Correlation between language features and location of origin

[jal]

Interesting hypothesis, but I have doubts it's more than coincedence: Why Did Humans Develop So Many Different Languages? (the title, as is so often the case, bares little to the actual content)


JAL

[Salmoneus]

Oh dear lord. They're blaming density of trees?

When was the last time you had a conversation with someone in an environment so densely forested that there were enough trees between you and the other person that they acoustically dampened you? Sure, maybe this works for songbirds - but songbirds try to be distinguishable from the other side of a valley. Humans try to be distinguishable to the person standing next to them. This model assumes that the basic, most common and influential form of language use is "yelling at people through a forest", which I don't believe it is.

Theoretically, the idea of temperature being relevent seems prima facie less implausible. I'd be skeptical, however, whether the small variations in human living temperatures - remember, in practice we're not going to be contrasting "middle of the sahara at noon" to "middle of siberia at midnight at midwinter", because people try to avoid the extremes, and avoid standing around having conversations in them in particular - could really have such a huge and differential effect on phonetics.

[M Mira]

Are they doing their research on until recently hunter-gatherer tribes? If not then their research has little value as long-settled people do not need a far-reaching language are gonna screw with their hypothesis a lot.

[Salmoneus]

Are they doing their research on until recently hunter-gatherer tribes? If not then their research has little value as long-settled people do not need a far-reaching language are gonna screw with their hypothesis a lot.

How much do even hunter-gatherers shout at each other while hunting and gathering anyway? One would think that gathering would require little co-ordination of troops, and that hunters would prefer to use hand signals or naturalistic call sounds, rather than just talking loudly at one another from a distance.

[WeepingElf]

Are they doing their research on until recently hunter-gatherer tribes? If not then their research has little value as long-settled people do not need a far-reaching language are gonna screw with their hypothesis a lot.

How much do even hunter-gatherers shout at each other while hunting and gathering anyway? One would think that gathering would require little co-ordination of troops, and that hunters would prefer to use hand signals or naturalistic call sounds, rather than just talking loudly at one another from a distance.

Yep. Talking loudly during a hunt is one of the best ways of scaring the venison away.

I concur with you on the misguidedness of this kind of "research".

[jal]

I concur with you on the misguidedness of this kind of "research".

Yeah, even though one of the guys involved seems a somewhat distinguished linguist. Perhaps they should've contacted an anthropologist.

I once saw a documentary on TV about the San, and they have a story that the language they speak contains so many clicks because if they speak with each other during the hunt, the sound it makes is more "natural" than speaking without clicks, and hence doesn't scare away the animals.


JAL

[vokzhen]

I once saw a documentary on TV about the San, and they have a story that the language they speak contains so many clicks because if they speak with each other during the hunt, the sound it makes is more "natural" than speaking without clicks, and hence doesn't scare away the animals.


JAL

I've heard that too, but not any research whatsoever that animals are actually less startled by clicks. Especially when there's still vowels and normal consonants interspersed. It just seemed like armchair speculation.

[Zaarin]

I once saw a documentary on TV about the San, and they have a story that the language they speak contains so many clicks because if they speak with each other during the hunt, the sound it makes is more "natural" than speaking without clicks, and hence doesn't scare away the animals.


JAL

I've heard that too, but not any research whatsoever that animals are actually less startled by clicks. Especially when there's still vowels and normal consonants interspersed. It just seemed like armchair speculation.

Nor why the languages of other hunter-gatherers--a category that includes the ancestors of everyone--didn't develop clicks.

[Richard W]

Nor why the languages of other hunter-gatherers--a category that includes the ancestors of everyone--didn't develop clicks.

Or lost them.

There is, of course, the suggestion that the speakers of Khoisan are partly descended from another distantly related group of hominines, much as we have neanderthal ancestors. I don't know how far north this influence could extend.

[----]

I once saw a documentary on TV about the San, and they have a story that the language they speak contains so many clicks because if they speak with each other during the hunt, the sound it makes is more "natural" than speaking without clicks, and hence doesn't scare away the animals.


JAL

I've heard that too, but not any research whatsoever that animals are actually less startled by clicks. Especially when there's still vowels and normal consonants interspersed. It just seemed like armchair speculation.

Anyone who's ever been within seeing distance of a wild animal before knows quite well that they are startled by pretty much everything.

[Abi]

I once saw a documentary on TV about the San, and they have a story that the language they speak contains so many clicks because if they speak with each other during the hunt, the sound it makes is more "natural" than speaking without clicks, and hence doesn't scare away the animals.


JAL

I've heard that too, but not any research whatsoever that animals are actually less startled by clicks. Especially when there's still vowels and normal consonants interspersed. It just seemed like armchair speculation.

Nor why the languages of other hunter-gatherers--a category that includes the ancestors of everyone--didn't develop clicks.

Are we talking clicks as phonemes proper or clicks as messengers of meaning? People in the US use clicks: horse commands, disagreement (usually in ghetto speak), agreement, onomatopoeias (gun noises, etc.).

[Jonlang]

This topic kinda reminds me of this video.

[Zaarin]

People in the US use clicks: ...disagreement (usually in ghetto speak), agreement...

Are you sure about that? I'm white middle-class, and to me a dental click marks disapproval or disagreement, not agreement...

[Tropylium]

The correlation might be onto something. The proposed form of causation is obviously bonkers.

It might be productive to rather consider an indirect approach along the lines of:
– what kind of societal organization do different biomes encourage?
– what degree of linguistic diversity (in terms of lineages) do different forms of societal organizations encourage?
– what degree of phonetic diversity do different degrees of linguistic diversity encourage?

There are productive results to be had also from the second and third points. Results like OP mainly seem to suggest that we also should not neglect point one in my list, although it may seemingly have nothing to do with linguistics.

I've seen a recent presentation by Blench that I think comes much closer to forming a sensible hypothesis of what's going on with correlations like "more trees ~ more vowels".

[jal]

I once saw a documentary on TV about the San, and they have a story that the language they speak contains so many clicks because if they speak with each other during the hunt, the sound it makes is more "natural" than speaking without clicks, and hence doesn't scare away the animals.

I've heard that too, but not any research whatsoever that animals are actually less startled by clicks. Especially when there's still vowels and normal consonants interspersed. It just seemed like armchair speculation.

Note that the "they" I referred to is the San themselves. And they don't have armchairs :). This is not a theory by Western linguists or the like, this is a San origin story.


JAL

[gach]

My problem with the similar papers by Everett (1, 2) has been a lack in the rigour of the statistical and data handling practices. I'd say that interpreting the correlations is premature if you can't be sure about the power of the methods or that the data are handled properly. I can't say much for the current paper, though, since I don't have access to its text.

It could be fun to write a reanalysis of some of these data with more care on doing the data analysis properly. There would just need to be some actual linguists co-authoring the paper so that it wouldn't be merely a random look at some data without any regard on what it's actually about.

Here's an example from my own field documenting how badly constructed research methods can have a severe effect on your results. Do things the wrong way and you'll end up introducing human biases which make you find the thing you were looking for even in cases of pure noise. The second sentence of the abstract tells it all:

However, numerical experiments with random input data show that most, if not all, of the observed longitude discrimination can be an artifact of the analysis method.

Go read the paper, it's a good cautionary tale. If you aren't at an institute that has access to the journal, go to arXiv, it's open for all.

I've seen a recent presentation by Blench that I think comes much closer to forming a sensible hypothesis of what's going on with correlations like "more trees ~ more vowels".

The presentation identifies two global waves of linguistic expansion (11-8000 BP and 5-3500 BP) leading to the current major language families. However, the available data (number of expansive language families) is sparse, dating language families is difficult, and the two proposed periods are quite close to each other compared to their lengths. The number of major expansions after 3500BP is also not insignificant. Based on this I'd say that the hypothesis of discrete global expansion waves is no more supported than a smoothly increasing expansion rate up to the present. Did the presenter address this in his talk?

[Tropylium]

I've seen a recent presentation by Blench that I think comes much closer to forming a sensible hypothesis of what's going on with correlations like "more trees ~ more vowels".

The presentation identifies two global waves of linguistic expansion (11-8000 BP and 5-3500 BP) leading to the current major language families. However, the available data (number of expansive language families) is sparse, dating language families is difficult, and the two proposed periods are quite close to each other compared to their lengths. The number of major expansions after 3500BP is also not insignificant. Based on this I'd say that the hypothesis of discrete global expansion waves is no more supported than a smoothly increasing expansion rate up to the present. Did the presenter address this in his talk?

As in, I've seen the slides, not the actual talk.

I don't think his argument actually requires a stratified structure for linguistic expansions though (it seems like a weak point to me). They key point is, IMO, the observation that language-family diversity correlates inversely to the age of human settlement. "Spread zones" get populated first; and since the geographic conditions have not changed, they then get progressively re-populated or at least reconquered, leading to modern-day relative uniformity. Meanwhile "refuge zones" are reached much later, and re-reached much more rarely — but still early enough that linguistic diversity has time to develop.

[----]

(usually in ghetto speak)

what the heck man

[gach]

They key point is, IMO, the observation that language-family diversity correlates inversely to the age of human settlement. "Spread zones" get populated first; and since the geographic conditions have not changed, they then get progressively re-populated or at least reconquered, leading to modern-day relative uniformity. Meanwhile "refuge zones" are reached much later, and re-reached much more rarely — but still early enough that linguistic diversity has time to develop.

Might be, I was mainly interested if there was anything more to the suggested global expansion phases since they sound like a quite sensational claim. I'll put them for now in the category of "say something speculative so that people remember your talk".

Still, I'm not entirely convinced that the data supports a claim that the age of human settlement is a required explanatory parameter in the problem. What exactly says that the frequency of expansions into the area isn't alone a sufficient parameter to describe the diversity data? This would actually be another nice model comparison problem to do.