Re: Why there is t- in German tausend "thousand"?

2013/11/14, gprosti <gprosti@...>:

>
>
> --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> <bhrihstlobhrouzghdhroy@...> wrote:
>>
>> 2013/11/13, gprosti <gprosti@...>:
>> >
>> >
>> > The question is *how* statistically probable it is that two words with
>> > the
>> > same meaning, which share five phonemes in the same sequence, are in
>> > fact
>> > two completely (historically) separate forms.
>>
>> *Bhr.: 100% probable.
>
>
> It's 100% probable -- i.e., physically or logically necessary -- that two
> semantically matching words with a long, matching sequence of phonemes, but
> one non-matching phoneme, must have no historical affinity whatsoever?

*Bhr.: if they irreducibly differ even in just one phoneme in the
root, especially word-initially, they ARE different by definition

>
> In case it's unclear, "historical affinity" here includes the possibility
> that the two words are dialectal variants, that one is a loan from the
> other, etc.

*Bhr.: Please find apt Upper German features in OHG in order to
explain t- for th-. Modern dialects are not enough, once there's old
evidence to the contrary

>
> Even if historical affinity doesn't include these cases, the probability is
> still not 100%, because the laws (= tendencies) of sound change are not laws
> of mathematics or physics.

*Bhr.: they are logical laws, otherwise one could not demonstrate
whether a given etymological hypothesis is correct or not

>
> It's not my fault, it's phonology. If two words
>> differ by a phoneme, they are minimal couple. If two words have the
>> same meaning, but different phonemes (even if just one: one is more
>> than zero), they are synonyms. If You want to underline the
>> statistically possible - although rare -
>
> Rare = a low probability.

*Bhr.: well, phonological irregularity is NO probability, so better
low than nothing

>
> case that they are synonyms
>> differing by just a phoneme, You can label them "paronymic synonyms".
>> Are paronymic synonyms possible, or are we obliged to convert them
>> into genetic cognates?
>>
>
> If
>
> - a set of words shares a non-trivial number of consecutive phonemes (again,
> I would say 5 is a non-trivial number)

*Bhr.: Four out of five are suffixes or second members of compound,
they are identical in both words in my etymology as well, so they
cannot be taken into consideration

> - they only differ in one phoneme

*Bhr.: The root is constituted by two phonemes, so they differ at 50%

> and
> - there is no clear process (morphological or otherwise) whereby the words
> could have been independently created in each language from separate
> components

*Bhr.: I can't understand. This is anyway not the case. A preverb
*dhu, the root *h1es-, a participle in *-nt- and the feminine ending
*-ih2 are trivial elements in Germanic Erbwortschatz

>
> then I would say the most probable explanation of these words' similarity is
> common origin (which, again, can involve loaning).

*Bhr.: Were it so easy, I think Kluge, Mitzka, Seebold &c. wouldn't
have had to wait for You in order to discover such a perfect solution.
In such cases, the solution can be either in something they didn't
know (like e.g. Hittite hapax tu-up-ra 'bound' or a kind of evil as a
comparandum for Zauber, as I've proposed some years ago) or in
something they couldn't admit, i.e. surely not diatopic variation or
loans, but - as You too seem completely incapable to accept - the
emergence of a couple of paronymic synonyms, yet Old Norse þúshund
(and Runic Swedish þūshundrað) vs. þúsund, Old Slav. tysęšti, tysǫšti
&c. show a similar opposition

>
> This is not the same as dismissing a dissimilarity between phonemes (t- vs.
> d-, etc.) entirely.

*Bhr.: I really can't understand. Who dismisses and who doesn't?

>
>> >
>> >>
>
>> >
>> > Haplology is a similar type of process: e.g., standard Spanish shows
>> > the
>> > simplification of *tenisista to "tenista", but no haplology of
>> > "narcisista"
>> > to *narcista.
>> >
>> >
>> *Bhr.: There's no haplology of *tenisista to tenista, rather there's
>> a readjustment rule quite like that of narcisista:
>> narcis-ismo > delete -ismo > add -ista > narcis-ista
>> tennis (< French ten-ez) > delete -is > add -ista > tenn-ista
>> (celt-a > delete -a > add -ista > celt-ista)
>>
>
> Proof?
>
>

*Bhr.: As proven by celt-ista, -ista does delete stem-final endings.
There's no word †narcis; tenista : tenis = narcisista : X, where X
surely doesn't end in -is like tenis