Halfway down the Danube

Brief notes on classic Armenian literature 3

The following may be the oldest piece of Armenian literature in existence. It's a fragment of an oral epic preserved by Moses Khorenats'i (Moses of Khorene), on the birth of the deity Vahagn:

Erkner erkin, erkner erkir,
erkner ew tsovn tsirani:
erkn i tsovun uner
zkarmik eghegnik:
end eghegan p'ogh tsukh elaner,
end eghegan p'ogh bots' elaner:
ew i bots'un vazer
kharteash patanekik:
na hur her uner,
bots' uner mawrus,
ew ach'kunk'n ein aregakunk'.

Heaven was in labor, earth was in labor,
the purple sea too was in labor.
Labor pangs in the sea seized
the little crimson reed.
Along the reed stalk smoke ascended,
along the reed stalk flame ascended:
and from the flames there leapt
a golden-haired little youth.
He had fire for hair,
flame he had for beard,
and his little eyes were suns.

The translation is taken from Calvert Watkins' How to Kill a Dragon: Aspects of Indo-European Poetics, which I believe I've recommended before, although most of the following will be taken from the work of James R. Russell. But look at that phonology! Armenia is definitely in the apostrophe Sprachbund.

Part of the unfamiliar look of Armenian to English readers derives from its linguistic development. Armenian, like English, Romanian, Serbian, and Albanian (but unlike Turkish, Georgian, or Chechen) is an Indo-European language. All Indo-European languages can be shown to be descended from the same (reconstructed) root language.

Armenian developed from Proto-Indo-European via a set of regular sound changes, which modified nearly all of the original sounds in the language, sometimes in striking ways. Perhaps the most famous of these changes is the great historical linguist Antoine Meillet's discovery that Proto-Indo-European *dw- (the asterisk means the word is a reconstructed form) shifted to Armenian erk-. [1]

There aren't that many Proto-Indo-European words that begin with *dw-, but one of them is the word for 'two': PIE *dwo, English two, Armenian erku.

There are other ways to get to initial erk- in Armenian. So initial erk- is not as odd as it might first appear to an English speaker.

The gh phoneme in the text comes from a roughly datable sound change. The character from which the text is transliterated originally represented a sound related to L. We know this from words borrowed by Armenian from other languages, like the name of Pilate, which became Armenian Pighatos. The sound change must have occurred after the borrowing, likely some time after the Gospel was first translated into Armenian.

Okay, background over. Now for some of Russell's notes. The following is taken from his article "Carmina Vahagni," originally published in Budapest; but since the original is a little hard to find, I am taking the text from his massive Armenian and Iranian Studies.

As to vocabulary, the Armenian song of Vahagn begins with two lines in pure Armenian, though subsequently loan-words from Middle Iranian appear: karmrik, vazer, kharteash, patanekik. [...] To emerge from a reed, Vahagn has to be very small, so the oft-encountered ending -ik is appropriately descriptive. Despite the Middle Iranian loans, these are probably the oldest words spoken in Armenian which we shall ever hear.

Arm. erk- corresponds to proto-Indo-European *dw-: Arm. erku, English two, Greek duo, etc. And the first two lines are made of two words each, descriping a pair; erk- is repeated in every word. [Russell elsewhere compares this to Germanic kennings. -- CY] The third element, water, comes in with a third line of three words (excepting the little ew 'and'): erkner [ew] tsovn tsirani. To introduce a new element, there is a new alliteration of the sound (ts) in tsov 'sea' and tsirani 'purple' (a characterisation which must recall Homer's wine-dark sea). The change represents the eternal mystery of generation: the conjunction of opposites to produce a third, the child, a being wholly new. It is striking here that the words used to describe the sea are all non-Indo-European (tsov is Urartean, from sue 'lake', cf. Georgian tba 'idem'; tsirani is areal, cf. Georgian ch'erami, Pashto chereby 'apricot'). From what are in the poem colorless abstractions defined only by their dualistic constrast to each other: sky and earth -- we move to purple, then in the next, very long line to the brighter red of the reed.

Reed is eghegn; in ancient Armenian this was pronounced with a deep L, not (gh) as today. L and R are related, as liquids, so when the verb erkn-er changes in this line to the nouns, erkn, it prefigures eghegn, the reed. The two next lines, in which Vahagn is being born, reverse the direction of the poem so far. There has been a movement down: heaven > earth > sea > reed growing in the sea (i.e., from its floor). Now smoke and flame ascend (elaner), and the upward movement is crowned by the springing out of the youth (vazer). This is swift and light, unlike the slow pain of labor. Note also the perfect parallelism of zkarmrikn eghegnik (six syllables) and kharteash patanekik (also six), the only pair of nouns in the song with epithets of more than one syllable, which makes them heavy and brings them out for comparison.

The last three lines describe the head of the newborn Vahagn: fiery hair (alliterative: hur her), fiery moustaches (bots' uner mawrus, with the verb uner shifted so bots', an explosive word for fire, can get the full stress of the start of the line, as hur did in the preceding line), and eyes which were little suns. Note that his eyes are not like suns, or fiery: they ARE suns. This is an ancient Indo-European homology: the eye sees, the eye is light, light is the sun, the sun sees, and so on. There are two eyes; we have moved them from the pairing, the repeated erk- of the beginning, in which the vast macrocosm is in labor, to the pairing in the microcosm, where the two tiny eyes of the reed-born god are two suns, an intricate parallel of two of one thing to two of another.

I love this sort of thing.

[1] Yeah, it's weird. Meillet thought that the *d- shifted to an *r- (think of a trill) and then the *-w- shifted to a *-gw-, which by a previously known sound law would change to a *-kw-. At some point a prothetic e- would have been added -- standard for Armenian, similar to Spanish -- and the *-w- lost. There are several other theories.

Posted by coyu at April 3, 2006 05:58 AM