# Is it possible that a language's number system is inherently wrong?

Is it possible due to mythological reasons or some other human error (what can be the reasons of those errors in that case) that a language's number is flawed for e.g. ->

Copainala Zoque number system

Rayon Zoque number system

In both cases (probably because they are neighbours/dialects) they miscount 8 and 9 as 9 and 10 respectively. (Or is this just some negligence on the linguist's side, which doesn't seem to be the case, as the miscounting is specifically indicated)

Also does anyone know of more examples like this?

• It's not clear to me what you think is actually wrong here. Any number is equal only to itself, as everyone knows. But what morphemes do may be completely unrelated. (Compare with an English word like "understand", which has nothing to do with standing or being neath under something.) Apr 6 '18 at 11:12
• These look more like dialects of a single language than "neighbors". My first suspicion would be that the linguist's postulated addition is incorrect -- that they have a different logic by which "three and six" really does result in "eight" where "and" stands for some other operation or relationship than addition. Apr 6 '18 at 11:22
• If I was able to postulate precisely how this works, I would post an answer. This is just a comment on my immediate intuition. Apr 6 '18 at 11:30
• Maybe related: In French, we commonly say «huit jours» (8 days) or «quinze jours» (15 days) to designate one or two weeks, but we almost never say «sept jours» or «quatorze jours», which sounds more exact for the same period. I’ve heard it was linked with the Roman ways of counting days (including the first and last day of the interval), but I guess it would qualify as “inherently wrong” for you. Apr 6 '18 at 14:06
• I've seen this kind of thing in PNG. It's normal that word-internal morphology is opaque to speakers, so a number word may shift meaning but still appear inside other number words as a morpheme implying the old meaning. I would suggest in these examples that 'tuhtaʼy' "6" has shifted meaning and at some earlier time meant "5". (Of course this assumes the data is connect). Apr 6 '18 at 22:00

I don't think a number system can be inherently wrong any more than gravity can be wrong; but a system can be confusing from a historical perspective. Numbers seems to be highly subject to reanalyses. From that number page, you can get additional examples of Mixe-Zoque numerals. These are the numerals 1-12 in Quetzaltepec Mixe, Juquila Mixe, Chuxnaban Mixe, Tontontepec Mixe, (Highland Popoluca), the latter parenthesized since they switch to Spanish for 6 and higher.

``````1: tuʼuk, tuʼuk, tuʔuk, toʼk, (tuum)
2: mæhck, mæhck, mææht͡sk, mæhck, (wɨsteen)
3: tɨɡɨɨk, tɨɡɨk, tɨkɨɨk, toohk, (tukuteen)
4: mahktaxk, mahtaxk, taʃk, maktaaxk, (maktasteen)
5: mɨɡɔxk, mɨɡɔxk, mɨkoʃk, muɡɔɔxk, (mosteen)
6: tɨduhk, tɨduhk, tɨtuuhk, tohtɨk
7: wɨxtuhk, jɨxtuhk, hʃtuuhk, wuxtohtɨk
8: tuktuhk, tuktuhk, tuktuuhk, toodohtɨk
9: taxtuhk, taxtuhk, taʃtuuhk, taaxtohɨk
10: mahk, mahk, maahk, mahk
11: mahktuʼuk, mahtuʼuk, maahktuʔuk, maktoʼk
12: mahkmæhck, mahmæhck, mamææht͡sk, makmæhck
``````

Compare to Rayón Zoque, Copainalá Zoque:

``````1: tumɨ, tumɨ
2: meca, meca
3: tukaʼ, tukaʼy
4: makšku, makškuʼy
5: mosaʼ, mohsaʼy
6: tuhtaʼ, tuhtaʼy
7: kuʼyaʼ, kuʼyaʼy
8: tukuduhtaʼ, tukuthtaʼy
9: makstuhtaʼ, makstuhtaʼy
10: mahkaʼ, mahkaʼy
11: maktumɨ, maktumaʼy
12: makwɨstɨhkaʼ, makwɨstɨhkaʼy
``````

The Zoque numbers for "8" are sufficiently similar to the numbers in the other languages that I could accept that these reflect the earlier situation, whereas "9" diverges substantially.

You can engage in the exercise of comparing the bits of these numerals and discern partial similarities across the languages. Zoque "9" makstuhtaʼ looks like it contains "4" makšku, but it also looks like it contains "10" mahkaʼ ("1 from 10"). Zoque "6" and "3" seem to have common elements. Zoque "12" seems to have the elements of "10" and "2", except that that "2" element itself is only found in Popoluca.

There is a common enough historical pattern across languages that adjacent numbers influence each other. If "8" tukuduhtaʼ looks like "3-6", and if Zoque "9" should have been takstuhta, then it is not surprising that Zoque "4" makšku could have influenced the change of takstuhta to makstuhta. There is an analogous influence in Russian (and maybe all of Slavic, I don't know) where expected nev'at' appears as dev'at', under the influence of des'at' "10", and similarly "8" vos'em was influence by "7" s'em.

• amazing as always and that probably answers my question. Though there is one thing, Quetzaltepec Mixe, Juquila Mixe, Chuxnaban Mixe, Tontontepec Mixe, (Highland Popoluca) all have "tuhk" at the end of 6-9 (and duhk where voiced) from the data we can maybe derive the fact that they are all just 5+1, 5+2, 5+3, 5+4 with tuhk representing 5 and toh, wux, tuk, tax (taʃk) being 1,2,3,4 in some dialect or other. Then, again we are stuck with 5+3=8 and 5+4=9 which is not the case with the dialects in the question Apr 7 '18 at 1:26

There are examples like this, and we all use that example every day.

Look how Arabic digits look like, compared to Devanagari, Gujarati, and other Indic scripts.

"૫ ૬ ૭ ૮" are "5 6 7 8" correspondingly, but they visually resemble "4 5 6 7".

Source: Wikipedia

• Oh I never thought about that! Apr 6 '18 at 13:29
• Darn, I didn't notice this correspondence until now! Apr 6 '18 at 13:38
• What's even more interesting is that the Gujarati 4 looks like 8, and the Devanagari 8 looks like 5 Apr 12 '18 at 18:40