The double zero grade *ǵʰi-m- is preserved in the compounds with numerals.

(de Vaan, Etymological Dictionary of Latin 2013: hiems)

E.g. *dwi-ǵʰim-os “two years old”, literally “of two winters” (en.Wiktionary: https://en.wiktionary.org/wiki/Reconstruction:Proto-Indo-European/%C7%B5%CA%B0%C3%A9y%C5%8Dm)

Unbelebtes *h1ésu-∅ kann zwar durch gr. εὖ, nicht aber durch altheth. šu oder ved. śu fortgesetzt werden, d.h. das Adverb *h1sú-∅ stellt keine erstarrte Kasusform dar. Vielmehr wurde es von der Komposition her, wo die Doppelnullstufe eines nominalen Vorderglieds durchaus am Platz ist (s. *du(-s-) Fn. 18), verallgemeinert.

(Dunkel, Lexikon der indogermanischen Partikeln und Pronominalstämme, Bd. 1, p. 305)

In short, Dunkel says *h1sú-∅ was generalized as *h1ésu-∅ from the double zero grade of some (“eines”) noun prefix.

If I understand correctly, this means that *dwi is zero grade, *ǵʰi-m- is zero grade, so the compound is double that. The parallel in Dunkel's notation *-∅ makes that very clear.

I just don't understand the implication, because it is usually enough to point out the single zero grade and the evidence for it.

  • Double zero grade in numeral compounds may be ominous, as if the full grade might be derived from metanalysis of a consonant cluster as zero grade in one of those words, requiring an alternative etymology for it, but there is no indication of that.

    • De Vaan's bracketing *-m- is obvious, not to me though.
  • Dunkel's analysis doesn't profit from double zero grade. I don't have a full copy, but there is no obvious reason that the zero grade would occur first in double zero grade.

    • FYI, note that Wiktionary lists *h1su- as uncertain under *h1es- and offers an alternative explanation from *h1wes-.

Am I missing something? I am hesitantly not posting this as a reference request because I don't believe it is a thing. Convince me I am wrong.

Following @Janus' comment, *-m- is zero grade of the Genetive suffix, full-grade **-ém-, o-grade **-ōm-.

But still, this seems diametrical opposed to Dunkel saying that "*h1sú-∅ stellt keine erstarrte Kasusform dar", precisely not fossilized in case form.

  • 3
    De Vaan is saying that *g̑ʰi-m- is a double zero grade in itself, without even considering the form of the numeral: he takes the base form as being nom. *g̑ʰei̯-ōm (full-grade root, long o-grade suffix), acc. *g̑ʰi-ém-m̥ (zero-grade root, full-grade suffix), gen. *g̑ʰi-m-ós (zero-grade root, zero-grade suffix = double zero grade). The double zero grade found in the PIE genitive was lost in the inflectional paradigm of the base word in Latin, where the zero-grade/full-grade combination was generalised from the accusative, but it survived in compounds with numerals. Nov 26, 2023 at 20:12
  • 2
    As for the Dunkel quote, I’m not sure I quite follow his implications either, but as best I can tell, he’s assuming that *h₁su- is from a combination of *h₁es- + *-eu̯- (≈ a u-stem to *h₁es-), and then saying that the adverb *h₁su cannot be a frozen case form of this u-stem, because there are no forms in that paradigm that have double zero-grade (in both the root, *h₁s-, and the suffix, *-u-) and no overt case ending. Therefore, he says, the standalone adverb must be extracted from compounds with the stress on the head noun, where *h₁s-u-CV́- is just what we’d expect. Jan 24 at 16:31

1 Answer 1


German seems to use different terms for Schwundstufe and Nullstufe, but English uses zero grade indiscriminately.

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.