Austronesian and Taiwan.

John Cowan sent me a link to Roger Blench’s paper (draft circulated for comment) “Suppose we are wrong about the Austronesian settlement of Taiwan?,” a fascinating attempt to upend the usual narrative. Here’s the abstract:

The current model of the prehistory of Taiwan assumes that it was first settled some 25,000 years ago by a population of unknown affinities, who reached what is now an island via a landbridge, at a time of much lower sea-levels. Some 5500 years ago, the Ta Pen Keng (TPK) culture, attested on the Peng Hu islands in the Taiwan Strait, apparently represents an incoming Neolithic population. Similar TPK sites are recorded around the shores of Taiwan in the centuries immediately following this. The pervasive assumption has been that these early settlers were the bearers of the Austronesian languages, which then diversified. If so, related Austronesian languages were formerly spoken on the Chinese mainland and these subsequently disappeared as a consequence of the Sinitic expansions. The indigenous Austronesian languages of Taiwan are claimed to reconstruct to a single proto-language, PAN, and from these reconstructions we can derive hypotheses about the lifestyle and subsistence of the earliest settlers.

This paper will argue that the single migration model is mistaken, and that it is not consistent with either the archaeology or the lexicon. If Formosan languages appear to reconstruct to a proto-language it is because they have been interacting over a long period, but they actually represent a continuing flow of pre-Austronesian languages from the mainland. Part of the evidence for this is the exceptional diversity of lexical items which are supposedly part of basic subsistence vocabulary.

Three phases of migration are distinguished, the TPK, the Longshan type culture and the Yuanshan, all of which originate on different places on the Chinese mainland. A further back migration from the Philippines may be responsible for the primary settlement of Green island and parts of the east coast, resulting in the present-day Amis population.

From the conclusion:

From this it follows that a single PAN cannot be reconstructed, in the sense of an apical ancestor, merely a Common Formosan (CF). The Formosanisms identified by Dyen and Blust are not PAN but rather local innovations. This explains why reconstructions of PAN phonology and grammar have always tended to be inconclusive. The flat arrays proposed by Blust and Ross would thus be a reflection of prehistory, although not in the sense originally intended.

It’s clearly written, and there are very useful maps, tables, and illustrations; I’ll be interested to see if the theory becomes accepted.

Comments

  1. —Unfortunately we know nothing of the languages the present Austronesian languages replaced. Since Taiwan has been inhabited for 25,000 years, it is likely that it was formerly extremely linguistically diverse, as is common with small bands of foragers. The oral traditions of several peoples, notably the Saisiyat and the Bunun, point strongly to a survival of ‘pygmies’ until fairly recent times.These populations could be parallel to the Orang Asli of the Malay Peninsula or the negritos of the Philippines. The Saisiyat people continue to hold ceremonies to thank the pygmies for their medical and agricultural [!] skills. That said, although we know Taiwan was long inhabited by foragers we have no evidence for their physical anthropology and the small size could be simply a gloss on their subordinate status.—-

    Were these vanished Taiwan pigmies humans? Or hobbits, like on Flores?

  2. This brings to mind the recent papers from Peter Norquest and Sean Downey, who argue for several new Proto-Malayo-Polynesian phonemes. One of them is equal to PAn *C, which had previously been assumed to have merged with *t everywhere in MP. Also, since most of their new distinctions fail to be found anywhere in Formosan, this means that the status of Malayo-Polynesian as a distinct subgroup is not on a very good standing after all. They propose that the Austronesian homeland could well have rather been located in Borneo, with Formosan representing a divergent sub-branch.

    (See e.g. Expanding the PAn consonant inventory.)

    This additionally reminds me of our current debate in Uralic studies over just how much of a special position Samoyedic deserves. It’s quite interesting to see that similar taxonomic problems are coming up elsewhere as well.

  3. David Marjanović says:

    Were these vanished Taiwan pigmies humans? Or hobbits, like on Flores?

    Read on to the genetics section. 🙂

    It’s quite interesting to see that similar taxonomic problems are coming up elsewhere as well.

    I’m reminded of the Altaic situation, the biggest difference being just that Austronesian happens to be generally accepted while Altaic is “already” controversial…

  4. @J: The Austronesian and Uralic situations also seem reminiscent of Sino-Tibetan, in which the the traditional analysis posits a Sinitic branch comprising the Chinese dialects, and a Tibeto-Burman branch comprising everything else. The latter is vastly diverse, and from what I can tell, its status as a monophyletic group is increasingly being called into question.

  5. Historical linguistics is certainly a lively field these days!

  6. The crux of Blench’s argument is the great diversity in Formosan, as compared to the relative uniformity of Malayo-Polynesian. That in itself is not enough to dislodge the standard model, it can just mean thet Malayo-Polynesian went hrough a bottleneck. Blench doesn’t explain to my satisfaction why diverse subsistence words, like ‘fish’, should be evidence against a common source for Formosan, while commonality of words for ‘seven’ does not constitute evidence for it (the distribution of numeral words in Formosan is interesting too, and Sagart has written about it, in the context of trying to link Austronesian to Thai.)

    Blench does not discuss grammar. He quotes Li on the “great diversity in the grammar of Formosan languages”, but does not refer to Ross’s reconstructions of the verbal system of Proto Austronesian, especially what he calls Proto Nuclear Formosan (i.e. Austronesian minus Puyuma, Tsou and Rukai). Attaching various innovations and retentions to particular historical and sociolinguistic events can and will take a lot of work, and I think Blench is much too premature in drawing a definite scenario.

    All the same, I think on impressionistic grounds his model goes in the right direction. Elsewhere in SE Asia, and even in Europe, clean speciation can no longer be taken as the default. I don’t know if his exact model will carry the day, but a simple tree model does not seem particularly attractive either.

  7. Great read!

    I noted with interest the reference to “Dixon 2002”; it’s not in the bibligraphy (no problem — drafts get fixed!) but I imagine the intended reference is “Australian Languages: their nature and development”. This book actually argues for a leveling/”equilibrium”-based model for Australian languages — which is firmly rejected by most other Australianists as far as I can tell. Nicholas Evans’s 2005 review, “Australian languages reconsidered: a review of Dixon (2002)” makes good reading, and Paul Black’s “Equilibrium Theory Applied to Top End Australian Languages” is a good practical exploration of some difficulties that arise in attempting to apply the model to the data. I wonder how much of an affinity Blench feels with Dixon’s model — the paper itself only vaguely references “pan-Australian lexemes”, which uncontroversially exist.

  8. I think Blench is much too premature in drawing a definite scenario.

    Hence “draft circulated for comment”. It’s a common type of communique from Blench, mainly intended to provoke discussion AIUI.

    Calling Austronesian a pseudo-language family created entirely by convergence is a bit extreme, I agree. Perhaps only some of the Formosan languages are what we’d be calling, in traditional terms, isolates with accumulated Malayo-Polynesian(-ish) influence, while others are actual archaic branches. Perhaps the Austronesian homeland has indeed been in the mainland and the various migrations in the archeological data simply demonstrate a “stepping stone” status for Taiwan. Etc. In any case it sounds like further exploration of what specific evidence of Austronesian there is and is not in various Formosan groups would be warranted. A better understanding of etymological typology would also help here, I’m sure. (As in, yes, ‘fish’ is usually a well-retained word, while ‘seven’ is known to be sometimes loaned widely. But does this hold in all historical sociolinguistic situations? I could imagine e.g. ‘fish’ being retained as a substrate loan, if the conditions are right.)

  9. It’s not so much that the word for ‘fish’ is unreconstructable, but that there are so many words for ‘fish’, all of them unreconstructable.

  10. Maybe they had only words for different kinds but none for “generic fish”? After all, most European languages didn’t have a word for “mammal” until scientific classification began, only words for different kinds of mammal.

  11. Middle English had “beast”, which correlated pretty closely with “mammal” if you grandfather out humans. OE “deor” ( > deer) might also qualify. (Actually in my admittedly limited experience, most claims that “the X have no generic word for Y” turn out to be misinferences based on incomplete or poorly understood data.)

  12. Also, mammals are not an intuitive group. Wild large mammals, domestic mammals and rodents are, and they usually do have common names. Fish, on the other hand, looks pretty much like rodents or domestic mammals – it’s impossible not to perceive them together.

  13. I dunno, I assume that “beast” would have included animals like crocodiles? The same way like “fish” included whales and (at least to some sources) beavers? My point is notr so much about the details, but simply that different languages and cultures slice and dice concept spaces in a different way, so maybe PAN did it differently to us?
    And having read Blench’s paper, I think he puts too much weight on the differences in what are, after all, culture words, which can be replaced. We know that the Germanic people have had agriculture for millennia, but look just at a German dialect map for the names for “corn (= generic sense)”, the different meanings of “Korn” (sometimes generic, but also different sorts like “rye” or “oats”), and the different words for “rye” – it’s nice when we can reconstruct a few of such words for PIE, but that’s luck, not something one should expect. He’s probably right to criticize the assumption that, when a word in PMP shows up in a few Taiwanese languages, it can be reconstructed for PAN (if that’s not caricature anyway), but if he cannot prove that some of the Taiwanese languages are not PAN at all, but something else that only came under PAN influence, how does his theory about several waves of immmigration invalidate the traditional trees where the Taiwanese families form branches of equal Status like PMP?

  14. David Marjanović says:

    The same way like “fish” included whales and (at least to some sources) beavers?

    I think including beavers and swamp turtles and so on was never intuitive for anyone, but deliberate rules-lawyering to allow a richer diet during Lent and for monks who had sworn to fast for the rest of their lives.

    We know that the Germanic people have had agriculture for millennia, but look just at a German dialect map for the names for “corn (= generic sense)”, the different meanings of “Korn” (sometimes generic, but also different sorts like “rye” or “oats”), and the different words for “rye” – it’s nice when we can reconstruct a few of such words for PIE, but that’s luck, not something one should expect.

    While your link doesn’t work, I’ve seen such a map and agree. The American meaning of corn is itself a stellar example – I’ve seen Americans suspect all kinds of fraud because that word is in the King James Bible!

  15. Stefan Holm says:

    Is there any safe pan-IE word for ’fish’? From what I understand ‘fish’ is only safely attested in Romance, Celtic and Germanic. ‘Lax’ (Swedish for Salmon) is attested in Germanic, Baltic and (east and west) Slavic. Some claim that both words can be reconstructed in PIE. If so ‘lax’ can’t have meant Salmon since there are no such in Anatolia, the Pontic steppe, Indus valley, Hindukush, Central Asia or wherever you might find some Urheimat.

    Russian ‘ryba’ for fish has cognates in other Slavic dialects. Though my Russian is a bit rusty I figure that Vasmer doesn’t give any clear etymology but mentions suggestions from others that it could be connected to both, ‘burbot’, ‘toad’ and – ‘blackbird(!)’.

    So the argument that there ought to be a common word for fish in a linguistic family at least doesn’t look fool proof. Compare if you like with Piotr’s discussion on his Language Evolution blog about the IE word(s) for ‘water’ – the habitat of the fish.

  16. The problem with the “salmon” debate is the same as with the “beech” debate – even if the original meaning was “salmon”, the Name may have been transferred to other species by the languages ina Areas where there were no salmons; but, as you say, it may also mean that the original meaning designated some other type of fish, and the meaning “salmon” was an innovation in Europe.

    On fish: this is is what Mallory & Adams have to say:

    The reconstructed vocabulary pertaining to fish in Proto-Indo-European is quite small, and even when words are reconstructable, the precise meaning may be quite ambiguous. It is an area of the Indo-European vocabulary where Asian cognates are so few that one cannot even reconstruct a generic word for ‘fish’ that meets our full requirements of Proto-Indo-European. The general word for ‘fish’ with the widest potential distribution is *pikˆskˆos ‘fish’ with cognates in Celtic (e.g. OIr īasc), Lat piscis, Germanic (e.g. NE fish), and Skt picchā– ‘calf of the leg’. The Indic is semantically far removed but is commonly justifed on the widespread folk association of the calf of the leg with the belly of a fish filled with roe. The word is generally derived from *pikˆ-skˆo- ‘spotted’ or the like, a derivative of *peikˆ- ‘paint, mark’, and the original referent is taken to be the ‘trout’ which, given its ubiquity across Eurasia, developed into the more general meaning of ‘fish’. …

    In the West Central region we have a generic word for ‘fish’, *dhgˆhuhx-, in Baltic (e.g. Lith žuvìs), Grk ikhthûs, and Arm jukn which exhibits an archaic shape that suggests it may have been the word for ‘fish’ in Proto-Indo-European but was replaced by other words on the extremities of the Indo-European world.

  17. In the West Central region we have a generic word for ‘fish’, *dhgˆhuhx-, in Baltic (e.g. Lith žuvìs), Grk ikhthûs, and Arm jukn which exhibits an archaic shape that suggests it may have been the word for ‘fish’ in Proto-Indo-European but was replaced by other words on the extremities of the Indo-European world.

    Mongolian *ǯiɣa– (Khalkha: ʒagas) is likely related (perhaps by borrowing). Lithuanian *ǯuw-i– comes particularly close.

  18. I assume that “beast” would have included animals like crocodiles?

    Beast is of course a borrowing, and it originally meant ‘animal’, displacing OE dēor (now specialized to deer) in that meaning. The OED1 (1887) shows it in various 13-14C uses, wider and narrower, which I have merged here:

    Ne schule ȝe habben nan beast bute cat ane. (1200)

    Ȝef ani were vnwriȝen & beast feolle þer in. (1200)

    Hwet medschipe makeð þe, þu bittre balefule beast! (1210)

    Beastes þat dumbe neb habbeð. (1220)

    […] þæt bittre best makede to bersten. (1220)

    Þe nywe forest […] he […] astored yt wel myd bestys. (1297)

    Men ne dorste […] wylde best nyme noȝt / Hare ne wylde swyn. (1297)

    Þe nedder […] was mast wis of ani best. (1300)

    Þar sal yee find an ass beist. (1325)

    Lord þou madest / boþe foul and best. (1360)

    Axest not me quod I, wheþir þat [man] be a resonable best mortel. (1374)

    He that hath vndirstonding, acounte the noumbre of the beest. (1382)

    A wounde feers and worst is maad in to alle that hadden the carecte or marke [L. characteram] of the beest. (1384)

    And woneden in wildernesse · among wilde bestes. (1393)

  19. Middle English Dictionary: bę̄ste).

    See in particular examples like the following:

    The bee … is a litel schort beste wiþ many feet. (beste translates Lat. animal)

    The same source (John of Trevisa’s translation of Bartholomaeus Anglicus’s De Proprietatibus Rerum) regards fisshe and foules as bestes, too.

  20. @ David: While your link doesn’t work, I’ve seen such a map and agree. Maybe even the same map? – I try again.

    @John, Piotr: Thanks for the examples! The meaning seems to have ranged from General “animal” to a more limited subgroup, as in Lord þou madest / boþe foul and best (1360). That last usage could be interpeted as “mammal”, but I think that would be anachronistic.

    @hat: Thanks for the formatting!

  21. I should have omitted 1374 and 1382, which are Biblical references to the Antichrist. The OED interprets 1360 as meaning ‘quadruped, or animal popularly considered as such’, i.e. excluding both humans and birds; indeed, of my list only 1384 explicitly includes humans. I don’t understand 1200b: is feolle the preterite of fall, or does it mean ‘hide’, as in ModE beast-fell?

  22. Preterite of “fall”. A little googling reveals that it’s from the Ancrene Wisse:

    Forþi wes ihaten on godes laȝe þet put were iwriȝen eauer, ant ȝef ani were unwriȝen ant beast feolle þerin, he þe unwreah þe put hit schulde ȝelden.

    “Therefore was it ordered in God’s law that pits were to be covered always, and that if any were uncovered and a beast fell therein, he that uncovered the pit must make repayment.”

    Since a bird or bee falling into a pit could easily get out again, I’m going to claim this one for “non-human mammals” too.

  23. Since a bird or bee falling into a pit could easily get out again, I’m going to claim this one for “non-human mammals” too.

    I’d assume that its meaning here is even more narrow, something like “livestock” – animals that belong to someone who could demand recompense.

  24. Preterite of “fall”

    Preterite subjunctive, to be precise.

  25. a fascinating attempt to upend the usual narrative

    Isn’t it sort of accepted without much thought that any reconstructed proto-language might not have been a real language of a real population, but a representative snapshot of a subset of related languages, with their internally-evolved differences and unique borrowings, which gave raise to the languages known much later?

    Even if the AN indigenous language of Taiwan emerged from several waves of migration to the island, by people some of whom spoke languages more related to other migration waves, and some less, there still must have existed a subset of relatively closely related languages of migrants to Taiwan whose languages were ancestral to Austronesian languages, regardless of hypothetical early borrowings from the pre-existing or concurring linguistically unrelated peoples of Taiwan. In this sense, PAN predating its radiation within and out of Taiwan exists. The only argument would be about the extent the Formosan languages derived from additional ancestors.

    Fox millet was domesticated on the mainland some 10,000 years ago, and started to dominate Taiwan economy island-wide by 4,000 years bp (when the island culture shows other parallels with Longshan culture of the mainland). Recent genetic data shows that Liangdao Man (8000 ypb, Liang island right off the coast of Fuzhou)) was substantially different from the extant mainland people, and intermediate between them and the indigenous Taiwanese (and ancestral to a large subset of the Austronesians from Madagascar to Bismarck Archipelago). It looks like Liangdao descendants radiated in Northern Taiwan 5 to 6 thousand years ago, and in Southern Taiwan, by 4,000 years ago.

    It just doesn’t look like archaeology, genetics, and linguistics are that much out of sync here. Other things (beyond the simple passage of time) might have influenced Formosan languages, but it’s inconceivable that there weren’t ancestors of Liangdao on the mainland, and that the antecedent languages of today’s AN weren’t spoken among them.

  26. Dmitry: What Blench is disputing is not the claim that the Austronesian languages are related, but the specific tree structure with what-we-call-PAN at the root, the Formosan languages in N branches below it, and Proto-Malayo-Polynesian as either the N+1th branch or an offshoot of the Eastern Formosan branch (which is spoken in multiple coastal regions, not just in the east). Rather, he claims that what-we-call-PMP is the root and that the Formosan languages are aberrant descendants of it, probably not all from the same branches, that have heavily influenced each other. In particular, the resemblance of Eastern Formosan (Blench speaks only of Amis) to PMP reflects the fact that it is a more recent arrival that wasn’t as affected by the Formosan Sprachbund.

  27. The more I think of this paper, the less impressed I am. I’m looking at the ‘fish’ words. From the Austronesian Basic Vocabulary Database, here’s a fairly complete view of generic ‘fish’ words in Austronesian, including Formosan (and in some other phyla too). Even though reflexes of PMP *hikan predominate in MP (i.e. non-Formosan), they are by no means universal. In some places the non-*hikan words may be loans from Papuan. In other places they might not be. Likewise, the Formosan words could be borrowings or substrate words; that still is not enough to exclude the standard tree model as a representation of the majority of the salient characters of the Austronesian languages, just as borrowings and shift do not exclude Yiddish from the Germanic tree.

    In the more in-depth Austronesian Comparative Dictionary, one of the Formosan words, reconstructed as PAn *Ciqaw has Formosan reflexes glossed ‘kind of tasty river fish’ and ‘fish (generic)’, with apparent cognates in MP glossed ‘goatfish’, a saltwater species. This makes me think that first, some of the Formosan words may have more semantically restricted cognates in MP languages; and second, that Formosa, with its large landmass, has large areas with many familiar freshwater fishes, and relying little on ocean fishes. If the founding MP population was coastal/islander, they would not have retained Formosan freshwater fish words, except for whatever had been adopted to fit saltwater species.

    A proper study of generic ‘fish’ words would have to compare generic words as well as names for individual species, of which there are many, in the Austronesian languages, of which there are also many. This is a very large undertaking, but is the more natural step to take before one claims that the Formosan abundance of words reflects an abundance of substrates.

    Overall, I think Blench would like to upend the current model based on very little linguistic work. That’s fine with a little-explored language family. In an enormously well-studied field like Austronesian, you have to do more than hand-wave away the great deal of detailed work which had been used to justify existing models.

  28. not the claim that the Austronesian languages are related, but the specific tree structure

    The way I see, the claim is much wider. Not just that there may have been a Sprachbund phase in the history of Formosan languages which could have made the appear closer, to the modern observer, than they once were (prompting a suggestion that the structure of that part of the AN tree may be somewhat deeper, and the reconstructed protolanguages, somewhat suffering from an error of inclusion).

    The claim is also that there was no migration of early AN languages from mainland China to Taiwan, but rather that the languages must have emerged in situ, from migrants of much earlier epochs. But given that the spread of cultural factors from the mainland, and spread and radiation of the genetic material from the mainland coast, and radiation of the subsequent languages, are well documented and appear to be contemporaneous and interrelated, I would think that it must take much more to dislodge the “usual narrative”.

  29. It’s precisely the radiation of the subsequent languages (assuming you mean “from Taiwan” rather than “to Taiwan”) that is in doubt if the claim of the Formosan languages to be basal is set aside. If in fact the Formosan languages are descendants, however aberrant and converged, of languages once spoken elsewhere in the Austronesian world, then the claim that all Austronesian settlement had to pass through Taiwan falls to the ground, and the Urheimat of the Austronesian (Malayo-Polynesian) languages could be in Borneo or the Philippines or who knows where.

  30. David Marjanović says:

    Maybe even the same map? – I try again.

    Definitely the same map! 🙂

    The text mentions that corn means “wheat” in southern England and “oats” in Scotland and Ireland. I wonder if that’s still current.

    Isn’t it sort of accepted without much thought that any reconstructed proto-language might not have been a real language of a real population, but a representative snapshot of a subset of related languages, with their internally-evolved differences and unique borrowings, which gave raise to the languages known much later?

    Apparently historical linguists differ a lot on this question. Ideally, a reconstructed protolanguage would be a real dialect (perhaps the only one, perhaps not) of a real language of a real population; sometimes, however, attempts to reconstruct such things run into problems biologists call “the species problem” and, more seriously, “incomplete lineage sorting”. I’m too tired to explain them right now 🙂

  31. Is there any safe pan-IE word for ’fish’? (…) So the argument that there ought to be a common word for fish in a linguistic family at least doesn’t look fool proof.

    That’s not quite the argument advanced. We have no particular reason to think that PIE speakers were predominantly fishermen, so there is no particular reason to expect a large body of reconstructible fishing terminology to turn up. On the other hand, we can reconstruct a decent amount of terminology referring to agriculture and animal husbandry, and hence we will suppose these to have been important in the PIE society.

    So it’s reconstructing fishing as an important sustenance method of Proto-Austronesians (if considered distinct from Proto-Malayo-Polynesians) which is going to look suspicious if supposed primary branches commonly fail to reflect original fishing terminology. Of course though, the criterion is “reflect in some form”, not “reflect unchanged”… and Blench’s survey does look a bit superficial in this respect. Semantic drift happens.

    As for how realistic reconstructed protolanguages are: we usually believe that they are decent approximations in phonology (whose development is relatively easy to sketch a chronology for) and grammar (slightly more difficult), but determining the age of lexical material can be more difficult. The kind of arguments that are usually advanced against e.g. dating words like ‘coffee’ to Vulgar Latin often stop working quite as well when the time difference between divergence and acquisition is shallower (closer to half a millennium than two), when the dialect continuum is less differentiated (allowing for etymological nativization between varieties), when the words have had a long time to organically evolve afterwards (which may end up disguising earlier variation), or when there is no outgroup or loangiver evidence for establishing if a word is an innovation or not in the first place.

  32. the Urheimat of the Austronesian (Malayo-Polynesian) languages could be in Borneo or the Philippines or who knows where

    to consider this hypothesis of later arrival of Austronesian languages to Taiwan, we would need to think of Sparchbund-like qualities of Taiwan in a very specific and unusual way. It is not inconceivable for a language to be so heavily influenced by its neighbors that it appears to have split off its ancestral group earlier than its actually happened. In most cases one should be able to pinpoint the actual sources of borrowings, grammatical influences, etc., and to make a correction. One can’t do it with Formosan languages, therefore one have to posit the existence of extinct (but once influential) language groups there. Sparchbund settings are generally expected to exert a unifying influences on the languages (and it is also known as “linguistic convergence are”), but Formosan languages are highly dissimilar so we would need to consider multiple Sprachbund-like melting pots of languages in multiple localities, involving multiple extinct language families. The conjecture becomes the more convoluted the more I think about it. Yet if we stick to the conventional viewpoint that Formosan languages have really ancient roots, perhaps overlayed by later-period borrowing which made them somewhat more similar (rather than dissimilar) to each other, then both their diversity and the pitfalls of proto-language reconstructions may be easy to explain. Occam’s razor, anyone?

    Meanwhile, archaeological characteristics of the earliest Neolithic sites of Luzon (such as pottery and polished adze types) are shared with, and postdate, Taiwan Neolithic. The accumulation of shared genetic innovation also places Taiwan at the root of the migration to the Philippines and beyond, and the timing is the same. So to hypothesize that the Austronesian languages reached Taiwan later, we would have to posit waves of migration in the opposite directions (people and technologies heading South, language heading North). Again it sounds too convoluted not to think of Occam’s razor?

    Sometimes, of course, the complicated explanations turn out to be true, but IMVHO it would take much more data before we may seriously consider it…

  33. Stefan Holm says:

    j:

    One neglected point is the economy. For the very long period when we were hunter-gatherers we lived in small, isolated communities. The number of humans on earth actually didn’t increase at all (except for geographic expansion). The ten thousand years or so old invention of agriculture allowed the population to slowly increase (farming can basically feed 30-50 times as many people on a given area as hunting-gathering). We thus started to live in bigger, more stable (resident) communities. The 200 years old industrial revolution finally made the number of us explode.

    The IE common vocabulary is, when it comes to economy, all post-agriculture including spoked wheel wagons etc. It must have emerged in a relatively stable and large society (or cluster of societies).

    The situation is different for hunter-gatherers. Depending on their habitat they must have been more diverse in culture. Hunting game in Africa, following reindeer herds around the polar circle, collecting fruits, berries, nuts and edible roots in a tropical forest, catching fish or seals along a sea or lake coast etc. are far more ‘multi-cultural’ than farming. And archeology tells us that we have always been opportunists, i.e. have eaten whatever there has been around to eat – and survived! (Something which the myriad of modern dietist experts might reflect upon).

    So, I think that (when it comes to vocabulary) among pre-historic people one should search for possible cognates among concepts which are common to all humans, independent of economic system: words for sun, earth, star, day, night, mother, father, hair, eye, hand, eat/food and so forth.

    That’s why I think ‘fish’ isn’t valid but ‘day’, ‘wheel’ and ‘cow’ are.

  34. Stefan Holm says:

    I wrote: I think ‘fish’ isn’t valid but ‘day’, ‘wheel’ and ‘cow’ are..

    Sorry – within the actual economic concept of course

  35. This brings to mind the recent papers from Peter Norquest and Sean Downey, who argue for several new Proto-Malayo-Polynesian phonemes. One of them is equal to PAn *C, which had previously been assumed to have merged with *t everywhere in MP. Also, since most of their new distinctions fail to be found anywhere in Formosan, this means that the status of Malayo-Polynesian as a distinct subgroup is not on a very good standing after all. They propose that the Austronesian homeland could well have rather been located in Borneo, with Formosan representing a divergent sub-branch.

  36. buran ramai says:

    “Fish” in Formosan call “vulaw” / “vulau” / “kuraw” / “pulaw” / “alaw” / “aelaw” ==> long word.
    “Fish” in Tai-Kadai call “pla” / “hla” ==> short word.
    “Fish” in Melayu and Indonesian call other name “ikan” ,BUT there is another word “pulau” or “Island”.
    “pu” + “laut” ==> “pulau” ,of which “pu” = arising and “laut” = sea in Melayu and Indonesian languages.
    “Fish” and “Island” share the common meaning ==> “something arising from the water (sea)”.
    Therefore, from the oldest to youngest, “pu” + “laut” => “pulau” == migration and split into => “vulaw” and “pla”.

  37. marie-lucie says:

    I must have missed most of this discussion, which started about a year ago.

    Skt picchā- ‘calf of the leg’. The Indic is semantically far removed but is commonly justifed on the widespread folk association of the calf of the leg with the belly of a fish filled with roe.

    Here is another example from the North Pacific coast: hoon ‘fish (generic, esp. salmon)’, derivative hoon’ix ‘calf of leg’ (oo = ‘open (low) o’, ix as German ‘ich’). The “association” is based on similarity of shape, that’s why it seems to be found in many cultures.

    calf : Is the ambiguity of “calf” based on some similarity, or is it just a coincidence due to phonological convergence?

  38. It’s uncertain, but the notion that the animal is so named by transference from the swollen part (the uterus) to what causes it to swell (the fetus) to the newborn animal is not a priori unreasonable. Certainly the two words have been the same since Proto-Germanic, so there is no evidence of any phonological convergence, unlike cleave ‘stick’ and ‘divide’, which were distinct as recently as Old English.

Speak Your Mind

*