Hallo conlangers!

On 23/01/2018 17:35, David McCann wrote:

> On Mon, 22 Jan 2018 21:57:07 +0100
> Jörg Rhiemeier <[log in to unmask]> wrote:
>
>> Also, Greenberg's proposal is based on poor method - compiling large
>> lists of lexical similarities *without* establishing regular sound
>> correspondences (this equally holds for his "Eurasiatic").
> So you'd reject Afro-Asiatic and Sino-Tibetan then? Neither of these
> have an agreed reconstruction of the proto-language and a list of the
> sound changes that create the daughter languages. And why do Uralicists
> have symbols for "any front vowel" and "any back vowel"? Because the
> changes there are not always regular, so the vowel cannot be
> reconstructed except for its function in harmony.

I wouldn't reject either Afro-Asiatic or Sino-Tibetan because there are
clearly cognate morphologies, even though lexical reconstruction may be
wanting. Indo-European was established just that way, too, as you will
certainly know. Lexical cognates came later. I would even be so bold to
say that the morphological resemblances between Indo-European and Uralic
are very, very, very hard to explain otherwise than by descent from a
common ancestor that may have been spoken perhaps 8,000 to 10,000 years
ago somewhere north of the Caspian Sea. Some other language families
also show such affinities, though at a lesser degree. They'd be more
distantly related. The morphologies of Afro-Asiatic and Dravidian,
though, are wholly different from this; and Bomhard does not reconstruct
a comprehensive morphological system of Proto-Nostratic but presents a
smörgåsbord of morphemes from which each family "picked" some and in
which each category is represented multiple times. I simply can't
believe that any language could have as many as six different
first-person morphemes.

I know that there are cover symbols in Uralic, too, which is of course
also an established family. There are even such cases in Indo-European,
where there are numerous words where we know that there was a laryngeal
but not which one. It is all a matter of *degree*, though. If you look
at Dolgopolsky's Proto-Nostratic, there are almost as many cover symbols
as there are "known" segments.

I see we disagree on these matters, and have disagreed in the past. I
would not say that the various proposed macrofamilies can't be true -
many of them indeed seem rather plausible - but I am unconvinced by most
of the proposals I have seen. We can leave it at that. And at any rate,
such proposals can be used as a source of inspiration for conlangs!

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1