18 Jan 2010

The hidden face

In prior posts, I've been babbling about all these meanings applied to *mar-, *mer- or *smer- by Indo-Europeanists and how skeptical I am of one particular set of meanings in this clumsy word pile, that revolving around the notions of 'shimmer' and 'shine'. I recently found an instructive account of this whole messy group of words, published in French: Skoda, Le redoublement expressif: Un universal linguistique. Analyse du procédé en grec ancien et en d'autres langues (1982), pp.120-2.

I strongly suspect that there lies an Aegean root underneath this and it has nothing to do with Proto-Indo-European at all. However with what little we know of the Aegean group (ie. Etruscan, Lemnian, Minoan, etc.) and with some even denying that such a group exists, etymologists don't seem to be getting very far here. Half-solutions everywhere; nothing tied together very well.

Let's cut to the chase and propose something bold and new. Let's say for a moment that there's a Proto-Aegean verb *amára 'to reflect, to shine'. Let's say that it remained unchanged in later Minoan. Let's also say that this same root however was reduced to Proto-Cyprian *mar from which we should expect the same in Etruscan, Eteo-Cretan, etc. Let's then run with this and say that there's a derivative *amária meaning 'reflection, shine' which is again preserved in Minoan while reduced to *mera in Proto-Cyprian. From 'reflection' come the notions of 'appearance' and 'face' in a similar fashion to that of Japanese 色 (iro) 'colour; appearance'.

What's the point of all my crazy theorizin'? Have I cracked? Probably, but damned if this doesn't give an interesting alternative solution to a plethora of poorly etymologized or unetymologized words in Greek using a fresh non-IE perspective:
  • ἀμαρυγή (amarugḗ) 'a sparkle'
  • ἀμάρυγμα (amárugma) 'a sparkle'
  • ἀμαρύσσω (amarússō) 'to sparkle, to glance'
  • μαραυγέω (maraugéō) 'to contract the pupil in light'
  • μαραυγία (maraugía) 'dazzling of the eyes'
  • μαρμαίρω (marmáirō) 'to sparkle, flash, glitter'
  • μαρμάρεος (marmáreos) 'flashing, glittering'
  • μάρμαρος (mármaros) 'sparkling stone; marble'
It's difficult from an Indo-European point of view to explain 1) the unusual reduplication, 2) the vocalism in -a-, and 3) the curious application of prothetic a- which Robert Beekes has identified (I believe, rightly) as a hint at a 'Pre-Greek' (ie. non-Indo-European) origin[1]. To me, the prothetic a- is not a genuine prefix but is due to the presence or absence of syncope between different Aegean dialects and time periods. If one is still skeptical, Hesychius was kind enough to add the gloss μέρα (méra) in his work Glossai which he equated directly with ὄμματα (ómmata) 'eyes'. The 'eye' was used in Greek to also refer to the face as a whole, as with ὤψ (ōps).

[1] Consult for example Beekes' Greek Etymological Dictionary under ἀμαρύσσω.


  1. Always nice to see someone work on a dictionary which I've poured hours and hours of work into. (I did some minor editing and wrote half of the bibliography, because Beekes didn't think it was necessary to keep a database for that :P)

    Anyway, to me it seems clear that this word family is indeed Pre-Greek. I did want to comment on the 'unusual' reduplication though.

    It does happen in more cases, which are more clearly Indo-European that there's a reduplication of multiple letters. In all cases I can think of though, it's a reduplication of a Laryngeal plus a Resonant.
    (ἀπ)όλλυμαι 'to die'

    Which Beekes etymologises to come from a root *h3elh1- 'to destroy'

    Compare with Hitt. hallanna/i- `to trample, destroy'
    TochB alāṣṣäṃ `is exhausted'

    So this is pretty much a different problem really, but at least the resonant part is consistent ;-). I do wonder if all HR- initial roots have this type of reduplication. I've never ever seen someone be absolutely clear on that, and I've not been motivated to find out myself ;-)

    I like the idea of prothetic a- being dialectal variation rather than a Prefix. because if it was a prefix we'd want to see it have a bit more functional load after all. Also the presence of several different prothetic vowels would seem to support this claim.

  2. Phoenix: "I did want to comment on the 'unusual' reduplication though."

    Taken individually, odd reduplication, the appearance of a prothetic a-, *or* a-vocalism of the root doesn't necessarily suggest non-IE origin, I realize. Taken *collectively* however, reduplication, prothetic a- *and* a-vocalism of the root makes the attempts to etymologize this as a genuine PIE root very painful to me. ;o)

    Concerning typical PIE reduplication now, one would expect that *h₃elh₁- is properly reduplicated as *h₃e-h₃elh₁- whereas **h₃elh₁-h₃elh₁- would break the general rules, although full reduplication might seem to some more aesthetic than the characteristic partial kind.

    However after the loss of laryngeals, *h₃e-h₃elh₁- /xʷe-xʷelh-/ would tend towards *ōl- and would no longer look like reduplication. At that point, it would be perfectly understandable that some dialects would adopt a new form like *ōl-ōl- or something similar simply because /l/ was the only consonant left to reduplicate.

    I don't understand how full reduplication, if existent in PIE, would fit into the grammar of the language.