Indian linguist Anvita Abbi writes for Scientific American (June 2023; archived) about the Great Andamanese languages she has been studying, and a fascinating group of languages it is:
When I first met Nao Jr., at the turn of the millennium, he was in his 40s and one of only nine members of his Indigenous group, Great Andamanese, who still spoke the idiom of his ancestors; the youngsters preferred Hindi. As a linguist with a passion for decoding structure, I had researched more than 80 Indian languages from five different families: Indo-European (to which Hindi belongs), Dravidian, Austroasiatic, Tibeto-Burman and Tai-Kadai. I was on the islands to document their Indigenous voices before they faded into whispers. What little I heard was so baffling that I returned many times in later years to try to pin down the principles undergirding Great Andamanese languages. […]
Great Andamanese, it turns out, is exceptional among the world’s languages in its anthropocentrism. It uses categories derived from the human body to describe abstract concepts such as spatial orientation and relations between objects. To be sure, in English we might say things like “the room faces the bay,” “the chair leg broke” and “she heads the firm.” But in Great Andamanese such descriptions take an extreme form, with morphemes, or meaningful sound segments, that designate different zones of the body getting attached to nouns, verbs, adjectives and adverbs—indeed, to every part of speech—to make diverse meanings. Because no other known language has a grammar based on the human body or shares cognates—words that are similar in meaning and pronunciation, indicating a genealogical connection—with Great Andamanese, the language constitutes its own family. […]
British officials had observed that the Andamanese languages were a bit like links in a chain: members of neighboring Great Andaman tribes understood one another, but those speaking languages at opposite ends of the chain, in North and South Andaman, were mutually unintelligible. In 1887 British military administrator Maurice Vidal Portman published a comparative lexicon of four languages, as well as a few sentences with their English translations. And around 1920 Edward Horace Man compiled an exhaustive dictionary of Bea, a South Andaman language. These were significant records, but neither cracked the puzzle the grammar posed.
Nor could I. Somehow my extensive experience with all five Indian language families was no help. One time I asked Nao Jr. to tell me the word for “blood.” He looked at me as if I were an utter fool and did not reply. When I insisted, he said, “Tell me where it is coming from.” I replied, “From nowhere.” Irritated, he repeated, “Where did you see it?” Now I had to make up something, so I said, “On the finger.” The reply came promptly—“ongtei!”—and then he rattled off several words for blood on different parts of the body. If the blood emerged from the feet or legs, it was otei; internal bleeding was etei; and a clot on the skin was ertei. Something as basic as a noun changed form depending on location. […]
The grammar I was piecing together was based primarily on Jero, but a look through Portman’s and Man’s books convinced me that the southern Great Andamanese languages had similar structures. The lexicon consisted of two classes of words: free and bound. The free words were all nouns that referred to the environment and its denizens, such as ra for “pig.” They could occur alone. The bound words were nouns, verbs, adjectives and adverbs that always existed with markers indicating a relation to other objects, events or states. The markers (specifically, a-; er-; ong-; ot- or ut-; e- or i-; ara-; and o-) derived from seven zones of the body and were attached to a root word, usually as a prefix, to describe concepts such as “inside,” “outside,” “upper” and “lower.” For example, the morpheme er-, which qualified most anything having to do with an outer body part, could be stuck to -cho to yield ercho, meaning “head.” A pig’s head was thus raercho.
This conceptual dependency did not always indicate physical attachment. For example, if the pig’s head were cut off for roasting, the marker t- for an inanimate object would be attached to er- to yield ratercho; it was no longer alive but still a pig’s head. The suffix -icho indicated truly separable possessions. For example, Boa-icho julu meant “Boa’s clothes.”
Just as a head, a noun, could not conceptually exist on its own, the mode and effect of an action could not be severed from the verb describing the action. Great Andamanese had no words for agriculture or cultivation but a great many for hunting and fishing, mainly with a bow and arrow. Thus, the root word shile, meaning “to aim,” had several versions: utshile, to aim from above (for example, at a fish); arashile, to aim from a distance (as at a pig); and eshile, aiming to pierce. […]
In the Great Andamanese view of nature, the foremost distinction was between tajio, the living, and eleo, the nonliving. Creatures were tajio-tut-bech, “living beings with feathers”—that is, of the air; tajio-tot chor, “living beings with scales,” or of the water; or tajio-chola, “living beings of the land.” Among the land creatures, there were ishongo, humans and other animals, and tong, plants and trees. These categories, along with multiple attributes of appearance, motion and habits, made for an elaborate system of classification and nomenclature that I documented for birds in particular. Sometimes the etymology of a Great Andamanese name bore a resemblance to the English one. For example, Celene, made up of root words for “crab” and “thorn,” was so named because it cracks and eats crabs with its hard, pointed beak. […]
Even “morning” and “evening” were relative, depending on who experienced them. To say, for instance, “I will visit you tomorrow,” one would use ngambikhir, for “your tomorrow.” But in the sentence “I will finish this tomorrow,” the word would be thambikhir, “my tomorrow.” Time depended on the perspective of whoever was involved in the event.
Much more at the link; this kind of thing makes me wish I’d stuck with linguistics. Thanks, Trevor!
Because no other known language has a grammar based on the human body or shares cognates—words that are similar in meaning and pronunciation, indicating a genealogical connection—with Great Andamanese, the language constitutes its own family.
The part no other known language has a grammar based on the human body is not really relevant to whether the language is an isolate, of course; as Abbi is most definitely a Proper Linguist, I don’t expect she meant to imply that it is.
I’m actually the proud possessor of a copy of her (very good) Grammar of the Great Andamanese Language, and very interesting it is too. I’m not totally sold on this idea that all this stuff about body parts makes the language absolutely unique; it’s basically a sort of incorporated classifier system, and the only part where it seems fully operational is in the grammar of inalienably possessed nouns (unsurprisingly) and in verb prefixes, where the relationship to body parts is mostly fairly transparent; but that can easily be paralleled in other language families. Apart from that, quite a lot is rather speculative, and certainty over many points seems unattainable given the moribund state of the language: this is something of a rescue grammar, sadly.
So I’m not convinced Great Andamanese is particularly unique, compared with all the other unique languages out there. Still very interesting, though.
certainty over many points seems unattainable given the moribund state of the language: this is something of a rescue grammar, sadly.
If the language is moribund, rescue is unlikely. The book must be a palliative morphine grammar.
I’m not totally sold on this idea that all this stuff about body parts makes the language absolutely unique
Oh, me neither, but I write that off as the understandable exuberance of a linguist trying to convey the wonderfulness of her chosen language group. Totally forgivable!
Exuberance or not – for those who like it a bit more sober: Bernard Comrie and Raoul Zamponi, ›Typological profile of the Great Andamanese family‹, Journal of South Asian Languages and Linguistics 4 (2017) 1, 55-83 [DOI 10.1515/jsall-2017-0002] or by the same authors A Grammar of Akajeru (open access).
“absolutely unique”
I don’t think it’s wrong to describe any given person as unique. Why not languages? (including typology – except maybe very close languages)
“The part no other known language has a grammar based on the human body is not really relevant to whether the language is an isolate, of course;” – borrowable, but when there are not any other languages with such a property, likely it is an isolate.
English DOES use content words where my language uses suffixes (…-berry, …-tree etc.), but mostly not body parts. English teapot, Russian chay-nik. -Nik really does the very same job as -pot (and there is some specialisation, like a dedicated suffix for berries).
I guess there must also be “a nosebleed” (UPD: I guessed right, but maybe I saw this word already). Such words have to do with that in Enlgish you (I assume) can say “I have/got a XY” or “my X is Ying” where in Russian you say “by me blood flows from nose”.
Is there other evidence of the Great Andamanese/Germanic link you’re proposing?
Why not languages?
Sure. There are hundreds of unique languages. If not thousands …
when there are not any other languages with such a property, likely it is an isolate
I think that this doesn’t follow at all. It’s a variation on the age-old error of relying on typology to determine genetic affiliation. More specifically, it’s pretty easy to identify rare or even possibly unique traits in an individual language which is by no means an isolate. I don’t, for example, know of any language apart from Kusaal that has six distinguishable enclitics with no segmental form whatsoever; but Kusaal has well over a thousand proven relatives (including some so close that the resemblance is obvious even to local people completely untrained in linguistics.) It’s about as far from an isolate as a language well can be.
Just recently we were discussing Udi endoclitics. Even if you stretch a point with the definition of “endoclitic” to make it as inclusive as possible, still I’ve never come across another language which is quite so hardcore about having honest-to-God bona fide unexplainable-away endoclitics. But Udi is not an isolate either …
A well-known example of typology not matching phylogeny is Celtic, where the initial mutations led to a lot of initial doubts about whether they should be classified as Indo-European or not.
“Is there other evidence of the Great Andamanese/Germanic link you’re proposing?”
need to read the
Englishgrammar first…P.S.”A well-known example of typology not matching phylogeny is Celtic,”
I’m not sure it is so unique cross-linguistically… (surely, the system as a whole (the part of the system shared by Brittonic and Goidelic, that is) is unique. Simpler things are not).
I don’t know of an initial-mutation system elsewhere which is quite as elaborate as in Insular Celtic, but there certainly are others (Mende and its close relations, for example.) Fulfulde (famously) is a kinda-example, but there it’s all in the morphology, rather than mixed in with the syntax the way it is in Celtic and Mende.
Mind you, with Celtic it wasn’t just the mutations that confused people: there was the whole VSO thing, conjugated prepositions and what have you. That still gets people wittering about imaginary substrates. It was quite late in the day that people realised that much of this tends to go together in one package typologically anyway, regardless of the actual origin of the language.
Czech ř.
I wrote about English -berry as a suffix because of “But in Great Andamanese such descriptions take an extreme form, with morphemes, or meaningful sound segments, that designate different zones of the body getting attached to …“. I don’t mean that Englihs is exotic or that Andamanese is not.
I just look for parallels almost mechanically.
The choice of “head” as the lexeme she pursues through a Great Andamanese declension in order to explain how these languages use “categories derived from the human body to describe abstract concepts such as spatial orientation and relations between objects”, is certainly interesting. For me, it raises more questions than it answers.
In such a system, I might have expected head was a core bodily concept that would have been used to define these other things. I’m surprised to learn it works the other way — that you need to know something spatially about a head (in relation to the rest of the body?) to even call it a head.
It does leave me wondering to what degree -er = outer is truly bodily, and not just spatial. Maybe “bodily” in the already somewhat abstract sense of “related to a physical thing” rather than bodily in the sense of “related to a (human) body”? But she clearly says “human body”.
And the fact that a head is an “outer” -cho rather than an “upper” -cho also has me flustered and wondering what I’m misunderstanding, how much is left out/oversimplified (which would be reasonable). But also wondering whether something is being misinterpreted in how this all works. It’s hard for me to consider “outer” as a defining quality of heads.
The choice of -cho, without ever defining it, adds to the confusion. I find myself wondering what “-cho” means, how it might help me understand ercho. And immediately, I come across her mention of -icho, meaning “separable.” Are -icho and -cho related? Probably not, since I’m even less drawn to the idea that Andamanese think of separability as one of the defining qualities of heads. But of all the forms she might have used for examples, did she need to use two that were nearly identical, right next to each other, and only define one of them?
It’s very possible that if I learned one of the dialects of Great Andamanese, I would understand that her description is spot on. But at this point, I’m left scratching my outer -cho.
Andamanese think of separability as one of the defining qualities of heads
That’s only seen in Sentinelese.
Did you even consider that there might be a Sentinelese reader here who would not take kindly to your jokes?
Edited – I reflect after a moment that my joke does tend to further marginalize them, but weighing the balance of funny vs. real-world impact, I’m content that the flaps of my wings will not cause a cyclone in the Bay of Bengal.
Jokes? I’m just reporting on my own fieldwo
Turner ✔
Eddyshaw ✔
Lukashenko ☐
More technical and pleasingly detailed discussions of Abbi’s work are here and here.
Turner ✔
Eddyshaw ✔
Lukashenko ☐
Enemies List. David seems to an expert on tv tropes. I find the site funny in small doses, but its exhaustive coverage of ephemera is exhausting. This is a job for Galactic Housewife Brain.
I see that Abbi has also written a paper on a Nicobarese language, coauthored with Vysakh R. I have never seen that single-letter surname before.
But you have seen Y. It’s not that different.
The link has “languagehat.com” in the URL; replacing it by “academia.edu” does not help.
Oops. How’d I do that?
https://www.researchgate.net/publication/339857949_Aspects_of_word_formation_processes_in_Luro_The_endangered_language_of_the_Nicobar_Islands
Stu, “Y” is my first name. “Y.” is my full name. “.” is my surname.
Thanks Y. The “Partonomy of human body and grammaticalisation in PGA” chart (the second one) helps. The grammatical classes themselves are already considerably more abstract than might be suggested by the idea of body-part classes. Abbi writes “It is generally seen that in the process of grammaticalisation lexical units are desemanticised and decategorialised.” I don’t know whether that statement is accurate, but it does help me understand what’s going on.
It would be fascinating to understand the boundaries between classes, since they’re not intuitive to (a westerner/me). Even if they’re partly “desemanticised and decategorialised,” they likely make some intuitive sense to Andamanese, and I’d be interested to understand how.
A little more explanation would have been helpful. I can understand why the mouth-part clitic might be part of the word for greedy. I’m still not clear on whether a-tutlup (greedy) has a host of related forms with tutlup and different clitics, carrying different meanings.
I’m not very comfortable with Abbi’s interest in using genetics and physiology as evidence for linguistic groupings. But this factoid is fascinating (if true) regardless of why it’s in the paper:
>Studies have shown that the Jarawas and the Onges have distinct physiological and genetic signatures from the Great Andamanese, like a low blood pressure prole, low body temperature, low pulse rate and very low frequency to absence of the B gene in the ABO blood group (Kashyap et al. 2004: 3)
As to Abbi’s idea that the body-part classification is archaic and “indicative of the early times when human beings conceptualised their world through their body and its divisions,” well… hopefully the linguistics is on a more solid basis.
And I guess in retrospect, the low blood pressure profile stuff may belong to the same category of science. It does make me worry a bit more about the quality of the linguistics, and whether the body-part classification stuff is depicted accurately.
The genetics she discusses is pretty archaic. Our ancestors in the early 2000s were fond of quoting genetic studies in linguistic ones.
Comrie and Zamponi’s paper has a nice diagram of the semantic evolution of one somatic affix in Akabea (e.g. mouth > food > food source > tree > canoe > seaman). They also say, “Somatic affixes constitute a typologically unusual feature, otherwise attested, to our knowledge, in various languages of the Pacific Northwest of North America (Wakashan, Salishan, Chimakuan families, and Takelma), the isolate P’urhépecha/Tarascan and the Totonacan family of Mesoamerica, and the Panoan family of South America…” For P’urhepecha, Paul Friedrich’s paper, Shape in Grammar, is excellent and deep, like everything he did.
The Friedrich paper is interesting. I’ll need to read it again to take more of it in.
The Comrie paper link also has languagehat.com as the domain. I wonder what is causing that?
It’s understood today that various Andamanese groups are closely linked by their overall genetic composition but their small sizes resulted in bottleneck effects which often cause big frequency differences between specific genetic factors (like the ones used 20+ years ago, before the whole-genome scans)
The Comrie paper link also has languagehat.com as the domain. I wonder what is causing that?
There’s no https to create a link.
“Partonomy of human body and grammaticalisation in PGA”
Or partonymy? Hmm, there may be a difference. A principled one, since the part–whole relations are less relevant here? At that location, in a section headed “Meronymy”:
The relevant fully Greek terms to contemplate (are there more?):
Compare mereology, etc., in philosophical talk.
Other links seem to be paywalled, but here’s a free link to the Comrie paper
>Or partonymy? Hmm, there may be a difference. A principled one, since the part–whole relations are less relevant here?
She uses partonomy, and defines class 4 (ut=) as bodily products and part-whole relationships.
‘I’m not sure it is so unique cross-linguistically…’
drasvi, what was at stake was not cross-linguistic distinctiveness as understood in the 21st century, but how unusual these languages seemed to linguists in the early 19th century. It’s not like Bopp or Rask could just turn to Iosad 2010 for a convenient overview of mutation around the globe. And it wasn’t increased typological knowledge that eventually solved the issue, but Bopp’s explanation of Irish mutation with reference to Sanskrit endings, which showed that however strange mutation might seem, it clearly developed within these Indo-European languages. I’m inclined to think those earlier 19th-century philologists did a pretty decent job of things on the whole, given what was available to them.
here’s a free link to the Comrie paper
Thanks, Ryan. (Good paper, as you’d expect. The verb root ellipsis thing is interesting: in fact, that seems rather harder to parallel cross-linguistically than the body-part business.)
@Nelson, my logic is that IF (I have no idea if she’s right!) a language is really unusual today, then:
– the scenario “it is an old feature, and it is not observed in other languages because of recent migration” becomes somewhat more likely.
– the scenario “it is a young feature that developed recently (even though surrounding langauges don’t have it)”
becomes somewhat less likely.
But of course, the fewer languages you have examined the less qualified you are to speak of what is “unusual”. People of 19th century were less qualified to rely on cross-linguistical uniqueness (to the limited extent it can be relied upon. We have a good reason to be humble here. Their reason was better).
On the other hand, their a priori assumption could be that the Celtic system must be something ancient.
You know a number of things about Celtic mutation (particularly that it is unusual among IE langauges).
So you have some a priori probability that this feature is “old”.
Now, in light of new information (is it unusual cross-linguistically?) you modify your probability somehow. I’m speaking about how we can modify it.
Part of the joy of comparative linguistics is that it can often shed light on how some very odd synchronic feature arose (in the process demonstrating that it isn’t very ancient.)
The Celtic mutations are a good example. So, in fact, are the Kusaal zero-form enclitics, the odd properties of which are pretty easy to explain as consequences of the loss of word-final short vowels, which is presumably recent enough that such oddities have not yet been levelled away. Harris’ account of Udi endoclitics also leans greatly on comparison to explain how things got that way, and Evans’ grammar of Kayardild (a language with a pretty good claim on being Weirdest Language Evah) has a whole section of comparative stuff which he says he put in because unless he had a diachronic explanation of the weirdness, he thought that “otherwise open-minded linguists” might have some trouble accepting it at all.
So perhaps any apparent association between a lsnguage being an isolate and a lsnguage being very unusual is an artefact of not being able to explain how it got that way.
DE, if a property can easily develop in many languages, you expect to see it in many languages.
If some property P is extremely rare/unique, then:
(a) it requires another rare property P’ as a pre-requisite – and we repeat the same about P’
(b) it takes time to develop
(c) levelling influence of neighbouring languages prevents it
(d) it is just rare:)
Regarding (b) : we know thousands languages. This sample is large and diverse when we are discussing properties that develop quickly (b), don’t have pre-requisites absent from our sample (a, c), and are not rare ((d) – of course, when a property is found in 1 language out of 10 000 we don’t expect to find it among our “thousands”). For something “slow” (10 thousands years) we need to count language families or areas to determine our effective sample size.
The option (d) is especially interesting to people who aspire to find biological or logical limitations on the space of all possible languages (or cultural…, but then we have to deal witht he space of possible cultures, and also with relativism). When I draw attention to b and c instead, I’m actually being sceptical.
Universals are informative.
They can be trivial (“no language has a word kjksjks in the sense “beer”) and they exist (some arrangements are clearly inconvenient).
When you find an universal (or almost universal), is it an element of the universal grammar, or absolutely arbitrary retention from proto-world or maybe it correlates with culture (or climate or…)? The suggestion that it can be a retention may seem totally shameless to you. But I think, assigning an universal to the “universal grammar” (unless we understand it just as “universally observed grammar”) is even more shameless. And what other options are there?
When someone says: “there are no words without vowels!” (indeed such words may seem ‘clearly inconvenient’ and impossible) and I can’t (in this case I can) point at numerous counterexamples, I can say: it is not necessary a restriction on the space of possible languages, it is not “clearly” inconvenient, it can be just a retention.
With an isolate, there is no way of telling whether anything is a retention or not. In fact, the question doesn’t even make sense with isolates. Retained from what, exactly?
You end up in the territory of all those people who solemnly inform us that Proto-Human had clicks because “click languages” are all so wonderfully primitive and hunter-gathery and that anyway it all follows from the genetics in some way. Or all those pretend-historical-linguistics papers that Nature has perpetrated, that explain to us how Bayesian Methods have Solved Historical Linguistics.
There aren’t any non-trivial absolute universals in human language (by “non-trivial”, I mean, things that aren’t obvious consequences of languages being used by creatures with our own particular physical limitations, so it is “trivial” that there are no natural languages which use gamma rays as their medium, for example.)
There are numerous very interesting statistical tendencies, some of which are so strong that they surely must mean something. Though It’s hard to think of any plausible meanings that have actually been identified. I mean, sure, languages with VSO word order are much more likely to have prepostions than postpostions, for example. From which it surely follows that …. well, what follows?
DE, I agree.
But when the list of possible explanations is “A or B” and you insist that B is silly, you are saying “A” (even if you intended to say “we don’t know”).
But it is never just “A or B”. I think it is Pirsig who says there are an infinite number of rational hypotheses that can account for a given set of phenomena (though Dr. Google is not helping me here).
From ZAMM:
#
“The number of rational hypotheses that can explain any given phenomenon is infinite.” (p. 139)
#
Found by Goggle search of “pirsig infinite number rational hypothesis”.
https://books.google.com/books?id=M69poeV1UhoC&pg=PA179&dq=%22For+every+fact+there+is+an+infinity+of+hypotheses
I think it is Pirsig who says there are an infinite number of rational hypotheses that can account for a given set of phenomena (though Dr. Google is not helping me here).
John doesn’t give his search terms. I took the search terms from his complaint.
This is a strategy I often use: when I don’t succeed at getting something done, I complain about it to myself in an exaggerated, whiney manner. Inspecting these complaints gives me new ideas about how to get the thing done. Usually after a few rounds of this I hit on a solution.
If at first you don’t succeed, whine, whine again.
@stu
This works if your problem is best solved by technical means. If not, you need to moan about it to your manager. Having to listen to your moaning is why she earns more than you.
Ad hoc ingenuity. Works for me.
The thing is to shake off cognitive constraints, aka obstacles épistémologiques [Bachelard]. As a dog shakes off rainwater. There are other techniques – I imagine autoflagellation is one such in matters of faith.
PP: this is not limited to technical matters, and managers are not a constraint. I get things done in despite of adversity. As Noe says, it’s ad hoc ingenuity – in which method can be found, as I will add.
Behind many a silk purse is a sow’s ear.