(Caveat: there may be IPA mistakes below!)
Here’s a question for any morphophonologically sophisticated readers out there, following up on something I noticed a few months ago:
As we all know, the adjectival participle in English is formed with the same -ed suffix as the past tense and verbal participles, and has the same allomorphs /t/, /d/ and /əd/, conditioned by the same phonological environments (modulo a few morphologically specified exceptions; see some of Dave Embick’s recent work).
But adjectives can be subject to a further morphological process in English: you can usually adverbialize a participial adjective with -ly:
But, in my judgment, when certain such adjectives have -ly added, their -t or -d allomorph of -ed has to become syllabic. Looking at *edly sequences that actually have entries in the OneLook.com metadictionary, I see things like
I noticed this when I made one up writing to a student, 'relaxedly', from adj. ppl. 'relaxed' /rə.lækst/; thinking about it, I felt I had to pronounce this /rə.læk.səd.li/, not /rə.lækst.li/
It’s easy to show that this is something about -ed, not something about the phonological shape of the stem to which -ly attaches, by comparing pairs like the following
blind~blindly /blaɪnd.li/ with resigned~resignedly /rə.saɪ.nəd.li/ (not /rə.saɪnd.li/)
and sound~soundly /saʊnd.li/ with renowned~renownedly: (Actually I don't really have a judgment on this one, but some dictionaries seem to be saying it has to be /rə.naʊ.nəd.li/, not /rə.naʊnd.li/.)
So the generalization so far is as follows, I think: when a /t/ or /d/coda allomorph of -ed follows a consonant, it is resyllabified with an epenthetic ə as /Cəd/ when -ly is attached.
Syllabic 'l', unsurprisingly, gets treated like a vowel, not a consonant, so when /d/ follows it, no epenthesis is required:
unbridled /ən.braɪ.dld/ unbridledly /ən.braɪ.dld.li/, not /ən.braɪd.ləd.li/
Syllabic ‘r’ similarly, when followed by a /d/ allomorph of -ed, doesn’t need epenthesis:
e.g. good-naturedly = /gʊd neɪ.tʃrd.li/ not /gʊd neɪ.tʃə.rəd.li/; also
bewilderedly, self-centeredly, laboredly, measuredly, ill-mannerdly
But consonantal -r, following a stressed vowel, does need the epenthetic vowel:
assured ~ assuredly: /ə.ʃʊ.rəd.li/; also
endearedly, preparedly, declaredly, deploredly
I think one could turn this into an argument that there really are two syllables in words like tired, inspired, retired, the second being a syllabic /r/, because they don’t need an epenthetic schwa when suffixed with -ly:
If they were like 'prepare', etc., syllabically speaking, then epenthesis ought to be required.
A similar thing seems to be going on with stems with stressed syllables ending in a diphthongized vowel, too (compare to unstressed ones like 'worriedly', on the right); seems like the offglide is treated consonantally in that it triggers resyllabification
avowedly /ə.vaʊ.wəd.li/ worriedly /wʊr.rid.li/ (not /wʊr.ri.jəd.li)
appliedly /əplaɪ.jəd.li/ hurriedly /hʊr.rid.li/
cowedly /kaʊ.wəd.li/ studiedly /stʌ.did.li/; also
Some judgments are a bit fuzzy, e.g. some stems ending in -z seem to be fine without resyllabification (though I can also resyllabify):
uncivilized /ən.sɪ.və.laɪzd/ uncivilizedly /ən.sɪ.və.laɪzd.li/ seems ok, although /ən.sɪ.və.laɪ.zəd.li/ is also fine to my ear.
So, I'm wondering if this is something about English (adjectival) -ed that has been described or analyzed somewhere before? And is it productive in most Englishes?