Thanks for your comment, Tavi.
That's entirely true that, though there is, IMO, certainly a relation between "to build" and "to tame", we might never by which it occured, and if it followed the same way as the French verb "dresser".
If I look at the LIV, we have:
for "bauen"/"to build"
Imperfective (ie. present): démH2-e
(uncertain) Desiderative : démH2-s
for "zähmen/to tame"
Imperfective : dm-né-H2
Perfective (aorist): démH2- (the "bare" form)
Desiderative : démH2-s (the same as above).
It's a pity that the LIV is not more precise about the form for "bauen".
If it is really an imperfective with a fixed accent and a thematic vowel (like the famous verb "bhéret/bhéront" = he bears/they bear), this implies that the perfective is *démH2 i.e. the bare form with "e" grade in the singular active and zero grade elsewhere, the same perfective than the form for "to tame".
This analysis seems to be confirmed by the fact that both verbs had a similar desiderative : démH2-s
This form entails that the perfective (ie. aorist) ought not to be a sigmatic one : **dêmH2-s/**démH2-s; this form would have been quasi-homophonous with the desiderative.
So, there is some probability that those verbs had indeed the same perfective.
However, as Tavi said, this does not imply an evolution similar to French "dresser".
For example, if we admit that dom- and demH2 were from the same root (implying **domH2 then), the form *dm-né-H2 could have meant something like "to give a *dom- to (an animal)" (cf: "domestiquer" in French...)
(cf. *yu-né-g = "to join; to put a yoke on")
In this case, we would have an evolution : *démH2-e (to build) > **domH2- / *dom- (house[hold]) > *dm-né-H2- (to tame = "to give a household to an animal)
Just a proposition...
Some time ago, I remember I read on the Net a study according to which the first element of Latin "materia" (whose first meaning was construction material, carpentry; cf. French "madrier" = "big pole", Portguese "madeira" = "wood") would come from *dmH2 > dma- > ma- (teria)
Olivier
> > A compound like *d�mH2pedom = "house floor" shows the original
> meaning of *d�mH2-, that I interpret as similar to the French verb
> "dresser" which can both mean "to put up" (ex: a pillar) and "to tame".
> >
> Once again, you're extrapolating to PIE something what happened in
> French. To say the least, this is very dangerous, but I have to admit
> that it's still possible for PIE *domH2- 'to tame' to be a
> SPECIALIZATION of *demH2- 'to build' instead of being homonymous.
>