12 Apr 2010

New info on 'lily'

After my last post just beforehand I ended up finding a slew of more 'lily' loanwords and relevant info. A new pattern is starting to reveal itself that adds to what I previously said. Previously I justified a Minoan reconstruction *léri (Lat lilium, Gk leirion) with *e rather than a possible alternative *ai and this is helped along by the Greek geographical name Lerna which may specifically attest to *e if meaning '(Place) of Lillies'. If this turns out to be off the mark, nonetheless ne'er a critic can question the power of my imagination.

But check this out now:
Despite displaying a more generalized meaning 'flower', the words may very well be related to 'lily' as many argue[1]. But exactly how? The curse here is the persistent ignorance and avoidance of establishing Egyptian vocalism in words. The chic thing to do is simply write the consonants and leave any mention of vowel reconstructions to stuffy academics who publish obscure works well beyond the hands of the general public, hidden somewhere in some journal perhaps rather than out in the open in general references which opt to err on such a safe side as to shun any informative position altogether on the etymology, assigning it nebulous catch-all phrases like "from Pre-Greek" or "from an unknown Mediterranean source". WHY??? What on earth is wrong with being accurate? Aside from Adolf Erman's casual attempt of *réret, I see no other mention online of what the vowels in this particular word may be. All these purely consonantal transcriptions do is mire everything in artificial mystery. Let's piece this together ourselves then.

In the Sahidic dialect of Coptic, the word has become ϩρⲏρⲉ with its feminine termination -ⲉ securely from Middle Egyptian *-at /-aʔ/. Its letter eta may be traced back ultimately to Old Egyptian *ū but, somewhere in the second millenium BCE, Loprieno educates us that it had already become [2]. This points me then to an original *ḥarūrat becoming *ḥarērat /ħə'ɾe:ɾəʔ/ at precisely the time we need to source the lurking Minoan term with just a slight modification: *aléri. I had missed the unaccented initial vowel so common in Minoan and which must reflect the Egyptian first syllable /ħə/. This then establishes the Hittite form as a borrowing directly from Minoan and which in turn implies that specialization to 'lily' came after this time. Lerna can now point to earlier *Alérina without problems considering Diktē < *Adíkituna.

(2010 Apr 13) Correction in "In the Sahidic dialect of Coptic, the word has become xrere [...]". Should be hrēre, a plain /h/. Thanks to fiosachd. And upon closer inspection, unaccented *a in Egyptian should be /ə/, not /a/. As I say, it's nice to be accurate.
(2010 Apr 14) Chuck Coleman corrects further: /hrerə/ (eta = short /e/ in Coptic). Upon reading further, this may indeed be more accurate and yet it dangerously strays us off topic. Ergo, all indication of Coptic phonetics in my above account has been eradicated in order to return to the main matters: 1) the etymology of 'lily', and 2) the vocalism of ḥrr.t 'flower' in the 2nd millennium BCE.

[1] Brown, Israel and Hellas - Vol. 3 (2001), p.46 (see link); Puhvel, Hittite etymological dictionary - Vol. 1 Words beginning with A (1984), p.33.
[2] Loprieno, Ancient Egyptian: A linguistic introduction (1995), p.38 (see link).


  1. Isn’t Coptic ϩ (hori) glottal (/h/), rather than velar (/x/) [= ϧ (ḫai)]?

  2. Ah yes, you're correct. Good eye. I'll change that.

  3. Thank you Glen, a really nice work from your side. I think I have to reconsider my opinion in the light of these new facts. Words like the Hittite alel look more than proof enough to accept that the ancestor of the word 'lily' had no diphthong in it.

    The only thing I wonder is the original meaning of RA3. If we apply the acrophonic theorem (with some flexibility), there must have been a flower-name beginning with *(a)rai-, *(a)lai- or *(a)lá-. Was it indeed Crocus? That flower has already had at least two other names (Krokos, Knakos). Or was it another lily-like flower?

    A mad though has just entered my mind: do you happen to know the origin of Persian/Turkish name for tulips: 'Lale' ?

  4. I had yet another thought. I wonder if alel is via Hattic influence (Hattic *alil ?) since there are many native nouns ending in -il in that language already and might have been contaminated before entering Hittite. Big unprovable assumptions, I know.

    Bayndor: "If we apply the acrophonic theorem (with some flexibility), there must have been a flower-name beginning with *(a)rai-, *(a)lai- or *(a)lá-."

    Why is it not possible that a symbol for re in Linear A becomes used for ra in Linear B?

    "A mad though has just entered my mind: do you happen to know the origin of Persian/Turkish name for tulips: 'Lale' ?"

    Note that Turkish lale violates vowel harmony, marking it as a loan, from the Persian word you cite. However, I can't yet track down the etymology of this Persian لاله (lâleh). Hmm...

  5. The first vowel in ϩρⲏρⲉ is not long, it is short /e/, as opposed to /ɛ/, which was written by ⲉ. The last vowel in Akmimic and Sahidic was pronounced /ə/. Thus, ϩρⲏρⲉ was pronounced /'hrerə/. In Boharic, the last vowel became /i/, a regular development. Boharic also sometimes shows an epenthetic initial /ə/. Following Peust, we can construct a "Paleo-Coptic" pronunciation of /ħrurə/ or a variant of the final vowel. (The final /t/ had been long lost.) Interestingly, Fayyumic sometimes substitutes /l/ for /r/. This is probably an expression of dialectical variation that was only first expressed in the New Kingdom. The standard writing pre-NK lacked /l/.

    Part of the confusion over vowel lengths comes from Greek, which distinguished /a:/, /e:/ and /o:/, which are, not coincidentally, reconstructed for Coptic. Only the Mesokemic dialect seems to have used Greek conventions for these vowels. Other dialects used doubling to indicate long vowels.

    See: Peust, Carsten, Egyptian phonology: An introduction to the phonology of a dead language.

  6. Chuck Coleman: "The first vowel in ϩρⲏρⲉ is not long, it is short /e/, as opposed to /ɛ/, which was written by ⲉ."

    I appreciate the added info but expecting a clearly non-IPA transcription to be laboriously exact despite being further linked to Crum's Coptic Dictionary where the original Coptic spelling may be read is just off-topic pedantry and will be henceforth deleted. Thank you for your respect.

    "Following Peust, we can construct a 'Paleo-Coptic' pronunciation of /ħrurə/ or a variant of the final vowel."

    You ignore my reference to Loprieno who explains not only that /u:/ > /e:/ in the 2nd millennium BCE with contemporaneous Akkadian evidence, but that Coptic eta is indeed a long vowel. This same transcription has been recommended by Joseph Greenberg who explains the double vowels as true vowel clusters in all dialects but Bohairic.

  7. I'm obligated to correct myself: "**recommended**" should rather be revised to "was recognized as commonplace in literature" since ironically he reasons thereafter why Coptic eta and omega must be short /e/ and /o/.

    Nonetheless, a "Paleo-Coptic" */ħrurə/ contradicts other things like Loprieno's Akkadian evidence of vowel change and Erman's *ḥréret while also disconnecting the link to 'lily' altogether. So forgive me if I'm confused and cautious. What concrete evidence decides the issue of vocalism in this particular word?

  8. Please allow me one final remark: the one regarding the origins of Linear B signs 'RE' and 'RA3'. These signs actually have a distinct ancestry. Linear B basically has 8 R-series signs: RA, RA2, RA3, RE, RI, RO, RO2, RU. All of these are identifiable in Linear A as well. The site of John Younger fails to give counterparts of Lin B RA3 and RO2, but it is quite obvious that they correspond to Lin A *122 (=logogram OLE) and Lin A *315 (a purely phonetic sign), respectively. RA3 is a rare sign in Linear B, and it only occurs as a logogram in Linear A: nevertheless, we can be certain that the two signs are the same, since Lin B also uses the phrase E-RA3-WA for olive-oil (Lin A simply uses RA3).

    What is more, these signs even have distinct Hieroglyphic ancestors. So far, I have been only able to certainly identify the RA2, RA3, RE, RO and RU signs, but I am essentially certain that the RA and RI signs are also present in Minoan Hieroglyphs (though we have issues at their identification, e.g. I am unsure if the Hiero "human leg" sign is actually RA or RI). RO2 is so incredibly rare sign in the Linear scripts, I would not be surprised if its Hieroglyphic ancestor would never be discovered.

    The RE sign is very common in Linear B, Linear A, and also in Hieroglyphs. The latter can be confirmed by the high frequency and typical terminal position of the "three-branched flower" sign, so it can only be 'RE'. On the other hand, there is the "single-flower" sign (clearly distinct from the former), considerably rare in Hieroglyphic Minoan, with a shape similar to Lin AB RA3. I have tried reading a few signets with the substitution of a value 'rai', 'lai', 'ra' or 'la', and they seem to sound reasonably well for a Linear A word.

    Based on the "no-homophones" rule of Linear B by John Chadwick, these signs must represent distinct sounds. The distinctions might not be easy to find, but these 'higher order' signs are generally assumed to be either different diphthongal versions, or those with an aspirated consonant. So far most linguists have mostly dismissed the possibility that distinction would have been based on vowel length.

  9. Bayndor: "Please allow me one final remark"

    Pleading is unnecessary. I'm only encouraging on-topic comments regarding:

    1) the vocalism of Ancient Egyptian ḥrr.t,
    2) the justification of any proposed vocalism,
    3) the transmission of the 'lily'/'flower' term.

    Quibbling about my Coptic transliteration despite being commonly published as such risks becoming obnoxious. Your comment however is relevant.

    "Based on the 'no-homophones' rule of Linear B by John Chadwick, these signs must represent distinct sounds."

    Fundamentally then, it appears your view rests solely on an unproven assumption.

    For me, I reason that since Linear B has disregard for syllable codae and shows clumsy representation of Greek consonant clusters, there must be a basis, perhaps a simpler rule or "seed" in Linear A, from which these Linear B rules have developed.

    I conclude that, in order to explain these simple facts, Minoan must have been a more syllabic language than Greek with only CV(C) as an allowable syllable shape and with valid codae restricted to resonants (/-w/, /-j/, /-n/). From this "seed" then, Linear B could adapt its less transparent system of spelling rules for a more consonant-rich language.

    I then conclude that Linear A more likely didn't distinguish between RAI and RA as you claim. This however doesn't negate the possibility that a symbol for RA (representing all of la, lai, lau, lan, ra, rai, rau and ran), if the proper Linear A value, could not have developed from a word containing original *-rai-. Yet there would be many other possibilities as well.

  10. Correction on the last part: "containing original *-lai-" ( as per your reconstruction *lairi).

  11. Turkish lale / Persian lāla looks like it could belong to the same family. The etymologic dictionary of Turkish I consulted (Sevan Nişanyan, “Sözlerın Soyağacı. Çağdaş Türkçenin Etimolojik Sözlüğü”, 4th issue, Istanbul 2009) states that Turkish lale is indeed a loan from Persian lāla “anemone or similar red-coloured flower”, and traces it back to a Middle Persian alālak “anemone”. This word looks non-IE in structure. If this descends from the same etymon as the other forms you discuss, it also shows short initial /a/ and it shows that, at least in the source language Persian loaned the word from, the vowel between the two liquids was long (perhaps /e:/?).