We cannot describe all the theories that have evolved or the cases that have been made for them, but we will examine four broad categories of theory related to language that have had a major influence. They are behaviorist, active construction of a grammar, neural con- nectionism, and social interaction.
Behaviorist Theories As appealing as behaviorism was in the early part of the 20th century, it has little credence as a theory of language acquisition. Basically, behaviorist theories take the position that children learn through imitation. They listen to the speech around them, imitate what they hear, and then through a system of reinforcement (i.e., being praised or rewarded for correct utterances and having errors ignored or corrected), they learn to discard their imperfect imitations. The problems with applying this theory to real children learning language are obvi- ous. First, children produce utterances they have never heard and, second, adults rarely respond to the form of the utterance. No theory of imitation can account for this and simi- lar utterances. See Chomsky’s Case Against Behaviorism for more on this topic.
CHAPTER 3Section 3.3 Theoretical Perspectives
The most serious flaw of imitation as a theory is that it cannot account for how children come to produce or understand novel utterances, whether in the way they pronounce words or in the way they inflect them, or in the sentences they produce. Pronunciation errors are generally attributed to children’s immature articulators (i.e., their physical inability to produce an exact replica of the adult form). Chloe’s lalo for yellow would be assumed to be caused by her articulators not being sufficiently well developed to produce two distinct consonants and two distinct vowels in the same word.
Other kinds of errors are more problematic. It is highly unlikely that Chloe ever heard anyone say, “Nana, you forgotted.” Yet she and all children her age regularly produce sen- tences they could not have heard from anyone else. Even if behaviorism could account for how these forms are created, the theory stumbles on the notion of reinforcement. There is overwhelming evidence that, in general, adults neither negatively reinforce flawed utter- ances nor positively reinforce correct ones. When I responded to Chloe with, “Yes, I sup- pose I did forget,” although I modeled the correct form, my response was to her mean- ing, not her imperfect syntax. This type of response is typical of adult-child interactions: Adults focus on the meaning that the child is trying to make, and any correction is geared toward helping to make that meaning clear rather than to correct the form. More impor- tantly, even if I had (foolishly!) tried to correct Chloe’s syntax, and if she had understood what I was trying to do, it would not have effected any change in her language. Children attend to meaning.
Chomsky’s Case Against Behaviorism Noam Chomsky pointed out the many problems with behaviorism in his 1959 review of Skinner’s Verbal Behav- ior. Although Chomsky’s criticism spoke to first language acquisition, the argument holds for second. Chomsky argued that behaviorism cannot account for the various facts of language acquisition, namely: • Young children acquire language at a remarkable speed. • Children are largely impervious to correction of form. • By the time most children are 4 or 5, they have an almost limitless capacity to understand and produce sentences they have never heard before. • The notion that reinforcement accounts for learning runs counter to experience—a child does not utter the word dog over and over in the presence of adults who reinforce that behavior.
Chomsky speculated that the abilities that account for all of these facts are largely innate, part of the cognitive endow- ment children have at birth. Chomsky, N. (1959). Review of verbal behavior. Language, 35, pp. 26–58. Skinner, B. F. (1957). Verbal behavior. New York: Appleton-Century-Crofts.
Noam Chomsky is an American philosopher, political activist, and linguist. In the 1960s and 1970s, he made an influential case for the existence of a language instinct.
Associated Press
CHAPTER 3Section 3.3 Theoretical Perspectives
The Active Construction of a Grammar Theory Active construction of a grammar theory contends that children use the speech they hear around them to construct the rules of the language by listening for and discovering pat- terns, hypothesizing about the rules that create those patterns, and then testing those rules in their own usage. They are, in a way, “little linguists.” Certainly, there are data which would appear to support such a theory. In the sentence Nana, you forgotted, Chloe had clearly detected that past tense is formed by adding an -ed. Other words that she used at this time give strong indication of the rule that she had created:
Learning Grammatical Morphemes In a longitudinal study of three children, Roger Brown (1973) revealed compelling data showing that the order in which children acquired grammatical morphemes in English was very consistent. Even allowing for slight individual variations, the results showed that children acquired 14 common morphemes in the following order: 1. Present progressive -ing 2. Preposition in 3. Preposition on 4. Plural -s 5. Irregular past, e.g., went 6. Possessive -’s 7. Uncontractible copula be 8. Articles a, the 9. Regular past -ed 10. 3rd person regular -s 11. 3rd person irregular, e.g., has 12. Uncontractible auxiliary be 13. Contractible copula be (It’s Mommy) 14. Contractible auxiliary be (He’s eating.)
Later, de Villiers and de Villiers replicated the study with 21 English-speaking children, and the order they obtained correlated very highly with the children in Brown’s study:
There is a consensus that this research has uncovered a crucial property of first lan- guage development, namely that the acquisition of a number of grammatical features follows an order that is largely the same across individuals. A plausible inference from this insight is that the underlying mechanisms determining this developmental order are identical across individuals and largely unaffected by external influences. (Meisel, 2011, p. 64)
Another compelling kind of evidence against behaviorism is the observation that children learning the same language tend to do so in remarkably similar ways. There is strong evidence, for example, that children acquire certain of the grammatical morphemes of English in a largely invariant order, as we see in Learning Grammatical Morphemes.
CHAPTER 3Section 3.3 Theoretical Perspectives
Sarah holded my doll.
She gived it to me.
I ated all my lunch.
(Chloe 4 years, 6 months)
With forgotted and ated, she had doubled up on her past tense, probably because she did not realize that the verbs were already in the past tense, or perhaps because she had heard the past tenses more often than the present. Eventually, she will discover that her forms do not match the language used by others around her, and she will refine her rules and try again. Chloe’s utterances were governed by her own grammar, which differs in significant ways but also shares important characteristics with adult grammar. Through the testing, modification, and editing of her rules, she will progress gradually, as a linguist does when constructing the rules of an unfamiliar language, until she achieves the same set of rules underlying adult speech.
Children’s language errors present no difficulty for the active construction of a grammar theory. In fact, those who hold this theory do not even see the imperfect forms as errors, but rather as developmental forms, those language forms, differing notably from adult forms, that children produce as they learn the language. They are seen as evidence that children’s learning is rule governed, as exemplars of rules in progress rather than as errors to be eradicated.
Connectionist Theories Connectionist theories hold that children learn language through neural connections in the brain, developing these connections by being exposed to and using languages.
Proponents of connectionist theories would concede that rules are useful for describing the language forms that children produce during the acquisition process, but they would also argue that children play no active role in actually creating those forms. These theo- rists recognize that children bring some significant degree of innate capacity to the task of language learning and would also concede that eventually children’s language is rule governed. Connectionists argue, however, that in the process of acquisition, children are not little linguists working out rules in their heads, but are more like little statisticians, computing probabilities and building neural connections in the brain based on their expo- sure to and use of language. With increased exposure, the connections are strengthened. (See Figure 3.1.)
CHAPTER 3Section 3.3 Theoretical Perspectives
Each brain cell, or neuron, has a sending fiber or axon (right), which is coated in myelin, and a receiving fiber, or dendrite (left). These neuron connections are the basis for connectionist theory.
Source: Hemera/Thinkstock
Two kinds of language learning are often cited in support of connectionist theories. One is how children learn what a word is. The second is their learning of the past tense. How does a connectionist believe a child learns what a word is? In a continuous stream of speech in which there are no overt demarcations between words, how is it that children learn where the word boundaries are? More specifically, how do children, hearing, “Once upon a time in a land far away, a little boy and a little girl lived in a cottage at the edge of the woods,” come to know which parts of that string constitute words? Over time, they will hear parts of that utterance in other contexts—perhaps in a sentence such as, “What is that little boy doing?” or “She couldn’t find her little boy anywhere.” Based on the evi- dence, they might conclude that littleboy is a word, but hearing little and boy separately in other contexts over time, they eventually create a mental representation for little and boy separately. Remember Chloe’s question, “Are we going to your ami today?” She had segmented Miami as my and ami, and according to this theory, she would have done so based on the statistical probability that whatever came after my was another word. In this instance, she was wrong, but connectionists hold that this was not because she had formed the right rule and applied it in the wrong context but that she had not yet had enough experience with Miami in other contexts to form the correct representation (i.e., the place where her nana lived).
Figure 3.1: A neuron
CHAPTER 3Section 3.3 Theoretical Perspectives
Connectionists also use children’s acquisition of the past tense to explain their theory of language acquisition. When children produce blowed as the past tense of blow, these theorists argue it is because they have calculated it to be the most likely past tense based on other words the child has heard, such as mowed, flowed, and glowed. The connection- ists may be right, but how do they account for the fact that some children will produce bringed as the past tense of bring while others will produce brang? If children are actively constructing rules, the theory would predict that they would produce bringed and later learn that there is an exception, brought. If, however, they are working from a probability or statistical basis, the theory would predict that they might produce either form—brang, analogous with rang and sang, or bringed because they had heard other regular past tense forms. In other words, connectionism would predict that some children will get it right and others get it wrong, and this appears to be the case, at least at first. Chloe’s forgotted might be evidence that exposure, or frequency of occurrence, plays a role. She had heard forgot often enough to internalize it as a verb form but had not yet recognized it as a past tense, so she added a regular verb ending because that is the form she had heard on so many other verbs. In fact, it would fit the pattern of spot/spotted, dot/dotted, or plot/plotted. Before any of these rules can be learned, children must first learn words. Let us consider how Sarah learned baby from this theoretical perspective.
One of Sarah’s first words was baby. How did she learn it? According to this theory, she would have heard people using the word baby in several different contexts—female infants, male infants, infants wearing snowsuits, pictures of infants, dolls, and even in reference to herself. Through multiple exposures to the word, she formed neural connec- tions of various kinds—to the word itself, to a mother holding an infant, to pictures, dolls, or possibly just the letter /b/ or people without much hair. Eventually, these connections would coalesce into her mental representation of the word baby. These connections are not all of the same strength, or so the theory goes. If she heard the word baby used more fre- quently in reference to a doll than to a human baby or other referents, then the connection between baby and doll would be stronger than the connection between baby and any other possible referent. If, however, she had heard the word used more often with reference to a human baby, she might well develop a neural connection between baby and human infants (see Figure 3.2).
CHAPTER 3Section 3.3 Theoretical Perspectives
According to connectionist theory, this is how Sarah would learn the word baby.
Source: Thinkstock
Social Interaction Theory Social interaction theory purports that children have an innate predisposition to acquire language and that they develop their own rules. This belief is shared with the active con- struction of a grammar theory. These theorists also believe that children acquire language as a result of interaction with more mature speakers of the language (i.e., older children and adults). On the face of it, this would seem to be obvious and self-explanatory. In fact, these theorists believe that children take a very active role by cuing “their parents to sup- ply them with the appropriate language they need” (Wilson & Mihalicek, 2011, p. 320). This is a truly interactive theory. On the one hand, children need the social environment to improve their linguistic and social skills, and on the other, the language-rich environment exists in large part because it is prompted by children. Social interactionists have observed that the speech older speakers use with children is specifically created to be more learn- able. They may have a point. Consider the following pair of utterances:
Figure 3.2: Example of connectionist theory
Hears: Baby
Sees
Creates Mental Representation
CHAPTER 3Section 3.3 Theoretical Perspectives
a. See the choo-choo? Look at how big it is! It’s going really fast, isn’t it?
b. There is in our immediate vicinity a locomotive, which, despite its impressive size, is traveling at great speed.
Few, if any, adults would use sentence (b) when speaking with an infant. Most people instinctively adjust their speech to what they perceive to be an appropriate level for a young child to understand. This adaptation is called child-directed speech (CDS). CDS has distinct characteristics:
• Different rhythm and pitch. The speech typically has a sing-song quality, and the speaker’s pitch is usually higher. • Simplified vocabulary. Rarely would we use choo-choo, doggie, or potty with adult speakers. • Shorter sentences. The difference between utterances (a) and (b) illustrate this distinction well. • Simplified structure. Shorter sentences usually have simpler structure than long ones. Again, the difference between (a) and (b) demonstrates this principle. • Particularly helpful responses. Adults are more tolerant of and likely to respond to infantile utterances that are barely comprehensible, if at all. For example, when Sarah referred to herself as Ah-Ah, simplifying the sound structure, adults did not say, “Who?” or “Whom are you talking about?” Nor did they try to explain to her that she really should be using “I” to refer to herself and not the third person. Instead, they acknowledged the meaning of her utterance and even began to refer to her in the same way.
Apparently, adults use CDS for two reasons, neither of which is that they are consciously attempting to teach their children language. One is to maximize the likelihood of mean- ingful communication, and the other is to express affection, which might explain why peo- ple tend to use similar speech when talking to their pets (Apel & Masterson, 2009).
There are three problems with social interaction theory: The first is that nobody is certain about how long children need to be exposed to CDS. The sec- ond is that they do eventually understand and use compli- cated sentences such as the one in the previous example; and the third is that CDS may be a cul- tural phenomenon—we do not know whether it is universal, although there is evidence from German, Mandarin, Russian,
All kinds of social interaction help children to learn language. This coach adapts her language to help her young players understand.
Larry Williams Associates/Corbis
duua27
theoretical_perspectives_theorists__chapter_three..docx
0
