Brief Lecture Notes for Unit 8
Part One: Language
What is language? Unless we have a precise definition of language, we can't make much headway in studying language development, nor can we answer questions like "Do infants have language?" or "Do nonhuman animals have language?" One influential approach is to define language in terms of six key components, the idea being that unless a form of communication passes all six tests, it isn't language in a full technical sense of the term.
1. Novelty: the ability to produce sentences you've never heard before, and/or to generate sentences you've never spoken before. Language, in other words, involves much more than rote memory. It involves an understanding of how rules can be applied to combine existing elements in new, yet meaningful and comprehensible, ways.
2. Structure: the fact that language follows defined rules. The use of language elements in a random, unstructured way is not language. Even if a speaker can't formally state what the rules are or are supposed to be (prescriptive rules), s/he is still following implicit or descriptive rules.
3. Arbitrariness: the fact that, in general, the sounds of language bear no logical (or necessary) connection to the meanings of language. This is why, if a given language is foreign to you (and dissimilar enough from the languages you do know), you can't figure out the meaning just by listening to the words.
4. Displaceability: the use of language to distance oneself from the tangible, concrete, immediate environment. This includes talking about the past, the future, contingent events ("perhaps it might rain tomorrow"), purely imaginary or hypothetical entities ("the crown prince of South Dakota"), theoretical abstractions ("historical inevitability"), and so forth.
5. Interpersonality: the deliberate use of language to convey information to someone else. (In contrast, many forms of nonlinguistic communication are primarily inadvertant and unintentional.)
6. Empirical acquisition: the fact that language is not innate (though the capacity to learn language may be), but is learned only through contact with a language community.
Given the connections between language, culture, and cognition, it should come as no surprise that people from different language communities tend to think differently. According to the Whorfian hypothesis, this is an inevitable concomitant of linguistic diversity: the words at our disposal determine what we are able to think, hence the thought-forms of different language systems are incommensurable. Click here for one unforgettable example of this. Translation from one language to another is a very difficult process, and word-meanings cannot easily be mapped from one language system to another.
Language can be studied at four different structural levels:
1. Phonemics is the study of individual speech sounds or phonemes. At this pre-meaning level, phonological rules govern what combinations of sounds can go together (which ones "sound right"). For instance, the words plet and glet sound as if they could be English words, but not the words mlet or tlet. Different languages have different phonological rules, which is one reason it is so hard to learn how to speak a foreign language without an "accent".
2. Semantics is the study of word meaning, or of the meaning units (morphemes) of language. Morphemes can comprise whole words, or parts of words: for instance, the word cats contains two morphemes, cat (meaning a small, furry creature superior to humans) and -s (meaning more than one). Note from this example that some morphemes (like cat) are capable of standing alone as words; these are free morphemes. Others (like -s) can only exist as parts of words, not as words in isolation; these are bound morphemes. Morphological rules indicate how morphemes can be combined to form words.
3. Syntactics is the study of sentence structure (syntax or grammar). There are two basic ways that the relationship between words in a sentence can be conveyed: one is by word order ("John loves Mary" means something different from "Mary loves John"), the other is by inflections or internal word changes (we say "he loves her" but not "her loves he"; we must change the words and say "she loves him" to convey the latter meaning). English relies about equally on these two ways of conveying structural relationships; other languages lean more heavily toward one or the other. The more heavily a language relies on inflections, the more synthetic it is: such languages usually have longer words and fewer words per sentence. Conversely, the more heavily a language relies on word order, the more analytic it is: such languages usually have shorter words and more words per sentence. For instance, an idea that needs four words to express in English ("I am an American") requires only two words in Russian, a more synthetic language ("Ya amyerikanyets" if you are male, or "Ya amyerikanska" if you are female), but nine words in Chinese, a more analytic language ("Wo shr i go mei guo di nan ren" if you are male, or "... nu ren" if you are female.) Some languages are so extremely synthetic that all sentences consist of one huge single word: such languages are called polysynthetic or agglutinative. The outward structure of a sentence (its surface structure) can be contrasted with the inward conceptual meaning (deep structure) that it is trying to convey; when the two do not match, these linguistic ambiguities are often a source of humor (e.g., "Hey, buddy, call me a cab?" "Okay, if you want... you're a cab.") Syntactic rules govern the grammar of speech.
4. Pragmatics is the study of real-world (social) uses of language. For instance, one pragmatic rule is that we do not normally ask questions to which we already know the answer, unless we are in a specialized social role (like a teacher). As a result, when someone asks what sounds like a request for information -- "Can you pass the salt?" -- we know that this is actually something else, namely, a disguised command (the use of the question form is to soften the command so it does not seem so much like a bald assertion of power). A special class of pragmatics has to do with verdictives, or words that (in the right context) actually change external reality (or, at least, social reality) by means of being said. For instance, when a legally constituted authority speaks the words "I now pronounce you husband and wife" to two unmarried persons of appropriate genders, they're married -- like it or not. (So be careful at the wedding rehearsal.)
How might we develop an adequate theory of how children learn language? Clearly, language learning is an active, not a passive, process. It is difficult if not impossible to account for language learning in behaviorist (see Units 1 and 6) terms for reasons such as these:
1. Parents do not consistently reward grammatical speech and punish (or ignore) ungrammatical speech; adults focus on the content of what children say, not the form in which it is said. Yet we all learn (more or less) to speak grammatically: how so, from the behaviorist viewpoint, if the appropriate response consequences are not present?
2. Child language is creative, not imitative: children are constantly making up words and expressions (neologisms) that they have certainly never heard adult models use. Such behavioral spontaneity is difficult to explain from a purely behaviorist standpoint.
3. Language develops in a regular, culturally invariant sequence of steps the world over, which suggests that this sequence is not the product of culture, environment, or training: it must in some sense be genetically hard-wired or have some more complex explanation that goes deeper than the simple behaviorist idea of learning through reinforcement.
4. The existence of a critical period in language learning also argues for a nativist component. If you have not learned a (first or second) language by age 10 or 11, you'll probably find it extremely difficult ever to become fluent, due most likely to neurological changes induced by the onset of puberty that greatly minimize the brain's flexibility in terms of generating new neural connections in the language areas of the brain (see Unit 3).
Some more technical observations also call into question the behaviorist account of language learning:
5. The phenomenon of phonemic drift is almost impossible to explain behaviorally. In the months prior to the onset of true language -- the so-called babbling or "baby talk" stage -- children initially can be heard to use all possible phonemes but gradually come to lose most of them, retaining only those that belong to the language community in which they live. Clearly adults are not differentially rewarding the use of different phonemes, yet this occurs. How so, unless infants are actively trying to "figure out the rules of the game of language" even before they can yet talk?
6. By the time children become able to speak in sentences (to put strings of words together, known as the telegraphic stage of language), it becomes clear that their language is heavy in so-called content words (concrete nouns, action verbs, and other key words that convey the essential meaning of adult sentences); the less essential function words (articles, linking verbs, and other language refinements) appear much later. How do children differentiate between content and function words? Certainly there is no credible behaviorist explanation for how they can come to do so.
7. The fact that preschool and early grade school children often overapply or overgeneralize the rules of language -- "we seed the sheeps and mouses and then runned home" -- shows that children apparently are on a search for the rules and regularities of language first, picking up only later (sometimes very belatedly) on the illogical exceptions.
Part Two: Intelligence
Views of intelligence can be compared and contrasted with respect to two dimensions: unitary vs. multivariate, and aptitude vs. achievement. Since the two dimensions are independent or orthogonal, this yields four possible views of intelligence.
In the unitary view, there is one basic kind of intelligence (general or global intelligence) that is applicable to all problem solving situations. Individuals differ quantitatively in terms of how much intelligence they have, and it makes sense to think of measuring intelligence in terms of a number (such as IQ). In the multivariate view, there are "multiple intelligences" that may be only very loosely correlated (perhaps even negatively correlated? -- it was said that Albert Einstein could not follow a simple movie plot) with each other: for instance, Sternberg's model, which asserts that there are three basic kinds of intelligence or problem solving ability (formal/abstract/verbal, concrete/pragmatic/practical, and social/emotional/relational). Click here for an exercise related to this idea: how many different kinds of problems (hence, how many different kinds of problem solving abilities or "intelligences") can you spot on this list? In this view, individuals differ qualitatively in terms of the type or kind of intelligence they have, or the nature of their "mental profile", hence it is meaningless to think of reducing the complexity of such individual differences to a single number. Note from the above (see also Unit 1) that the unitary view is inherently nomothetic, while the multivariate view is idiographic.
Aptitude refers to a built-in or pre-existing capacity for learning (note the connections to the nativist view, see Unit 1). In contrast, achievement refers to a learned set of skills or knowledge (note the connections to the empiricist view, see Unit 1). Hence, aptitude testing attempts to measure capacity for learning prior to the fact, while achievement testing attempts to measure success at learning after the fact.
Historically, the notion of IQ was rooted in the unitary aptitude model: it was an attempt to quantify children's general pre-existing capacity for learning, or their "intellectual potential", representing it as a single number. Initially it involved a ratio between a child's mental age or MA (determined by her/his score on the test relative to the average performance of children of various ages) and her/his chronological age or CA (as noted on the birth certificate). Multiplying this ratio by 100 (and dropping any fractional remainder) yielded the IQ score: hence when MA = CA (the child was average for her/his own actual age), IQ = 100. IQ scores above 100 (allowing for statistical measurement error) meant above average intelligence, as with a child of 9 who did as well on the test as an average 12-year-old (IQ = 12/9 x 100 = 133). IQ scores below 100 meant below average intelligence. On most IQ tests, some 90% of the general population would be expected to score between 70 and 130 (the IQ of the average college professor is about 130).
The notion of IQ testing can be critiqued on a number of grounds as follows:
1. The existence of labeling effects, in which the existence of a label (such as an IQ score) causes others to treat a person in a certain way, thus reinforcing the label. When a person does this to her/himself ("buying into the label"), this is known as a self-fulfilling prophecy effect.
2. Questions related to culture fairness, since children whose culture of origin is the same as that of the test developer might be presumed for various reasons to have an unfair advantage; the actual intellectual potential of children without this advantage might likely be underestimated to an unknown, and perhaps unknowable, extent.
3. Non-cognitive influences on test performance such as personality and motivational factors: for instance, children with an overly deliberate or overly impulsive response style will likely do less well on the test as compared to those with a moderate, optimal style.
4. The questionable test-retest reliability of IQ scores due in part to statistical measurement error; small differences, particularly, likely represent mostly "noise" in the data and are thus easily overinterpreted.
5. The underlying unitary aptitude assumptions of the IQ testing enterprise itself: certainly, those who do not share those assumptions are likely to question the validity or usefulness of the entire enterprise.