Jakobson’s remark about a possible connection between Russian čët and četýre is discussed in Blažek (1999: 212-213) and especially in Greenberg (2001). Both authors mention earlier, more sketchy treatments of the problem, and they both add more Slavic material to the Russian words originally listed by Jakobson (which were čët, čëtka ‘even number’, četá ‘pair, union’, and čeť ‘quarter’). Blažek also notes an interesting potential cognate in Ossetian, an Indo-European language spoken in the north-central Caucasus (Ossetian is the only living descendant of the Northeast Iranian languages once spoken by the Scytho-Sarmatian inhabitants of the Eurasian steppe belt). The word in question is cæd ‘pair of oxen yoked together’, as if from Proto-Iranian *čatā (the Digor dialect of Ossetian has preserved a more conservative disyllabic form of the word, cædæ).
Blažek does not follow up Jakobson’s suggestion (presumably because he favours a different etymology of ‘four’, proposed by Schmid 1989; see pp. 213, 215, 331 in Blažek’s book). Greenberg, however, regards it as convincing and develops it further. Like Blažek, he considers the predominantly South Slavic *četa ‘troop, military unit’ (hence Serbo-Croatian Četnici ‘Chetniks’) to be part of the word-family of čët, and tries to explain the accentual difference between the end-stressed word četá (< *četa̍) in Russian and the root-stressed South Slavic forms – Bulgarian čéta, Serbian/Croatian čȅta, Slovene čẹ́ta (< *čèta) – in order to defend their common origin.
According to Greenberg, the word ‘four’ is derived from the root *kʷet- meaning ‘two’ extended with a multiplicative suffix, so that *kʷet-wor- means ‘(two) groups of two, twice two’. Greenberg also speculates that Proto-Indo-European *kʷotero- ‘which (of two)?’ (Greek póteros, English whether) contains the same root. This is hardly a good idea, since there is no compelling reason to question the straightforward standard analysis of *kʷo-tero- as the interrogative pronoun *kʷo- plus *-tero-, the IE suffix of binary contrast. The semantic gap between ‘two’ and ‘military unit’ is bridged by Greenberg as follows: Slavic *četa originated as the collective (in *-ah₂) of a word meaning ‘two, pair’, and ‘multitude of pairs’ evolved into ‘troop, group, band (of soldiers)’.
|Arranged in pairs|
There are serious problems with this derivation. First, (East/West) Slavic *četъ means ‘even number’, not ‘two’ or ‘pair’, while, on the contrary, the supposedly collective četá can mean ‘pair’ in Russian (beside some related meanings: ne četá, accompanied by a dative, means ‘not on a par with, superior to…’). What appears to be its exact cognate in Ossetian means ‘pair of oxen’, not, say, ‘herd of cattle’. Furthermore, while it’s true that the semantics of Russian četá covers not only ‘pair’ but also ‘troop’ (the latter attested already in Old Russian), we are probably dealing with a lexical merger between a native East Slavic word and a borrowing from Church Slavic (Czech četa ‘platoon’ is likewise a South Slavic loan, as are, ultimately, a number of similar “wandering words” in various neighbouring languages – Romanian, Hungarian, Albanian, and even Turkish). The non-attestation of intermediate meanings like ‘double column (of soldiers)’ makes it hard to justify the derivation of ‘troop’ from ‘pair’. Since the semantic difference is combined with a formal difference (conflicting accentuation), the etymology simply falls apart. It seems reasonable to conclude that the contrast between *četa̍ and *čèta is old and distinguishes two words of different origin (notwithstanding their merger in Russian). [See this comment, however.]
Jakobson’s final hypothetical relative of ‘four’, čeť ‘fourth part (of land), quarter’ (Old Russian četь ~ četъka), is in all likelihood a popular truncation of četverť (~ četvertka) < Proto-Slavic *četvьrtь ‘quarter’ < *kʷetwr̥-ti-, a noun corresponding to the widespread ordinal *kʷetwr̥-to- ‘fourth’. It is of course related to ‘four’, but in a rather trivial manner.
Etymological dictionaries often attempt to connect četa (in either sense) with the Slavic verb *čьtǫ (inf. *čisti) ‘count, reckon, read’, derived from PIE *kʷeit- ‘notice, recognise’. This verb has produced numerous derivatives in Slavic (e.g. *čislo ‘number’); some of them may be accidentally similar to members of the čët group both in form and in meaning, e.g. Old Czech čet ‘count, quantity’ (Modern Czech počet, with a prefix). Note, however, the gen.sg. čtu ~ čta. The disappearing root vowel reflects Proto-Slavic *ь (a reduced vowel continuing earlier short *i in the weak form of the root, *kʷit-). Despite their deceptive similarity, Russian četá (or čët) and Czech čet have different etymologies.
If we remove all the false or dubious cognates, we are left with just the initial material: *četъ ‘even number’, *četьnъ ‘even (of numbers)’ and *četa ‘pair’ – a word-family securely attested in East and West Slavic. We can safely add the Ossetian word (isolated in Iranian, as far as I know, but a perfect match for *četa, semantically and formally). There’s no evidence that the original meaning of the morpheme *čet- was ‘two’; nevertheless, it seems to have had something to do with arranging things in couples. Typologically, the Slavic “odd/even” terminology is parallel to what we have seen in Greek and Sanskrit, even if different lexical roots are involved. If so, one could expect *čet- to be semantically close to the familiar Indo-European roots *h₂ar- ‘fit together’ and *jeug- ‘yoke, connect’. I shall therefore tentatively assume that *čet- continues a verb root like *kʷet-, with the approximate meaning of ‘combine into pairs’. Let’s see if we can work from here – next time.
Václav Blažek. 1999. Numerals: Comparative–etymological analyses of numeral systems and their implications. Brno: Masarykova Univerzita v Brně.
Marc L. Greenberg. 2001. “Is Slavic četa an Indo-European archaism?”. International Journal of Slavic Linguistics and Poetics 43: 35-39.