This has cropped up in another thread but I think it deserves a thread of its own.

Does it make any sense saying that language is an instinctual, or partly instinctual, behaviour, as the title of Steven Pinker's book implies? And if yes, what in this complex behaviour can be meaningfully chalked down to instinct? What would this help to explain (or muddle)?

Pinker argues for the existence of an "innate grammar machinery" in the brain. The best corroboration for his claim that I can think of comes from research on the development of creole languages, particularly the work of Derek Bickerton. But I think it is safe to say that the ways the known languages organise their material, as soon as you step out of the Indoeuropean family best known to most linguists, differ so much that any such machinery would have to operate at a level of abstraction that makes it difficult to see how it would be practically useful, let alone how it would have evolved ...

It could be argued that something so basic to humans "must" have a basis in their biology. But does the fact that all human societies engage in language really mean that this complex symbolic practice must be an "instinct"? What does it even mean that a behaviour is "instinctual"? The concept of "instinct" needs some clarification, but I'm counting on more knowledgeable members to chime in here (hint, hint, Mr. Samsa!). As for "language", it refers both to a structured system of linguistic signs (a natural language such as English, or Arabic, or Malgasy, etc.) and to the cognitive ability to produce and use such a system. Obviously it is this latter sense that applies here, and equally obviously all we can observe are languages in the first sense (their structure, functioning, development and acquisition); but it is so tempting to slide from one sense to the other in the debate that I think a reminder of the distinction can't hurt.

Yeh din toh aata hai ek din jawaani mein...

I am rooting for Pinker! Post inbound!

natselrox wrote:Yeh din toh aata hai ek din jawaani mein...

Er ... Lan nga wax?

I am rooting for Pinker! Post inbound!

katja z wrote:
Pinker argues for the existence of an "innate grammar machinery" in the brain. The best corroboration for his claim that I can think of comes from research on the development of creole languages, particularly the work of Derek Bickerton. But I think it is safe to say that the ways the known languages organise their material, as soon as you step out of the Indoeuropean family best known to most linguists, differ so much that any such machinery would have to operate at a level of abstraction that makes it difficult to see how it would be practically useful, let alone how it would have evolved ...

God I hate the creole argument.. Can you explain to me how it's supposed to demonstrate universal grammar? Most explanations I've seem tend to amount to: "If people had to learn language, then they wouldn't try to simplify it and make it easy to understand". In other words, they assume that people wouldn't develop a simple form of grammar unless it was built into their brains - which seems absolutely ridiculous to me. Anyway, I like this rejection of Bickerton's work:

However, extensive work by Carla Hudson-Kam and Elissa Newport suggests that creole languages may not support a universal grammar, as has sometimes been supposed. In a series of experiments, Hudson-Kam and Newport looked at how children and adults learn artificial grammars. Notably, they found that children tend to ignore minor variations in the input when those variations are infrequent, and reproduce only the most frequent forms. In doing so, they tend to standardize the language that they hear around them. Hudson-Kam and Newport hypothesize that in a pidgin situation (and in the real life situation of a deaf child whose parents were disfluent signers), children are systematizing the language they hear based on the probability and frequency of forms, and not, as has been suggested on the basis of a universal grammar.[2][3] Further, it seems unsurprising that creoles would share features with the languages they are derived from and thus look similar "grammatically."

(From here: http://en.wikipedia.org/wiki/Universal_ ... _languages).

katja z wrote:It could be argued that something so basic to humans "must" have a basis in their biology. But does the fact that all human societies engage in language really mean that this complex symbolic practice must be an "instinct"? What does it even mean that a behaviour is "instinctual"? The concept of "instinct" needs some clarification, but I'm counting on more knowledgeable members to chime in here (hint, hint, Mr. Samsa!).

The concept of "instinct" has had a troubled history. It's taken on a number of different meanings over the years, but it essentially it was used to mean 'innate behaviors'. Due to the vagueness of the concept, it's been replaced by the concepts "reflex" and "fixed-action patterns". Taken together they represent what people usually mean when they talk about 'instincts'. However, now that they've been rigorously defined, it becomes quite clear that language in no way can possibly be viewed as instinctual.

This obviously doesn't mean that we don't need structures in our brains that make us capable of developing language though. This is the confusion that has led to Pinker's books.

Mr.Samsa wrote:
God I hate the creole argument..

Sorry, I had to include it in order to be fair and balanced . I'm far from swayed by this argument myself, but from what you've written about it, I think you hate it for partly the wrong reasons. It's not as stupid as you seem to think (mind you, this is not saying that it is correct!).

Can you explain to me how it's supposed to demonstrate universal grammar? Most explanations I've seem tend to amount to: "If people had to learn language, then they wouldn't try to simplify it and make it easy to understand". In other words, they assume that people wouldn't develop a simple form of grammar unless it was built into their brains - which seems absolutely ridiculous to me.

I don't know which explanations you've seen, but the original idea is exactly the other way round. Creole languages are not "simple" in any sense. On the contrary, these are full-blown languages with their own complex grammar which developed from so-called pidgin languages - these are the simple languages you had in mind. The crucial differences between creoles and pidgins, are the level of complexity and the function - while a pidgin is a simplified form of linguistic communication between different-language communities (typically for limited purposes, such as trading, so it doesn't express the whole of social reality, since it doesn't have to), a creole is formed when such a pidgin language is taken over as a mother tongue (i.e., the first language) of a new generation. At that stage, the ex-pidgin typically develops very rapidly, its rudimentary grammatical structure acquiring a high level of complexity typical of "normal" natural languages. All of this is well-known and observed. This development stage is really fascinating - within one generation, the new speaking community will have fashioned a tool to serve its own communication needs in all domains of social life. You can't blame Bickerton for being impressed. I am.

It doesn't seem too far-fetched to suppose that the development of a well-structured creole from the "soup" of linguistic bits and pieces can tell something about how language developed. Especially because the new grammar is fully developed only when a new linguistic community of native speakers of such a language is formed. Of course, the situation is different here than at the mythical beginning of language, in that the input comes from speakers of several other natural languages, and also the whole of the (social) environment is now such that it not only demands, but presupposes language. So inventing one if there's none available seems like a reasonable strategy

But what Bickerton's claims really hinge upon - and here things get more problematic - is the perceived structural similarities of creoles. Now I haven't read Bickerton's books, so I don't know what his research corpus was. If he only used the Caribbean creoles (I'm guessing here, but I've seen his work quoted in relation to those) which have very similar histories, not only in terms of the language material they are based on but also the social context of their development, then the rebuttal you quoted certainly stands:

it seems unsurprising that creoles would share features with the languages they are derived from and thus look similar "grammatically."

(From here: http://en.wikipedia.org

If, on the contrary, he can show structural similarities between these and creoles based on languages from completely different language families, say a Chinese-based creole, this would get more interesting - although it still needn't mean that there is anything innate to grammar, it could be that there are just so many ways of effectively organising linguistic material, especially when you take into account the constraints of language evolution.

Anyway, I like this rejection of Bickerton's work:

However, extensive work by Carla Hudson-Kam and Elissa Newport suggests that creole languages may not support a universal grammar, as has sometimes been supposed. In a series of experiments, Hudson-Kam and Newport looked at how children and adults learn artificial grammars. Notably, they found that children tend to ignore minor variations in the input when those variations are infrequent, and reproduce only the most frequent forms. In doing so, they tend to standardize the language that they hear around them. Hudson-Kam and Newport hypothesize that in a pidgin situation (and in the real life situation of a deaf child whose parents were disfluent signers), children are systematizing the language they hear based on the probability and frequency of forms, and not, as has been suggested on the basis of a universal grammar.[2][3] /wiki/Universal_grammar#Presence_of_creole_languages).

I'm familiar with this argument and I like it too. It makes a whole lot of sense and fits with a large amount of observation. Unfortunately, it doesn't address the production of structural complexity at the transition stage to a creole language, which is the key ingredient of Bickerton's hypothesis.

The concept of "instinct" has had a troubled history. It's taken on a number of different meanings over the years, but it essentially it was used to mean 'innate behaviors'. Due to the vagueness of the concept, it's been replaced by the concepts "reflex" and "fixed-action patterns". Taken together they represent what people usually mean when they talk about 'instincts'. However, now that they've been rigorously defined, it becomes quite clear that language in no way can possibly be viewed as instinctual.

Thanks. I know you've explained this in other threads but I think it's useful to have it here as well. If I understand this right, language is too complex a behaviour (or set of behaviours) for these more rigorous concepts to apply. But, playing the devil's advocate here for a moment, could FAPs be responsible for aspects of basic language acquisition? (I'm not out to torture you, I'm trying to get enough understanding that I can present a convincingly criticism of Pinker&co for a linguistic/literary studies crowd). Then I suppose there's the whole wretched question of how such specific behaviours would have evolved.

This obviously doesn't mean that we don't need structures in our brains that make us capable of developing language though. This is the confusion that has led to Pinker's books.

And not just grammar, either, which is what Pinker (inspired by ... you know who ...) concentrates upon. Symbolic representation, with only arbitrary and conventional links between the signifier and the referent, is a quite an impressive feat in its own right. I haven't read much about the history of that though (apart from the speculations that Neandertals were capable of it).

I'd like to follow this a bit further. What would these structures in our brains be that make us capable of developing language? Or would it be more interesting to follow the history of the material culture and social interactions of H. sapiens and look for the clues on the development of language there?

Yesterday I borrowed Pinker's The Language Instinct to have a closer look at it, but I've just received the reviews of a paper that I only expected to get in two months' time, along with a request to rewrite and resend it asap. Damn editors, they have a knack of hitting you over the head with a deadline just when you think you can have some quiet time to yourself!

katja z wrote:

Mr.Samsa wrote:
God I hate the creole argument..

Sorry, I had to include it in order to be fair and balanced . I'm far from swayed by this argument myself, but from what you've written about it, I think you hate it for partly the wrong reasons. It's not as stupid as you seem to think (mind you, this is not saying that it is correct!).

And I think I haven't explained myself too well..

katja z wrote:

Can you explain to me how it's supposed to demonstrate universal grammar? Most explanations I've seem tend to amount to: "If people had to learn language, then they wouldn't try to simplify it and make it easy to understand". In other words, they assume that people wouldn't develop a simple form of grammar unless it was built into their brains - which seems absolutely ridiculous to me.

I don't know which explanations you've seen, but the original idea is exactly the other way round. Creole languages are not "simple" in any sense. On the contrary, these are full-blown languages with their own complex grammar which developed from so-called pidgin languages - these are the simple languages you had in mind. The crucial differences between creoles and pidgins, are the level of complexity and the function - while a pidgin is a simplified form of linguistic communication between different-language communities (typically for limited purposes, such as trading, so it doesn't express the whole of social reality, since it doesn't have to), a creole is formed when such a pidgin language is taken over as a mother tongue (i.e., the first language) of a new generation. At that stage, the ex-pidgin typically develops very rapidly, its rudimentary grammatical structure acquiring a high level of complexity typical of "normal" natural languages. All of this is well-known and observed. This development stage is really fascinating - within one generation, the new speaking community will have fashioned a tool to serve its own communication needs in all domains of social life. You can't blame Bickerton for being impressed. I am.

When I referred to creole as "simplifying it and making it easy to understand" I was specifically talking about grammar - however, looking back at my sentence I worded that terribly. I meant to indicate that by devising a simple form of grammar, it simplified a lot of their conversation and made things easier to understand; simplified in the sense that there was a structure and rule to follow in order to convey specific information. Of course creole languages are more complex than pidgin, but the grammar used in creole is pretty simple.

And I certainly don't blame Bickerton for being impressed, I think those who aren't impressed by findings like this aren't worth considering being called scientists, but my complaint is more at his "surprise". It seems obvious to me if that you stick a bunch of people together with only a rudimentary form of communication, then they'll inevitably form grammatical norms.

katja z wrote:It doesn't seem too far-fetched to suppose that the development of a well-structured creole from the "soup" of linguistic bits and pieces can tell something about how language developed. Especially because the new grammar is fully developed only when a new linguistic community of native speakers of such a language is formed. Of course, the situation is different here than at the mythical beginning of language, in that the input comes from speakers of several other natural languages, and also the whole of the (social) environment is now such that it not only demands, but presupposes language. So inventing one if there's none available seems like a reasonable strategy

Definitely true but I imagine the two situations (the creation of creole languages and creation of original language) would make for some interesting comparisons.

katja z wrote:But what Bickerton's claims really hinge upon - and here things get more problematic - is the perceived structural similarities of creoles. Now I haven't read Bickerton's books, so I don't know what his research corpus was. If he only used the Caribbean creoles (I'm guessing here, but I've seen his work quoted in relation to those) which have very similar histories, not only in terms of the language material they are based on but also the social context of their development, then the rebuttal you quoted certainly stands:

it seems unsurprising that creoles would share features with the languages they are derived from and thus look similar "grammatically."

(From here: http://en.wikipedia.org

If, on the contrary, he can show structural similarities between these and creoles based on languages from completely different language families, say a Chinese-based creole, this would get more interesting - although it still needn't mean that there is anything innate to grammar, it could be that there are just so many ways of effectively organising linguistic material, especially when you take into account the constraints of language evolution.

Indeed, the bolded part is generally the approach I take - or rather, I think it's the best explanation until it is ruled out.

katja z wrote:
I'm familiar with this argument and I like it too. It makes a whole lot of sense and fits with a large amount of observation. Unfortunately, it doesn't address the production of structural complexity at the transition stage to a creole language, which is the key ingredient of Bickerton's hypothesis.

What is the transitional stage? Is that the between generations finding where the grammar isn't fully formed until the new generation comes along? Surely that's explained by simple learning; otherwise known as "old dog, new trick".

katja z wrote:Thanks. I know you've explained this in other threads but I think it's useful to have it here as well. If I understand this right, language is too complex a behaviour (or set of behaviours) for these more rigorous concepts to apply. But, playing the devil's advocate here for a moment, could FAPs be responsible for aspects of basic language acquisition? (I'm not out to torture you, I'm trying to get enough understanding that I can present a convincingly criticism of Pinker&co for a linguistic/literary studies crowd). Then I suppose there's the whole wretched question of how such specific behaviours would have evolved.

Too complex to have evolved? Not quite, no. That would be a very creationist-like argument

Complex behaviors can easily come about through evolution (just look at the wasp that controls the mind of it's prey and rides it back to it's nest to lay eggs in it) and it is feasible that something like language could develop in an organism through means like this, but the problem is that it is completely unlike any other innate behavior we have. For example, if language were a FAP then it needs to meet the criteria of FAPs which generally include: 1) presence in every member of the species, 2) they're automatic and irresistable, 3) elicited by a specific stimulus in the environment, and 4) occur without fail every time the stimulus is presented. So for language to be an "instinct", or specifically a FAP, it needs to meet all of these criteria. Because of arguments for "critical periods" (where if you don't learn something like language you never can) then I think we could be generous and grant option 1 - although I think the evidence for it is much weaker than for any other FAP we know about. Criteria 2, 3 and 4 are basically nonsensical when discussing language though - what stimulus can we possibly think of to elicit language, without fail every time? Arguably initiation of verbal contact might work, if we ignore the fact that it does not meet criteria 2 and 4 by any stretch of the imagination (i.e. me saying "hello" to you does not force you to respond verbally in some way).

So I think language as a FAP is not really a valid line of thought. We could try to argue that language might be a new form of "instinctual" behavior, unlike any other we know of, which requires years and years of learning before we grasp the basics and it's not automatic or triggered by any kind of stimulus, and manifests in different ways, forms and topographies... But then every single thing we do is an "instinctual" behavior. Watching tv, reading the paper, learning instruments, driving a car, etc. (And language certainly isn't a reflex either, as those stimulus-response actions are usually just electrical signals that don't even reach the brain. This would make our "language centres" rather redundant)..

katja z wrote:

This obviously doesn't mean that we don't need structures in our brains that make us capable of developing language though. This is the confusion that has led to Pinker's books.

And not just grammar, either, which is what Pinker (inspired by ... you know who ...) concentrates upon. Symbolic representation, with only arbitrary and conventional links between the signifier and the referent, is a quite an impressive feat in its own right. I haven't read much about the history of that though (apart from the speculations that Neandertals were capable of it).

I'd like to follow this a bit further. What would these structures in our brains be that make us capable of developing language? Or would it be more interesting to follow the history of the material culture and social interactions of H. sapiens and look for the clues on the development of language there?

The main structures that are argued to be necessary for language are things like Broca's area and Wernicke's area, both controlling very specific aspects of language. Depending on where the damage is, you can get cases where a patient is still able to understand language but their speech production is messed up - they think they're speaking perfectly normally, but all that's coming out is garbage. You ask a man how he feels today, and he'll say something like, "The lion greened snowflakes" with a pleasant smile on his face. Or you can get people with perfect language and syntax, but who can't understand anything you say to them. Language generally is a global process through, you require multiple (and non-specific) areas of your brain to use it.

The thing that needs to be noted is that these areas aren't solely used for language, they control other functions too. It's speculated that they began as structures to comprehend body language and eventually took on the role of language as we developed it - this supposedly explains why other apes have brain areas which are analogous to Broca's and Wernicke's areas. This is also used as evidence for the Gestural Theory for the origin of language.

katja z wrote:Yesterday I borrowed Pinker's The Language Instinct to have a closer look at it, but I've just received the reviews of a paper that I only expected to get in two months' time, along with a request to rewrite and resend it asap. Damn editors, they have a knack of hitting you over the head with a deadline just when you think you can have some quiet time to yourself!

To be fair, I bet you have more scientific value in a single paper than Pinker has in his entire career..

Mr.Samsa wrote:
When I referred to creole as "simplifying it and making it easy to understand" I was specifically talking about grammar - however, looking back at my sentence I worded that terribly. I meant to indicate that by devising a simple form of grammar, it simplified a lot of their conversation and made things easier to understand; simplified in the sense that there was a structure and rule to follow in order to convey specific information. Of course creole languages are more complex than pidgin, but the grammar used in creole is pretty simple.

Um, I still think that you're looking at this the wrong way around, what's important about the creoles is precisely the unexpectedly high degree of complexity that seeminlgy comes "ex nihilo", not their structural simplicity (which is relative, they keep on refining the structure; at least, they do up to a point you'd expect in a language that doesn't have a strong written tradition - since you can get extremely complex syntax, with a large amount of subordination, only in written communication ... )

And I certainly don't blame Bickerton for being impressed, I think those who aren't impressed by findings like this aren't worth considering being called scientists, but my complaint is more at his "surprise". It seems obvious to me if that you stick a bunch of people together with only a rudimentary form of communication, then they'll inevitably form grammatical norms.

Definitely true but I imagine the two situations (the creation of creole languages and creation of original language) would make for some interesting comparisons.

Certainly, and these would probably be far more relevant than those between language acquisition in children and the origin of human language ... In so far I think Bickerton may be onto something.

:nod: Indeed, the bolded part is generally the approach I take - or rather, I think it's the best explanation until it is ruled out.

Great minds and so on

katja z wrote:
I'm familiar with this argument and I like it too. It makes a whole lot of sense and fits with a large amount of observation. Unfortunately, it doesn't address the production of structural complexity at the transition stage to a creole language, which is the key ingredient of Bickerton's hypothesis.

What is the transitional stage? Is that the between generations finding where the grammar isn't fully formed until the new generation comes along? Surely that's explained by simple learning; otherwise known as "old dog, new trick".

I'm not sure I understand the second sentence, but the transition I was referring to is the transition from pidgin to creole, when the language becomes the mother tongue to a new speaking community and acquires an astounding level of structural complexity in the process, very fast (much faster than had been thought languages could change). And this new grammar wasn't learned from the previous, pidgin-speaking generation; it's certainly based on their speech, but goes far beyond it, very fast. The point (I think) is that it is difficult to imagine such a complex structure being elaborated and accepted as convention so quickly in a collective, non-directed process (there's no grammarian sitting down and creating the rules!). So our biology gets blamed for it.

Too complex to have evolved? Not quite, no. That would be a very creationist-like argument

That was some terrible wording on my part! It's sometimes difficult because most of my knowledge of evo biology is very recent and I haven't had time to digest it yet. I meant that language was too complex to qualify as a FAP, and I think my use of the word "complex" wasn't too rigorous either, I was thinking of all the different situations, ways and goals of linguistic communication, basically what you too seem to refer to a bit further on:

So I think language as a FAP is not really a valid line of thought. We could try to argue that language might be a new form of "instinctual" behavior, unlike any other we know of, which requires years and years of learning before we grasp the basics and it's not automatic or triggered by any kind of stimulus, and manifests in different ways, forms and topographies... But then every single thing we do is an "instinctual" behavior. Watching tv, reading the paper, learning instruments, driving a car, etc.

The main structures that are argued to be necessary for language are things like Broca's area and Wernicke's area, both controlling very specific aspects of language.

These sound like something I should know more about. I'll try the Wikipedia. Tomorrow

Language generally is a global process through, you require multiple (and non-specific) areas of your brain to use it.

Mmm, there might be another difficulty right here. It's difficult for many people to think of language as not being something very specific (since it seems to be so specific to humans), as something that is "just" the result of many other processes - as if this detracted from its value, that it is not something biologically unique that humans have.

The thing that needs to be noted is that these areas aren't solely used for language, they control other functions too. It's speculated that they began as structures to comprehend body language and eventually took on the role of language as we developed it - this supposedly explains why other apes have brain areas which are analogous to Broca's and Wernicke's areas. This is also used as evidence for the Gestural Theory for the origin of language.

Yes, I knew that. Thanks to a certain Mr. Samsa you might know

katja z wrote:Yesterday I borrowed Pinker's The Language Instinct to have a closer look at it, but I've just received the reviews of a paper that I only expected to get in two months' time, along with a request to rewrite and resend it asap. Damn editors, they have a knack of hitting you over the head with a deadline just when you think you can have some quiet time to yourself!

To be fair, I bet you have more scientific value in a single paper than Pinker has in his entire career..

Um, is that a compliment for me or a lash out at Pinker? I don't really, to be fair, but then I don't claim that I'm doing science. (It's a paper on literary translation where there's a power/prestige dissymetry between the source and target languages ).

katja z wrote:
Um, I still think that you're looking at this the wrong way around, what's important about the creoles is precisely the unexpectedly high degree of complexity that seeminlgy comes "ex nihilo", not their structural simplicity (which is relative, they keep on refining the structure; at least, they do up to a point you'd expect in a language that doesn't have a strong written tradition - since you can get extremely complex syntax, with a large amount of subordination, only in written communication ... )

katja z wrote:I'm not sure I understand the second sentence, but the transition I was referring to is the transition from pidgin to creole, when the language becomes the mother tongue to a new speaking community and acquires an astounding level of structural complexity in the process, very fast (much faster than had been thought languages could change). And this new grammar wasn't learned from the previous, pidgin-speaking generation; it's certainly based on their speech, but goes far beyond it, very fast. The point (I think) is that it is difficult to imagine such a complex structure being elaborated and accepted as convention so quickly in a collective, non-directed process (there's no grammarian sitting down and creating the rules!). So our biology gets blamed for it.

Sorry, my explanations have been terrible (a consequence of writing all my posts to you at 4am..). I do know what creole language is, so any confusion in my posts above is due to my seeming inability to explain myself, rather than a misunderstanding of the subject. My knowledge of the area isn't vast, obviously, but I do understand the fundamental concepts.

The problem I have with the description of creole languages, and thus the ensuing explanations, is what you've touched on above - the idea of "ex nihilo" of complexity. Even though there isn't a grammarian sitting down creating the rules, this doesn't stop the fact that people still need an effective set of rules in order to communicate. I haven't looked at the data on this, and I'm not sure if what I'd be looking for has even been recorded, but I'd imagine that when the transition begins, there would be a large increase in "grammatical rules" - but they'd be varied and overly specific or overly general in some cases. Then, through use, certain rules will be culled off or selected depending on certain factors; the main one would probably be pragmatism, but there would also be other drift factors involved like maybe poor fidelity in passing on certain information etc.

In other words, the grammar would work as a kind of "meme" and spread through the community. (Obviously, the concept of "meme" is not very scientific, but the fundamental principles behind it are well-documented and I'm just using the term to get the general idea across as most people understand the basic idea of what a meme is and does).

I also think part of the problem is poor observation. I bet the people who notice the ex nihilo presentation of complex creole language also think that babies just spontaneously start speaking without any previous input or training..

katja z wrote:

Too complex to have evolved? Not quite, no. That would be a very creationist-like argument

That was some terrible wording on my part! It's sometimes difficult because most of my knowledge of evo biology is very recent and I haven't had time to digest it yet. I meant that language was too complex to qualify as a FAP, and I think my use of the word "complex" wasn't too rigorous either, I was thinking of all the different situations, ways and goals of linguistic communication, basically what you too seem to refer to a bit further on...

Sure, sure. I bet you're really just a creationist in disguise, aren't you?!

But yeah, complex isn't quite the right word because the form and complexity of language as we currently know it could be a FAP, there is nothing stopping evolutionary processes developing something like that (although it would probably be quite a difficult thing to produce and be a highly inefficient and wasteful exercise on evolution's part) but the problem is simply that it doesn't behave like an innate behavior does. It shares none of the necessary characteristics that it should, and to try to bend the definition of "innate behavior" results in a completely meaningless concept.

katja z wrote:

The main structures that are argued to be necessary for language are things like Broca's area and Wernicke's area, both controlling very specific aspects of language.

These sound like something I should know more about. I'll try the Wikipedia. Tomorrow

Yes, you probably should.

katja z wrote:

Language generally is a global process through, you require multiple (and non-specific) areas of your brain to use it.

Mmm, there might be another difficulty right here. It's difficult for many people to think of language as not being something very specific (since it seems to be so specific to humans), as something that is "just" the result of many other processes - as if this detracted from its value, that it is not something biologically unique that humans have.

Indeed.. I love the animal language studies, but people always try to rubbish them when they don't understand what's going on. Like one of my favourite studies is one where the experimenters train pigeons to discriminate between grammatical and ungrammatical strings in artificial languages, and they became very good at identifying grammar in novel "sentences" in a very short amount of time. But people look at it and say it's not language because they aren't speaking it, or because they can't write etc. People come up with ridiculous ideas to hold on to silly concepts sometimes.

katja z wrote:

The thing that needs to be noted is that these areas aren't solely used for language, they control other functions too. It's speculated that they began as structures to comprehend body language and eventually took on the role of language as we developed it - this supposedly explains why other apes have brain areas which are analogous to Broca's and Wernicke's areas. This is also used as evidence for the Gestural Theory for the origin of language.

Yes, I knew that. Thanks to a certain Mr. Samsa you might know

Ah yes, I think I've seen him around. Devilishly handsome, if I recall correctly.

katja z wrote:Um, is that a compliment for me or a lash out at Pinker? I don't really, to be fair, but then I don't claim that I'm doing science. (It's a paper on literary translation where there's a power/prestige dissymetry between the source and target languages ).

Still probably more scientific value than Pinker!

But it was meant to be a two birds with one stone deal; compliment you and disparage Pinker.

Mr.Samsa wrote:
Sorry, my explanations have been terrible (a consequence of writing all my posts to you at 4am..).

I must say then that you're surprisingly articulate at 4 a.m. Practice that a lot, don't you?
Sorry if I "explained" some obvious stuff. It might not have been completely useless though - there may be others reading this thread

The problem I have with the description of creole languages, and thus the ensuing explanations, is what you've touched on above - the idea of "ex nihilo" of complexity. Even though there isn't a grammarian sitting down creating the rules, this doesn't stop the fact that people still need an effective set of rules in order to communicate. I haven't looked at the data on this, and I'm not sure if what I'd be looking for has even been recorded, but I'd imagine that when the transition begins, there would be a large increase in "grammatical rules" - but they'd be varied and overly specific or overly general in some cases. Then, through use, certain rules will be culled off or selected depending on certain factors; the main one would probably be pragmatism, but there would also be other drift factors involved like maybe poor fidelity in passing on certain information etc.

Language change in languages that are only (or mostly) spoken is notoriously hard to observe, and in the case of Caribbean creoles, which is where creole linguistics begins, the most interesting period is in the past centuries when nobody really bothered with this anyway. As for our time, with the media, travel etc., as well as official language policies, education etc., the contacts and influences between languages are played out in a situation that is very different from what it must have been for most of our history. So even if you did observe the "birth" of a new creole in vivo, you'd have to be very careful with any generalisations.

In other words, the grammar would work as a kind of "meme" and spread through the community. (Obviously, the concept of "meme" is not very scientific, but the fundamental principles behind it are well-documented and I'm just using the term to get the general idea across as most people understand the basic idea of what a meme is and does).

Language evolution is one of the areas where the concept of memes and memetic evolution seems very relevant. I haven't really looked into this in any detail, but it's tempting. Too much interesting stuff, too little time in life!

I also think part of the problem is poor observation. I bet the people who notice the ex nihilo presentation of complex creole language also think that babies just spontaneously start speaking without any previous input or training..

This is very probable. Not all people who have thought about language are used to rigorous observation and thinking. And there are still a lot of myths around in the humanities and social sciences. I fell for some of them in my time, but I have seen the light.

Indeed.. I love the animal language studies, but people always try to rubbish them when they don't understand what's going on. Like one of my favourite studies is one where the experimenters train pigeons to discriminate between grammatical and ungrammatical strings in artificial languages, and they became very good at identifying grammar in novel "sentences" in a very short amount of time. But people look at it and say it's not language because they aren't speaking it, or because they can't write etc. People come up with ridiculous ideas to hold on to silly concepts sometimes.

I remember that study. Fascinating. But yes, this is very much about holding onto some comfortable ideas (or maybe just about habits of thought that have become, er, second nature - although this is only an excuse for laypeople).

Ah yes, I think I've seen him around. Devilishly handsome, if I recall correctly.

Ah, so you do know him

katja z wrote:

Mr.Samsa wrote:
Sorry, my explanations have been terrible (a consequence of writing all my posts to you at 4am..).

I must say then that you're surprisingly articulate at 4 a.m. Practice that a lot, don't you?

Yeah this is when I do all my work.. I like to sleep through most of the day but the problem is that I still get tired when it gets to 3-4am. Stupid circadian rhythm!

katja z wrote:Sorry if I "explained" some obvious stuff. It might not have been completely useless though - there may be others reading this thread

No not at all! I didn't mean it like that. I just didn't want to keep frustrating you by saying, "Yeah I understand it, it's just blah blah blah", then having you reply with, "No, that's not quite right, it's blah blah", and me saying, "Yeah, blah blah" and you saying, "No, blah".

katja z wrote:Language change in languages that are only (or mostly) spoken is notoriously hard to observe, and in the case of Caribbean creoles, which is where creole linguistics begins, the most interesting period is in the past centuries when nobody really bothered with this anyway. As for our time, with the media, travel etc., as well as official language policies, education etc., the contacts and influences between languages are played out in a situation that is very different from what it must have been for most of our history. So even if you did observe the "birth" of a new creole in vivo, you'd have to be very careful with any generalisations.

Good point. Whilst they may be different factors going on (like the social issues associated with slave trade), I think the fundamental pattern the process takes should be the same though.

katja z wrote::nod: Language evolution is one of the areas where the concept of memes and memetic evolution seems very relevant. I haven't really looked into this in any detail, but it's tempting. Too much interesting stuff, too little time in life!

I know what you mean. The concept of the meme is a troublesome creature though. As an analogy, it's fine, but as a scientific concept it is sorely lacking in substance. Like I say though, the general process that it's trying to describe is a real phenomenon, it just doesn't work the way Dawkins and co thought.

katja z wrote:This is very probable. Not all people who have thought about language are used to rigorous observation and thinking. And there are still a lot of myths around in the humanities and social sciences. I fell for some of them in my time, but I have seen the light.

To be honest, I think some parts of the humanities and social sciences are entirely composed of myths..

katja z wrote:

Ah yes, I think I've seen him around. Devilishly handsome, if I recall correctly.

Ah, so you do know him

Cor yeah, he always gets all the ladies! He's so dreamy... *sigh*

Mr.Samsa wrote:

katja z wrote::nod: Language evolution is one of the areas where the concept of memes and memetic evolution seems very relevant. I haven't really looked into this in any detail, but it's tempting. Too much interesting stuff, too little time in life!

I know what you mean. The concept of the meme is a troublesome creature though. As an analogy, it's fine, but as a scientific concept it is sorely lacking in substance. Like I say though, the general process that it's trying to describe is a real phenomenon, it just doesn't work the way Dawkins and co thought.

As an analogy, it's very good, and offers useful ways of talking about language evolution (but I haven't been convinced by attempts to apply this concept to material culture).

To be honest, I think some parts of the humanities and social sciences are entirely composed of myths..

Sadly, you are right. But let's leave this for a different thread ... or maybe not

Cor yeah, he always gets all the ladies! He's so dreamy... *sigh*

Poor you. Life really isn't fair, is it.

ETA: Where's natselrox? He promised to play Pinker's advocate but he hasn't been near this thread since.

I'm watching.

natselrox wrote:I'm watching.

From under your chair? I didn't realise this looked so dangerous!

katja z wrote:But what Bickerton's claims really hinge upon - and here things get more problematic - is the perceived structural similarities of creoles. Now I haven't read Bickerton's books, so I don't know what his research corpus was. If he only used the Caribbean creoles (I'm guessing here, but I've seen his work quoted in relation to those) which have very similar histories, not only in terms of the language material they are based on but also the social context of their development, then the rebuttal you quoted certainly stands:

it seems unsurprising that creoles would share features with the languages they are derived from and thus look similar "grammatically."

If, on the contrary, he can show structural similarities between these and creoles based on languages from completely different language families, say a Chinese-based creole, this would get more interesting - although it still needn't mean that there is anything innate to grammar, it could be that there are just so many ways of effectively organising linguistic material, especially when you take into account the constraints of language evolution.

Assuming that the similarity covers all creoles there still exists an explanatory option: the cultural similarity of pidgins. AFAIK, most pidgins were born for trading purposes and everyone's conceptual world consisted more or less of operations of trading. So, if the original languages weren't linguistically similar and the corresponding pidgins were just simplified subsets of their origins but the creoles show more linguistic similarity than both of the former phases, what other logical conclusions could be drawn?

If there were an innate, evolved language faculty and the similarity of creoles were based on that faculty, wouldn't creoles then be kind of universal stem languages? If that were the case, wouldn't it be noticed?

Rilx wrote:
Assuming that the similarity covers all creoles there still exists an explanatory option: the cultural similarity of pidgins. AFAIK, most pidgins were born for trading purposes and everyone's conceptual world consisted more or less of operations of trading. So, if the original languages weren't linguistically similar and the corresponding pidgins were just simplified subsets of their origins but the creoles show more linguistic similarity than both of the former phases, what other logical conclusions could be drawn?

Ummm, yes, pidgins are typically trading languages. But don't forget that a) many, probably most pidgins don't develop into creoles and b) many of those that gave rise to creoles, and to some of the most significant creoles at that, were not trading languages - specifically, the creoles of the Caribbean and the Indian Ocean arose among the slave populations in plantation colonies. It's also not entirely clear to me how the similarity in the social function of trading pidgins would translate into grammatic similarities.

If there were an innate, evolved language faculty and the similarity of creoles were based on that faculty, wouldn't creoles then be kind of universal stem languages? If that were the case, wouldn't it be noticed?

Sorry, you've lost me there. What's a stem language?

katja z wrote:Ummm, yes, pidgins are typically trading languages. But don't forget that a) many, probably most pidgins don't develop into creoles and b) many of those that gave rise to creoles, and to some of the most significant creoles at that, were not trading languages - specifically, the creoles of the Caribbean and the Indian Ocean arose among the slave populations in plantation colonies.

Well, pidgins which didn't give rise to creoles are not significant in this case. I see "trading" in broader sense, including production, shipping, etc - where pidgin-speakers were used as labor force.

It's also not entirely clear to me how the similarity in the social function of trading pidgins would translate into grammatic similarities.

That was my key point. Language - grammar included - reflects culture. Language is not innate as a "language instict" and no "language aquisition device" as an organ in brains exists. That's why the social function of pidgins can translate into grammatic similarities. If you follow my logic, I conclude that instead of being instinctual, language is subject to social functions. I.e., I turn Bickerton's logic upside down.

If there were an innate, evolved language faculty and the similarity of creoles were based on that faculty, wouldn't creoles then be kind of universal stem languages? If that were the case, wouldn't it be noticed?

Sorry, you've lost me there. What's a stem language?

If language were an evolved instinct, all languages would have diverged from a hypothetical "stem language". IMO Bickerton reasons that creoles are this kind of languages; because of similarity they are not developed from pidgins, pidgins only catalyze their rise from the "pure instinct".

Rilx wrote:

katja z wrote:Ummm, yes, pidgins are typically trading languages. But don't forget that a) many, probably most pidgins don't develop into creoles and b) many of those that gave rise to creoles, and to some of the most significant creoles at that, were not trading languages - specifically, the creoles of the Caribbean and the Indian Ocean arose among the slave populations in plantation colonies.

Well, pidgins which didn't give rise to creoles are not significant in this case. I see "trading" in broader sense, including production, shipping, etc - where pidgin-speakers were used as labor force.

Mmm, I see what you're saying, but we still have two different situations - in one, speakers are cut off from their original linguistic communities and a pidgin is the only communication tool they have (think plantation colonies), in the other, the pidgin is an additional communication tool for specific purposes, but at least part of the speakers' social life is done in other languages.

It's also not entirely clear to me how the similarity in the social function of trading pidgins would translate into grammatic similarities.

That was my key point. Language - grammar included - reflects culture.

This is very vague. How does language reflect culture*? What in culture is reflected by which aspects of language? It is trivially true that there's a strong link between the two - as in, language encodes, in its vocabulary, the information (in a broad sense) relevant to a particular culture. But grammar? Can you give me an example of a basic grammatic feature in English that can be explained from the culture, not from its linguistic evolution from the (hypothetical) Indo-European?

*A more precise wording would be that language is part of culture, it's not something that comes after culture to reflect it.

Language is not innate as a "language instict" and no "language aquisition device" as an organ in brains exists. That's why the social function of pidgins can translate into grammatic similarities. If you follow my logic, I conclude that instead of being instinctual, language is subject to social functions. I.e., I turn Bickerton's logic upside down.

Agreed, but I would add linguistic evolution (see above). Social environment is a definite factor, but it can only act on available linguistic material (both vocabulary and structures - of course, in pidgins, the latter are reduced to a minimum, which is what makes the development of creoles so interesting).

If there were an innate, evolved language faculty and the similarity of creoles were based on that faculty, wouldn't creoles then be kind of universal stem languages? If that were the case, wouldn't it be noticed?

Sorry, you've lost me there. What's a stem language?

If language were an evolved instinct, all languages would have diverged from a hypothetical "stem language". IMO Bickerton reasons that creoles are this kind of languages; because of similarity they are not developed from pidgins, pidgins only catalyze their rise from the "pure instinct".

Good point, yes, logically they should point to a hypothetical "base" language. There's just a small practical problem - nobody is clear on what such an original language would have looked like (higher - or possibly deeper - levels of analysis in generative grammar are so abstract that I don't know if one could generate any useful predictions from them), so we don't know what we should notice, even if it were staring us in the eye.

Going a bit OT: As for the "original" language, we don't know how soon (or late) it diversified, and closer to the topic here, what its level of grammatic organisation was when this happened. So we don't really know what in our languages is the common memetic heritage, what are the results of convergent evolution etc. Reconstructions can only go so far into the past, and as far as I know, further for vocabulary than for grammar. We also know that grammar can change very rapidly (look at classical Latin and then at modern French and compare, for example, their system of tenses; most of the change happened in a bit over a thousand years, practically overnight!). Frankly, I've no idea how an "innate grammar machinery" is supposed to keep up - unless it refers, very trivially, to our ability to learn to recognise and produce grammatic sequences, which is obviously true but not very helpful.

Just reading Louis-Jean Calvet's fascinating book on linguistic ecology (tr. in English as Towards an Ecology of World Languages, part of the first chapter available on Google Books) and I found this passage on language evolution and the principles and parameters hypothesis*:

The fact that ancient languages, those monuments of complications and irregularities, show a regular tendency to become more regular (to "simplify" themselves), and that the same tendency is shown when a language spreads, suggests that there is, at the origin of languages, a large proportion of improvisation. We have the clear impression that communicational practices have, in the course of centuries, "polished" imperfect products. But the "principles and parameters" hypothesis is not very good at explaining diachrony or the obvious fact that languages change: was the innate grammar modified, did the parameters vary, or do they actually lead to evolution?

(pp. 20-21, my bold)

*The "principles and parameters" framework is a more rigorous formulation of the idea of universal grammar; the idea is that the innate universal principles define all possible grammars and "exposure to language merely triggers the parameters to adopt the correct setting" (wikipedia). The problem, of course, is that this implies that all possible grammars were defined in advance ... um, but by whom? Did Mother Nature have a plan for us?

The criticism section in the linked wikipedia article is very good as well, highly recommended!

Language evolution theorist, Terrence Deacon notes that it is logically problematic to consider language structure as innate, that is, as having been subject to the forces of natural selection, because languages change much too quickly for natural selection to act upon them.

It also mentions that various critics have pointed out that there are, in facts, conspicuously few grammatical universals. So what is this about the universals? I had suspected for some time that there might be confirmation bias at work here, and it would seem that my suspicion was not completely unfounded ... Calvet mentions an amusing example where Pinker saw a common principle, gender, at work in English and Kivunjo. Such a finding would be interesting indeed, since the languages belong to two completely distinct language families (Indo-European and Bantu, respectively). But Calvet shows that this perceived similarity comes from the fact that certain phenomena in Bantu languages had already been baptised "gender" by other Anglophone linguists (a less confusing name for this is "noun classes"), so this is a case of "semantic artefact" rather than an expression of one and the same universal principle regulated by universal parameters. (Pp. 16-17 in the English edition, but p. 17 is unfortunately not included in the preview on Google Books).
(Calvet himself gives only one linguistic universal - the linearity of language, and that is for the simple reason that we can only produce one sound at a time.)

My English teachers always told me that English came sort of instinctively to me. When I had to use different tenses especially or expression. I watched a lot of cartoons (Cartoon Network) before I started to learn English, in high school, and stuff just came to me without knowing the rules. It's still that way with tenses. I just say it and most of the time it's correct. I don't care about rules. They're just there to confuse people and to mess with what they already know.

katja z wrote:

Rilx wrote:That was my key point. Language - grammar included - reflects culture.

This is very vague. How does language reflect culture*? What in culture is reflected by which aspects of language? It is trivially true that there's a strong link between the two - as in, language encodes, in its vocabulary, the information (in a broad sense) relevant to a particular culture. But grammar? Can you give me an example of a basic grammatic feature in English that can be explained from the culture, not from its linguistic evolution from the (hypothetical) Indo-European?

*A more precise wording would be that language is part of culture, it's not something that comes after culture to reflect it.

As to "reflect", I'm not satisfied in my wording but I'm not satisfied in your suggestion either. I just can't find better word.

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.