In short : what's the final -t in déḱm̥t?

Full details : The Proto-Indo-European root for ten is traditionally defined as déḱm̥/déḱm̥t(ᵃ). The final -t may be analyzed as a casual ending, e.g. as a "collective form"; Szemerényi describes déḱm̥ as derived from déḱm̥t(ᵇ).

The final -t is required by various languages, like OCS desętĭ and Lithuanian dešimt(ᵃ). Likewise, according to Don Ringe, the Proto-Germanic form for ten is derived from déḱm̥t, surfaced as déḱm̥d(ᶜ), hence tekun(ᵈ) leading to Gothic taíhun(ᵉ).

My question : I would like to know the details behind this -t : if it's a collective form ("marking collective number (group of ten)"), is there any other example of a PIE word using -t as a collective form? The only way I know to create collective PIE forms is (for direct cases) by adding the ending -h₂ to the stem.

Any help, any reference would be appreciated !

related question here.

(a) See e.g. (Szemerényi) Introduction to Indo-European Linguistics, 8.5.2 p.223

(b) (Szemerényi) Introduction to Indo-European Linguistics, 8.5.3 p.226 Reconstruction:Proto-Indo-European/déḱm̥ (Wikipedia)

(c) Don Ringe, From Proto-Indo-European to Proto-Germanic, 2.2.4(iv) p.20

(d) Don Ringe, From Proto-Indo-European to Proto-Germanic, 3.2.2(i) p.81

(e) Don Ringe, From Proto-Indo-European to Proto-Germanic, 3.2.2(ii) p.87

  • 2
    Don'get confused with Slavic (and possibly Baltic?) languages. Our numerals from 5-10 are derived from ordinals of the original numerals and all end with -t because of that (also they have rather unusual syntax because of this).
    – Eleshar
    Commented Nov 26, 2016 at 11:53
  • @Eleshar : thanks, I didn't know. Is there any online reference describing the way OCS numerals have been made ?
    – suizokukan
    Commented Nov 26, 2016 at 14:43
  • 2
    I can try looking up something but it is fairly easily visible even in modern Czech - you have penkwe/pemnk-to-s - pě-t; sweks/sweks-to-s - šes-t; septm/septm-m-os - sedm (cf. hebdomos); h3ektó/h3ektow-o-s - osm; h1newn/h1newn (cf. nin-th) - devě-t (probably from earlier nevě-t); dekm/dekm-m-os (cf. ten-th) - dese-t
    – Eleshar
    Commented Nov 26, 2016 at 16:01
  • Looking at my comment, I was obviously incorrect - not all the mentioned numerals end with -t (but all developed from the ordinals - this is also evidenced by the strange syntax, numerals 2-4 work like regular adjectives and are followed by plural; 5+ is behaves like a head of NP in Nom/Acc where it is followed always by plural genitive). Reg. the developement though, note also the principle of paired rhyming in eg. Czech: tři/čtyři; pět/šest, sedm/osm, devět/deset.
    – Eleshar
    Commented Nov 27, 2016 at 11:55
  • 1
    In Phrygian dekmoutais means 'tenth', not plain ten (10). Unfortunately, we lack the word for plain 10, which makes it hard to pin point the function of -t.
    – Midas
    Commented Jan 4, 2017 at 13:12

2 Answers 2


The only language to preserve both is Sanskrit: dáśa- 'ten' ~ daśát- f. 'a set of ten; decade'.

A similar form is Lith. dešimtìs which reflects a Balto-Slavic consonant stem, as shown by gen.pl. dešimtų̃, gen.sg. dešimtès, dat.sg. dešimti:

Ponas … pasirodo dešimti savo apaštalams. 'The lord appeared to his ten apostles.' (Bretke, Postilla II, 35)

Furthermore, the OLith. and dial. indeclinable form dešim̃s probably represents a fossilized nom.sg. *deḱm-t-s. The Lith. form is therefore identical to the Sanskrit one. The same formation is found in Old Prussian dessimpts (II) 'ten'. Standard Lithuanian dẽšimt could be from *deḱm with -t- on the basis of this form (dial. dešim̃ 'ten' is actually attested).

Gm. *tehun has restored -n- on the basis of ordinal *tehunþas. PIE *deḱm should have given *tehu with a lost *-m, a Verner variant of which, *tegu is preserved in the suffix fif-ty, OE -tig, OHG -zig.

It appears that IE *deḱm was an indeclinable numeral like *penkʷe, *sueḱs, etc., while *deḱm-t- was a derived feminine t-stem, possibly meaning 'a group or set of ten', cf. the Sanskrit meaning, which developed into a cardinal number in Baltic.

The t-suffix is also found in the decades, cf. *penkʷe-dḱomt- > Skt. pañcāśát- f., Gr. πεντήκοντα 'fifty'.

I am not aware of any cases where a t-suffix has a collective function, however the function of the suffix here is rather nominalizing, i.e. *deḱm 'ten' vs *deḱmts 'a ten'.

  • 1
    Why the difference in the reflexes of *ḱ in the two Sanskrit words?
    – TKR
    Commented Oct 21, 2018 at 17:25
  • 1
    Sorry, I accidentally copied the Proto-Indo-Iranian reconstruction rather than the Sanskrit form. Corrected. Commented Oct 21, 2018 at 18:00

It depends if you're ready to believe it means 'two hands' in which case the 't' is the last letter of the root of *komtos > PG *xamðaz > NE 'hand'.

  • lol " *komtos > PG *xamðaz " is a common, standard and still current Abanian for legs not hands. Commented Nov 29, 2016 at 4:05
  • If it is from the root for hand, it is a suffix. Because the word for right hand came from the root for "right" dec-.
    – Anixx
    Commented Jan 4, 2017 at 3:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.