Language Theory and the bilingual child

[YAUTIAPR]

The Sapir-Whorf hypothesis:
New surprising evidence and how it could apply to Puerto Rican children.

Here is an article that should be of interest to teacher who wonder why Puerto Rican children that use two languages often have problems in both Spanish and English

Edited by Dr. Iris Santos Rivera
Centre for Philosophy of Social and Natural Sciences, London School of Economics, London

ABSTRACT
A key question Puerto Rican linguists is whether the Sapir-Whorf hypothesis is true. While easy to formulate, it has proved hard to establish or disprove. The reason is that such research needs unusual circumstances --ideally two culturally identical groups using language that differs in one way which affects a testable cognition. Let us assume we are comparing Puerto Rican Spanish in Puerto Rico with Puerto Rican Spanish in the USA. But because researchers have failed to find such circumstances or make such comparisons on a scientific criteria that is constant and testable, the Sapir-Whorf hypothesis has remained controversial. A natural experiment however which fulfils the criteria for showing the existence of the Sapir-Whorf hypothesis has recently been discovered. Research upon it shows that language has a strong impact upon cognition. Thus we can conclude that differences between languages can effect cognition as claimed by Sapir and Whorf. Can we apply this to the difference between academic achievement between Puerto Ricans on the Island and Puerto Ricans in the USA?

Linguistic Relativism

An important question in anthropology since the mid part of this century has been the validity of linguistic relativity as put forward by Edward Sapir and Benjamin Lee Whorf. Was the latter, for instance, correct to suggest that 'every language is a vast pattern-system, different from others, in which are culturally ordained the forms and categories by which the personality not only communicates, but also analyzes nature, notices or neglects types of relationship and phenomena, channels his or her reasoning, and builds the house of his or her consciousness' (Whorf 1956:252). Does language have that power to mould cognition?

The idea of linguistic relativism has been far easier to state however than empirically evaluate. New approaches to this particularly when they are unexpected and offer methodological advantages must therefore be welcomed. Here I report a piece of research from developmental psychology which might go overlooked by anthropologists which pertains upon how far language can determine cognition.

The Role of Developmental Psychology

Developmental psychology is a possible source of evidence since linguistic relativism is both an anthropological (between culture) and a psychological (within culture) question. The following two problems for instance are closely related.
Do people in different cultures analyze the world in different ways linked to differences in the vocabulary of their language? Do Puerto Ricans change their world view when they come to the USA to live and raise their children?
Does an individual's analysis of their world link to their particular acquisition of their language's vocabulary? Does the particular acquisition of a second language have a negative impact on the individual once he or she leaves Puerto Rico?
Does cognition diminish with the use of two languages? and Is this true for Puerto Rican children in a US situational culture?

Answers to one will have baring upon the answer to the other. If a strong linguistic effect upon cognition is found across cultures then we can infer the cognition of an individual is likely to be shaped by their language. Likewise, if within a culture a strong cognition link to language is found among individuals then cross cultural differences in cognition (if languages differ) could be expected.

The opportunity to test linguistic relativism within culture has not been used however because languages differ cross-culturally while within culture people usually share much the same linguistic experiences. Thus people within culture are unlikely to experience (even if they could arise) cognitive differences derived from language -- but they should (if they exist) stand out cross-culturally. Further, within culture any cognition- language differences which might be found (for instance, linked to different employment) would strongly correlate with confounds such as education and class. Separating the effects of these from language is likely to be impossible. In contrast, cross-culturally domains such as color terms offer the possibility of observable cognitive differences. Example: Some cultures do not distinguish between the color blue and the color green. Given this greater ease of studying the cognition-language interaction cross-culturally the within culture examination of linguistic relativism has been ignored.
The cross cultural route, however, is not without problems. First, considerable theoretical ambiguity exists as what the Sapir-Whorf hypothesis says. Are we to restrict linguistic determinism solely to referents which are independent of culture or are we also to include those where language as part of culture shapes human experience? Most investigators have limited interpretation to domains which are not linked strongly to cultural experience such as color perception. This however weakens the cultural implications of the link as proposed at least on some occasions by Whorf between language and cognition.

Ambiguity also concerns the strength of the claim being made. For instance, Brown (1976:128) argued Whorf had proposed two types of claims.

I. Structural differences between language systems will, in general, be paralleled by non-linguistic cognitive differences, of an unspecified sort, in the native speakers of the two languages.
II. The structure of anyone's native language strongly influences or fully determines the world-view he will acquire as he learns the language at home, in a monolingual environment.
The first is trivial but makes no specific claims (what amounts to 'an unspecified sort of non-linguistic cognitive difference')? The latter is interesting but untestable at least directly. As Kay and Kempton noted (1984:66) 'Until a technique is developed for assessing the world view of a people independently of the language they speak, no direct test of [linguistic determinism] II is possible'. Thus research has been done by testing an implication of linguistic relativity (Kay and Kempton 1984): if language determines the world-view of an individual then cognitions in certain domains, such as color, should also vary, (if language terms covering them also differ). It is upon testing this claim that modern evaluation of the Sapir-Whorf hypothesis has mostly been made. In general, the resulting research using color terms particularly after Brent Berlin and Paul Kay (1969) has failed to provide support. Hence, as John Carrol (1994) has recently concluded 'In general, the Sapir-Whorf hypothesis has come to be regarded as either unconformable or incorrect. As of this writing, most linguists and psychologists believe that evidence offered in its support is flawed'. One can teach a person to distinguish between the colors blue and variations of the color green, for example.

Nonetheless it is possible language could shape cognition.

First, language terms do shape some aspects of cognition such as memory. The ability to discriminate colors is affected when storing memories (Garro 1986; Lucy and Shweder 1979; 1988). However this is not surprising since memory in such cases uses language based storage and so would be expected to be affected by the codability of available language terms. The testing of the Sapir-Whorf hypothesis needs a cognition not directly using language but nonetheless developmentally shaped by it.

Second, the choice of testing the Sapir-Whorf hypothesis upon color has been done less in regard to its being an ideal domain for testing this theory than the convenience offered by standardized and portable Munsell color chips. However we might expect color perception to be fixed by the neurophysiology properties of retinal cones and so immune to the effects of language. For instance, the short wavelength-sensitive cones in the retina responsible for a sense of blueness do not contribute to luminosity (Eisner and MacLeod 1980). As a consequence, luminosity -- whiteness -- is entirely derived from the summation of inputs from the medium and long wavelength- sensitive cones which give us yellows and reds. Thus how the brain senses blue colors will link with lack of luminosity -- black and that of yellows and reds, with whiteness. Thus it is not surprising that Berlin and Kay (1969) found that if a language had two terms they distinguished between a color group centered upon reds, yellows and white and one upon greens, blues and black. That if they had three, they divided the first group into yellow and red term opposed to a white one. The failure to find an effect of language upon color thus reflects less a failure of language than the dominance of the neurophysiology of retinal cones upon color term differentiation.

That color terms in retrospect were unlikely to show a language effect upon cognition does not prevent such effects existing in other domains that are less neurophysiologically fixed. Recent work by Melissa Bowerman of the Max Planck research group for cognitive anthropology suggest this might be so. She and her colleague Soonja Choi at San Diego State University have shown in the cross-cultural comparison of Korean and English that language shapes spatial semantic development (Choi and Bowerman 1991). Even skeptics such as Paul Kay note these findings are 'very impressive' (Ross 1992:12). However her work only links differences in language to differences in spatial semantic development. It is not clear whether this also effects spatial cognition (rather than how people talk about space).

Preferably, we would like to test a specific and isolated change in language terms with a specific test of cognition that is not mediated (at least in an obvious way) by language. The problem here is finding (i) two groups of people that differ only in regard to experience in language terms, (ii) a suitable cognition and (iii) a means of testing it. Moreover, (iv) we would like a cognition which concerns part of a person's view of the world known to be culturally variable.

Sapir-Whorf and The Learning Disabled Child

In this paper I shall argue a 'natural experiment' has occurred which allows for a within culture test of linguistic relativity. Two groups in western society have been found to exist that are identical except that one grows up with a linguistic experience of reference terms the other does not. Thus while they share culture, they do not share the same language terms. Further, a cognition discovered in the 1980s concerning how people perceive a key part of their world and not mediated in an obvious way by language has been tested upon them. I shall first describe the 'natural experiment', then the cognition involved, how it is tested , whether it is effected by language and the implications of this for the Sapir-Whorf hypothesis.
The 'natural experiment' concerns the situation of those born deaf or becoming deaf shortly after birth raised by hearing parents (Peterson and Siegal 1995). Their deafness potentially might effect their cognitive development so the comparison group with which they are compared are not hearing individuals but those born deaf and raised by deaf parents.
Deaf children raised by hearing parents (but not deaf children raised by deaf parents) experience a linguistic non-exposure to terms referring to nonconcrete references, in particular those linked to mental states such as beliefs, desires and feelings. This is a result of their hearing parents being incompetent in the language needed to discuss them. A deaf child needs to be communicated with in sign language but their parents only know spoken language. This, of course, does not stop their parents communicating with them to a limited degree since some sign language is easy to acquire in regard to concrete things because they can be pointed at. What is difficult for such parents to learn is to communicate about abstract entities like minds which unlike concrete ones cannot be indicated except through using signs. The result is that hearing parents do not talk to their child about mental states -- intents, thoughts and feelings thus limiting their conversations to concrete things. If this was not enough, deaf children fail to find compensating mind-talk when they go to school. Here, the overriding concern of teachers is to teach correct sign language (and lip reading) not to discuss feelings and thoughts (which might happen in the more intimate relationship of a child and its family). Moreover the opportunity provided by deaf children among themselves is limited as they are in the same 'boat' of not being exposed to mind related talk.
Lack of exposure to reference terms about mental states is important. Normally, mind-related talk about mental intangibles, at least in individualistic cultures, is a key topic of conversation between parents, children and their friends. There is no reason to believe that this is not also the case between deaf and their deaf children. Thus we have a situation of two groups which are very alike except that one (deaf children raised by hearing parents) is raised with linguistic terms concerning mental states and another (deaf children raised by deaf parents) is not.
A cognition exists concerned with mental states but which itself is not mediated by mental state terms -- theory of mind (Baron-Cohen, Leslie and Frith 1985; Happ‚ 1995). This cognition concerns the ability to attribute mental states to others --to take on their perspective and so see things from 'their shoes' in terms of their intents, beliefs, desires. Cognitively it is a 'second-order representation' reflecting the ability to metarepresent what is known about a situation from another's viewpoint in contrast to that of our own one. Moreover, such a skill is one of immediate and unexplained intuition rather than one of reasoned inference -- we can put ourselves in another shoes without an immediate sense of why.
Its presence in children is tested experimentally by a 'game' involving a doll called 'Sally'. The experimenter shows a child Sally putting a marble in a box and then they take Sally out of the room. While she is gone, the experimenter hides the marble in a basket. The experimenter brings back Sally and then the child is asked where Sally will look for her marble, Children without being able to explain why know that she will look in the box where she put it rather than in the basket where they saw it later moved.

Other Learning Disabilities

This skill as shown by successful performance on this task in retarded people such as those with Downs syndrome (Baron-Cohen, Leslie and Frith 1985) appears not to be related to intelligence. The only group known to fail it are autistic individuals (Baron-Cohen, Leslie and Frith 1985; Happ‚ 1995). Either they fail to answer correctly tests like the above or they do so correctly but in a nonintuitive way which includes an explanation (Happ‚ 1995).
In the Sally doll test and theory of mind we have a cognition which links to mental state terms, is testable yet is intuitive rather than linguistic (at least to the degree people cannot verbalize what they do). Moreover theory of the mind is a cognition which concerns how people cognate an important part of their world -- the minds of others. Further, various researchers have suggested cultures differ in references to mental states and even notions of self (for instance, Lee 1952; Lienhardt 1985). Thus it seems the ideal cognition together with the existence of the above two groups of children raised with and without mental state terms with which to test the Sapir-Whorf hypothesis. If correct the Sapir-Whorf hypothesis would predict that deaf children lacking experience in mental state terms (born to hearing parents) compared to those with them (born to deaf ones) should have impaired cognition concerning minds as measured by failing like autistic people the Sally doll test.

Two Australian psychologists Candida Peterson and Michael Siegal (1995) in the course of researching theory of mind have done this. They did two experiments of the Sally doll test using slightly different wordings. In experiment 1, they tested 11 deaf children of hearing parents and one deaf child of deaf parents aged between eight and thirteen. Only one deaf child of hearing parents answered correctly in the box; the others like autistic people judged where they saw the marble put. In a second experiment with slightly different wording six passed and seven did not (an additional deaf child of deaf parent also passed). The results of the second experiment might suggest theory of mind competence but their performance in fact was slightly worse (though not statistically significant so) than that of a group of twenty autistic children given the Sally doll with this slightly different wording. These are dramatic findings. No other group, has been identified apart from autistic people and those under four, which fail this task. That includes as noted above retarded people with Downs syndrome.

The failure of these children it should be noted does not link to deafness itself: the deaf children of deaf parents answered correctly. Like all researchers, they tested for other explanations and obviously they were asked by a fluent sign speaker.


Here therefore we have direct evidence that language moulds cognition. We need more experiments with Puerto Rican children from the public schools of Puerto Rico to compare them with Puerto Rican children from the public schools in the USA. Then we can see if the following is true for these children. Two groups of otherwise identical and normal children have been raised in linguistic environments which differ in one specific domain, and found to show a specific cognitive difference. Thus we can infer that at least in some cases language specific cognitive development exist and thus at a cross-culture level language might potentially shape cognition as suggested by the Sapir-Whorf hypothesis.

____________________________________________________________

It is my opinion that teachers can use this information to justify either position, Puerto Rican children are alingual, they cannot function in either language in US schools or in Puerto Rico, or Puerto Rican children can be gifted and talented because of the bilingual/bicognitive and bicamaral brain. Language is moulded by culture. Culture is moulded by language.

Yautia