Maxime Papillon, On 24/10/2010 05:28:
>> From: [log in to unmask]
>>> 2010/10/22 And Rosta<[log in to unmask]>
>>>> IMO English 'sp,st,sk' clusters are actually /sb,sd,sg/, so if the /s/ were
>>>> lost, you'd end up with /b,d,g/.
>> The default obstruent 'sonants' are /b, d, g, v, ð, z/. Each has a 'sharp' version, /b♯, d♯, g♯, v♯, D♯, z♯/, or (notationally equivalent) /p, t, k, f, θ, s/. In any cluster of obstruents, either none is sharp or only the first is sharp. 'St' is /sd/: the obvious alternative analyses to 'st' being /sd/ are (i) /st/ and (ii) /zd/ (or with archiphonemes, /ST/, amounting to much the same analysis). Word-initially, grounds for choosing between /sd/ and /zd/ are meagre, but elsewhere, e.g. in _east_, it is clearly /sd/, not /zd/, because the sharp /s/ triggers prefortis clipping, and there is a contrast with /zd/ (cf _eased_). Arguments against /st/ are as follows. 'Onset' /p,t,k/ are aspirated, but you don't get aspirated [p_h, t_h, k_h] following an obstruent. 'Coda' /t/ is realized [?] in many accents, but you don't get [?] following an obstruent. And _mist_ and _missed_ (_rift_ and _riffed_, etc etc) are homophonous, which is to be expected if they are /misd/
>> d /mis+d/, where /+d/ is the _ed_ suffix.
> Maybe it's me, but I don't see what is the advantage of this
> description over the down to earth "/t/ is realized as |t| after /s/
> and as |t_h| elsewhere" kind. Is it more comprehensive, or does it
> provide a description of English phonotactics at a lesser algorithmic
> entropy?

The latter, if "lesser algorithmic entropy" is "greater simplicity".

>To me, it seems to be merely an alternative -perhaps as
> valid, but surely more difficult- description of the same thing, with
> no advantage over the traditional description that I can see.

To my eyes, my analysis is difficult only in that it involves a departure from orthodoxy; departures from orthodoxy are always cognitively taxing for those who understand and embrace the orthodoxy. (I intend that as a factual observation, not as a rhetorical argument, and not directed at you.)

The advantages of reanalysing "obstruent + p/t/k" as "obstruent + b/d/g" (where p/t/k are special sharp subvarieties of default b/d/g) are that it spares you having to state an allophony rule for the plosive in that postobstruental environment and it spares you having suspicious-looking morphophonemic rules to convert the form of _-ed_ to /t/ postobstruentally (and likewise for _-s_ to /s/). And there are no countervailing complications, as far as I can see: /sd/ is a clear winner over /st/, with no weighing up of pros and cons needed. But I'd be only too delighted for you to try to convince me otherwise, though perhaps list etiquette dictates that follow-up discussion should go off-list.