Re: Does every language family contain one with "ma-" "da-" "ta-" words for parents?

On Wed, 10 May 2006 16:33:34 +0100, Chris Bates
<chris.maths_student@...> wrote:

>> A "ta-" word is or begins with "ta-" or
>>"ata-" or "tVta-" where V is a vowel.
>>
>>
>You may want to modify that. Basque has aita for father, which while
>similar does not fit any of those patterns. Perhaps you should have ta-
>or Vta- or -tVta?
Yes, you're probably right.
The problem is to find out the exact hypothesis actually proposed by the
linguists who propose it.
I'm hoping Paul can get me or us started on a literature search to find out.

>=========================================================================
On Wed, 10 May 2006 16:49:26 +0100, Chris Bates
<chris.maths_student@...> wrote:

> I missed the bit at the bottom where you were talking about disputed
>isolates. But it seems to me that by that criterion we can pretty much
>dismiss all counter examples, since for every isolate out there some
>nutter has proposed that it is related to another language with little
>real evidence.
No; I said "_serious_ doubt or dispute".  The trick, obviously, will be to
define "serious"; but a few isolated "nutters", as you call them, won't
qualify.

>I think this question is probably unanswerable in the
>absolute sense, since I doubt that we'll ever all agree on all the
>links, given that good reconstruction can only really go back so far,
>and without reconstruction lexical item similarities can be explained
>away by chance or borrowing at some point.
Well, you could always claim, even with an "isolate", that even if it
doesn't have any "ma-" or "da-" or "ta-" words for "mother" or "father", it
must have had some extinct relative that did.  So you are right, in a way,
in saying it's "unanswerable".

I am saying, however, that if the hypothesis is that "every _extant_
language family contains an _extant_ member that etc.", and if we wish to
_disprove_ it, the _easiest_ way is to investigate every member of the
_smallest_ families first.

I admit that permanently ruling out extinct languages and languages whose
family-membership is in serious doubt or dispute, will leave room for doubt
at the end.  The point is that it is my opinion that such languages could
be left til last, since their probative value will be dubious either way.

> I also question the value of the proposed universal. What value is
>there in claiming that a member of every language family has such a set
>of words? I simply don't see how it advances our knowledge of
>language... if what we want to say is that most languages contain such
>pairs of words, then why not simply phrase it that way and ignore
>genetic relations when doing the study?
I am not saying the proposed universal necessarily has value.  I am merely
proposing a way to test it -- a way I think is close-to-best.  I happen to
doubt the hypothesis a lot.  I think that if it is true, it must be true of
each of the first 72 languages and 5 families I mentioned; and it should be
possible to check each of them; and it should be a lot easier to check all
of them, than to check all the languages in a big family such as Niger-
Congo.

As for its value -- as Benjamin Franklin said, "Of what value is a baby?".
Most people would question the value of linguistics as a whole; many
linguists question the value of conlanging; and some linguists question the
value of other linguists' subspecialties.  Let those who refuse to ever do
anything that has no value, just ignore the question.

>Unless, of course, we're trying to prove some barmy proto-world theory
If you look again you'll see I specifically ruled out "Proto-World" as an
object of this investigation.  If "World" is accepted as a "language
family", the investigation is easily concluded by showing there is even one
language with "mama" for mother and "dada" for father.

>(forgive me for not having read what led up to this post),
Not your fault -- three of the posts in this thread were off-list.

>but in that case I don't see why we're insisting that every family must
>contain such words.
Originally I questioned Paul Bennett's Sun May 7, 2006 4:10 pm message on
the thread "Re: Common words for man & husband, woman & wife" that "the
ma / da pattern is found in every documented language family on the
planet".  He responded that he had said "every documented language
_family_", rather than "every documented _language_".  I am responding that
there are many very small documented families; so it should be easiest to
start with them.

[snip]

Thanks.

eldin