25 February 2012

Greek hēméra 'day' (updated)


In Greek mythology, Hēméra was the primeval goddness of the day. From this word and Armenian awr 'day', Indo-Europeanists such as Mallory-Adams1 reconstruct an IE protoform *h2ēm-ər- 'heat (of the day)', which I link to Semitic *ħamm- 'to be hot; warm'2, with the voiceless pharyngeal fricative ħ corresponding to the "laryngeal" h2. 

The IE protoform belongs to what I call "IE B"3, corresponding to the "IE A" protoform *səm-/*səm-ro- 'summer'4, whose initial s- is a consequence of the sound shift I call Fournet's Law, by which a post-velar fricative becames fronted to an alveolar articulation point.
1 J.P. Mallory & Q.D. Adams (2006): The Oxford Introduction to PIE and the PIE World 
2 Also cognate are Hurrian am- 'to burn' and (possibly through an Etruscan intermediate) Latin amāre 'to love', amor 'love'.
3 Mostly represented in Eastern languages, mainly Greek-Armenian and Indo-Iranian, but occasionally also Celtic.
4 The ablaut form *sem- usually quoted in dictionaries isn't attested anywhere.


  1. This hypothetical Fournet's Law you are speaking about, is not found at all in Tyrrhenian,
    I didn't say otherwise.

    and it is very dubious it even existed elsewhere. You are continuing to speak about this item? Well, you have to provide a complete list of roots. The few I saw in the Yahoo list seem to be all but convincing.
    Apart from the above example, there a few more:

    *seh2l- 'salt' ~ NEC *q’eɦlV (~ -ɫ-) 'bitter' (here we need an intermediate language where the initial stop became a fricative )

    *sap- 'thorn' (Hittite sapi-kkusta, sepi-kkusta 'needle', Gaulish *sapo- 'fir', Welsh syb-wydd 'pine', Germanic *saf- 'reed') ~ Latin abi-ēs 'fir tree' ~ Altaic *sjó:phì 'thorn, thorny bush'

    *sel- 'dwelling, settlement' (Russian seló 'village', Latin solum 'ground, floor') ~ Kartvelian *xl- 'to dwell, to live'

    Don't forget also the Latin doublet sorbus ~ arbutus.

  2. Besides these cases there're also the ones where s- arises from a former palatal, but in these cases we've got *y- as the counterpart:

    *sa(n)k- 'to sanctify' (Latin sacer, sanctus) ~ *yag´- 'to honour, to worship' (Greek hágios)

    *sem- 'one' ~ *yem- 'twin'

    *seh2-n- 'healthy' (Latin sānus) ~ *yak- 'to cure' (Greek ákos)

    It looks like Fournet's Law corresponds to an older IE layer/proto-language than the one where the initial is lost.