Rolf Noyer: Noun Compounding Question (reply to Andrea Zukowski)

[Martha McGinnis]

If what you mean is the "mice-eaten"/"teeth-marks" phenomenon, the DM
analysis is that /mays/ and /tiy0/ are stem allomorphs of the vocabulary
items 'mouse' and 'tooth'.  As such, we may find them in morphological
circumstances that do not involve syntactico-semantic pluralization per se.
However, explicit appearance of "level 3" "regular inflection", i.e. plural
/-z/ is in fact the spell-out of a plural morpheme.  Thus we do not expect
to find /-z/ in morphosyntactic situations where there is no NP as a whole
that is number-marked, as will be the case in noun-incorporation cases such
as the synthetic compounds like _rat-eater_.

Lieber in her dissertation makes similar arguments about the
"Verbindungsvokal" in German examples like _Hund-e-muede_ "dog-tired": here
the /-e/ cannot possibly be an inflectional morpheme but is part of a the
e-suffixed stem allomorph of the root 'Hund' (in more DM-ish parlance).

Does this help?

Rolf



>Hi all.
>
>I've got some Williams syndrome data on "rat-eater" compounding (they obey
>what Kiparsky characterizes as Level Ordering), and I'm wondering if anyone
>knows of any DM analyses of the kinds of facts that Level Ordering was
>designed to capture.  Do you know of any?
>
>Andrea