Solving Linear A.

Andrew Trounson writes for Pursuit about the ongoing quest to decipher Linear A:

How do you go about deciphering the script of a wholly different language that was lost more than 3,000 years ago? Linguist and archaeologist Dr Brent Davis says it’s like walking out on a tightrope anchored at just one end and supported by nothing but thin air, hoping you find something to stop you from falling. […]

Dr Davis, a lecturer in Archaeology and Ancient Egyptian at the University of Melbourne, is one of only a handful of people around the world to have made any significant headway on solving Linear A in the last 50 years. He established for the first time the word order of the language as being Verb-Subject-Object, like ancient Egyptian. So rather than ‘Minos has a minotaur’, a Minos would write ‘has Minos a minotaur’. […]

It was while he was doing his PhD at University of Melbourne that Dr Davis began to make real headway on solving Linear A. By establishing the word order of the language, linguists can identify the function of a word in a sentence just from its position. It’s like finding a key word in a massive crossword puzzle.

“The definite word order in English is Subject (S)-Verb (V)-Object (O), as in the phrase John likes cats. And we know that about 97 per cent of human languages are either in this form or S-O-V (John cats likes) or V-S-O (Likes John cats).” But when Dr Davis looked at other Bronze Age languages of this period in the region, none were like English. They were either S-O-V (like early Greek and Sumerian), or V-S-O (like ancient Egyptian). He guessed Linear A was likely to have one of these two word orders.

He then applied this framework to a series of inscriptions that appear on Minoan offering bowls. To put it simply, he found that the words on the bowls tended to recur in what was obviously a formula, except for the second word in the inscription, which was always different from bowl to bowl. His guess was that this word was probably the name of the person (the subject) making the offering. If correct then Linear A was likely a V-S-O language. That was confirmed when he found the Linear B sign for ‘olives’ (which had been borrowed from Linear A), occurring after the name as the object of the phrase. The repeated start of the phrase was therefore a verb, like “gives”, yielding the phrase gives Yasumatu olives, or in English [order] Yasumatu gives olives.

Davis points out that to actually decipher the language, more signs are needed:

“Discoveries are still being made, so I’m optimistic, but what we really need to find is a palace archive, which is where we are likely to find enough Linear A to finally decipher it.”

Thanks, Trevor!

Comments

  1. I’m intrigued to see what people who know more than I do make of this.

    Biblical Hebrew, formal Arabic and “ancient Egyptian” are V-S-O. Is V-S-O widespread among ancient Semitic languages or Afro-Asiatic languages? I see that Akkadian was S-O-V, and that this distinguishes it from South- and Northwest-Semitic. Sumerian was also S-O-V.

    I assume from Hindi, Hittite, Latin and Greek that S-O-V is likely to be a characteristic of Proto-Indo-European. Is that accurate?

    Does Davis’s discovery constrain the search for a language behind Linear A by making a Semitic language much more likely? Or does word order flip so often that it doesn’t help exclude anything, though it may still help guide the effort to decipher individual words or word-endings?

  2. Trond Engen says:

    Interesting. I’ll read the article (when I get to my laptop!), but until then, I don’t think the “name second” word order equates S2. A name in a drinking formula could juster than well be that of a god, and the subject be unexpressed, at least by name. That means verb – benefactor – patient or V IO DO.

  3. ə de vivre says:

    VSO order often happens when the pronouns of an SVO language cliticize to the verb then become affixed. IIRC, the NENA languages (re)acquired their VSO order this way. I also have vague memories of discussions about Proto-Semitic (and/or Western Semitic) word order being explained this way.

    But Welsh also became AUX V O this way, so VSO doesn’t itself mean Afro-Asiatic or Semitic.

  4. David Eddyshaw says:

    It seems very unlikely that Linear A is Semitic. Cyrus Gordon suggested this a long time ago, but seemed to feel that his work was done simply in making the suggestion. The fact is, that if it were Semitic, you’d expect a lot more progress to have been made with the language by now: the ancient Semitic languages resemble each other closely enough that you’d expect numerous characteristic patterns to have come to light already: the decipherment of Akkadian was much eased by the fact that it was Semitic.

    It’s generally believed that the Akkadian word order is due Sumerian influence. VSO certainly is widespread in Afroasiatic, and this probably reflects the (extremely ancient) protolanguage; however, most Chadic languages are SVO, and Cushitic languages SOV; the modern Semitic languages of Ethiopia have become SOV likewise presumably because of Cushitic influence.

    One thing this demonstrates, of course, is that basic word order isn’t particularly stable over time, and it’s not a good predictor of linguistic affiliation at all.

    Latin was SOV; all its descendents are SVO. Insular Celtic is VSO; continental Celtic was SOV.

    In West Africa, most Mande languages have a typologically extremely unusual subtype of SOV order where adverbial elements of all kinds (even including indirect objects) follow the verb; this order is shared by most Songhay languages, but not even the most enthusiastic Lumpers believe that Songhay and Mande are in any way closely related*. Moreover, Timbuktu Songhay has the SVO order of neighbouring Fulfulde.

    Gur languages, like most West African Niger-Congo languages, are SVO; but the Senoufo languages are SOV, just like neighbouring Mande.

    There are, however, strong though not invariable correlations between VSO word order and other syntactic features (like the conjugated prepositions of insular Celtic and Semitic); so if it is firmly established that Linear A was VSO that might indeed help with further analysis.

    *Actually, this is not quite accurate; there have been (very stupid) suggestions that the resemblances between neighbouring Songhay and Mande somehow prove that the vast and very-far-from-fully-demonstrated Niger-Congo and Nilo-Saharan phyla are themselves related; and rather more respectable (though still far from mainstream) ideas that Songhay is a sort of Mande creole relexified with Berber.

  5. The fact is, that if it were Semitic, you’d expect a lot more progress to have been made with the language by now

    Yes, that’s my take on it as well.

  6. David Eddyshaw says:

    However, I’d concede that, given that VSO is relatively uncommon, and the Afro-Asiatic languages show VSO a good deal more often than most other groups do, if Linear A has VSO order it would increase the chance of it turning out to be (non-Semitic) Afro-Asiatic; perhaps a whole new branch. There’d be nothing surprising about that geographically. It wouldn’t help much with decipherment, unfortunately: Afro-Asiatic is too diverse.

    There were almost certainly a lot more isolates around four thousand years ago than nowadays, though. No particular reason a priori for Linear A to turn out to be related to anything we know.

    Thinking about it, if Linear A, which is a syllabary, did represent an Afro-Asiatic language, and that language showed the kind of ablaut-as-extreme-sport you’d expect of a four-thousand-year old Afro-Asiatic language, the writing system would do a grand job of making the morphology hard to recognise (not like Greek with its well-behaved final flexions, as famously recognised by Michael Ventris.) That might partly explain why the decipherment has been so intractable (though I think there is also the fact that there is nowhere near as much material as with Linear B.)

  7. David Eddyshaw says:

    Even though all Hatters will have seen this already, somebody has to link it:

    https://xkcd.com/2151/

  8. There’s danger in extrapolating from the grammar of formulas to that of the standard language; Gandalf tea Wednesday doesn’t represent a normal English sentence type. But of course even if this only establishes that the syntax of these particular formulas is VSO, that’s still a very useful result.

  9. Trond Engen says:

    The article wasn’t much more than you quoted. What strikes me most on rereading is this (my emphasis):

    That was confirmed when he found the Linear B sign for ‘olives’ (which had been borrowed from Linear A), occurring after the name as the object of the phrase.

    The repeated start of the phrase was therefore a verb, like “gives”, yielding the phrase gives Yasumatu olives, or in English Yasumatu gives olives.

    Just such an offering of olives in a goblet has been found, preserved at the bottom of a sacred Minoan well.

    Doesn’t this suggest that Linear A was a combination of hieroglyphic script and syllabary, presumably that some glyphs had both a lexical and a syllabic phonetic reading? And, possibly, that the Minoans took the script and used the syllabary, either with their original values or their own based on transltion of the sign name? These are certainly not new ideas, but having them supported would be very helpful.

  10. David Marjanović says:

    Hurritic was SOV, Basque is topic-comment-verb. Nobody is telling about Hattic, and I can’t find anything on the Caucasian languages right now.

    There were syllable signs and word(-root) signs in Linear B, and for that matter in cuneiform and modern Japanese, while hieroglyphs can represent 1, 2, or 3 consonants and, with a diacritic, a word.

  11. Trond Engen says:

    There were syllable signs and word(-root) signs in Linear B

    Oh, right, I should have remembered.

    If root signs in linear B correspond to words used without added derivational morphology in Linear A, that’s suggestive of the morphological structure of Minoan. I really should get hold of his paper. Or something later. I see it’s from 2014.

  12. David Eddyshaw says:

    According to Johanna Nichols, Ingush is

    verb-final but with frequent verb-second order in main clauses (with prefixes and first elements of compound verbs left in clause-final position.)

    which sounds vaguely familiar.
    Hewitt says that in Georgian “the neutral orderings are SOV and SVO. Arguments can be put in focus by being placed first or last in the clause.”
    Colarusso says that Kabardian is “a rigid verb-final language.”

  13. David Eddyshaw says:

    If root signs in linear B correspond to words used without added derivational morphology in Linear A, that’s suggestive of the morphological structure of Minoan.

    Up to a point; but it could just reflect the alterations necessary to write a foreign language in a script originally developed for Minoan. Early Sumerian texts omit much of the agglutinative morphology; the progressive increase in writing out of affixes seems to go hand in hand with ongoing language death, presumably because native speakers of Sumerian didn’t need the help, but Akkadian scribes did. (Unlike me, ə will Actually Know about this.)

  14. David Marjanović says:

    which sounds vaguely familiar.

    It does…!

  15. David Marjanović says:

    Also from 2014 is a book with this abstract (farther down the author’s Academia page):

    Inscribed Minoan stone vessels are ritual gifts that index their dedicants’ intention that both their gift and their name should survive permanently at the place of dedication. These vessels contained offerings, yet the vessels themselves were also offerings, serving as permanent records of a ritual act. These rituals were most likely communal, incorporating group feasting and drinking. The seasonality of these rituals suggests that they were focused on the cycle of life: fertility, birth, death and renewal. Offerings left with the vessels suggest that these rituals also addressed other, more personal concerns. As for Linear A itself: the language behind the script appears to contain a fairly standard phonemic inventory, though there are hints of additional, more exotic phonemes. The morphology of the language appears to involve affixation, a typical mode of inflection in human languages. The presence of significant prefixing tends to rule out PIE as a parent language, while the word-internal vowel alternations typical of Afroasiatic verbal inflection are nowhere to be found in this script. In the end, Linear A appears most likely to represent a non-IE, non-Afroasiatic language, perhaps with agglutinative tendencies, and perhaps with VSO word order.

  16. David Eddyshaw says:

    More Caucasian: Diana Forker’s Hinuq grammar says that in main clauses the overwhelming majority are SV or APV (it’s ergative) but all logically possible orders are attested or at least elicitable, with “pragmatic” differences which she doesn’t much go into; however, in subordinate clauses verb-final order is “almost obligatory.”

    In languages where varying of the basic word order is possible, it seems to be cross-linguistically very common for such variation to be limited or absent in subordinate clauses. Not surprising when you start thinking about it, if word order changes are used to mark focus, foregrounding and such.

    Actually, this is potentially relevant to this Linear A issue: even if the examples have been correctly identified as (say) “Gave Minos olives”, it does not at all follow that this is the unmarked word order of Minoan (as TR has really already said.) The VSO order could be focussing the object (“What did Minos give?”) or the subject (“Who gave olives?”); if so, the unmarked order might be pretty much anything, even boring old SVO.

    Still, you’ve got to start somewhere.

    It’s just occurred to me that we have no reason to assume that the first word is in fact a finite verb at all. It might be (for example) “gift” or “offering.” Then we have a possessum-possessor NP construction, a null copula, and a complement: “The offering of Minos – olives.” Latin word order, pretty much.

  17. Kabardian is “a rigid verb-final language.”

    Kardashian, on the other hand, is a soft tushy-final language.

    =======

    “Is it really true that OVS in particular sentence types indicates a truly OVS language?” asked John.

  18. David Eddyshaw says:

    The offering of Minos – olives.

    In fact, the word order is clearly Russian.

    It is only the archaic character of the writing system which, by omitting flexions which could be readily supplied by a native Russian speaker, disguises the fact that the second word is in fact in the genitive.

    @JC:

    Clearly not, replied David.

  19. David Eddyshaw says:

    The example “Has Minos a minotaur” is poorly chosen. Now that it is proven that Minoan was Russian, this should be recast as “To Minos exists a Minotaur.”

    Verbs used like English “have” are not as universal as Andrew Trounson supposes.

  20. David Eddyshaw says:

    Kardashian, on the other hand, is a soft tushy-final language.

    Those Kashubian-Jenners are just famous for being famous.

  21. John Cowan says:

    Well, also for the smallpox vaccine.

  22. Stu Clayton says:

    the conjugated prepositions of insular Celtic and Semitic

    German zum, zur, am, vom look like conjugated prepositions, don’t they ?

  23. Stu Clayton says:

    Dammit, confused conjugation and declination again.

  24. I’d like to point out that one thing is a fixed formula, and another is the general, or prevailing order of words in a language.

    My language (a dialect of Croatian or Bosnian or Serbian) is always described as SVO. But, depending on the verb, the word order in some sentences will be actually VS, or even (e.g. when talking about pain) OVS.

    Not to mention that it lacks some typical traits of SVO languages…

  25. A few comments some may find pertinent if not interesting…

    1-A possible reason why Linear A is yet to be cracked (I believe I once made the same point here at Casa Hat on the subject of the Harappan inscriptions): perhaps the Linear A corpus involves a script used to write more than one language. I do have a precedent in mind: the Iberian, Tartessian and Celtiberian languages were all written using (variants of) the same (Paleohispanic) script, and early attempts at decipherment were hampered because researchers assumed the inscriptions were likewise written in a single “Paleohispanic” language. Only later research determined that the paleohispanic corpus was multilingual, with the Celtiberian portion the only one to have been deciphered.

    2-Speaking of Celtiberian: David Eddyshaw, I believe Celtiberian is the only Continental Celtic language whose syntax is consistently verb-final: Gaulish seems to be predominantly SVO, whereas for Lepontic and Galatian we simply have too little data. Of course, this strengthens your point: Among known Celtic languages we have some that are SVO, others that are SOV, and others that are VSO.

    3-David Eddyshaw, Hat: Even if for the sake of argument we accept that the Linear A corpus consists of a single language, I could easily believe the language is Semitic. For decades Carian was believed to be a non-Indo-European language: it is in fact Indo-European (Anatolian, actually). This fact was long obscured by very unusual sound values which had been given some letters of the Carian alphabet: perhaps, in like fashion, the sound values of the shared symbols of Linear B and of Linear A differed more than we would like to think, obscuring whatever Semitic features might otherwise be detectable today.

    4-Whatever one might think of my points above, I must echo TR and Daniel N: the grammar of a formula needn’t be the grammar of a language: this danger should be especially acute when (as in this case) the corpus is dominated by writings of a religious nature.

    5-Something the article does not mention is the fact that Linear A was not the first script to be used in Crete: that honor goes to Cretan Hieroglyphs. Whether Linear A and Hieroglyphs were different scripts used to represent the same (non-Greek) language or different scripts used to represent different languages (perhaps either or both scripts represented more than one language, as pointed out above) remains, I think, wholly unknown.

    6-Unfortunately, the fact that linear A was in contact with an even more obscure script complicates matters, as contact between different languages and scripts is liable to create complications: I have in mind the logograms in Hittite, where words were written out in cuneiform as Sumerian or Akkadian words to represent native Hittite words. If Cretan Hieroglyphs and Linear A originally represented separate languages, perhaps some similar features of Linear A orthography (words spelled out like their equivalent in the “Cretan Hieroglyph language” to represent “Linear A language” words) account for Linear A remaining this opaque today.

  26. David Eddyshaw says:

    For decades Carian was believed to be a non-Indo-European language

    Fair point; however, Indo-European is much more diverse than the Semitic of four milllenia ago (and Carian turned out to belong to the most aberrant branch of Indo-European at that) so the late recognition of its true affiliation is much less surprising than if Minoan turned out to be Semitic.

    Having said that, my argument that the Linear A writing system, which is surely a syllabary where it is phonetic, would be calculated to thoroughly obscure the morphology of any ancient Afro-Asiatic language, certainly applies to Semitic itself.

  27. Stu Clayton says:

    “Calculated” = apt ?

  28. David Eddyshaw says:

    Pretty much. “Apt” is not as refined, though.

  29. And “doggone determined” would be even less refined!

  30. January First-of-May says:

    As far as phonological values go, Linear B (and its successor the Cypriot syllabary) is a lot like hiragana; this had been taken as suggesting that whichever language the predecessor of Linear B (probably Linear A) was originally invented for was likely to have had a phonology not unlike that of Japanese (and thus very much unlike Russian, English, or Greek) – i.e. mostly CV syllables and no phonemic voicing.

    Dunno what this possibility means about the assorted hypotheses – though of course it’s perfectly possible that Linear A as we know it was not the original language of that syllabary either.

    Now that it is proven that Minoan was Russian, this should be recast as “To Minos exists a Minotaur.”

    “At Minos”, rather. “To” would be a different preposition.

    And Hebrew would have “Exists to Minos a Minotaur” (…I think; not sure if the word I translated as “exists” actually means that, but I’m confident about the “to”).

    It is only the archaic character of the writing system which, by omitting flexions which could be readily supplied by a native Russian speaker, disguises the fact that the second word is in fact in the genitive.

    It doesn’t really have to omit flexions; maybe the second word is in the genitive and we don’t know that because those names had not been attested in any other form, and/or because their genitives happened to be the same as said other form (say, the accusative: “the offering of Minos – olives” and “at Minos exists a Minotaur” would both have had Minosa in Russian, despite being in different cases).

    IIRC, the Linear B decipherment was significantly aided by words being attested in multiple case forms, which allowed the decipherers to guess that the difference was probably only in the final vowel (a guess that ultimately turned out to be correct).

  31. David Marjanović says:

    The idea that the default word order of Linear A texts was VSO was already suggested here; latest update: August 2019.

    A general introduction to “the Aegean pre-alphabetic scripts” (including the two of Cyprus) is here.

    Gaulish seems to be predominantly SVO

    Maybe Alesia was in all of us all along!

    This fact was long obscured by very unusual sound values which had been given some letters of the Carian alphabet: perhaps, in like fashion, the sound values of the shared symbols of Linear B and of Linear A differed more than we would like to think

    That’s unlikely in this case. The Carian alphabet appears to be derived from the Greek one by turning the letters cursive and then back to monumental, sometimes ending up with the existing shapes of other letters and thus with the false friends you mention. Between Linear A and B, many logograms are shared, and so are entire names. Also, you can trace the sign mu from Linear B through Linear A back to Cretan hieroglyphs, where it’s the head of a bull; ra is the head of a dog, and ma that of a cat (easily identifiable in some of the many Linear A versions).

    the corpus is dominated by writings of a religious nature

    There are lots of writings of a bureaucratic nature. But those have a fixed two-dimensional layout – they’re pretty well understood nowadays, which lets us reconstruct the economy but helps us with barely more than two words (because all the goods are written with logograms).

    Cretan Hieroglyphs

    Pretty clearly related to the Linears like the Chinese characters on oracle bones are to Chinese seal script: they’re more recognizable picture versions of the same characters.

  32. David Eddyshaw says:

    January will receive suitable credit when I publish my monograph on the Russian nature of Minoan. These are all excellent points.

    Actually, the point about genitives is spot on. Ogham inscriptions are the same, though there at least we have enough information from other sources to see what’s going on.

    Good reminder about the Linear B decipherment: and in fact if Linear A encoded a Semitic language (of that antiquity) you’d expect an even more transparent situation, with the singular cases going -u nominative, -a accusative, -i genitive.

    I suspect the intractability of the decipherment in fact is as much as anything to do with sheer paucity of source material (as the original article implies.) If Etienne is right that Linear A texts are actually in more than one language that would make the situation even worse, of course.

    I don’t think the nature of the syllabary necessarily tells us much about Minoan, even if it was devised for that language (and even if there was only one Minoan language.) Cuneiform is basically a very bad way of writing even Sumerian; these are after all first attempts at devising writing systems, and it’s not surprising that they leave a lot to be desired from our point of view. Only the unsung hero(ine) who created Egyptian hieroglyphs had enough linguistic insight to abstract away from the syllable (admittedly s/he was presumably helped along by the peculiar morphology of Afro-Asiatic, but even so.)

  33. David Marjanović says:

    mostly CV syllables

    The Luwian hieroglyphs consist only of V & CV syllables except apparently for a sign for final -r. There’s no evidence that they were invented for another language, and from what we can tell about Luwian from cuneiform texts and other Anatolian languages it had plenty of closed syllables.

    I wonder if this restriction was a deliberate decision to keep the number of signs manageable.

  34. David Eddyshaw says:

    After all, the writing systems didn’t have to be good; they just had to be good enough.

    And that means good enough for professional scribes, who had a perverse incentive for the system not to be readily usable by muggles (Exhibit A: Pahlavi, for which this is surely the only rational explanation.)

  35. David: Luwian hieroglyphs were indeed almost certainly invented for Luwian, but there is some linguistic evidence that they were created in a Luwian-Hittite bilingual environment (I can give references if anybody asks).

    A much more blatant example of a phonologically inadequate script consists of various paleohispanic scripts which, despite being of Phoenician origin (with Phoenician script having separate symbols for several voiceless and voiced consonants) and despite being used to write a language (Celtiberian) which had separate voiceless and voiced stop phonemes (as is shown quite clearly in Greek and Latin transcriptions of Celtiberian words), failed to differentiate voiced and voiceless stops. And in this case this limitation definitely cannot be explained by a desire to keep the number of signs manageable.

  36. ə de vivre says:

    mostly CV syllables and no phonemic voicing.

    Huh, not having phonemic voicing seems to be an areal feature of the non-Semitic Bronze Age Middle East: Sumerian, Elamite, Hurrian—and Hittite lost it. If Hittite could lose the distinction, a Semitic language could too, but without a 3-way stop distinction or an u/i/a case system, I don’t think VSO word order is very convincing evidence of Semitic affiliation.

  37. January First-of-May says:

    Exhibit A: Pahlavi, for which this is surely the only rational explanation.

    Very early Pahlavi (the so-called Inscriptional Pahlavi) was pretty decent, IIRC. Book Pahlavi, on the other hand, is basically the Pahlavi version of Sütterlin.

    (Admittedly I’m not sure when or how did the random Aramaic logograms show up.)

  38. “Please give a big welcome to tonight’s opening act, the Random Aramaic Logograms!”

  39. David Eddyshaw says:

    (Supporting Chadwick and the Neutrons.)

  40. these are after all first attempts at devising writing systems, and it’s not surprising that they leave a lot to be desired

    Interesting that you view writing systems in terms of “progress”. The lack of “progress” (stuck at an earlier stage) of the Chinese writing system was a common observation in the colonial period. I don’t think it is fashionable any more, but fashions change…

  41. David Eddyshaw says:

    I don’t think it commits one to the Whig view of history to regard alphabetic writing as an advance on cuneiform.
    What I had in mind, though, was actually a change during the history of cuneiform itself in the direction of fuller representation of speech sounds specifically, even in writing Sumerian, and of course a fortiori when the system was adapted to new languages. On that metric there was certainly “progress.”

    This judgment does conceal an assumption about what writing is actually for, admittedly. I am not committed to the view that practical utility is the only worthwhile criterion (I am no Mairist.) I would still regard it as quite desirable though …

    I think you could make an argument that the Chinese system has actually regressed since Li Si’s reforms: the relationship between writing and speech was probably a good bit less opaque at that point.

    Other systems have been made worse (in the utilitarian sense, at any rate) over time, either by the sheer entropy of not keeping up with sound changes (like our own dear English spelling, and Chinese, and many others) or sometimes on purpose. It must have been relatively straightforward for a bright native speaker of Egyptian to acquire literacy at the beginning of the Middle Kingdom; by the New Kingdom, not so much at all.

  42. not having phonemic voicing seems to be an areal feature of the non-Semitic Bronze Age Middle East: Sumerian, Elamite, Hurrian—and Hittite lost it.
    Whether Hittite lost the distinction or not depends on whether you assume that voicing was distinctive in PIE. That certainly is the traditional view, but there are models of the PIE consonant system being proposed where the phonemic distinctions were tense / lax / glottalised and features like voicedness or aspiration as known from e.g. Greek or Sanscrit are explained as purely phonetic or as later developments.

  43. Athel Cornish-Bowden says:

    I drew to attention of someone at sci.lang who knows about Linear A (more than I know, anyway) to this thread. He described Davis’s work as “a lot of foam and no beer”. I won”t quote any more, because he knows how to get here and express his views himself if he wants to.

  44. David Marjanović says:

    I think you could make an argument that the Chinese system has actually regressed since Li Si’s reforms: the relationship between writing and speech was probably a good bit less opaque at that point.

    Yep, Baxter & Sagart say before him (roughly) it was pretty close to a syllabary with determinatives.

    the sheer entropy of not keeping up with sound changes (like our own dear English spelling, and Chinese, and many others) or sometimes on purpose.

    “aCHe, agHast, aiSle, aLmond, ancHor, arbOUr, bliGHt, bUild, bUry, cauGHt, (musical) cHords, coLonel, couLd, crumB, deliGHt, dHow, endeAvOUr, dingHy, fOetus, foreiGn, gHastly, gHerkin, gHost, ginkGo, glamOUr, glisTen, hauGHty, iSland, lacHrYmose, limB, lisTen, misdemeanOUr, neighbOUr, numB, postHumous, Ptarmigan, QUeue, redouBt, rHumb, rHyme, roWlocks, sCent, Scissor, sCythe, sovereiGn, spriGHtly, thumB, tongUE, wHelk, Whole, Whore. All the capitalised letters are spurious, and often they were deliberately added as “improvements” by incompetent etymologists.”

    (I added the Y in lachrymose because that’s only “justified” if the word switches from Latin lacrima to Greek dákryma halfway through. And not all of the spurious letters were added specifically in English, but most were. …And I’m not sure about neighbour.)

  45. PlasticPaddy says:

    Re listen, I think in old-fashioned Hiberno-english (and maybe some English or Scots dialects?) “list” can be used. Compare fast and fasten.

  46. He described Davis’s work as “a lot of foam and no beer”.

    Sad but not surprising; thanks for checking with someone who actually knows!

  47. David Eddyshaw says:

    The expert in question believes that most of Linear A is in Hurrian, and also that Carian is a dialect of Hurrian, the identification as Anatolian being “entirely bogus.” He has identified Uralic elements in Na-Dene and various indigenous languages of California.

  48. Stu Clayton says:

    Please don’t leave us Fußvolk in breathless, clueless speculation ! Are you “suggesting” (as the cowardly newspapers now say, where I would say “saying”) that the expert in question is a foamhead too, or that the beer is bitter ? We like to see effective punches, even though they may be below the belt.

    I would be dazzled, for merely formal reasons of self-referentiality, to learn that the e. in q. is yourself.

  49. David Marjanović says:

    that the expert in question is a foamhead too

    Yes. I think what’s on Wikipedia should be enough to make clear that Carian is an Anatolian language and nowhere near Hurrian, which is not IE.

  50. I’m confused. Which expert? The one whose output is claimed to be foamy or the one making that claim?

  51. Stu Clayton says:

    The claimants are the experts in response. They are not the expert in question. Or so it would seem.

  52. Stu Clayton says:

    what’s on Wikipedia should be enough

    Now Wikipedia is being invoked as another expert in question and/or response ! My standpoint is: when I know zilch about a subject, I am in no position judge whom to “believe” among those who claim to know more than zilch.

  53. David Eddyshaw says:

    I would be dazzled, for merely formal reasons of self-referentiality, to learn that the e. in q. is yourself.

    Alas, I am no expert in Hurrian. If I were, I too might be inclined to see Hurrian everywhere.

    Though I wonder if a Niger-Congo affiliation for Linear A has been given proper consideration?
    The very sensible page on Linear A that David M linked to above says that the language seems to have a lot of prefixing. And it (maybe) had lots of CV syllables.
    Hmmm … got it! Bantu!

    (It is actually quite common in Bantu languages for the word order, which tends to be free-ish but to gravitate towards SVO, to go VS in order to focus the subject. Must stop now before I’m tempted to take the suggestion seriously and become the linguistic equivalent of a Crazy Cat Lady.)

  54. David Eddyshaw says:

    @Stu:

    OK. I was honouring AC-B’s reticence, but the identification did not take Cowan-like levels of google-fu; so, from his many papers on academia.edu, the merits of which I lack the expertise to assess, I was struck by (among others):

    https://www.academia.edu/40243933/Chechen_reasons_why_Nakh_is_Indo_European

  55. Oh dear.

  56. Stu Clayton says:

    Alas, I am no expert in Hurrian

    But, along the lines suggested by DM, you could easily become one after consulting the WiPe. The English and (slightly more detail-luxuriant) German articles contain a whole lot of stuff that should be instructive to those in the know about this kind of stuff.

  57. Stu Clayton says:

    Chechen reasons

    In what language is “chechen” a numeral ? The paper doesn’t have the word in its title. Seven Brides for Seven Brothers ?

  58. David Eddyshaw says:
  59. Stu Clayton says:

    From there: annuitants of the ongoing Indo-Europeanist fraud

    It’s good to learn that somebody is profiting from this business. To me it has always seemed a brotlose Beschäftigung.

    This guy, or someone of similar allure, ruffled feathers here a few months ago. You can usually recognize the unemployed by the vigor with which they attack pensioners.

  60. PlasticPaddy says:

    Chechen seems to be the author’s stand-in for proto-nakh. Perhaps he should have included more Batsbi reflexes. Some of the correspondences are ingenious, e.g. hair and “fur, sheep-hide”.

  61. Stu Clayton says:

    Some of the correspondences are ingenious, e.g. hair and “fur, sheep-hide”

    Ingenious-seeming perhaps, depending on what one already knows. Spanish piel “[human] skin”, pellejo/a “fur/hide [of an animal]”. The first from Lat. pellis, the second from Lat. pellicula, according to the RAE.

  62. David Eddyshaw says:

    Kusaal uses the same word for animal hair and human body hair, and a completely unrelated word for human head hair. Makes sense (more or less), I suppose.

  63. David Eddyshaw says:

    This is a review of his work with Allan Bomhard purporting to show that Hurrian is Indo-European:

    https://www.academia.edu/345706/Review_of_The_Indo-European_Elements_in_Hurrian_by_Arnaud_Fournet_and_Allan_R._Bomhard_2010_

  64. Athel Cornish-Bowden says:

    Seven Brides for Seven Brothers

    I once mentioned elsewhere that one of my ancestors was the tenth of ten brothers. Someone said that as he wasn’t his own brother and he only had nine brothers he couldn’t be the tenth of ten brothers. I thought that was taking pedantry to an extreme, and said that by that logic the film should have been called Seven Brides for Six Brothers. Thoughts?

  65. John Cowan says:

    Perhaps the Kusaal words originally meant “where the head lice are” and “where the body lice are”. I always snicker softly to myself when I remember yet again that human pubic lice (Pthirus pubis) are most closely related not to human head/body lice (Pediculus humanus) and chimp/bonobo lice (Pediculus schaeffi), but to gorilla lice, and they can’t jump. How did they get there? The world wonders…

  66. Stu Clayton says:

    By sleeping on the same leaf bed as gorillas used the night before. Today motel sheets are an important vector of pubic lice, their pubic transportation system.

  67. Stu Clayton says:

    Seven Brides for Six Brothers and a Relative

  68. >The claimants are the experts in response. They are not the expert in question. Or so it would seem.

    Did it really seem this would answer anything for anyone who didn’t already know the answer? It’s seems deliberately opaque yet at the same time thoroughly muddled.

  69. David Eddyshaw says:

    Interestingly (well, to me, anyway) the Kusaal word for “animal” (of the large warm-blooded non-human sort) is bʋnkɔnbʋg, which looks like a transparent compound “hairy thing”, whereas one core term for “human being” (as opposed to animal) is saal, which looks like a transparent use of the adjective stem “smooth” in the “human” a/ba noun class: so “naked big mammal” (or “naked ape” if you must.)

    However, all the “looks like” above is because I’ve been bitten by this sort of tempting derivation before, and discovered that the relevant words must have been completely distinct historically and have fallen together through regular phonetic mergers.

    One that I’m still not altogether clear about is ya’am/yam “gall bladder; common sense”, where the dictionaries of Western Oti-Volta languages duly say of the second meaning “(as being the seat of understanding): common sense, wisdom.” Unfortunately the reflexes outside Western Oti-Volta strongly suggest that these were two completely distinct words which have fallen together as the result of regular sound changes, so that if the belief that the gall bladder is the seat of wisdom was not simply invented by Western lexicographers to “explain” the phenomena, it’s a genuine belief that arose as a result of regular phonological changes.

    I was thinking about this sort of thing because of a recent mention in a different thread of Bruce Chatwin (OK, fish in a barrel, I know.) I remember a list he came up with somewhere of (I think) Araucanian vocabulary, where he was attempting to divine titillating exotic Araucanian mystical connections between concepts which seem very different to us; somewhat undermined by the fact that he evidently had no concept of (or perhaps interest in) the fact that many of his supposed homophones were almost certainly illusions created by inadequate transcription.

  70. Stu Clayton says:

    Deliberately opaque yet at the same time thoroughly muddled

    Thanks! That’s exactly what it was intended to be. My work is done.

  71. David Eddyshaw says:

    Seven Brides for Seven Brothers

    In Sanskrit, the dual pitarau “two fathers” means “parents.”
    OK, that’s not immediately relevant, but when else am I ever going to have the chance to say it?

    Perhaps slightly more relevant: in Kusaal, dɔl means “go with, accompany (in a subordinate role)”; this is incidentally what “follow” usually means in West African English: it doesn’t mean that you are (necessarily) trailing after someone. You can “follow” someone from in front. It just implies that you are basically the sidekick.

    However, you can perfectly well say Ba dɔl taaba “they go together” (“they follow each other.”)

  72. Stu Clayton says:

    You can “follow” someone from in front

    It’s called a front tail. Seriously.

  73. David Eddyshaw says:

    Actually local etiquette more or less requires that if you’re the sidekick you should go first. I expect it’s so that you get to step on the snakes/land mines/whatever, rather than the boss. Makes sense.

    The Kusaal personal name Ana’ab “Anaba” is literally “Chief”; but what it actually means is “afterbirth” (because a chief leaves the house after his retainers.) It’s a name given to the sole survivor of a pair of twins.

  74. PlasticPaddy says:

    https://en.m.wikipedia.org/wiki/Pierson%27s_Puppeteers
    Clearly the Kusaal are descended from space aliens and proof that WE ARE NOT ALONE!

  75. David Eddyshaw says:

    Are we not all descended from space aliens?
    (On second thoughts, it may be better if you do not answer that question.)

    The neighbouring group to the immediate north of the Kusaasi, the Bisa, are said to believe that they are descended from crocodiles, and that consequently crocodiles will not eat them if they fall in the water. A Kusaasi friend with a rigorously Galtonian approach to such matters told me that he had sadly been unable to persuade any Bisa to subject this belief to controlled experiment.

  76. In Sanskrit, the dual pitarau “two fathers” means “parents.”

    And so does mātarau “two mothers”. And the “two Indras” are Indra and Vāruna, the “two heavens” are heaven and earth, “two mortars” are a mortar and pestle… This IE “associative/elliptical dual” survived into Latin — Castor and Pollux can be referred to as the Castores — and into Greek, where the early Homeric use of Αἴαντε “the two Ajaxes” to mean “Ajax and Teucer” appears to have been misunderstood by later bards, who duly wrote the lesser Ajax as a sidekick (à propos) into scenes where he never belonged.

  77. Trond Engen says:

    The “expert” who must not be named (but whose name can be gleaned from David’s links) has a long history on the Internet. Little is of substance, none of it is pleasant. How he got Bomhard to put his name on a shared paper is beyond me.

    When it comes to the morphology of the Minoan language(s), I concede all points made above. Still, as David E. says, you have to start somewhere, and repetitive patterns in the restricted and formulaic setting of inscriptions on drinking vessels seems a very good place for just that.

  78. David Eddyshaw says:

    What Trond says …

    The weakest point in this seems to me to be the assumption that the first word in this formula has to be a finite verb; but it may be (for example) that in similar contemporary nearby cultures dedication formulae are almost always expressed with finite verbs, so that that would be a reasonable assumption. (I’ve no idea if that’s so myself.)

    If the language were really shown to be consistently VSO, it certainly would be a step forward quite apart from the question of linguistic affinity: it would be extremely likely to have prepositions rather than postpositions, for example, and to have noun phrases where the head usually comes before dependent modifiers and such.

  79. David Marjanović says:

    Castor and Pollux can be referred to as the Castores

    That’s more likely a Greek loan (as the names themselves are, with the usual l ~ d confusion): it is the dual Κάστορε in the original.

    The “two heavens”, though, did survive in Latin: at least that’s the only explanation for why the plural of neuter caelum is the masculine-looking caeli. Now if only I could remember where on academia.edu the paper is.

    How he got Bomhard to put his name on a shared paper is beyond me.

    I was going to say!

  80. PlasticPaddy says:

    Re caeli
    Has anyone suggested a fossilised dual (compare viginti)? Old Latin caelus could even be back-formation. I agree there are many other Latin gender benders which would have to have a different explanation.

  81. That’s more likely a Greek loan

    Yeah, good point. There’s also Cereres, which in a cult context refers to Ceres and Persephone; I don’t know if there’s a Greek parallel for that. (In any case of course these are plurals rather than duals, so not a perfect continuation of the PIE use.)

  82. David Marjanović says:

    Has anyone suggested a fossilised dual

    That’s what the paper says that I can’t find.

  83. that’s the only explanation for why the plural of neuter caelum is the masculine-looking caeli

    Well, it could have originally been a regular plural of caelus (attested in Ennius); the plural of this word is pretty rare and poetic anyway. And caeli doesn’t mean “heaven and earth”, as it should if it was an associative dual.

  84. David Marjanović says:
  85. David Eddyshaw says:

    You can never have too much Hattic.

  86. David Eddyshaw says:

    From a series of reviews of works on Hurrian by Alexei Kassian, obligingly produced for me by academia.edu (well, it’s much preferable to the time it spent under the misapprehension that I was interested in Mormon studies of American prehistory) comes a brief note on the same AF/ARB work, which ends:

    The book under review will be interesting for lovers of linguistic curiosities.

  87. PlasticPaddy: Re listen, I think in old-fashioned Hiberno-english (and maybe some English or Scots dialects?) “list” can be used. Compare fast and fasten.

    This may be most familiar outside these dialects from the song “The Patriot Game,” as sung here by Liam Clancy. (Clancy, like practically all performers, has edited the original lyrics to take out their worst excesses of violent nihilism.)

    Come all ye young rebels, and list while I sing,
    For the love of one’s country is a terrible thing.
    It banishes fear with the speed of a flame,
    And it makes us all part of the patriot game.

  88. Reading Kassian’s article on Hattic, discovered that at least one Russian word is likely of Hattic origin.

    киноварь (cinnabar) from Ancient Greek κιννάβαρι, of Middle Eastern origin, ultimately might derive from Hattic kinawar (copper).

  89. I feel one ought to distinguish degrees of faultiness of English spelling. Apportioning blame is such fun! A folk etymology is more forgivable than a pedant’s miscorrection. OE hlystan “hear” != hlysnan “listen” but the ME conflation is venial, not venal.

    I don’t see how “QUeue” works; if the original is “cue” than it should be “QUEue”, but even that I disagree with; it’s the whole spelling that’s been replaced with the French (contrast quire > choir != chœur).

    Arguably, “queue” is a regular English spelling (“qu” = /k/ + “eu” = /ju/ + “e” = ∅); compare “liqueur”.

  90. January First-of-May says:

    The problem with “qu” (as actually pointed out elsewhere in JBR’s linked spelling rant) is that the Latinate qu = /kw/ conflicts with the Romance qu = /k/ (before front vowels), and Old English didn’t have either.

    Apparently Spanish gets around the same ambiguity by using cu for /kw/ and qu for /k/ before front vowels, which I didn’t know before. It’s a bit ironic that English mostly goes the other way.

  91. John Cowan says:

    the ME conflation is venial, not venal

    Joseph Wright advising the young Tolkien: “Go into folk etymology, lad; there’s money in it.”

    (Actually he said “Celtic”, but wevs.)

    Tolkien also records this conversation with Wright:

    Years before I had rejected as disgusting cynicism by an old vulgarian the words of warning given me by old Joseph Wright.

    “What do you take Oxford for, lad?”

    “A university, a place of learning.”

    “Nay, lad, it’s a factory! And what’s it making? I’ll tell you. It’s making fees. Get that in your head, and you’ll begin to understand what goes on.”

    Alas, by 1935 I now knew that was perfectly true. At any rate as a key to dons’ behavior. Quite true, but not the whole truth.

Speak Your Mind

*