Slavic and Baltic numerals

[Chrzaszcz Saproksyliczny]

Is there a historical relationship between words for 9 and 10, at least in Slavic and Baltic languages? Why are the initial syllables the same in 9 and 10?
Proto-Slavic: 9 devętь ; 10 desętъ ; Modern Lithuanian: 9 devyni ; 10 dešimt

 

[pollohispanizado]

From Wiktionary:

Дев'ять:

Ultimately from Proto-Indo-European *h₁néwn̥. Compare Latvian deviņi, Lithuanian devyni, from Eastern Baltic *dewin, ultimately of the same Indo-European root. The initial 'd' in Eastern Baltic and Slavic has sometimes been explained as dissimilation, or by alliteration to *desętь (“ten”) (compare a similar alliteration that may have occurred in Proto-Germanic between *fedwōr (“four”) and *fimf (“five”)).

Десять:

From Proto-Balto-Slavic *deśimt, from Proto-Indo-European *déḱm̥t. Cognate with Lithuanian dešimt, Latvian desmit, Old Prussian dessimpts.

 

[Sobakus]

Good question! Alliterative contamination in numerals is a well-known phenomenon. Here's what Trask’s Historical Linguistics (3d. ed) has to say about it:​

Numerals appear to be particularly prone to contamination, probably because they are very often used in sequence while counting. The Latin numeral for ‘nine’ would have been *noven if the word had developed regularly, but the classical form is novem, influenced by the following decem ‘ten’. Latin quinque ‘five’ should be *pinque, but the numeral has initial /kw/ due to the influence of quattuor ‘four’. Conversely, English four should be whour, but was influenced by the /f/ in five (as were all the Germanic languages). The Russian and Lithuanian numerals for ‘nine’ should have been *nevyni and *nevjat’, respectively, but the forms are devyni and devjat’, again influenced by the following desimt and desjat’ ‘ten’. The original Basque bederatzu ‘nine’, preserved in the east, has become bederatzi in most dialects under the influence of the preceding zortzi ‘eight’.​

​

Apart from the Basque one, which is new to me, these examples are standard Eurocentric textbook material, but I'm absolutely positive there are dozens upon dozens of parallel examples in languages from other parts of the world, and it'd be neat if someone can share a few of them.

 

[ahvalj]

In Baltic, n- is retained in the Prussian newīnts “ninth” (the word for “nine” is not attested) and, in Slavic, perhaps in the Czech nevěsil and Serbo-Croatian nevesilj for devęsilъ. By the way, -ev- should have become -av- in Baltic and -ov- in Slavic, so this e must have been influenced by the words for “ten” as well.

It is often suggested that the original Indo-European word for “six” was *u̯ekʲs and that the initial s- penetrated from the word for “seven” (with resulting variants *se-, *su̯e- or even *kse- and *ksu̯e- across languages; the Slavic and East Baltic word šestь~šeši~seši implies *ksekʲs->*kṣekʲs->*ṣeś-, Prussian again retains usts~uschts “sixth”).

The modern South-Western and North-Western Iranic word for “eight” (for example, the Tajik ҳашт/hašt) acquired its h after the word for “seven” (Tajik ҳафт/haft).

 

[danielstan]

Apart from the Basque one, which is new to me, these examples are standard Eurocentric textbook material, but I'm absolutely positive there are dozens upon dozens of parallel examples in languages from other parts of the world, and it'd be neat if someone can share a few of them.

Another Eurocentric example in Romanian:

rom. șase < lat. *sess (sex)
has the same ending with the next numeral in row:
rom. șapte < lat. septem

The normal evolution of lat. sex in Romanian would have been *șes or *șe
(see also rom. șes "plain" < lat. sessum)

 

[ahvalj]

A special case in the form of Baltic and Slavic collective numerals. In Proto-Indo-European, there were thematic outcomes for "three" and “four”: *trei̯om~*troi̯om from *trei̯es (compare the Sanskrit trayam) and *kʷetu̯erom~*kʷetu̯orom from *kʷetu̯ores (compare the Sanskrit catvaram). In Balto-Slavic they have given e. g. the modern Lithuanian treji : Russian трое/troje and ketveri : четверо/četvʲero, and hence -j- has spread to “two” producing dveji : двое/dvoje (back in PIE: compare the Sanskrit dvayam) and -er- spread to the numerals above “four” producing penkeri : пятеро/pʲatʲero, šešeri : шестеро/šestʲero and so on (the Russian сторица/storʲiʦa implies the former existence of even *съторо/sъtoro).

 

[Zec]

A special case in the form of Baltic and Slavic collective numerals. In Proto-Indo-European, there were thematic outcomes for "three" and “four”: *trei̯om~*troi̯om from *trei̯es (compare the Sanskrit trayam) and *kʷetu̯erom~*kʷetu̯orom from *kʷetu̯ores (compare the Sanskrit catvaram). In Balto-Slavic they have given e. g. the modern Lithuanian treji : Russian трое/troje and ketveri : четверо/četvʲero, and hence -j- has spread to “two” producing dveji : двое/dvoje (back in PIE: compare the Sanskrit dvayam) and -er- spread to the numerals above “four” producing penkeri : пятеро/pʲatʲero, šešeri : шестеро/šestʲero and so on (the Russian сторица/storʲiʦa implies the former existence of even *съторо/sъtoro).

If I may digress, these collective numerals are an excellent example of how actual speakers of a language are much less "rigorous" when doing morphological (re)analysis than linguists studying it. My linguist's instinct was that this sort of extension of -er- is implausible: it's obvious from *kʷetwores that the -er- in *kʷetwerom is a part of the root, how could it be extended to other numerals? But that's obviously what happened! This sort of relaxation of plausibility criteria would be really helpful when dealing with some other problems of historical morphology, e.g. the Latin -bā- imperfect.

To get back to the topic (sort of), isn't Spanish cinco instead of *cinque another case of this phenomenon? cuatro - *cinque > cuatro - cinco.

 

[ahvalj]

True, but such innovations are mostly introduced by kids who aren't able to analyze the language etymologically: they learn a pattern and then tend to apply it throughout. Some of the innovations survive till these kids' adulthood and become the new norm. Consider the French ordinal numerals in -ième, with the suffix extracted from the likes of centēsimum>centième and then applied to everything possible.

I agree with cinco.

 

[ahvalj]

Considering the Latin imperfect: it has been noticed many times that the stems and the principle of formation are mirrored in the Slavic imperfect:

-ā-bā- : -a-ax-​

-ē-bā- : -ě-ax-​

-i-ē-bā- : -*jě>ja-ax-​

C-ē-bā- : C-ě-ax-​

That is, if the producing stem ends in ā or ē, the auxiliary verb (that became the imperfect suffix) is added directly to it, if the stem is thematic or ends in i/ī, a suffix ē is added between the consonant and the imperfect suffix. This all suggests that the imperfects in both groups are remade from some inherited construction.

The suffix itself in Latin apparently is an ā-preterite of *bʰeHu- “to be”, that is *-bʰHu-ehₐ->*-bʰu̯ā->-bā- (in eram survived without the auxiliary verb); in Slavic it is a thematic aorist of some verb *ax- of unknown origin (pad-ě-ax-ъ : pad-ъ; pad-ě-aš-e : pad-e; pad-ě-ax-ově : pad-ově; pad-ě-ax-ǫ : pad-ǫ).