Rilx wrote:
I think that we must compare different languages and cultures to find features that can be explained from culture. The essential feature of English compared to some other language is its poor morphology. While another language uses few long words consisting of a root and morphemes, English forms the expression by several words. English and Eskimo languages are near the opposite ends of the scale, and I think that my native language (Finnish) is nearer the Eskimo end. I assume you know the nature of Eskimo languages.

What cultural differences that kind of difference in languages can explain? I think that forming an expression by several small words - which also means different grammar - grrrr.. reflects an extremely diversified culture, complex and hectic, with high science and technology. If almost every word is a sentence, there are very few sentences in the language; the culture is extremely static.

IMO this kind of evolution, towards multiple shorter words, is not basically linguistic. It's culture-driven, and language follows.

Sorry, but for this claim to have any value, you would need to show that the opposition analytic/synthetic languages maps reliably onto "dynamic" vs. "static" cultures (whatever criteria and timescale you use for that.) You should also be able to show that in the history of one society. Just some food for thought, while English is a relatively analytic (isolating) language, Chinese is even more isolating; several West African languages are highly isolating as well. Synthetic languages, on the other hand (the ones you link to "static" culture) include many European languages (from Finnish to French) and Japanese (as well as Sanskrit, Quechua ...). Furthermore, the isolating structure of English is the result of an evolutionary trend in that language that we can follow back to the Middle Ages.

I wonder: Is there something that can be rescued from all of Chomsky´s and Pinker´s production?
About Chomsky, it´s clear that he didn´t understand a word of Skinner and behaviorism, that his defense of nativism and mentalism is based on fallacious arguments, and that his "language acquisition device" is a speculation without evidence.
The proposal of universal features of grammar is an interesting issue, but is there some good evidence about this? I would expect to find evidence from a corpus of examples taken from very different languages, but instead of that, I only find some speculative categories taken from a few set of indoeuropean languages (mostly english), and derived from "grammatical judgements" of descontextualized imaginary examples made by the theorist himself (without any empirical study of the diversity shown by language-users in real contexts).

Good questions. One of the main problems with Chomsky is that he takes as the subject of his study the "ideal" speaker-listener in a homogeneous language community - and these caricatures of speaker and language community can't begin to account for all that's interesting about language. It's a bit as if you decided to study genetics but ignore variability

I know there has been some work done on non-Indo-European languages, but I'm uncomfortable about the fact that this uses Chomsky's theory that was formulated beforehand, based on a much more limited sample of languages. Confirmation bias is a very real danger here.

About universal features of grammar, I have a confession to make. I mentioned Bickerton's work on creole languages. The basis is the pidgin-creole cycle, which I'd understood was a well-accepted model in linguistics. However, from my more recent readings, it would seem to be quite controversial, which takes away what I mentioned in the OP as the best evidence for Pinker's claims.

ETA: There was an interesting thread just about Chomsky some time ago in Psychology and Neuroscience (linky). Maybe we can revive it, I'd be interested in your input on the points that were raised there.

Palindnilap has some interesting ideas on Universal Grammar, and I think he had a mathematical proof that it's a valid idea or something.. Hopefully he'll stop by with more info.

Mr.Samsa wrote:Palindnilap has some interesting ideas on Universal Grammar, and I think he had a mathematical proof that it's a valid idea or something.. Hopefully he'll stop by with more info.

I think he said it provided a useful formalisation for computer modelling of language. It was in the Chomsky thread I've linked.

katja z wrote:

Mr.Samsa wrote:Palindnilap has some interesting ideas on Universal Grammar, and I think he had a mathematical proof that it's a valid idea or something.. Hopefully he'll stop by with more info.

I think he said it provided a useful formalisation for computer modelling of language. It was in the Chomsky thread I've linked.

Ah thanks!

I've discussed it with him a number of times but I thought we had that particular conversation over at RDF (and I think we did have it there too?). Brain too frazzled.

Mr.Samsa wrote:

katja z wrote:

Mr.Samsa wrote:Palindnilap has some interesting ideas on Universal Grammar, and I think he had a mathematical proof that it's a valid idea or something.. Hopefully he'll stop by with more info.

I think he said it provided a useful formalisation for computer modelling of language. It was in the Chomsky thread I've linked.

Ah thanks!

I've discussed it with him a number of times but I thought we had that particular conversation over at RDF (and I think we did have it there too?). Brain too frazzled.

It seems that he´s arguing that there´re innate constraints for grammar learning. That´s ok, but that´s not "Universal Grammar", and that´s not a criticism to Skinner, behaviorism, or empiricism.
For a more inductive approach to universal features of language, see:
http://typo.uni-konstanz.de/archive/int ... x.php?pt=1
http://en.wikipedia.org/wiki/Joseph_Greenberg

I didn't bother to read this thread, some this point might have already been made, but... evidence for a language instinct:

98% of languages are either Subject-Object-Verb or Subject-Verb-Object (i.e., in 98% of languages, the subject precedes the verb) - evidence that language develops as a product of how our brain thinks about things.

Children are highly likely to make grammatical errors saying things such as "runned" as opposed to the correct "ran." They have never heard "runned" before, but instead can construct new words as opposed to behaviorist theories from the likes of Skinner that we simply acquire language through conditioning of the classical and operational sort.

There is a "critical period" of sorts, between the ages of 2-7, where language acquisition is extraordinarily rapid, and people who learn languages during this time period (i.e., native speakers) are almost always more adept than people who learn the languages later in life. fMRI also show that these individuals use different parts of their brain to speak the language. fMRI has also shown that Broca's and Wernicke's areas are active when deaf people sign.

People who are deaf naturally develop their own sign language even if not instructed about developing one. Also, it was African children who developed creole as a language, not adults - showing that, while adults surpass the intellectual capacity of children in almost every area, language is one of those few exceptions where they do not (in fact, I can't think of another one off the top of of my head...).

To me, this says that the ability to acquire language is something innately human - not any particular language, but a simply an ability for a language of some sort.

The might of Samsan shall bury us alive!

MillsianUtilitarian wrote:I didn't bother to read this thread, some this point might have already been made

The points have been made and subsequently debunked..

MillsianUtilitarian wrote:but... evidence for a language instinct:

98% of languages are either Subject-Object-Verb or Subject-Verb-Object (i.e., in 98% of languages, the subject precedes the verb) - evidence that language develops as a product of how our brain thinks about things.

Questions to ask:

If it was innate, why not 100%? (Barring physical or mental deformities which would explain the discrepancy).

How do you separate this commonality out from being a product of an innate function, and it being a function of language having common roots? i.e. Why do you think it is that the more remote a tribe is, the less "universal" their language is?

MillsianUtilitarian wrote:Children are highly likely to make grammatical errors saying things such as "runned" as opposed to the correct "ran." They have never heard "runned" before, but instead can construct new words as opposed to behaviorist theories from the likes of Skinner that we simply acquire language through conditioning of the classical and operational sort.

Uh, simply no. Skinner and behavioral psychology's theories on language predict that children must necessarily go through this phase. It's called stimulus generalisation. If children did not say "runned", then it would disprove a lot of Skinner's ideas on language. This suggestion is like saying evolution predicts that we should find a crocoduck. It's actually the nativists that should be predicting perfect understanding of grammar without a learning curve - if grammar is innate, then why do children make mistakes that need to be trained out of them?

MillsianUtilitarian wrote:There is a "critical period" of sorts, between the ages of 2-7, where language acquisition is extraordinarily rapid, and people who learn languages during this time period (i.e., native speakers) are almost always more adept than people who learn the languages later in life.

Irrelevant - this is true of all things. That is the age when your brain is essentially on fire forming connections with anything it comes into contact with.

MillsianUtilitarian wrote:fMRI also show that these individuals use different parts of their brain to speak the language. fMRI has also shown that Broca's and Wernicke's areas are active when deaf people sign.

So? Are you saying that if language was learnt and not innate, then we would use some other organ other than the brain? Keep in mind that Broca's and Wernicke's areas are not simply "language centres" - they perform a number of other tasks, like symbolic representation and are used in perceptual tasks. Also keep in mind that language is a global process and requires multiple parts of the brain.

You've probably read Chomsky, or perhaps Pinker, on this subject and unfortunately they don't know what they're talking about. For more information check out this review of Chomsky's review where the author demonstrates, in laughable detail, how ridiculous Chomsky's criticisms were. Apparently his reputation for being a bit of a "language creationist" doesn't apply solely to his ideas, but also to his tactics.

MillsianUtilitarian wrote:People who are deaf naturally develop their own sign language even if not instructed about developing one. Also, it was African children who developed creole as a language, not adults - showing that, while adults surpass the intellectual capacity of children in almost every area, language is one of those few exceptions where they do not (in fact, I can't think of another one off the top of of my head...).

Language is an immensely useful tool and can easily be developed by anything with vocal chords and hands; we should expect these results from a learning account of language.

MillsianUtilitarian wrote:To me, this says that the ability to acquire language is something innately human - not any particular language, but a simply an ability for a language of some sort.

It doesn't explain why other animals can pick it up though, even when they don't share a recent evolutionary link to us or comparable brain structures..

Mayak wrote:The might of Samsan shall bury us alive!

And to make it worse, there's also Katja in this thread who seems to be a bit of an honorary behaviorist, and Seeker who appears to know more about behaviorism than I do.

Mr.Samsa wrote:
How do you separate this commonality out from being a product of an innate function, and it being a function of language having common roots? i.e. Why do you think it is that the more remote a tribe is, the less "universal" their language is?

I would say the last sentence there is rather unfounded?

Uh, simply no. Skinner and behavioral psychology's theories on language predict that children must necessarily go through this phase. It's called stimulus generalisation. If children did not say "runned", then it would disprove a lot of Skinner's ideas on language. This suggestion is like saying evolution predicts that we should find a crocoduck. It's actually the nativists that should be predicting perfect understanding of grammar without a learning curve - if grammar is innate, then why do children make mistakes that need to be trained out of them?

Interestingly, some mistakes cannot be accounted for by analogy to observed behaviours (that is, not like 'past tense is formed by -ed, so run -> runned). Placement of negative particles apparently is one of these, where there basically is a hierarchy as to where the negative particle will go in different stages. I don't remember how the hierarchy goes, but nowhere can the child observe the negative particles going there, and it's surprisingly universal for not being observed anywhere outside this context.

@ MillsianUtilitarian: Mr Samsa has already done a splendid job and you can find more in the Chomsky thread (I linked to it yesterday). Just to add to a couple of points ...

Mr.Samsa wrote:

MillsianUtilitarian wrote:but... evidence for a language instinct:

98% of languages are either Subject-Object-Verb or Subject-Verb-Object (i.e., in 98% of languages, the subject precedes the verb) - evidence that language develops as a product of how our brain thinks about things.

Questions to ask:

If it was innate, why not 100%? (Barring physical or mental deformities which would explain the discrepancy).

How do you separate this commonality out from being a product of an innate function, and it being a function of language having common roots? i.e. Why do you think it is that the more remote a tribe is, the less "universal" their language is?

Unsurprising. There are only so many ways of organising linguistic information, and some are more practical than others. You just might pile up all the particles and articles for the whole sentence in front, but it would't result in very efficient communication. This is like saying that because every human group has figured out the most useful sequence of actions for, say, hunting, or growing crops (those that do engage in agriculture, anyway), these actions must be innate.

Also, VSO and VOS patterns are found in a number of languages, esp. in the Americas, as the unmarked word order, and in more languages as a marked word order, or in restricted grammatical contexts.

MillsianUtilitarian wrote:People who are deaf naturally develop their own sign language even if not instructed about developing one. Also, it was African children who developed creole as a language, not adults - showing that, while adults surpass the intellectual capacity of children in almost every area, language is one of those few exceptions where they do not (in fact, I can't think of another one off the top of of my head...).

Language is an immensely useful tool and can easily be developed by anything with vocal chords and hands; we should expect these results from a learning account of language.

Add to that that deaf children live in a social setting which is completely moulded by language. Human social interactions presuppose language, and it is this state of affairs they are born into. They can observe symbolic communication going on all the time all around them, and people (beginning with their parents) trying to communicate with them in other ways, in order to compensate for their inability to use spoken language (which is why, in the absence of a standard signing language, any family with a deaf member will develop a limited set of signs to manage everyday communication - it is these diverse sets that the Nicaraguan deaf community, which is probably what you had in mind, used to base their sign language on).

Also, see my yesterday's post where I touch upon the pidgin-creole cycle. I think I read some very good criticism of this hypothesis (which I had presented myself in the OP) in Louis-Jean Calvet, an excellent sociolinguist (my favourite, in fact). If you're interested and can read French, I'll try to remember which book it was.

MillsianUtilitarian wrote:To me, this says that the ability to acquire language is something innately human - not any particular language, but a simply an ability for a language of some sort.

It doesn't explain why other animals can pick it up though, even when they don't share a recent evolutionary link to us or comparable brain structures..

You forgot your bike analogy, Mr. Samsa.

Mayak wrote:The might of Samsan shall bury us alive!

And to make it worse, there's also Katja in this thread who seems to be a bit of an honorary behaviorist,

MillsianUtilitarian wrote:To me, this says that the ability to acquire language is something innately human - not any particular language, but a simply an ability for a language of some sort.

It doesn't explain why other animals can pick it up though, even when they don't share a recent evolutionary link to us or comparable brain structures..

They can't, though. Any claims to the contrary you've heard is likely a myth. The nearest any animal gets is reacting to single words.

(k, I admit, there's some syntactic or paratactic structure to bees' dancing language)

Zwaarddijk wrote:

Mr.Samsa wrote:
How do you separate this commonality out from being a product of an innate function, and it being a function of language having common roots? i.e. Why do you think it is that the more remote a tribe is, the less "universal" their language is?

I would say the last sentence there is rather unfounded?

I think it refers to the fact that what is found outside the big important families, esp. IE, tends to be thought of as a quaint exception ...

Interestingly, some mistakes cannot be accounted for by analogy to observed behaviours (that is, not like 'past tense is formed by -ed, so run -> runned). Placement of negative particles apparently is one of these, where there basically is a hierarchy as to where the negative particle will go in different stages. I don't remember how the hierarchy goes, but nowhere can the child observe the negative particles going there, and it's surprisingly universal for not being observed anywhere outside this context.

Any examples? I'm not certain I understand the bolded bit. You are probably thinking of the hierarchy of the sentence structure, with recursive embedding of phrases? This recursive embedding can certainly be observed in practice.

I don't remember the details, but something like:
Children first tend to negate by putting the negative particle first, then they go on to putting it before the verb (or something, I seriously don't remember the details), then it progresses through a few more until it settles at the point that the language they are speaking has settled for. If the language does usually negate by an initial the negative particle, the child won't make any obvious mistakes, of course.

Zwaarddijk wrote:I don't remember the details, but something like:
Children first tend to negate by putting the negative particle first, then they go on to putting it before the verb (or something, I seriously don't remember the details), then it progresses through a few more until it settles at the point that the language they are speaking has settled for. If the language does usually negate by an initial the negative particle, the child won't make any obvious mistakes, of course.

I'd have to think about this a bit more, but just off the top of my head, negation is difficult because depending on where you put the negation element, you negate just one bit of the sentence (one phrase) or the whole sentence. I suppose until a child has mastered the recursivity I mention above, they can't generalise rules for where to put the element of negation (this is complicated by the fact that "no" is just the tip of the iceberg, there are negative pronouns and adverbs and so on to go with it, often requiring further grammatical adjustments of the sentence), so maybe the best strategy turns out to be to stick it into the most obvious place in the sentence where it also won't interfere with the rest of the sentence structure? Just speculating, of course.

katja z wrote:

Zwaarddijk wrote:I don't remember the details, but something like:
Children first tend to negate by putting the negative particle first, then they go on to putting it before the verb (or something, I seriously don't remember the details), then it progresses through a few more until it settles at the point that the language they are speaking has settled for. If the language does usually negate by an initial the negative particle, the child won't make any obvious mistakes, of course.

I'd have to think about this a bit more, but just off the top of my head, negation is difficult because depending on where you put the negation element, you negate just one bit of the sentence (one phrase) or the whole sentence. I suppose until a child has mastered the recursivity I mention above, they can't generalise rules for where to put the element of negation (this is complicated by the fact that "no" is just the tip of the iceberg, there are negative pronouns and adverbs and so on to go with it, often requiring further grammatical adjustments of the sentence), so maybe the best strategy turns out to be to stick it into the most obvious place in the sentence where it also won't interfere with the rest of the sentence structure? Just speculating, of course.

obviously - I'll need to try and find some sources on this, really.

Animals can only acquire a rudimentary understanding of language - the smartest nonhuman apes knowing about 200 words, for instance. Saying animals can simply pick up a language is either false, or you meant what you said differently (what exactly did you mean?). Also, all infants begin assuming that pronouns can be dropped at the beginning of sentences, but later, after more exposure to languages that require them (like English as opposed to Spanish) will begin adding redundant pronouns. Sorry for the short reply, I typed this out on my phone. I hope to get into more depth tomorrow!

Zwaarddijk wrote:

Mr.Samsa wrote:
How do you separate this commonality out from being a product of an innate function, and it being a function of language having common roots? i.e. Why do you think it is that the more remote a tribe is, the less "universal" their language is?

I would say the last sentence there is rather unfounded?

How so? It seems that whenever we find a new tribe that is supposedly so remote from civilisation that it has never had contact with any modern cultures, their language structure is so different that it requires a massive reworking of ideas like universal grammar to incorporate it, or in the case of the Piraha tribe, we have to accept that their language falsifies the notion of universal grammar.

This, to me, suggests that any ideas of "universal grammar" or common innate elements across cultures is a result of a shared root, rather than some innate issue. And, as Katja added, a result of the fact that there are only so many possible variations on language structures.

Zwaarddijk wrote:

Uh, simply no. Skinner and behavioral psychology's theories on language predict that children must necessarily go through this phase. It's called stimulus generalisation. If children did not say "runned", then it would disprove a lot of Skinner's ideas on language. This suggestion is like saying evolution predicts that we should find a crocoduck. It's actually the nativists that should be predicting perfect understanding of grammar without a learning curve - if grammar is innate, then why do children make mistakes that need to be trained out of them?

Interestingly, some mistakes cannot be accounted for by analogy to observed behaviours (that is, not like 'past tense is formed by -ed, so run -> runned). Placement of negative particles apparently is one of these, where there basically is a hierarchy as to where the negative particle will go in different stages. I don't remember how the hierarchy goes, but nowhere can the child observe the negative particles going there, and it's surprisingly universal for not being observed anywhere outside this context.

I'm not sure what you mean? Why would we need an analogy to observed behaviors to account for it? It's not like Skinner suggested that we sit down with a kid and get him to repeat the word "runned" followed by a candy bar each time he does it...

Or, if you like, think about it this way: Skinner's ideas on language are pretty well accepted in academia, and are the basis for the most successful language therapies around at the moment; do you really think that they would be so successful if it required very intelligent people to accept a position which cannot account for such a common observation like children saying "runned" and "drinked"?

Zwaarddijk wrote:

MillsianUtilitarian wrote:To me, this says that the ability to acquire language is something innately human - not any particular language, but a simply an ability for a language of some sort.

It doesn't explain why other animals can pick it up though, even when they don't share a recent evolutionary link to us or comparable brain structures..

They can't, though. Any claims to the contrary you've heard is likely a myth. The nearest any animal gets is reacting to single words.

(k, I admit, there's some syntactic or paratactic structure to bees' dancing language)

MillsianUtilitarian wrote:Animals can only acquire a rudimentary understanding of language - the smartest nonhuman apes knowing about 200 words, for instance. Saying animals can simply pick up a language is either false, or you meant what you said differently (what exactly did you mean?). Also, all infants begin assuming that pronouns can be dropped at the beginning of sentences, but later, after more exposure to languages that require them (like English as opposed to Spanish) will begin adding redundant pronouns. Sorry for the short reply, I typed this out on my phone. I hope to get into more depth tomorrow!

They aren't "claims" I've heard, they are research I've read and studied. It's true that animals don't have language at the level we do, but there are obvious reasons for why this is (i.e. our massive culture which provides our children with intensive language training from the moment they are born until they die). Despite this, we still have evidence of animals doing spectacularly well at picking up aspects of language; for example, pigeons understand grammar and successfully discriminate between grammatical and ungrammatical sentences after around 2 months of training. Compare that to children, who still struggle into their teenage years (and often beyond that). We also have other animal feats in this area, like chickadees outperforming humans in their understanding of recursion, apes using vocal or sign language at a meaningful level, parrots and dolphins understanding grammatical order to perform tasks etc.

So if the question is whether animals can use language at the level we do, then the answer is clearly "no". If the question is whether animals can understand and use language, then the answer is clearly "yes". We have so much evidence on this that it really is unarguable at this point, so whilst there are still questions over the details, the one concrete conclusion we can come to is that language definitely isn't a human-only ability.

I am pretty sure the animal language stuff you posted is pure bunk.

Try a search for animal language on language log ( http://languagelog.ldc.upenn.edu/nll/ ), iirc there should be a fair share of information there.