> I am not trying to be contentious, Piotr, by insisting on this[snip]
> point. But I think the explanation offered by Rob is likely to be
> the better hypothesis, namely, that because of the existence of the
> non-baby talk *máº´r.-, when the baby produces /ma(ma)/, 'mother'
> responds encouragingly, reinforcing the baby's use of /mama/ to
> summon mother and her breasts. If 'mother' were a
> fictitious /dzali/, she would theoretically respond to /da(da)/,
> eventually perhaps /dza(dza)/; and this would become the baby-talk
> equivalent in this 'language'.