4 Mar 2010

My sweet honey bee

The Indo-European word for 'honey', I have to confess, has always bothered me. Technically *mélit displays proper form with full-grade in the accented syllable and zero-grade in the unaccented and it also is supported by reflexes in Celtic, Germanic, Latin, Greek and Anatolian languages. Yet...

Douglas and Adams inadvertently uncovers a problem with this irreproachable hypothesis at the bottom of page 262 of The Oxford introduction to Proto-Indo-European and the Proto-Indo-European world (2006): "The noun *mélit is found widely in the West and Centre (e.g. OIr mil 'honey', Lat mel 'honey', NE mildew [< *'sweet sap'], Alb bletë 'honey-bee', Grk méli 'honey', mélissa 'honey-bee', Arm melr 'honey', including Anatolian, e.g. Hit militt- 'honey') and has one Iranian cognate in the form of a reference to melition, a drink of the Scythians."

Naturally we should ask: Why is the distribution of this word west and center? And you've probably surmised that I'm thinking of a possible Aegean explanation for what appears to be nothing more than a wanderword. A more valid PIE root for 'honey' might best be sought in the more-fully-attested root *médʰu, normally assigned the value of 'mead', a wine made from fermented honey.

What I'm pondering on is the possibility that only the Anatolian forms for 'honey' are truly Indo-European but which represent some innovative derivative from a native root. From there, Anatolian *mélit would be borrowed into Aegean as *méli with other related borrowings like *malítu 'sweet' (cf. Britomartis) which would in turn be borrowed into Greek. By the onset of the 1st millenium BCE, Proto-Cyprian *meli 'honey' and *mlitu 'sweet' would yield Old Etruscan *mel and *mliθ via early syncope. Once the Etruscans arrived in Italy, the early Latini could have easily borrowed the Etruscan term, thus Latin mel. These terms, along with their Etruscan-derived runes, could then have also travelled eventually to the Germani and among the Celts.

That would certainly explain the distribution a lot better than the standard Indo-European-based theory.

10 comments:

  1. Are there any other words that can be safely said to have been loaned into either Italic and Celtic from Etruscan?

    The trouble I have with this is the fact that the *melit- root is attested in Irish apparently displaying the proper sound changes, which means that the borrowing must be quite old if one did occur.

    ReplyDelete
  2. I agree, the word looks strange. It may have a zero-grade in the second syllable, but having monomorphemic polysyllabic words in Indo-European is weird to begin with.

    Maybe there's a nominal suffix *-it but this is then not at all common. But I believe there's some evidence for it in Sanskrit.

    It's a romantic notion that *médʰu means 'mead' as it would provide an insight into the Indo-European culture to a point that we even knew what they drank. But just because it seams 'nice' to reconstruct an alcoholic beverage is not enough evidence to support a claim that it existed. In a majority of the languages the word simply means 'honey', and in a large amount of languages where it means 'mead' it continues to have the meaning 'honey'.

    Also the directionality is difficult. I don't find it likely for 'mead' to develop into 'honey'. That's like expecting 'wine' to take on the meaning 'grapes'.

    Therefore *médʰu probably (also) meant 'honey'.

    As a side note have you ever heard of the idea that 蜜 'honey' might be a loan from Tocharian mit 'honey'? I always found that a pretty exciting idea.

    In Latin the word is quite problematic. The form mel, gen. mellis points to all kinds of stuff, but not to an ending *-it. De Vaan does not really take a stance on this matter. But considers the PIt. form to be *meli.

    I've wondered more than once if *melit isn't somehow related to *médʰu after all.

    But the shift is of course very problematic. Latin has some instances of *d > l (e.g. lacruma 'tear' compare: Gr. δάκρυ 'tear'. But obviously none of the other Indo-European languages had such a shift, and on top of that we're talking about *dʰ rather than *d.

    Kloekhorst proposes that this root had ablaut though. One finds two forms in Hittite militt- besides malitt- where the latter could point to a zero grade. This zero grade form is attested in the dative, but not in the genitive, which is odd.

    Also the development of an accented to i is not very usual.

    So much to wonder about. It really is a difficult word. It would be really nice if you could actually find cognates in Aegean languages.

    ReplyDelete
  3. Ethan Osten: "The trouble I have with this is the fact that the *melit- root is attested in Irish apparently displaying the proper sound changes"

    You never bothered reading Douglas/Adams, Encyclopedia of Indo-European Culture, p.271 under *mélit: "[...] while OIr mil (gen. melo) is an i-stem, analogically refashioned after the u-stem 'mead'."

    A 'refashioned' root is not a root "displaying the proper sound changes". Please try not to make it up as you go along.

    ReplyDelete
  4. What about the Latin "meles", the word for badger? I've not seen any etymologies offered for it anywhere, but it seems quite plausible to link it to this root. After all, there IS a species of badger from Africa and Asia called the "honey badger" that habitually raids bee hives for the sweet stuff, and I would be surprised if the European species (Meles meles) didn't at least occasionally do the same.

    ReplyDelete
  5. Phoenix: "Maybe there's a nominal suffix *-it but this is then not at all common."

    Yes, (*)*-it isn't typical of PIE. I don't even recall other roots with it off the top of my head, do you?


    "As a side note have you ever heard of the idea that 蜜 mì 'honey' might be a loan from Tocharian mit 'honey'?"

    Yes, surely it is. This says to me that a) this root is ancient and b) that it initially referred to 'honey' rather than its derived alcohol.


    "Also the directionality is difficult. I don't find it likely for 'mead' to develop into 'honey'."

    Agreed.


    "I've wondered more than once if *melit isn't somehow related to *médʰu after all."

    If *mélit is only an Anatolian innovation, it may be either through native derivation or through substrate like Hattic. One crazy scenario I thought up is that *medʰu wanders its way into Pre-Hattic where a further change of -t > -l and derivation yields a borrowed Hittite millit-. On the other hand, there must be simpler derivations within Anatolian to consider first before this resort.

    ReplyDelete
  6. Glen Gordon: "You never bothered reading Douglas/Adams, Encyclopedia of Indo-European Culture, p.271 under *mélit: "[...] while OIr mil (gen. melo) is an i-stem, analogically refashioned after the u-stem 'mead'."

    A 'refashioned' root is not a root "displaying the proper sound changes". Please try not to make it up as you go along.
    "


    And you clearly don't know anything about Old Irish; the genitive of /all/ i-stems is refashioned after the u-stem nouns. Cf. http://www.archive.org/stream/conciseoldirishg01pokouoft#page/62/mode/2up/search/declension

    ReplyDelete
  7. Ethan Osten: "And you clearly don't know anything about Old Irish; the genitive of /all/ i-stems is refashioned after the u-stem nouns."

    If you feel personally threatened by freely talking about new ideas, obviously you should take a break from the computer.

    The only thing that matters on Old Irish mil is the answer to the following question, an answer that we're still not hearing from you:

    WHAT SHOWS *CONCLUSIVELY* THAT THE CELTIC FORMS ARE **NOT** BORROWED?

    We already know how they *could* be from an inherited *mélit but we don't know how they *can't* be a borrowing. So tell us something new. Help us eliminate the possibilities.

    Keep in mind the linguistic contact between eastern Celts and the Etrusco-Rhaetic peoples as far back as the early 1st millenium BCE.

    ReplyDelete
  8. Beyond better judgment, I've decided to respond to only the portions of your post, Ethan Osten, that aren't testing my patience with ludicrous denials and are actually potentially informative for everyone.

    Ethan Osten: "Let us ignore, for the moment, that as the man proposing the theory that this root is a borrowing the onus is on you to prove it."

    Yet *mélit lacks any eastern cognates, a very damaging fact considering the widespread distribution of honey. Therefore there can rationally be no such root at the PIE level until evidence is broadened. This is YOUR onus, otherwise constantly insisting on the existence of *mélit and artificially pitting it against the Aegean alternative is an obnoxious opinion.

    "Given that this root is attested in Irish, and given the generally accepted separation of Insular Celtic from the rest by the 6th century, we can probably assume that this root was probably in common Celtic before that date, as contact between the two branches was limited."

    "Limited" is subjective rhetoric and does nothing to disprove their historical contact in the Alps with the Raeti.

    "The reflexes we have are OIr mil, Welsh mēl, Breton mel; these, along with Irish milis "sweet," point to a root *meliss, where the -s (instead of t) is representative of a common sound change already complete in Proto-Celtic."

    Yet Proto-Celtic is generally dated only to circa 1000 BCE give or take a few centuries, just as the Etrusco-Rhaetic are first making their way into Italy. The chronology is sound.

    "Note also that this root displays Grimm's law, changing t into þ; this means, regardless of the chronology you put Grimm's law in, that the introduction of the root had to happen well before the break-up of common Germanic."

    Again, Proto-Germanic appears to be contemporaneous with Proto-Celtic. The chronology remains sound.

    "So we either have to assume that, somehow, a *melit- root must have somehow been transmitted through *melis Celtic territory, that the Etruscans were (inexplicably) in Scandinavia for some reason at some point, or that Germanic already had *melit-."

    Or simply that the Etrusco-Rhaetic language group, already present in the Alps by the early first millenium BCE, were in contact with the Proto-Celts who were in turn in contact with the Proto-Germani.

    The myths, rituals and writing systems obviously transferred in this direction through the Etrusco-Rhaetic speaking peoples into the Celtic and ultimately Germanic populations. So objecting to a word also travelling this same trodden path is rather, well, Sisyphean of you.

    ReplyDelete
  9. You did not put up my post to which you are responding. That is irresponsible and unsporting. Please amend this.

    Glen Gordon: "Yet *mélit lacks any eastern cognates, a very damaging fact considering the widespread distribution of honey. Therefore there can rationally be no such root at the PIE level until evidence is broadened. This is YOUR onus, otherwise constantly insisting on the existence of *mélit and artificially pitting it against the Aegean alternative is an obnoxious opinion."

    Mr. Gordon, before we get any further, I would really appreciate it if you stopped treating discussions of your theory on the level of obnoxious personal attacks. If you don't want critical discussion, that is your right, but you ought to make that clear instead of treating criticism as odious.

    The existence of the root *melit is presently the communis opinio. Regardless of whether that common opinion is right or not (and we could probably have a spirited debate on the qualifications of various Indo-Europeanists) the fact remains that, as with any other theory which explicitly goes against the relatively reasoned opinion of acknowledged experts, the burden of suspicion rests more on the innovating theory. This is not a commentary on the strength of either theory, merely on the fact that there usually needs to be a reason to switch from a common opinion to an individual one.

    The fact that *melit is not attested in the eastern branches (by which we mean Indo-Iranian and Tocharian, presumably; Balto-Slavic is rather in-between) is not evidence for anything in itself; that is an argumentum ex silentio, which I'm sure you're aware is a logical fallacy. It is not at all unusual for a given root to be missing in several of the main branches, and odds dictate that a certain number of those roots will be missing from the eastern branches. This is inevitable, and as such, does not constitute an actual argument unless there is some compelling reason to argue that absence is specifically reflective of some other circumstance.

    Nor does the fact that honey is widespread really mean anything, for that is no reason the root could not be replaced or changed. Since we're on honey, it's quite certain that the Slavs were constantly in contact with bears; it did not stop them from losing the original root in favor of medved. These are interesting facts only if another compelling reason can be found.

    You may find this argumentation difficult to work around, but that does not mean it is odious or dilatory; this is simply the way that historical linguistics works. You can hardly fault me for following at least some established methods here.

    Glen Gordon: ""Limited" is subjective rhetoric and does nothing to disprove their historical contact in the Alps with the Raeti."

    You seem to misunderstand me; I don't mean to imply that the Celts had no contact with the Etruscans, quite to the contrary. I merely mean that the Insular Celts had limited contact with Continental Celts, which means the root must have been introduced before that stage, ie. Proto-Celtic. Nothing more.

    Glen Gordon: "Yet Proto-Celtic is generally dated only to circa 1000 BCE give or take a few centuries, just as the Etrusco-Rhaetic are first making their way into Italy. The chronology is sound.

    Again, Proto-Germanic appears to be contemporaneous with Proto-Celtic. The chronology remains sound.
    "


    The Etruscans did not push into the Po Valley and beyond until the 7th century, according to the archaeological evidence, and given that Raetic inscriptions are not found until the late fifth century, it's unlikely that they were there before then either. By this time Proto-Celtic has already broken up, and Insular Celtic has already split off.

    At any rate, the -t -> -s change in Proto-Celtic is apparently a fairly early one. This would be, at best, threading a needle.

    ReplyDelete
  10. Ethan Osten: "You did not put up my post to which you are responding."

    No website is obligated to post your comments. House rules. My judgments are final. Don't like house rules? Then leave the house or don't be surprised if people find you offensive.


    "[...] I would really appreciate it if you stopped treating discussions of your theory on the level of obnoxious personal attacks."

    Strawman. By mischaracterizing my objection to YOUR views specifically as a general hostility towards all opposing views is precisely how you're offensive and trollish. Final judgment, end of discussion, let's move forward.


    "Regardless of whether that common opinion is right or not [...] the fact remains that [...] the burden of suspicion rests more on the innovating theory."

    Ad populum. Any rational person could care less how you justify communis opinio. The argument against *melit is about burden of proof, not about how many people publish the same crap.


    "The fact that *melit is not attested in the eastern branches [...] is not evidence for anything in itself"

    Burden of proof. Convenient argumentation but truly pathetic nonetheless.


    "Nor does the fact that honey is widespread really mean anything, for that is no reason the root could not be replaced or changed."

    Occam's Razor. Nothing obligates us to believe that the root was replaced or changed because nothing obligates us to believe it's a valid PIE root.


    "You may find this argumentation difficult to work around, [...]"

    After at least four identifiable logical fallacies one after the other, you're done.


    "I merely mean that the Insular Celts had limited contact with Continental Celts, which means the root must have been introduced before that stage, ie. Proto-Celtic."

    You misunderstand me. Your subjective assessment of "limited contact" is meaningless. Either there is or there isn't contact between the Celtic groups. You admit that there is. Therefore your statement has no weight or relevancy.


    "The Etruscans did not push into the Po Valley and beyond until the 7th century, according to the archaeological evidence, and given that Raetic inscriptions are not found until the late fifth century, it's unlikely that they were there before then either."

    You exaggerate. Woodard in The ancient languages of Europe (2008), p.142 explains the limit of what is known about the directions and movements of the Etrusco-Rhaetic peoples by stating clearly that "A decisive judgement is not currently possible." (see here). We know however from linguistic evidence that Etruscan and Rhaetic, as explained by Woodard, can be traced back to a common 'Tyrsenic' proto-language dated to the end of the 2nd millenium BCE.

    ReplyDelete