12 Nov 2009

Minoan inscription HT 104

I find it very sad that there's so little decent conversation about Minoan artifacts despite ample information available online. John Younger has gone to great trouble to detail all sorts of Minoan artifacts and their inscriptions (link here) yet I haven't seen much active discussion in the blogosphere about it. Why?? This should be a fascinating topic for any linguaphile to explore! Maybe the world needs a giant Paxil. Personally, I've been noticing a lot of interesting patterns and lately I've been looking at HT 104. So I thought I'd share my speculations for the sake of promoting constructive conversation.

This is the apparent accounting record made by a scribe in HT 104 (after formatting into a tidy table) that is also found on John Younger's Linear A website:


Considering the pattern of the numbers, there's no doubt that kuro is the Minoan word for 'total' since it's used so regularly for other sums in many other documents. I figure it's a borrowing specifically from Ugaritic *kullu (kll) since this is a common root for 'whole, all' among all Semitic languages. The change of l to r is likely indicative of a lack of Minoan /l/, making it an areal feature shared with Middle Egyptian to the south. Now what about the other items here? Are they names? Commodities? What?

One thing that excites me here is TA-PA. In Linear B script (ie. Mycenaean Greek), TE-PA is the word for 'heavy rug'[1], a commodity. If we presume that the Greek word has been borrowed from Minoan, we might theorize an underlying noun *tapiya 'heavy rug' (cf. TA-PI in ARKH 1.a.1). So this probably represents a header describing what sort of objects are being tracked.

Between the header and the total, we have three names, all ending in -TI. While Miguel Valério interprets such endings as ablatives meaning 'from', I recognize the Etruscan inessive postclitic -θi 'in'. I expect that many may likely mutter a skeptical "so what?" to this as-yet unproven connection. Yet, if we allow ourselves to explore for the sake of argument we get what appear to be Anatolian placenames declined in a common locative case:

Dakusene-ti'in Cape Lekton' (cf. HT 103.2: DA-KU-SE-NE)
Idu-ti'in Mount Ida' (cf. PK Za 18: I-DA)
Padasu-ti'in Pedasos'

The numbers then would presumably represent weights or prices of the material divvied out, signalled by symbol {505}.

[1] Chadwick, The Mycenaean world (1976), p.152 (see link).


  1. Chadwick in the passage you reference says "te-pa ... can not be equated with a later Greek word, though it bears a remarkable resemblance to 'tapes' "carpet".
    I'm curious why it can't; perhaps you can explain.
    BTW, American Heritage Dictionary (under "tapestry") says 'tapes' is "perhaps of Iranian origin".

  2. Robert Castleden on page 75 of Mycenaeans (2005) likewise states that the Mycenaean word "is similar to the later Greek word for carpet, tapes" (see link). The word "similar" here gives me the impression of an equally noncommittal opinion like that of Chadwick. I presume the conservative etymologist's fear stems from the change in vocalism if they were indeed related.

    As for the AHD and the claim of Iranian origin, I'd like to see that backed up with a clear source. What Iranian word? And how was the word transmitted? What happened to the Mycenaean word then? It all smells fishy, doesn't it? Yet it's not the first time I've stepped on etymological landmines like this where dictionaries show a bunch of different solutions to the origins of the same word, no matter how contrived and provably false (eg. the origin of Latin persōna as well: From an idle connection to a likely mistranslated Etruscan phersu? Or more rationally from a native derivation of personare 'to sound out, to sound through'?).

    So I frankly don't have a clue what Iranian word they're referring to and I'll dare propose that they don't either, merely ascribing an Iranian source due to that region being a historical and noteworthy epicenter for quality woven goods.

    I think a benefit to reasoning it a Minoan loanword is that it explains the odd variations in Greek, both tapēs and dapis, which as a pair suggest that the source lacked voicing contrasts in plosive phonemes. Indeed, Minoan appears to lack voicing contrasts in plosives.

    To further add an even more complicated but tempting twist on this idea is the hypothesis that Cyprian dialects (ie. Etruscan, Lemnian, Rhaetic, Eteo-Cretan & Eteo-Cypriot) which survived Minoan into the 1st millenium BCE, no doubt would've contained similar cognates that further interfered with later Greek forms. Contemplate that! ;o)

  3. According to the Online Etymological Dictionary, the Persian root in question is tâb- (the two forms it gives are both formed from the same present stem, which is older than the past or infinitive). This is the same root that, via the New Persian past participle tâfteh, gives us the word "taffetah". Nowadays it usually means "burn" or "glow", but Dekhoda makes it clear that it originally mean "twist", especially with regards to cords. A brief examination of the Old Persian corpus doesn't turn up any likely antecedents, and I couldn't find any potential descendants of a pre-New Persian participle to push the age back a bit farther, which I would really like given the time depth we're talking about here. The time depth may indeed be an insurmountable problem for any Iranian etymology, given that the Aryans only started arriving in Iran around 1000 BC; I would still be somewhat incline to go with it if the association with woven goods could be pushed back to Old Iranian, but I don't think that's the case.

  4. The more I look at this word, the more I feel like I'm unintentionally pulling the string out of a badly woven tapestry, pardon the pun.

    Any time I see an author avoid clarity by claiming something vague like "probably of Iranian origin" without committing to even a reconstructed form of a hypothetical antecedent I get a little irritated because I feel that I'm being "strung along", so to speak, by a tall tale that only benefits the author by putting up a false image of his or her sureness at the expense of readers who need factual explanations.

    In this case, this Iranian etymology doesn't explain much at all. In fact, it only lures us into even more silly questions and conjecture. First off, what is the Persian form from which we get Greek dápis or tápēs? We as readers are forced to guess: Maybe *tāptān or *tāptah? Who knows? Apparently not the theorists themselves, it seems, and again I feel like I'm being made to labour on their vague theories. Irritating.

    And where did the *-t- in the presumed Persian form disappear to when it entered Greek? I've already also asked, "When did it enter Greek?" but again the theorists force us to fill in the blanks they created in the first place. Perhaps during the rise of the Persian Empire, we may surmise. Yet this chronology doesn't explain the remarkable coincidence of Mycenaean TE-PA nor the potential Minoan equivalent TA-PA. Too many disconnected factoids, not enough clarity.

    It seems rather that the reverse direction of borrowing has happened. Instead of attributing the Persian forms to an otherwise "lost" causative *(*)tāpayati as is the general vogue which, to add further unnecessary complication, is further alleged to be an "analogical remodeling" of a form based on zerograded *tmp- from PIE *temp- 'to stretch' (cf. Pokorny), why don't we try something else that doesn't require as many suppositions?

    From Minoan TA-PA ~ TA-PI (*tapiya ?), we obtain Mycenaean TE-PA (and quite possibly other variants) which yields the later Greek forms, directly or indirectly via continual Pelasgian influence.

    Anyways, by the 5th century BCE, we know that the Persian Empire sweeps into Greece. From a Greek diminutive τάπητιον tapētion, we obtain the Persian forms with *-t- without weaving imaginary causative forms. One benefit here then is that the Persian forms are now based on attested words, not theoretical forms and it also makes sense why it is that the Minoan and Greek forms show simpler forms without -t- since they're then recognized as the mother of them all.

    Oh my, this has been a long ramble, hasn't it? I really need to get a life, hahaha ;o)

  5. "First off, what is the Persian form from which we get Greek dápis or tápēs? We as readers are forced to guess: Maybe *tāptān or *tāptah?"

    I can tell you for certain that it was neither one of those. As I mentioned earlier, in Middle Persian the entire verbal system turned itself inside out. Part of this involved reanalyzing the past participle as the present via a passive construction and various Pahlavi shenanigans. Thus, the heteroclitic verb car-tanaiy (present stem kunau, past stem kar- or car) had the past participle kr̥-ta-, which yields New Persian kard-an (present stem -kon-, past stem kard-). This transmogrificational nightmare ended no more than a century or two before the Arab Invasion; I shudder to think what Persian would look like if it had still been going on.

    "an otherwise "lost" causative *(*)tāpayati as is the general vogue"

    A Google search for "tāpayati" turned up yet another bit of our tapestry. It turns out that it is attested, in Sanskrit, and with the meaning "burn, glow, radiate heat", exactly the same as the usual New Persian meaning.

    I'm starting to suspect that what we're looking at here is actually three (or possibly two, or even one) separate roots. One of them carries the meaning "glow" (or, in traditional Indo-Europeanist parlance, "shine") and is, to the best of my knowledge, specific to Indo-Iranian. Another one means "twist" in Indo-Iranian and is a semantic development from PIE "stretch". The third one is your *tapiya, which then somehow ends up as NP tâfteh (which I doubt is an independent creation since taffeta is far too fine a fabric to be explained by a verb meaning "to twist into cords"; if the name for it were to be a participle we would expect it to be **bâfteh, "woven"). It occurs to me that there is also a chance that the Minoan word is a loanword from Hittite, since the items in question were apparently imported from Anatolia and the word would nicely be explained by *temp-, if, of course, the phonological changes were appropriate. The word is not to the best of my knowledge attested in Hittite (which you would expect it to be if it was an export), but it is still a possibility. Another (remote) possibility is that the glowing tap- is derived from the twisting tap- via the Aryan goddess Tabiti, who apparently was identified by the sunlike pattern of her hair (note that I have nothing to do with that site, despite part of my username appearing in the URL). I do however feel that that would be a bit of a stretch, so to speak.

    "I really need to get a life, hahaha ;o)"

    If we had lives, we wouldn't be linguists ;)

  6. Let's cut this to the quick. I see 3 'threads' in this topic, so to speak:

    1. PIE *tep- 'to be hot'
    2. PIE *temp- 'to span' (cf. *ten-)
    3. Greek tapētion 'carpet, rug' (< tapēs)

    Meanings of 'shine' and 'burn' in the Persian verb are transparently from *tep- so in ONLY those senses must taftan be inherited from PIE. Despite Pokorny's claims, I don't see anything other than the Greek and Persian forms that hint at a textile usage behind *temp-. This leaves only Greek tapētion from which we also get the Latin reflex with medial -t-.

    Given that and considering the paradoxical meanings of both 'shine' and 'twist' in the same Persian word, I especially feel the Greek and Persian words can only be explained one way or another by way of a historical borrowing of 'carpet, rug' originating squarely from Greek tapētion which, through reanalysis and back formation in Persian itself of the foreign term yielded a new native meaning of 'that which is TAPE-ed' where this spliced *tapeta- would be assumed to be a participle meaning 'woven' via an imaginary participle in *-ta-. Thus the foreign term is naturalized as a native word meaning 'that which is woven'. This then would have surely effected a secondary meaning in pre-existing taftan.

    There is no causative *tāpayati meaning 'twist' and which is derived from *temp- despite unexpected lengthening; we find only one derived from the well-attested root *tep- with the meaning of 'burn, shine, glow'. There is no textile usage of *temp- surely because the tap- forms in Persian and Greek have nothing to do with it; it's very easy to ascribe *m as the source of a here, but there simply is no proof of this from what I see so far. Finally, I see no Hittite loan as you suggested nor do I see why it must be so.