CONtract/conTRACT

[Arnold Zwicky]

two additions to my previous posting on forestress...

1.  i have no reason to think that forestress in noun-noun compounds
(which is, after all, the default) has anything to to with the
latinate-prefix alternations or the dialectal stress retractions.

2.  the afterstressed pronunciation of "ice cream" has an obvious
source, namely the stress pattern of the modifier-noun phrase "iced
cream".  presumably, the stress pattern was carried over even after
the expression was reanalyzed as a noun-noun compound (which i think
it certainly is today).  the forestressed pronunciation of "ice cream"
then has at least two possible (and not contradictory) sources:
regularization to the default pattern, and replacement from the
stressing predicted by the rhythm rule in very frequent expressions
like "ice cream cone".

(saying that "iced cream" lost its final /t/ by "final t/d deletion"
and then was reinterpreted as a noun-noun compound doesn't say
anything about other phrases involving past-participial modifiers.
all i'm saying is that this particular fixed expression has changed,
and is now a noun-noun compound for speakers of english generally.
there are certainly other such phrases that have changed for some
speakers, or, possibly, have both structures for some speakers.)

arnold (zwicky at csli.stanford.edu)