Language and human nature

Author: Ray Jackendoff
Date: 1994


1   The fundamental arguments


1.2   The argument for mental grammar

Eventually, all normal human children are able to speak his or her local language.

Claim: Parent teach grammar to their children. Refutation:

  • Not parents. Parents teach words to their kids, but language learning is not entirely the result of teaching words. (e.g. "from")
  • Not school. Many non-literate societies. Certain grammatical patterns are taught (e.g. preposition can't end a sentence) but it's violated all the time. (That rule is based on notions of Latin and Greek.) Further, may apparent rules are _not_ taught. (e.g "manu-goddam-facturer" uses a rule that specifies where words can be injected)

Since adults are not consciously aware of the principles of mental grammar, they certainly cannot explain them to children- the most they can do is supply examples of patterns (in the form of grammatical sentences) or corrections to ungrammatical sentences (which do not explain rules, but provide correct form). This means children must figure out the patterns of the language.

How does a child construct his or her mental grammar?

Children are probably no more conscious of the patterns than adults. For one, they don't formally learn grammar until after they have learned how to speak.

Where does that leave the learning of language?

Conclusion: We can acquired unconscious patterns unconsciously, with little or no deliberate training. (Arguably, "learning" might not be the right word.) We might draw a parallel to learning how to skip or ride a bike.

1.3   The argument for innate knowledge

The way children learn to talk implies that the human brain contains a genetically determined specialization for language.

Jackendoff describes the inability of highly trained linguists to duplicate the feat of a unconsciously aided child the Paradox of Language Acquisition. (Neither have computer scientists.)

There are three steps involved in escaping the Paradox:

  1. What the child ends up with is a mental grammar that is completely inaccessible to consciousness. Hence adult linguists can't figure out the principle to mental grammar by just looking into their minds.
  2. A substantial part of the language-learning process is also unconscious, so linguists can neither directly observe it nor ask children about it.
  3. To overcome this difficult, we suppose that children have a head start on linguistics: children's unconscious strategies for language learning include some substantial hints about how a mental grammar ought to be constructed.

Perhaps language acquisition is like reproduction; people figure to how to reproduce, but it takes centuries to understand the mechanism.

More technically, the claim is that all of us as children come to the task of language learning equipped with a body of innate knowledge pertaining to language. Using this knowledge, children can find patterns in the stream of language being beamed at the them from the environment, and can use these patterns as a mental grammar. Because this innate knowledge must be sufficient to construct a mental grammar for any of the languages of the world, linguists call it Universal Grammar or UG.

This leads to three important questions for research:

  1. What do children know (unconsciously) about language in advance of language learning? What is Universal Grammar?
  2. How do they use Universal Grammar to construct a mental grammar?
  3. How do they acquire Universal Grammar?

We focus on now on (3): How could there be such as thing as innate knowledge (knowledge that is not learned)?

  • We have to remember that UG is as unconscious and inaccesible to introsepction as the final mental grammar we use.
  • "Innate" is also used loosely, as it's not clear its present immediately at birth. It could develop afterward.
  • Development of UG seems to follow a biological timetable. Children being acquiring grammatical patterns sometime toward their second birthday.
  • UG is not learned; UG makes learning possible.
  1. amounts to: How can knowledge or cognitive organization be available to a child before learning? How could there be such a thing as innate knowledge?

The outlines of a mechanism behind innate knowledge are available. Two components are involved:

  1. The determination of brain structure by genetic information
  2. The determination of mental functioning by brain structure

1: Although there is plasticity in the brain's physical organization, there is good reason to believe that substantial aspects of this organization are genetic. Why should we suppose that our brains acquire their fundamental structure through learning rather than genetic inheritance? 2: The way we think is partly constrained by the way our brains are built. UG might be rephrased as saying that children have a certain "way of thinking" that enables them to construct a mental grammar, given appropriate inputs in the surroundings. The hypothesis is that this "way of thinking" is a consequence of the physical organization of the brain which is determined by genetic structure. The mechanism for acquiring innate knowledge is genetic transmission, through the medium of brain structure.

We can call this hypothesis the Genetic Hypothesis. It leads to range of issues.

It says the ability to learn language is rooted in our biology, a genetic characteristic of the human species, like opposable thumbs and a pelvis adapted for upright stance.

It hardly seems outlandish to think there might be such a specialized organ for language given the specialization of observable structures in animals - a bat's sonar, an elephant's trunk.

According to the Genetic Hypothesis, innate knowledge is determined by brain structure, so it is present only when the supporting brain structures are present. Development of the physical structure is not complete at birth. The physical growth of various parts of the body follows a predictable timetable (e.g baby teeth, puberty).

  • Need concept of innate knowledge to solve the Paradox of Language Acquisition
    • that children can unconsciously learn a mental grammar on their own, while linguists as a community can't figure out how the mental grammar works. Innate knowledge would give children a head start
  • Why does Paradox of Language Acquisition arise? Because children don't learn language by being taught, rather they must figure out a MG that gives them the patterns for forming sentences. How do we know children must be figuring out a MG? Because that's what they have to end up with as adults in order to account for their ability to speak and understand an unlimited range of sentences.

### Questions about innate knowledge

The hypothesis of a genetically determined Universal Grammar has provoked various degrees of disbelief since it was proposed by Chomsky.

TODO: Come back to this

### Conclusions

Mental Grammar = Innate Part (Universal Grammar) + Learned Part

We leave it as an empirical question to determine how the mental grammar is parceled out between innate and learned parts.

Three basic criteria are involved:

  1. If the language in question is different from other languages in some respect, the child must be able to acquire this difference, so it must fall into the learned part.
  2. If certain aspects of all languages we have examined are alike, these aspects are _likely_ to fall into the innate part. Of course, there is always the possibility that they are alike purely by accident. In practice, this can be checked out by examining more languages, preferably unrelated ones.
  3. Poverty of the stimulus: Suppose there is some aspect of language that children couldn't possibly figure out from the evidence in the speech they hear around them. Then this aspect can't be learned; it has to fall under the innate part of the language.
  1. inspires debate on what sorts of evidence children are capable of using.

Innate part of language = Part due to special purpose endowment for learning + Part due to general properties of the mind

We should try to minimize the part due to special purpose endowment for learning. Such an endowment demands an evolutionary jump since the time we diverged from the apes, and we would like to think that the jump was not too extraordinary. (It must be though, to account for the Paradox of Language Acquisiton.)

Recall our initial question: What does human nature have to be like in order for us to be able to use language? Two more answers, having to do with the nature of learning, have emerged from the Argument for Innate Knowledge.

  1. The learning of language isn't just a passive "soaking up" of information from the environment. Language learners actively construct unconscious principles that permit them to make sense of information coming from the environment. These principles make it possible not just to reproduce the input parrotlike, but to use language in novel ways. What is learned comes as much from inside the learner as from the environment.

2   The organization of mental grammar


3   Evidence for the biological basis of language


4   Mental capacities other than language