Discuss results - consonants in lexicostatistics - M and B

M and B consonants
Posted by: Mehmet on Jul. 19, 2015, 15:42
This work is great and exciting. It's nice to use computer, programming, c#, alogirthm, math, etc.. for such area.
I realised that you use the term "Related consonant match" for example for -K and G consonants. But you don't regard consonants -B and -M releated.
It obviously seems that -M and -B are also releated consonants. At least it seems to me, as a turkic language speakers, obvious. (I can see the releation between -B and -M also in local dialects.)
For example:
English: I - nose
Turkish: Ben - burun
Kazakh: Men - muryn

Maybe, the system you use for sound correspondence, works only for indo-european languages. This is a guess but I hambly recommend you to research different approach for consonant releations.
Good bye.
Posted by: Mehmet on Jul. 19, 2015, 18:28
SH and -TH Sounds
I got the points for correspondence of sound in the link below:
Sound correspondence
You have given the following points:
S - TH : 73.78
S - SH : 93.88

TH - SH : 0 (I think, TH should be not zero because TH is related to S and also SH is related to S. TH should be related to both S and SH)
For that reason, in comparison Hebrew and Arabic, the result was affected negatively.
If we look at following example, it is understood that those words is closely related and also be understood that -TH and -SH sounds are releated. If we research more Arabic and Hebrew words, we can find more example, I think. We need someone who knows both Arabic and Hebrew.

Posted by: Vincent on Jul. 20, 2015, 21:28
Hello Mehmet,
Thank you very much for your input. I see you have a perfect understanding of the system. The issue with TH - SH is right, I will check my source when I come back from holiday. I try to take only sound correspondences into account which are universal or at least don't bring more exposure to chance than potential cognate recognition. There are many "local" correspondences like B - M which are such a case that bring more exposure to chance overall than advantages for some language families. One of my biggest "frustration" so far is not finding a link between Japanese/Korean and Altaic languages and this is probably due to the limited sound change rules which are in use so far. But with suggestions like yours, the system will be improved step by step... On the other hand, the link between Uralic and Turkic languages gets identifies with a fair probability that the resemblances are not due to chance, the same with Chechen and Basque. But of course,there are still many limitation - every result is a balance between signals which are due to chance and signals which are due to cognacy.
Thanks again for your founded feedback and feel free to suggest any idea in the future. From my side, I will inform you when I make the changes (TH - SH).
Posted by: Mehmet on Jul. 28, 2015, 21:56
Hello Vincent,
Thank you for your response and good comments.
I realised something else.. You use Turkish word "ruzgar" for wind but the word "ruzgar" is a loan-word from Persian language. So it's not suitable to use this word. I recommend you to use Turkish word "yel" which is synonym of ruzgar/wind.
It would be more meaningful for tracking evolution of the word.
And one more little thing. Arabic word for death "موت" is sounded like MAWT not like MUT.
I hope your holiday is/was great. :)
Posted by: Vincent on Jul. 29, 2015, 10:07
Dear Mehmet,
Thank you for this, it is very important indeed... I could detect only the Arabic loanwords in the Turkic word lists as I have learned some Arabic myself but I have zero knowledge of Persian... This change will bring the Uralic and Turkic languages a little bit closer in the system ("L" in common) ;-)
I will probably find time to make these changes next week...
Two interesting experiences by the way:
- Re. Loanwords Arabic -> Turkish, I have become aware of them mostly because of my knowledge of... Croatian. Serbo-Croatian/Bosnian have numerous Turkish loanwords still today and some of these were originally Arabic loanwords in Turkish...
- Re "موت" - you are right, the "Mawt" transcription is more exact. One anecdotal thing about this stem: "-M-T-" and/or "-M-R-" / "-M-R-T" is quite strongly represented in Indo-European, Semitic and Malayo-Polynesian for "Death" or "Murder", so in the broader semantic cluster around the concept of dying/killing. Do you known some "M/T", "M/R" or similar stems in Turkic languages in the same semantic cluster or any experience with this?
I would be pleased to learn more about your fields of interest/research as you seem to have a very broad knowledge...
Many thanks & Best regards
Posted by: Mehmet on Aug. 3, 2015, 16:03
Hello Vincent,
I don't know any stems similar to "M-R or M-T" for dying/killing concept in the Turkish language.
See you
Blog author: