Of pioneers, pilgrims and parsnips

In today’s Australian, Paul Frijters and Tony Beatton take a swipe at my ‘double standards’ in talking about ethnic diversity. Andrew Norton is kind enough to defend me (doing a far better job than I would have done myself). In principle, I agree with Frijters and Beatton’s conclusion:

What matters is that social scientists think critically about statements in support of truisms, just as they do about statements at odds with contemporary social norms.

But I’m not sure they live up to their own standards. In the opening paragraph, they say:

One of the most important contributions a scientist can make is to successfully question opinions that seem self-evident and obvious to the public. Once it was commonly accepted in the West that the world was flat and that the heart was the residence of the soul.

Where’s the evidence that the soul does not reside in the heart?

This entry was posted in Universities. Bookmark the permalink.

43 Responses to Of pioneers, pilgrims and parsnips

  1. Leon says:

    I especially love this paragraph:

    “The best guess of anthropology is that hunter-gatherers such as the Aborigines had a life expectancy of about 30 years. Average expectancy in Aboriginal surveys now is close to 60 years. On the face of it, that’s a doubling of the length of life, related to Western medicine (vaccinations and the like) and Western-style public services (such as pure drinking water, relatively safe transport, and basic housing).”

    Apparently, results suggesting a 17-year difference in life expectancy are skewed, as many are self-identifying as Aboriginies without having descended from them.

    However, for those among the cohort who actually have Aboriginal blood and are dying young, the “thrift gene” idea – usually hypothesised to explain high levels of type 2 diabetes and obesity – is tacitly implied. In considering this huge difference in lifespan, they suggest that there is an obvious positive which all the PC patrollers have missed: “we” should be most proud of the fact that our Western culture has in fact doubled Aboriginals’ longevity.

  2. John Quiggin says:

    And the claim about received belief in the earth being flat is untrue for any period in which the term “West” was remotely meaningful. The sphericity of the earth has been known since classical times, and claims to the contrary are 19th century fabrications, a point that’s been made many times.

    Pot, meet kettle.

  3. Peter Whiteford says:

    They also say “Why don’t we bemoan also the loss of diversity within Anglo-Saxon heritage? Now you may have only one box on the census for speaking English, yet the whole notion of Englishness is a medieval invention and hides the fact that until 300 or 400 years ago this was a far from homogenous identity. The War of the Roses testifies to that. The cities of what we now call England were thick with linguistic and cultural diversity.”

    As far as I recall, the War of the Roses was mainly led – from opposing sides – by first, second and third cousins, who all in different ways claimed descent from Edward III. While there were undoubtedly deep structural issues underlying the conflict – France had similar long-running conflicts, for example, I’ve never seen any historic account of the War of the Roses that relates the conflict to linguistic or cultural diversity.

    Moreover, the most recent research shows that most of the British are not actually Anglo-Saxons. See http://www.prospect-magazine.co.uk/article_details.php?id=7817

  4. Paul Frijters says:

    nice to see so many people trying to show how wrong me and Tony were about several statement. Quite probable, but lets see. I think Leon is being supportive, so thanks. John is overlooking the fact that we say ‘Once in the West’. We indeed use the word West for regions that did not always used to be called the West but dont apologise for that since this is a common use of language: one would say ‘the world 1 billion years ago’ even though there was no such word around at that time. ‘Once’ would for instance include antiquity where, if my high school memory serves me right, the world is not usually depicted as a round ball. I know that amongst the elites in many Western societies more enlightened views held sway, but the word ‘commonly’ means the majority of the population. I can dig up other ‘once’ if John’s serious though.
    Peter makes the obvious point that leaders in revolts close to each other are related, just as the monarchs in the first world war were close relatives. Just as in the first world war, closely related leaders do not mean unity. The point was thus that there was plenty of internal disunity in the UK and the war of the roses testifies to it.
    Peter’s point that much of the genetic heritage in the UK is not particularly Angel or Saxon probably relates to the recent suggestions that the male genes derive from the Angels (+ vikings) and the Saxons whereas the female genes are more kelt than once thought? True, but this not only underscores how diverse the UK really is, but also misses the obvious point that people can call themselves Anglo-Saxon without being genetically descendent from either Angels or Saxon. Such things hold for many regions with high population turnover where newcomers adopt the name of the incumbents.

    Andrew, any sign of that baby yet?

  5. Peter Whiteford says:


    Just to be a pedant – my original training was in History – my point is that while there was a lot of disunity behind the Wars of the Roses – self evidently – I’m not sure that there is evidence that the disunity was related to any sort of linguistic or cultural diversity. One commentator has noted : “It was precisely because England’s government was so centralized and the crown so powerful that nobles were so concerned about who controlled it.”

  6. I was struck by Tony’s thesis topic – imprecision in social science data. Sounds like endless fun – criticising other researchers!

    I was about to make a somewhat learned comment about the distinction between ethnicity, nationalism, patriotism and the nation state, but in the spirit of the original article, I’ll be a pedant instead.

    The point was thus that there was plenty of internal disunity in the UK and the war of the roses testifies to it.

    My understanding, Paul, was that the UK came into being with the abolition of the Scottish Parliament and the union of the crowns of England and Scotland in 1707.

    Which is interesting not just for historical reasons, but for political sociologists because the 300th anniversary of this event looks like it might coincide with an SNP electoral victory in Scotland.

    Anyway, if we were to have a serious discussion about ethnicity, language, nationalism, history and sociobiology perhaps we could start from an accurate premise.

  7. Paul Frijters says:

    The original article was never about 100% accuracy and explicitly says that total accuracy is not usually possible in social science.
    What we call for is equal effort in thinking through whether we as social scientists agree with common opinion, both when that common opinion is for or against something. We also make the case that its in a sense the responsibility of social scientists towards their society to stand up and be counted when they see other social scientists ‘going along with the stream’ without thinking critically whether that stream make sense or not but by their actions perpetuating false beliefs. If we’re not going to make a stand for the best guest our science can come up with, then who else will? It is my belief that all the commenters above share this opinion.

    As to the two comments above: we commonly say ‘Australia 500 years ago’ even though what we now call Australia was not called so in that period. By the same token, we can say the UK or the West X hundred years ago even when these regions were not then called thus. And the claim about the War of the Roses stands: in that time it mattered which region you belonged to and there would have been a strong identification, and now it doesnt matter anymore.

    Tony’s thesis indeed should be a lot of fun. It will be about the inherent difficulty in social science to measure any variable with accuracy and hence the inevitability of having to live with rough guesses.

  8. Claire says:

    Let me be pedantic too – simple average life expectancy is very misleading among hunter-gatherer communities because infant mortality is usually really high. It’s not the case that “Aboriginal people are living twice as long” – kids who survive to five or so tend to make it to adulthood. There tends to be another demographic dip around 40, and people who get past that tend to get to an average of 70 or so.

  9. Leon says:

    Sorry Paul, I wasn’t being supportive, although I think the tone of my post should have been less facetious and more constructive.

    The reason I disagree is because it seems to me that a 17-year drop in average lifespan when considering any self-identified group seems like great evidence for some kind of problem.

    Also, it seemed to me that the two attempts at routing the PC police (of whom I am not a huge fan, mind you) were a bit contradictory: firstly, suggesting the respondants’ gene pool was diluted by non-Aboriginal genes, then suggesting that because of a preexisting genetic disposition, things aren’t as bad as they seem.

    Also, I’m skeptical of appeals to a thrift gene because they seem like a way of presuming the insolubility of these problems – in my view a counter-productive way of thinking.

    Again, apologies for the attitude.

  10. parkos says:

    AL, from your original article:

    “SINCE the time of European settlement, Australia has been shaped by immigration. Successive waves of newcomers from Europe, the Americas, Asia, the Pacific and the Middle East have enriched Australia in many ways. From a purely economic standpoint, immigration supplements our labour market with much-needed skills.”

    Whilst I appreciate you are trying to generate discussion or ways of thinking about issues regarding contemporary multiculturalism, I think you make the basic error that most Australians and people who think about Australia make.

    Migration to Australia happened in recent centuries before European settlement, and made political, technological and cultural changes. Whilst browsing a copy of a Singhalese newspaper at a Sri Lankan lunch bar, it was brought to my attention just how many place Australian aboriginal names have a direct equivelant in Southern India. More than the 50 named I would presume. For example, there is a Karratha in Western Australia facing the Indian Ocean and one in India.
    I could go further into Anthropology textbooks and research from Europe which mention and depict that Indigenous people in WA look more typical of other peoples in the Indian Ocean region than their equivelants on the Pacific side, but I will leave it there.

    Furthermore, I dont believe the three books I perused for a few hours yesterday which were reputed to be the oldest books written in Australia (from 1790 worth $45,000), are indeed the the first manuscripts written on this continent.

  11. Claire says:

    parkos, with the greatest respect, that placename stuff you quote is bullshit. Singhala is Indo-European. It was trendy to relate Australian languages to Dravidian (e.g. Tamil, Mayalam, etc) in the 19th Century, but it’s completely unsupported.

    Also, there’s been a great deal of research done on both micro-evolutionary variation (the old biological anthropology topics) and DNA markers in Aboriginal communities, and the current research implies that there was a single Australian migration. The variation in Aboriginal groups that’s seen today is a combination of micro-acclimatisation (e.g. northerners have darker skin than southerners, even leaving intermarriage aside) and recent intermarriage/rape (e.g. between Papuan groups in Arnhem land and CYP, rape and intermarriage by/with Europeans in the SouthEast and Queensland, and with Asians in the pearling areas, like NW WA).

    If you want some reputable references email me.

  12. parkos says:

    Claire, it was a Singha newspaper, it did not say that each of the names were Singhalese. There are pre-Tamil/Singha people in Ceylon.

    Placenames are what they go by in Europe and America, why should Australia be any different?

    It was not exactly cut off by sea as you have pointed out, Melville Island has been an area of interaction with Indonesia long before Europeans arrived. There are many shared words in the languages. Europeans did not invent the boat as you seem to be implying.

    To dismiss centuries of European anthropolgy as “trendy” & “bullshit” shows the depth of your approach.

    Whilst I will not deny you some trend towards genetic unity amongst indigenes, I would point out that Peter Kershaw, palynologist, has found evidence of regular burnings amongst pollen samples from 120,000 years ago in Australia, which is presented as evidence of human settlement longer than 40,000 years ago.

    All humans DNA is mixed,

  13. parkos says:

    My point being that a bunch of emancipists sitting around discussing how Indians should have to jump through hoops to migrate to Australia is not really justifiable and as the world’s population grows will it be enforcable?

  14. John Quiggin says:

    Paul, you may want to read Wikipedia’s take. The sphericity of the earth was discovered around the third century BC, at which time most of what we now call “the West” was occupied by Bronze Age Celtic tribes, whose opinions on the matter are unknown. At any time when there has existed anything that could reasonably be called Western civilisation, educated people have known that the earth is round.

    The point being made is that if you want to demand exact accuracy from people’s opening and concluding remarks, rather than repetition of conventional wisdom, you ought to adhere to this yourself, rather than relying on memories of what you heard in high school.

    Alternatively, and more reasonably, you ought to extend to others the charitable reading you naturally want for your own writing.

  15. Claire says:

    Sorry Andrew, I’m temporarily hijacking your blog to talk about prehistory 🙂

    parkos, I didn’t mean that toponymy research is bullshit, simply that similarities between names in Sri Lanka and Australia don’t a priori tell us anything about prehistoric migration.

    “Placenames are what they go by in Europe and America, why should Australia be any different?”

    Three reasons – one is that placenames aren’t by any means the only thing that “they” go by when reconstructing prehistory. It’s only one of many pieces of evidence, and isn’t usually the primary one. Systematic sound correspondences is taken as primary, for example. What you conclude from place names alone is often quite dodgy. My favourite example of this is upstate New York. Imagine you have no records of the settlement patterns in the area and 1000 years in the future have to work out who was here just by the place names.

    Secondly, Australia’s different because preferences for toponym structure are to a certain extent cultural. Place naming conventions vary by language/culture. There are a lot of English place names named after people (Sydney, Melbourne, Fitzroy River, Boston, Wellington, etc), prominent features in the landscape (Table Mountain, the Blue Mountains, Black Mountain, etc), or permanent settlement structures (Warwick, Greenwich, Norwich, Tottenham, Whitby, etc). Australian Aboriginal place names differ a lot depending on the part of the country, but they are often named for culture hero actions (e.g. Oodnadatta “he had a shit here”) or they’re unanalysable. “The land is a map” (Pandanus Press, c2003) has a lot of information about this and “Placenames old and new” will be out in March (I have a paper in the second book on the structure of Bardi toponymy). Also, you have to factor in that a lot of Aboriginal names are not recorded with their original sounds – there were a lot of mistranscriptions and misunderstandings about what the scope of the name was. “Canberra” is a good example of that. The best guess seems to be that it was originally ngaparri (with ng like in “singer”).

    Finally, sometimes groups will give their own names to places, and sometimes they’ll use the names from the people there already. Aboriginal groups lean towards the latter, whereas English speakers lean towards the former.

    It was not exactly cut off by sea as you have pointed out, Melville Island has been an area of interaction with Indonesia long before Europeans arrived. There are many shared words in the languages. Europeans did not invent the boat as you seem to be implying.

    Where did I say that? I’m not denying loan words or a lack of isolation, I was arguing against your claim of Indian settlement in NW Australia. It’s not just Melville Is, most of the coastal Northern Territory was part of the greater Makassar Emirate at one stage, and there are over 800 loans from Makassarese and Buginese into Yolngu Matha. But they were short visits for trade, not immigration. And Makassar interaction on the Arnhem coast doesn’t tell you anything about possible Tamil interactions in the NW Kimberley.

    To dismiss centuries of European anthropolgy as “trendy” & “bullshit” shows the depth of your approach.

    Your words, not mine. My claim of bullshit was restricted to your specific assertion that place names in Sri Lanka provided evidence for settlement in NW Australia pre European settlement. It’s precisely because my approach to linguistic prehistory is deeper than superficial phonetic similarities between two items whose meaning is uncertain that I’m calling your claim bullshit.

    Whilst I will not deny you some trend towards genetic unity amongst indigenes, I would point out that Peter Kershaw, palynologist, has found evidence of regular burnings amongst pollen samples from 120,000 years ago in Australia, which is presented as evidence of human settlement longer than 40,000 years ago.

    Yes, but a) it’s highly debatable that it was caused by humans; b) there’s no evidence that if it is evidence for humans, they are the immediate ancestors of today’s Indigenous Australians, and c) the 40,000 year date has been bunk since we got beyond carbon dating.

  16. conrad says:

    Sorry to continue the hijacking, I think the example about place names is poor. I’m not sure about the Aboriginal debate, but phonetic differences in place names can tell you a lot, especially about Aboriginal/other language interactions. Its easy to see why

    1) All languages have an inventory of sounds. Some Aboringinal languages have heaps of consonants.
    2) If a new group turns up, then the place names from the old language need to somehow represent the place names of the new language (THis isparticularly so if the new language guys get later assimilated into the old language guys)
    3) Because the new names contain only a subset of the sounds of the old language (probably consonants if you are thinking about Aboriginal languages), the place names will tend to be represented in the old language using only a subset of sounds that are available. THis means some sounds will get used more than others, simply because they overlap the new language more. In terms of Aboriginal languages, lots of consonants won’t be used.
    4) This will cause the distribution of names in the old language to change depending on how much influence the new language has. Thus comparing the distirbution of sounds in place names vs. the average of the language gives us some indication of the interaction between the new language and the old.

    Lets take the New Yor example. If the French had done a decent job, then lots more places would have French-like English assimilated names. Hence sounds that exist in English but not French would not be represented as much as in places where the French didn’t get. Thus, we wouldn’t find as many names with clusters like -str (that don’t exist in French) in New York vs. places the French hadn’t been, because many of the names would have been assimilated and no assimilated names will contain the offending sound (-str). Thus the sounds of place names can tell us something, even if there is a lot of variance in the data.

  17. Claire says:

    1) Actually, English and Walmajarri (for example) have roughly the same number of consonants. Walmajarri has more places of articulation but English has more manner contrasts. Walmajarri has 3 vowels and English has about 14, though, depending on how you count.

    2) presupposes that the new group doesn’t just use their own new names. Upstate NY has more names of the “Ithaca” type than the “Irondequoit” type. And it presupposes enough language shift to use new place names, but not enough to learn new phonemes. All sorts of things happen in loan phonology.

    3) I don’t follow. Loan phonology mapping is complex. For example, “Aboriginal” palatal stops are realised in place names as s (as in Suggin Buggin), ch and j (Jindabyne); /ng/ as “c”/”k” (Canberra) or “n” (Nockatunga, Ngun(n)awal).

    4) doesn’t follow because it proposes too many arbitrary stages that aren’t reflected in language contact more generally, particularly in a place like Australia where the language names of Indigenous origin don’t always belong to the traditional language of the area.

    Your conclusion actually illustrates my point well. Language contact outcomes are incredibly diverse, and depend on all sorts of factors. Place names are only one piece of evidence.

  18. conrad says:

    Its no doubt language contact produces diverse outcomes, and its no doubt that languages may use their own place names after a while. Hovever, there are statistical regularities in the way loan words are assimlated (well known and well understood) and statistical regularities in the way languages change due to language contact (well described but not so well understood). If that wasn’t the case, every instance of language contact would produce something completley new, which of course never happens.

    Here’s a few examples of predictable and analyzable behavior. None of them are be-all and end-all evidence of two groups living in the same place, but they certainly are evidence:

    – morphemes get changed (e.g., New York vs. Nouvel York).
    – Sounds get added to clusters to break them up (typically schwa)
    – Sounds get replaced with others (typically those closest too them) in some sort of closest-to-closest mapping
    – Various constructions can get borrowed

    These are all predictable to some degree and hence if group A turns up, influences the language of group B, then gets displaced (possibly leaving no genetic evidence), they can still leave an influence on the language.

    Here’s an example not from places, but illustrates my point. A wide range of fruits only became available in places like Hong Kong in relatively recent time (mainly from English). These were assimilated multiple ways, some of which don’t represent the native language and some which do. For instance, strawberry is extremely hard to say so people us si-do-bo-lei. This we word can recognize from the orignial, but a lot of the time things get changed enough that its not possible (the origin of the word fei is unknown, and people guess it comes from the English “fare”, because it is the easiest way to represent it, but no-one is really sure). However, even if the sounds changed enough that we didn’t know it came from English (and this is common), it would still be possible to guess it was once a loan word because (a) 4 syllable words in Cantonese are very rare (b) most Cantonese words have related morphemes; (c) the name itself is atypical for fruits. The same is true for the word “plum” = bou lum. If we didn’t know this was an assimilated loan we would still guess it is. If I looked through all the fruit names, I would find lots of these sorts of examples, and I could conclude a fair few English people had gone to Hong Kong or traded with them etc.

    Here’s an example of a productive idiom which went from Cantonese to English. Most people are not aware of its origin but use it. This of course represents one end of the spectrum. You have “Hou loi mou gin = long time no see” in Cantonese and people use a word-for-word translation in English. However, this has become productive in many places in the English speaking world (Long time no eat, long time no X), yet it is only grammatical because of its idiomatic status. Thus if I was digging around databases I would recognize it as something which English has loaned, and if found more examples of similar patterns I could conclude with some probability that English people had met Cantonese ones.

    Here’s an example of Australian speech, which probably comes from Cantonese particles. Australians are one of the few English speaking groups that use “only” at the end of sentences. “Ten dollars ONLY”. This is of course extremely common in most languages that use sentence final particles a lot. If I wondered about the origin of this, I could guess that it is from all those Cantonese people that lived in Australia for the last 200 years (the particle is -zaa). Whether this is true or not I don’t know, but I could follow it up looking for similar patterns. Again, this would tell me that some sort of contact may have occured.

  19. Yobbo says:

    Claire, bullshit is Parkos’ specialty.

  20. Damien Eldridge says:

    I thought there were only five vowels in English (a,e,i,o and u). What are the other nine?

  21. Paul Frijters says:

    I’m with Claire on the Aboriginal issue. There’s basically no evidence for habitation before 40,000 years and its mightily unlikely that 80,00 years of prior habitation would have left no skeletons. Also, as far as my memory of the Antropology textbook allows, I thought the only remote human groups related very distantly to the Aboriginals were some small group living on the Andaman islands, but I could well be wrong in this.

    John, I dug up another ‘once it was commonly believed that the earth was flat’ for you since you persist to pretend to be an historian and ‘know’ that there was never a time when it wasnt commonly believed that the world was round. It relates to the story of Columbus and I got it from Wikipedia as you suggested:

    Following Washington Irving’s myth-filled 1828 biography of Columbus, it became common supposed knowledge that Columbus had difficulty obtaining support for his plan because Europeans believed that the earth was flat.[2] In fact, few people at the time of Columbus’s voyage, and virtually no sailors or navigators, believed this.[citation needed] Most agreed that the earth was a sphere. This had been the general opinion of ancient Greek science, and continued as the standard opinion (for example of Bede in The Reckoning of Time) until Isidore of Seville misread the classical authors and stated the earth was flat, inventing the T and O map concept. This view was very influential”

    Hence there was a ‘once commonly believed’ and indeed that view was challenged and eventually driven out by scientists, which is exactly what my original article claimed.

  22. parkos says:

    Thanks for your take on the infobahn Claire.

    Well Paul “Vlams Friters” (Flemish hot chips with mayo),
    you come to Australia and instantly you know more about settlement patterns and archaeological evidence than Kershaw, the head of the top rated geography and enviro department in Australia. You have a brilliant mind indeed.

    Yobbo, stick to physical combat and dont back down when the cops move you on big boy.

  23. Claire says:

    Paul, that’s not what I said. The fact that most dates (until fairly recently) were arrived at by carbon dating, and carbon dating has a horizon of about 36,000 years, and low and beyond most of the earliest dates clustered around that date, should set off alarm bells. There are thermo-luminescence dates for earlier. When I teach my Australian languages course, I usually use the c. 65,000 dates as examples of the oldest probably reliable dates for settlement.

    Conrad, I’m not sure what your field is, but I’m guessing it’s not linguistic prehistory, right?
    “Hovever, there are statistical regularities in the way loan words are assimlated (well known and well understood) and statistical regularities in the way languages change due to language contact (well described but not so well understood).”

    really? This is my field and I don’t think they’re really well understood at all beyond a catalogue of the possible outcomes. It’s a start but we aren’t anywhere near close to a predictive model that nuanced enough to use where we don’t have other data. For example, you mention schwa insertion in clusters. Another outcome is cluster reduction. No one knows why you sometimes get one, sometimes the other, and sometimes both. Which one isn’t predictable, and cluster reduction (through either mechanism) also occurs as a regular sound change without language contact. Therefore if you find a language which shows this change, it’s not possible to conclude that it was a contact variety (or shows substrate influence).

    The parts you’re vague on “some sort of mapping”, “morphemes get changed” are precisely where predictivity falls down and why using language to reconstruct prehistory is such a tricky business.

    You can’t conclude from your fruit example that speakers of Cantonese were in contact with speakers of English. They might have borrowed the word through a third language. Trade terms are especially diffusable. English “orange”, for example, is from Arabic, but through Spanish and then French. We’d guess it’s a loan from correspondences, but there are many steps between identifying the loan and tracing the contact conditions that led to the borrowing.

    Damien, there are five “letters”, but many more distinctive vowels.
    Ignore the spelling and say these words aloud (these are standard Australian English; the list is different for other dialects):
    bit, beat, bet, bait, bat, but, (ro)bot, boat, boot, put, Bert, bar, bite, beaut(y), bout.

  24. conrad says:


    First of all Parcos is just a nuisance, so listen to Yobbo. Back to the point, if we must be rude and patronizing to each other, then I guess like most language typologistics you probably don’t understand statistics and what can be learnt from them. Just because you might be stuck in a world of describing things in nice discrete symbolic terms, doesn’t mean that everyone else doesn’t use use statistics to understand language.

    There are there are numerous papers looking at what and what cannot be predicted from the statistics of language (although I never said it wasn’t tricky and no-one said it was the be all and end all of evidence for various things), like how the European languages evolved, although of course they’re controversial. Also no-one claimed it wasn’t tricky — it has of course generated lots of debate over time. A good example that generated a lot of controvesy to do with at the time to do with statistics and lanuguage is

    “Language Tree Divergence Times Support the Anatolian Theory of Indo-European Origin” Nature, 426, (November 27) Gray & Atkinson

    and I could put some more down looking at language contact in particular.

    What is reasonable well known, however, depsite your claim, is loan word assimilamation in some languages like my example, i.e., How English words get into Cantonese. There is a string of papers on this, and the better descriptsion start here: Yip, M. (1993). Cantonese Loanword Phonology and Optimality Theory. Journal of East Asian Linguistics 2.3, 261-293). If you think it isn’t a relatively predicatable process then you’re simply wrong.l

  25. Claire says:

    huh? fwiw, I don’t claim to be a statistician but I do use statistical methods in my work (e.g. in acoustic phonetics).I also have some familiarity with computational phylogeny, e.g. the sort of stuff that Ringe, Warnow and Taylor and Grey and Atkinson do. But garbage in, garbage out still applies. For example, Ringe, Warnow and Taylor’s model has as axiomatic that language change only occurs during acquisition, which isn’t true. That stuffs up their language contact model. For critiques of Grey and Atkinson, see some of the refs at languagelog.
    I’ve read Yip’s paper and there’s Steriade’s p-map, and so on. Perhaps we are claiming different things. You were pretty vague about the evidence you were citing “morphemes get changed, sounds get replaced, constructions get borrowed”. Even identifying a “same” versus “not same” in a test of lexicostatistics is non-trivial. You don’t have to be a statistophobe to feel that it’s a good idea to understand some of the basic properties of the data you’re trying to do statistics on. No?

  26. Sacha says:

    This discussion is very out of what I know, but I just thought I’d mention that Dawkins mentions in “The Ancestors’ Tale” that statistical techniques are often used to attempt to determine probabilistic branching trees in genetics and possibly (from memory) the development (including branchings) of languages.

  27. parkos says:

    conrad “First of all Parcos is just a nuisance, so listen to Yobbo.”

    Claire is all yours Conrad, she talk too much already my ears hurting, I am going to serve up chinee chilli noodle from under the Hills Hoist with a kookaburra loincloth and get me some poon action that way.

  28. conrad says:

    Obviously I’m not going to describe things like how a set of rules using an OT type paradigm works with parts of language I assume most people haven’t a clue about here.

    Also, you can like or loathe the statistical techniques for predicting language change as much as you want. In case you have some be all and end all type argument against entire branches of lexico-statistics versus individual models, I’m sure LI or Language would love to publish it, and I’d love to read it too for that matter (let alone just detecting that langauge contact has occured, which is what the origninal argument was about). Anyway, I think this is getting far off topic from anything that this blog is about.

  29. Claire says:

    off topic? I thought economics was about anything economists talked about.
    McMahon and McMahon “Language Classification by Numbers” (OUP) isn’t a terrible summary of the problems, and it’s got references to the standard critiques. This is not new.

  30. David Marjanović says:

    About that Language Log page I must say that the dating method used there is commonly used in molecular biology, and the method to get the shape of the tree (cladistics) is standard in biology as a whole. There is no requirement for a character to have only two states; DNA always has 5 states for every character (A, C, G, T, or gap). I don’t see why there should be a fundamental difference between phylogenetics in biology and phylogenetics in linguistics.

    That said, molecular dating (no matter by which method) is very sensitive to the choice and accuracy of the calibration points and a few other things. I’ve done it myself…

  31. Jesse says:

    “Where’s the evidence that the soul does not reside in the heart?”

    That people with artificial hearts are not soulless zombies?

  32. J Snyder says:

    “Where’s the evidence that the soul does not reside in the heart?”

    Heart transplant patients don’t resurrect the soul of the donor?

  33. Mike Rock says:

    Where’s the evidence “That people with artificial hearts are not soulless zombies?”


  34. Jesse says:

    On the other hand, if you eat someone’s heart still beating right out the socket, you DO gain all their powers and memories. So the evidence is inconclusive.

    Seriously, what’s odd about the quote is that the author seems to have confused a straightforwardly disprovable empirical statement with some sort of transcendent spiritual issue. This is goofy. If there is a soul that resides in a particular sector of the body — a question I express no opinion on here — that sector is obviously the brain, not the heart, as is shown by a moment’s contemplation of the key distinction between a heart transplant and a brain transplant, or between an artificial heart and an artificial brain.

  35. Jesse says:

    “Where’s the evidence “That people with artificial hearts are not soulless zombies?””

    Where the evidence that you’re not? All right, that settles it. Voight-Kampff tests for everybody.

  36. dusty bohl says:

    ‘The soul resides in the heart,’ is just too goofy to believe, yet that’s where a lot of people seem to find it. If you’re having trouble, try other places like your good foot or papa’s brand new bag. If you still can’t find it, get up off of that thing in case you’ve been sitting on it all along.

  37. Cheryl says:

    “Where’s the evidence that the soul does not reside in the heart?”

    The Japanese traditional practice of seppuku (hara-kiri, or self-disembowelment with a sword) is based on the belief that the soul resides in the belly.

    There are plenty of Japanese words outside the martial context that evidence the location of the soul in the belly (hara). See discussion at http://www.e-budo.com/forum/archive/index.php/t-9747.html.

  38. Jack says:

    “Where’s the evidence that the soul does not reside in the heart?” ~ Andrew Leigh

    Where’s the evidence that the soul DOES reside in the heart? So what is a soul anyway? Who has evidence that it exists, and evidence of exactly where it exist?

    To make a claim that something exists first REQUIRES evidence/proof on the one making the positive claim — to prove a negative of anything, such as the one Andrew requests, is an illogical exercise.

    But of course it’s the religious who believe a soul exists, and that it can’t be seen by humans. Well isn’t that just something? Reminds me of this quote:

    “The invisible and the non-existent look very much alike.” ~ Delos B. McKown

    And even Thomas Jefferson didn’t take kindly to religious claims of things immaterial such as the “soul”:

    “To talk of immaterial existences is to talk of nothings. To say that the human soul, angels, god, are immaterial, is to say they are nothings, or that there is no god, no angels, no soul. I cannot reason otherwise: but I believe I am supported in my creed of materialism by Locke, Tracy, and Stewart. At what age of the Christian church this heresy of immaterialism, this masked atheism, crept in, I do not know. But heresy it certainly is.”-Thomas Jefferson, letter to John Adams, Aug. 15, 1820

  39. LoafingOaf says:

    Where’s the evidence that the soul does not reside in the heart?

    Gee, I dunno. Maybe, for example, the fact that if you remove someone’s heart and replace it with another person’s, or an animal’s, or with a man-made pump, the person is still the same person and, if he had a “soul” (whatever that is), he has the same “soul.”

  40. David Marjanović says:

    Voight-Kampff tests for everybody.

    What’s this?

  41. Leon says:

    It’s from Blade Runner (see this.)

  42. David Marjanović says:


  43. Claire says:

    But David, there are multiple phylogenetic methods for tree estimation in biology, and it’s not clear which ones are appropriate for linguistics. UPGMA, Neighbourhood Joining or ML, for example? (Warnow’s shown that UPGMA does badly, but not why it does badly.) Dating’s a separate question.

Comments are closed.