [tied] People of the Rivers - Thought #3
>*k:al-axwa "female-in-law" consisting of *axwa
> "brother"
and *k:al-, presumably "woman".
Isn't that the sort of "compounding" that
you find in Semitic? "word-God"
(accent producing a reduction in
vocalism of the first element to what is
called a "construct" form) = "word
of God". So here, "woman-of brother".
Peter