Language Acquisition



This section of your review sheet covers the area of language acquisition. In broad strokes, it can be broken down into two parts. The first reviews general notions of how language acquisition proceeds. That is, it reviews theoretical positions that people have staked out regarding the acquisition process. The second part will focus more closely on particular landmarks in the acquisition process--landmarks that seem to be reached by all people regardless of the language being acquired.
 


Language as Innate


 


One of the themes that we have returned to numerous times throughout the course is that language is, as Pinker states rather concisely, an instinct. The idea is that humans are genetically predisposed to learn language (though as we've seen, this doesn't mean we're predisposed to learn any particular language). If we want to say that language is an instinct, however, it is important that we draw up a set of criteria regarding what we mean when we call something an instinct. Basically, saying that something is an instinct is a way of saying that something is an innate behavior, a.k.a. a biologically controlled behavior. This is different, for example, from saying that something is a learned behavior, like driving a car or playing the clarinet. Eric Lenneberg has drawn up a set of criteria for determining whether something is a biologically determined behavior.Here are the main points, and some musings about how we can see that language seems to fit the bill. I list them here as they are given to you in Language Files, (8th ed.) p. 259.

  • 1) The behavior emerges before it seems necessary.
  • 2) Its appearance is not the result of a conscious decision.
  • 3) Its emergence is not triggered by external events (though the surrounding environment must be sufficiently rich for it to develop adequately).
  • 4) Direct teaching and intensive practice have relatively little effect.
  • 5) There is a regular sequence of "milestones" as the behavior develops, and these can usually be correlated with age and other aspects of development.
  • 6) There is likely to be a critical period.
  • Language development in humans seems to exhibit these properties, some of which we'll look at a little more closely below. Briefly, let's note why.
  • 1) The behavior emerges before it seems necessary. Language does appear to emerge before it is necessary to the extent that we begin to develop and use it before we need it to fend for ourselves, i.e. while our every need is still being taken care of by our primary caregiver(s).
  • 2) Its appearance is not the result of a conscious decision. Children don't decide to start acquiring language. Language simply develops in them. Compare this to deciding whether or not to join the track team, or deciding whether or not to take an elective, or deciding whether or not to learn golf.
  • 3) Its emergence is not triggered by external events (though the surrounding environment must be sufficiently rich for it to develop adequately).What this means is that language doesn't emerge because adults suddenly decide that it's time for "language school" or organized "language play sessions". The only thing necessary is that kids live in a world of language. What I said in class, I think, is that we can think of it metaphorically as if we're born ready to sing and need simply to be exposed to the song.
  • 4) Direct teaching and intensive practice have relatively little effect. This is true. Correcting children and drilling them on grammar won't make them learn language any faster. Compare this to say, piano lessons. Without the drills and intense lessons, it is very difficult to learn to play the piano.
  • 5) There is a regular sequence of "milestones" as the behavior develops, and these can usually be correlated with age and other aspects of development.We'll look at this more below, but in broad strokes, this is true. Kids acquire language systematically, and cross the same basic milestones at roughly the same time regardless of the language that they are speaking.
  • 6) There is likely to be a critical period. This means that there is likely to be a period during which language acquisition is possible and after which language acquisition becomes highly unlikely, if not impossible. The idea is basically that there is a window of opportunity for language acquisition. This notion is generally accepted in one form or another by most linguists, though there is much debate regarding the issue. Your book notes that there are thought to be TWO critical periods. The first is hypothesized to last from birth to about two years old. During this time, it is hypothesized that if the child is not exposed to language at all, the child will never be able to gain "native" mastery of a language. The idea is that during the first couple of years of life, the language acquisition process has particular consequences for brain development and if the period is missed, the brain will never develop the same structures later. The second "critical period" has more to do with the issue of second language acquisition. Here, the basic idea is that after puberty, it is impossible (or nearly impossible) to learn a second language with "native" mastery. This is why, for example, that second language learners never seem to shed their accents.
  • Hard, definitive evidence for the critical period hypothesis is hard to come by. Why? Well, imagine how well suggesting an experiment in which babies were isolated at birth for 1, 2, 3, 4, 5, 6, 7, 8, 9, 10 ... years would go over with human subjects committees on campuses and in other research centers around the world. But there have been some cases in which children have been discovered after long periods of linguistic isolation. Language Files mentions two such cases: 1) the case of Genie who was isolated until she was roughly 14 (we saw her in class on a video) and 2) the case of Isabelle who was isolated from spoken language (though may have been exposed to sign language) until the age of six and a half. Genie was never able to acquire language with anything even resembling the fluency that native speakers exhibit. By contrast, Isabelle caught up to her peers within about two years. Though these cases are not definitive evidence, they are suggestive and have been taken as evidence that there is a critical period. Note that Isabelle's case would seem to go against the idea that there is a first critical period that ends at two years, unless, of course, she was exposed to sign language and thus underwent the kind of brain development that is hypothesized to occur during that stage. Your book is not clear on all the details here, but the question would be an important one to ask. One final note that does seem to support at least the second critical period hypothesis is that young children do not seem to have great problems acquiring multiple languages if they are brought up in a bilingual or even trilingual environment. This contrasts with the difficulty that adult language learners have when trying to acquire a second language, even if they are immersed in the language that they are trying to learn. This difference is predicted by the critical period hypothesis, but is somewhat of a mystery if we don't assume that there is a critical period for language acquisition that ends sometime around puberty.
  • Theories of Language Acquisition

    Okay, we jumped the gun a bit above and previewed the punch line by saying that language is an instinct and not a learned skill like playing the piano or riding a bicycle. But how do children acquire language? Your text book mentions three general approaches to the problem, i.e. three general theories of language acquisition. These are:

    Which one do you think is the best? Given that we started the course off talking about how grammar emerges in children, i.e. how children seem to be hard wired to be little grammar building devices, you probably can guess that the one we're going to argue in favor of is the Active Construction view. But let's briefly consider each so that we can get a sense of how each model views the problem and what each model predicts.

    Imitation Theory

    Imitation theory basically says that children learn grammar by memorizing the words and sentences of their language. Before attacking this view, let's make sure we recognize that language acquisition MUST involve a lot of memorizing. There's simply no way to get around this problem. Clearly, children must hear the words (or see the signs) of their language in order to go about committing them to their mental dictionaries, and clearly English children learn English because they are getting English input, but memorizing alone is not adequate. Here are some problems with the imitation perspective:


    Reinforcement Theory

    The basic idea here is that children learn to speak like adults because they are taught to do so by being praised and otherwise rewarded for doing things right. At the same time, they are helped because parents "correct" them when they make mistakes. There are two major flaws in this theory.

    The bottom line is this. If reinforcement theory is really at the root of language acquisition, why does it seem to a) not even be what parents do most of the time, and b) hardly effective, if effective at all, when used for correction?
     

    The Active Construction of a Grammar Theory

    This brings us to our third (and preferred) candidate. As Language Files puts it, this theory maintains that children "invent" the rules of grammar for themselves. Now, this is not to say that the grammar that they "invent" isn't based on the language around them. But, the crucial point is this. Only this approach explains how children do things like produce incorrect past tense forms such as "goed" or " hitted", or how they generate novel sentences that they have never heard, or why they seem almost impervious to correction from adults.

    Coming back to the question of forms such as "goed", the issue is actually doubly interesting. Why? Because kids seem actually to acquire some irregular past tense forms and use them pretty early. Then, they often seem to "lose" them, coming out with things like "goed". And then, after a while, they start using the correct form again. What's happening? This is what seems to be the case. At first, the child does memorize a form like "went", by hearing it and associating it with some activity and committing it to her or his lexicon. Then, however, the child's grammar gets to a point where a regular "rule" of past tense formation is solidified. At this point, the child over generalizes and applies the rule to the word "go" to get "goed" and this regular pattern overwhelms the irregular form. Finally, the child realizes that "went" is an exception to the past tense rule and again modifies her or his grammar accordingly by marking "go" as a verb that doesn't undergo the regular process of past tense formation and by reactivating "went" as the past form. Note that imitation and reinforcement have a tough time accounting for this type of phenomenon. Basically, in fact, they can't.

    Finally, though Language Files does not mention it here, think about how this approach makes creolization seem natural and expected. Think about children growing up with speakers of a pidgin language as their primary caretakers and thus as their primary source of linguistic input. If children are hard wired to be grammar builders, creolization makes perfect sense. They develop a grammar out of the language input they have around them. If imitation theory is right, however, then we expect that creolization should never happen. Children should simply acquire the pidgin they are exposed to by repeating what they hear.

    Important Aspects of Acquisition

    In this section, we'll cover some general aspects of phonetic/phonological, morphological, syntactic, and semantic development in child language acquisition. (See how these basic core areas keep cropping up!).

    Phonetic acquisition

    One of the first noticeable aspects of language acquisition is development in the phonetic realm. We know from our discussion of phonology in general that learning a language must involve learning to make sense of the speech signal by figuring out, for example, what the phonemes of one's language are and how these phonemes are realized as allophones in different contexts. This task is actually just the tip of the iceberg, but it is important to recognize that the child must develop a phonological grammar from the data that she or he is exposed to. Acquisition starts very early (perhaps in the womb!). Babies actually vocalize right from the time they are born, especially by crying. Within a couple of weeks, babies produce vocalizations that are limited to gurgling and cooing. This gurgling and cooing gradually changes into what linguists call the babbling stage at about six months of age.

  • Babbling: the stage when babies produce non-meaningful sequences of consonants and vowels. These are most often in the form of CV syllables repeated over and over, such as [mama] and so forth. In a way, the babbling stage can be thought of as a period of time in which the baby is practicing the difficult sequences of coordinated gestures (jaw and tongue and velum movement, along with voicing control) that are necessary for normal speech production. These seem easy to us, because we are grown up (at least some of us!) and have already acquired our language and speak with ease, but the gestures necessary for speech production are, in fact, quite intricate. At the babbling stage, there is no clear association of sound and meaning on the part of the baby. That is, the baby doesn't seem to be saying or trying to say particular "words". What is interesting to note is that at the babbling stage the baby does seem to be picking up on the rhythm and intonation of the language that it is acquiring.
  • Phonological acquisition

    Babbling gives way to a stage in which words do emerge. When babies come out with their first words, there appears to be a link between sound and meaning that is missing at the babbling stage. This happens at around 18 months of age. Babies seem at this point not to be aware of phoneme level differences in words. That is, they seem to memorize single words as chunks and not pay much attention to phonemic differences. So, the word for "bottle" might be both "ba" and "da". In adult speech, [b] and [d] are allophones of different phonemes and can be found in many minimal pairs. Babies aren't aware of this at first. They haven't developed a grammar that makes this distinction at this point.

    Though we didn't go over it in much detail in this class, babies proceed from this initial "first words" stage to a stage in which they begin to develop an awareness of phonemic contrasts, i.e. of the important differences signaled by [b] as opposed to [d] in a language such as English. So, they begin to acquire particular sounds. Interestingly, they don't get all their sounds at once. Some sounds are more difficult to master than others. Bilabial sounds like [m] and [p] are among the first sounds mastered. Consonants like [l] and [r], which have a large number of phonetic similarities, are the most difficult to master. In fact, some children have persistent problems in English with their [l] and [r] sounds well into childhood. In terms of syllable structure, consonant clusters are harder to master than simpler CV syllables, and children thus acquire CV structures before they get clusters such as [str] or even [st].

    Another thing children tend to do is shorten longer words. Examples are when they say 'nana' instead of 'banana' or 'tato' instead of 'potato'. What's interesting is that these deletions aren't random. Kids don't just shorten any word in any way. As your book points out, what gets deleted is an unstressed syllable that immediately precedes a stressed vowel. It is as though the baby has cued in on the stressed and most important syllable of the word and so starts the word off with that. In fact, many English words have this pattern, where the main stress falls on the first syllable. (For example, all of the words in the last sentence have stress on the first syllable!) What's interesting is that although babies don't have the grammar of the adult, they ARE developing a grammar simultaneously. It's not simply that they are incompetent and can't say 'banana'. Rather, they vary from adult speech in a systematic way. This again points to support for the "Active Construction of a Grammar" theory.

    Morphological and Syntactic Acquisition

    At the same time that kids are busy acquiring phonemic distinctions and working hard on their consonant clusters, they are also acquiring morphology and syntax. At the one-word stage that we referred to above, the first words a child utters are usually familiar objects or people in their world. Pretty soon they also start to pick up on activities and come up with words like verbs and other important concepts, such as "mine" and "gimme". (Isn't it interesting how some of the core properties of human nature emerge right away!!) Anyway, from a syntactic perspective, the one word stage is called the holophrastic stage.

    After the one word or holophrastic stage, children enter the Two-Word stage. This stage tends to combine important words with crucial semantic relationships. Your book lists a bunch of examples:

    Note that at this stage, the grammar is not the same as the adult grammar (this is a problem, remember, for imitation theory. How come children go through this stage if they don't get such sentences as input?) However, even though the word order isn't the same as the adult grammar, it is relatively consistent. And they control their intonation patterns more. This is more evidence for the active construction position. Here we have a picture of little kids as little grammar builders. The grammar may be simple at this point, but it still looks like a grammar and not like chaos. The two word stage, as you can see from the examples, has the feeling of being telegraphic. Babies go through this stage sometime between 18 and 24 months, depending on when they begin the one word stage.

    After the two word stage, there isn't really a three word stage. Babies' speech is still telegraphic, but as time goes by, they gradually start to move beyond just using content words (like nouns and adjective and verbs without the inflected endings) and begin to acquire function morphemes like inflectional affixes. Again, these affixes seem to be acquired in a regular fashion, with some appearing before others in case after case. For example, in English speaking children, the progressive suffix -ing, as in "go-ing" appears much earlier than the 3rd person -s suffix, as in [walk-s].  The big picture is that this regular order of acquisition of function across English speaking children (children of other languages exhibit similar effects in their development) speaks well for the Active Construction of a Grammar theory.

    Your book focuses on three aspects of morphological and syntactic development that are useful to review. These are, the development of plurals, of negatives, and of interrogatives. (see pp 273-274) Here are big points to bear in mind.

    Plurals. The plural suffix is acquired pretty early on in the game. But, as is the case in the acquisition of the past, children go through a period when they regularize irregular plural forms, in essence, overgeneralizing the use of the regular plural. Why is this important. Again, it points to the fact that kids are acquiring language through building a grammar, not through simple rote memorization. Again, this is evidence for the Active Construction position.

    Negatives. The acquisition of negatives also shows evidence of grammar building in children. First, children just seem to slap no down in front of a sentence to negate it, as in "no I drink juice". Anybody who has ever been around children, in fact, knows that the word NO appears with alarming frequency in their speech. Anyway, they start refining things by inserting negative words in their sentences, as in "baby no sleep" or "baby can't drink". But, at this point, they don't seem to be conscious of the fact that "don't" is actually morphologically complex, consisting of a contraction of "do" + "not". The same goes for things like "can't" and "won't". We know this because the auxiliary verbs "can" and "will" and "do" aren't acquired until later on! Over time, children refine their use of negation, gradually closing in on the adult grammar. One of the last things to be perfected in the use of words like "anybody" and "anything" in sentences like "I don't want anything" which are often realized as "I don't want something" in children's speech.

    Interrogatives. Acquiring interrogatives is also a systematic business that seems to proceed through some regular stages. At first, children seem to ask questions by making use of rising intonation, e.g. more milk?, Mommy go?, etc... At about three years old, children begin to use auxiliaries in yes-no questions, and they put them in the right order, e.g. "can I have more?" But, at this point, they still don't show adult order in Wh-questions. They generally use regular declarative order after the Wh-word, as in "Mommy, why you are mad?" instead of "Mommy, why are you mad?" This is interesting, because at this point they have undergone a rapid growth in vocabulary and are in control of the word order inversion necessary for yes-no questions. So, what does this point to? Once again, we see that kids are going through stages of grammar development. They are not simply imitating what they hear. IF this were so, they would not go through a stage where they use the Wh-word at the beginning of the question but fail to invert the subject and the auxiliary verb--something that we know they CAN do since they DO this in yes-no questions.

    Semantic Acquisition

    We'll end our discussion of child language acquisition with some quick notes on semantic acquisition. Of course, when children hear a word for the first time, they have no idea what it means. Think about getting off a plane in the middle of Nigeria and having someone walk up to you and loudly declare "Yetunde!" Well, babies are a lot dumber than we are, so you can imagine the task that they have. (By the way, Yetunde is the name of a friend of mine from Nigeria...so, even if it's a name, at least it's a real word...what language? Yoruba. Okay, back to the task at hand.)

    The literature on how children acquire meanings is actually quite extensive and various theories have generated debate for years. But, we do know that children basically have to make guesses about what words mean (remember that they have to do this while they are also working on the task of identifying individual words in the speech stream). They make their guesses from being exposed to words that are used in particular contexts. A common case is that a child is exposed, for example, to a poodle in the street and sees mommy or daddy point to it and say "doggie". Well, even if the kid associates "doggie" with that object, there are a number of routes that the child can logically take. The child might see a squirrel and say, "doggie", clearly having taken the word "doggie" to mean something more general that it does in the adult grammar. Or, the child might see a dog, say a labrador, and not say anything, failing to associate the word "doggie" with anything beyond poodle-ness. This trial and error process seems to be how children acquire word meanings. Though it might seem cumbersome, your book points out that children tend to have a vocabulary of about 14,000 words by the time they are six! Did you think that a six year old had that large of a vocabulary?

    The two concepts that I've illustrated for you with the "doggie" example have names. One is called overgeneralization, while the other is called underextension.

    Overgeneralizations. Overgeneralizations are cases in which a child gives a word a broad range of meaning that it has in the adult grammar. So, in our example above, the child incorrectly generalized the word "doggie" to include squirrels and perhaps other furry animals that make up part of his or her world. Your book gives the example of a child who overgeneralized the word "fly" to include specks of dirt, dust, small insects, and bread crumbs. What is important to note here is that overgeneralization is not simply random. Children overgeneralize along reasonable semantic grounds. The kid who overgeneralized "fly" was certainly keying in on the importance of size. All of the things are similar, speck-like entities.

    Underextensions. Underextensions have the opposite effect. In this case, it might help to think of them as cases in which the child is being overly cautious. The child associates a word with a smaller set of meanings or objects than it has in the adult grammar. I like the case that your book notes of the child who called a ball a ball only when it was under the sofa. One of the explanations for underextension that people have proposed is that children, especially very young ones around two years old, can't really tell the difference between inherent properties of things and accidental ones. In the case of calling a ball a ball only when it is under the sofa, the child is failing to realize that being under the sofa has nothing to do with the inherent "ball-ness" of the ball. Being under the sofa is an accident, i.e. not a quality of the ball at all. Older children do exhibit underextension as well, especially when dealing with a new item that may or may not be a member of a category whose boundaries are not overly clear. So, if I point to an olive and say "that's a fruit", you may not agree. You've essentially underextended the word "fruit" since an olive is, indeed, a fruit. I had this experience when I was learning Spanish. For me, as an English speaker, I don't generally include insects in the category of "animals". I realized that this wasn't true of Spanish when a friend of mine once discovered a rather large beetle lounging in his hair and jumped up screaming "Un animal!" Of course, I started looking around for a large rat, mostly due to the look of pure terror on his face. For him, beetles and other insects enter into the category animal. For me, they didn't. Moral of the story...that's a case of underextension which in this case took place in my acquisition of the word "animal" in Spanish. So, I now have one "animal" in my English lexicon and a different one in my Spanish lexicon.

    Okay, this wraps up our review of Language Acquisition. Don't forget to review the milestones in File 9.4  (8th ed.) in your book. They give a nice encapsulated view of various stages of language acquisition, along with milestones for other types of motor development in children. For anyone interested in developmental issues, both linguistic and motoric, this is a handy chart.