In PIE we have

du̯is twice

du̯iteros second

du̯oi̯os twofold

du̯eiplos double

etc, with the root du̯ei̯-

At the same time we have:

du̯eiros fearful

du̯eisos hated

with seemingly the same root.

Is there any etymological or semantic connection?

P.S. After making this post I found that de Vaan postulates the following: *dus 'into two > bad'

  • 3
    Very interesting question. What would the concept be for such a connection? What are the possibilities that this is a coincidence?
    – Midas
    Feb 28, 2014 at 11:31
  • 1
    It's hard to evaluate these PIE reconstructions, or determine whether the same root is involved in all of them, without seeing the actual comparative data from the attested languages in each case. Scholars differ in their reconstructions, and probably not everyone would agree that all of the PIE forms you list really existed, so you might get better answers if you gave the evidence for each form.
    – TKR
    Mar 1, 2014 at 21:35
  • @Midas there are known parallels: the dis- prefix which means "wrong". Compare disfunction, disorder, disorientation while it originated from the same root for 2.
    – Anixx
    Mar 28, 2014 at 9:21

2 Answers 2


This connection is possible, perhaps even plausible, as long as you remember that Proto-Indo-European is not a real language but a theoretical construct. Everything about it is conjecture.


Thing is, the PIE prefix meaning "bad" was *dus- from the root *dwe-, just as "two" was *dwṓh1 from the root *dwo-. Their similarity is probably just coincidental.

I cannot find any explanation for the form *dweysos as that doesn't fit the participle pattern (*-ónts *-ntés *-óntih2 *-ntyéh2s *-mh1nós) nor for *dweyros as it doesn't fit an adjective derivation pattern (*-yos *-iskos etc.) Check your sources, please, then tell me: there might be something I'm missing here.

  • Root du̯e? What made you think such root could exist? Vowels are prohibited at root end in PIE.
    – Anixx
    May 1, 2014 at 11:08
  • At the very least, the roots dwo- (two), kwṓ- (dog), dra- (to run), ghowro- (fear), solwo- (entire), médhu (honey, mead) and nogwó- (nude) end in a vowel; dwe- is just as likely as they are to exist.
    – Darkgamma
    May 1, 2014 at 14:07
  • There's no root kwṓ-, there is cu̯en- (cu̯ōn "dog"). The root dra- is twice impossible because PIE had no a. (so the root could be dreh2-), ghou̯ros has the ending -os (and r is suffix, adjectives in -ros are common), solu̯os has the ending -os and u̯ is suffix (adjectives in -u̯os are also common), u in médhu is resonant (counted as consonant in PIE when considering the root structure), and most likely the root was medh- because the resonant could not be further from the kernel vowel than dh, with -u- as suffix, -os in neğnos is ending (and -n- is suffix).
    – Anixx
    May 1, 2014 at 14:27
  • 1
    I don't know to what theory you subscribe, but Indo-European as known today is realised to have had /*a/. I believe only Beekes (and by extension Lubotsky) is expressly against its inclusion. The vowel in *dra- is short, therefore there can't have been a laryngeal after it; ḱwṓ- (what is that /c/?) was a hysterokinetic noun - it's reflected without that final /n/ you cite in Hittite, Celtic, Tocharian and Slavic. Even if you exclude all those other roots (validly or not), you're still left with *dwo- and *ḱwṓ-
    – Darkgamma
    May 1, 2014 at 15:25
  • 1
    @Anixx, it doesn't have to be either/or; there are some occurences of /a/ that just can't be explained with a laryngeal
    – Darkgamma
    May 3, 2014 at 7:27

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.