30 Dec 2009

A deeper source of Cretan Britomartis

Solinus spoke of Βριτόμαρτις Britómartis as a native Cretan name for Artemis, the Greek goddess of moon and hunt, which he claimed had underlyingly meant virgo dulcis 'sweet maiden'[1] and Hesychius doubly equates his Cretan gloss βριτύ with Greek γλυκύ 'sweet'.

Apparently though, from what I've dug up so far, this is all that's ever written on the subject of the etymology of her name, an important question that's tossed aside for airy interpretations of local mythology to fill up the latest tome weighed by the pound. Out of desperate curiosity, I consulted Wikipedia to see if any of those busy bees had found just a smidge more than the status quo but predictably the groupThink swarm proved once again worthless. There's far more to the origins of this name but the following exploration is curiously absent in any book I'm aware of on the subject despite being, I believe, highly illuminating.

The first element, brito-, may remind us of the Greek example of ἄμβροτος 'immortal' from Proto-Indo-European *n̥-mr̥tós 'non-dying' showing how easy it is for /m/ to strengthen to /b/ before another resonant. In light of Hittite militu- '(honey)sweet'[2], a characteristically Indo-European u-stem adjective derived from milit- 'honey', there should be no doubt where the first element comes from. The second part of the compound, -marti-, is contrastingly sourceable to Assyrian mārtu 'daughter, girl', a purely Semitic feminine form of māru 'son, boy'. These linguistic connections complement the already well-known Anatolian and Near-Eastern influences on Crete.

[1] Solinus, Polyhistor, 11.8.
[2] Puhvel, Hittite Etymological Dictionary, version 6 (2004), p.155: miliddu-, maliddu 'honeyed, (honey)sweet'.

Paleoglot: My sweet honey bee


  1. Another really exciting topic: since - in this instance - we have a Minoan word with almost 100% certainty, and the ancient Greek authors are also kind enough to give us the meaning.

    As for brito-, the really interesting thing is not only its shared origin with the PIE *mel(it)- stem, but its variation across Greek dialects. We also know a variant of the theonym in the form of brizo. This hints at the existence of a fricative in an earlier form (*britso-?).

    I must confess, this post has given me the idea to search for this stem in the Linear A corpus. But I had some trouble in finding a value for the initial 'b-'. Minoan scripts had no sign for the consonant 'b' (it likely had not existed at all), so I tried substituting P- , M- and even Q- (KW-). I could find a single word that is faintly similar, namely QI-RI-TU-QA on HT23. The only problem with this word is that this isn't a personal name, but rather a name of some agricultural product (the next line lists SA-SA-ME = sesame). Yet it is not necessarily a problem, if the commodity has something to do with "sweets".

  2. "We also know a variant of the theonym in the form of brizo."

    Given the direct Cretan gloss βριτύ and related Hittite militu-, there's no sense or need to theorize a fricative here.

    The name Βριζώ comes from βρίζω 'to be sleepy, slumber, nod'. If there is a relationship to the 'honey-sweet' word in any way, it's through wordplay; corruption of the original fricative-less word would have surely occurred then.

    I would expect the word in Minoan, if existent, to be written out simply as *MA-RI-TU.

    "[...] so I tried substituting P- , M- and even Q- (KW-)."

    From a phonological point of view, it's not sensible to assume that Minoan Q was the same value as the Greek labialized velar stop. Likelier values are [x] and/or [q].

  3. Maybe I should add more for clarity. For me, there are two branches in Aegean: Cyprian and Minoan. I've suggested earlier on my blog that Minoan T may have been lenited to [ts] before [o] and [u] (similar to what we find in Japanese), something I'm still considering. Whatever the case, this lenition doesn't appear to happen in Cyprian however because I haven't found trace of it in later Etruscan.

    So I interpret the Cretan gloss as a Cyprian term and am doubtful that the Minoan language survived as anything other than a religious language by the Classical Period. I have a general impression that it was Cyprian dialects that dominated Crete by the 1st millenium BCE.

    I would recommend reconstructing therefore Proto-Cyprian and later Eteo-Cretan *mlítu. In my point of view, Eteo-Cretan is not a direct ancestor of Minoan.

    Back to Minoan, one might presume further that a hypothetical *MA-RI-TU could be pronounced [mə'ɾitsu] whereas Cyprian dialects would have preserved the [t] recorded by Hesychius. If the name Brizo were old enough, perhaps it is indeed inspired by Minoan and this expected fricative.

    There's tonnes to consider here. Thanks, Bayndor. My head's churning now and will have to investigate Brizo further.

  4. I have a question about the liquid - the honey words you discuss elsewhere all have /l/ while the britu you discuss here has /r/. Do you assume that Minoan (or Cyprian) did not distiguish the liquids? Or is that a variation between individual Aegean languages?

  5. "Do you assume that Minoan (or Cyprian) did not distiguish the liquids?"

    While I used to believe that Minoans didn't distinguish /r/ and /l/, I now think they did and that it was purely a matter of the script which may have been in part inspired by Egyptians which we know didn't distinguish the two sounds (although they eventually did in Late Egyptian and Coptic).

    I purposely reconstruct *ml- in this Eteo-Cretan word because of its connection to Hittite. Since ml- is an impossible word-initial cluster in Greek, Hesychius apparently opted for βρ- as the best possible approximation. In the beginning of March 2010, I elaborated further on this.