Re: [tied] Re: the true nature of

On Lachmann's Law

The question whether Lachmann's Law (Lat. factus : a:ctus) is analogical
or phonetic is VERY easily answered if one remembers that diachrony takes
time (that's why it's named that way, by the way). That there is no
*pervasive* lengthening before, say, suffixal /t/ in roots ending in
mediae is clear from the counterexamples which are very solid: fissus,
sessus, scissus. Interestingly, tussis 'cough' has short u, while
the ptc. tu:sus 'hurt, struck' has /u:/. This indicates that archaisms do
not have the lengthening, for 'cough' is a good sign that the association
with the verb was severed.
Lachmann's lengthening thus befell those words which, after the
assimilation of (de)voicing, as *ag-to-s > *aktos, were *rederived* with
restoration of the root shape as known from other inflectional forms, in
casu *ak-to-s -> *ag-to-s. From that point on, what happened was
*phonetically regular*, i.e. *the change is a sound law* by which the
extra dash of sonority of the /g/ was imparted on the preceding vowel
which kept it even after /gt/ was assimilated to /kt/ the second time
around, the preceding vowel being now longer than in words without that
history.
Now, even a sound law only works on material the language has, so if
participles like *setstos, *fitstos (morphologically *sed-to-s, *fid-to-s)
were not changed to **sedstos, **fidstos, there just was no input to yield
**se:sus, **fi:sus, and they do not turn up. Not suprisingly *H1d-tó-s,
however realized, was restored, and so *edstos surfaces as e:sus.
Lachmann's Law is a phonetic event in the prehistory of latin working
on forms brought about by a previous analogical restoration.

Jens