Linguistic awareness and development in young children: a study of word awareness

Abstract

The present research investigates linguistic awareness and development in young children. It sets out to test some specific hypotheses concerning the nature of the relation among the development of linguistic awareness, language acquisition, social sensitivity and cognition. Specifically, the focus is on words - the awareness of word order, the awareness of the phonologically arbitrary nature of the word and the awareness of the word as a unit of language. A series of experiments was designed to try and tap into children's possible judgemental skills, segmentation skills, analysis skills and explanation skills, all of which can be said to be behavioural manifestations of linguistic awareness. The experiments often pull together previously separate lines of inquiry - for example, the two homonym studies tap into both the awareness of words and the awareness of ambiguity as the experiments were designed to expose the arbitrary nature of the word.

Data were collected both longitudinally and cross-sectionally. The longitudinal study was a nine - month investigation of two -year -old English- speaking children living in Edinburgh. The children were visited in their homes for approximately one hour per session. The first two visits were separated by only a week, in order to verify that the measure of linguistic ability (the median number of syllables per utterance) calculated from the first session was accurate, and did not underestimate the abilities of the children due to such factors as shyness, time of day and the like. All other sessions occurred at monthly intervals, except for a four -month period during which time comparison data were collected from bilingual two -year -olds in Montreal (see Appendix A). Initially there were ten children under observation, but by the final session the number was down to seven. The longitudinal study was necessarily exploratory in nature because very few studies in the literature had tested children as young as two years of age. The literature, however, was the base from which questions and ideas were generated and which provided techniques that could be tested out on the two -year -olds. The various tasks which were given to the two -year -old children included two which tried to elicit judgements about grammaticality; one which tried to determine if the children could segment the spoken stream into discrete units - a "word block" game requiring the children to drop one block into a bucket for each word of a given sentence; one which asked the children to construct a word from a given starting sound; one that asked class inclusion questions; one that noted children's responses to hearing a different language; and one that examined the children's notion of a word by asking them to put two words together to make a compound word. All of the sessions with the two-year-olds were audio recorded and later transcribed. Approximately fifty -five hours of tape recordings were collected.

The second way of collecting data was through systematic experimentation using slightly older children. Ideas for the experiments arose from both the observations of the two -year -old children and from studies in the literature. The subjects were nursery school children aged between approximately three and five years. We are grateful to the administrators of three of the nursery schools in Edinburgh and to the parents of the children attending these schools for allowing us to go in and play games with the children. All of the experiments were pilot- tested, usually, although not exclusively, with the children who attended the nursery school at the Department of Psychology of the University of Edinburgh.

Because research into linguistic awareness is so diverse, the background literature has not been reviewed in this introduction. Instead, the relevant research reviews occur at the beginning of the sections to follow, as each section deals with a different aspect o f linguistic awareness. Within each section, the o rganization consists of a critical review of the elevant research, followed by the two -year -old children's performances on some of the same tasks, followed in turn by a novel experimental investigation which was designed to address some of the issues and concerns that arose from the examination of previous research.

The first section to follow deals with the awareness of grammaticality. This is the area of linguistic awareness that has up to now generated the most research. Techniques abound that have attempted to e licit judgements (as behavioural manifestations of linguistic awareness) from children about the semantic and the syntactic acceptability of various sentence types. We applied two previously -used techniques to elicit judgements about word order in a replication attempt, using the ten two -year -old children. This was followed by an experiment that used a new procedure we developed to elicit judgements and explanations about the syntactic acceptability of sentences from nursery school - aged children. The second section deals with children's awareness of lexical ambiguity, which reflects the phonological arbitrariness of a word. The next section is an examination of children's awareness of the word as a unit of language. The results of the experiments are then brought together and discussed in accordance with the theoretical framework in a concluding section.

The studies that follow set out to test some specific hypotheses that were motivated largely by the disparity between various theoretical viewpoints. Ultimately, we wanted to see if the diverse theorizing could be brought together, and to show that the views might not be at odds with one another, but could in fact be in accord with one another. For instance, while some researchers have proposed a relationship between linguistic awareness and cognition, and others between linguistic awareness and language development, we hypothesize that there will be a relation among cognitive development, language acquisition and linguistic awareness. The theoretical basis for this hypothesis stems from recent work being done on cognitive development which suggests that the intellectual skills of young children are not context-free, but context-dependent (Donaldson, 1978). One kind of intellectual skill is the ability to decentre, to take account of another person's perspective. Among others, Hughes (1975) has shown that given an appropriate task, preschool children can be shown to be not as egocentric as Piaget suggested. What such researchers are trying to do, is to determine in which contexts children are likely to use their skills. Given that the testing room situation is one of social interaction between an experimenter and a child, it seemed appropriate to look at the ability to decentre from a social cognitive approach as well. Light (1979) devised a task of social sensitivity in a role- taking situation that can be said to bridge the gap between the social and individual aspects of cognition. Inherent in a task of this nature is a test of children's language comprehension and as such, it also provides an indication of linguistic development.

We propose that the link between social sensitivity and linguistic awareness is that both involve the ability to dissociate the actual from the hypothetical. In the case of role- taking, the link is the ability to treat oneself as an object; in linguistic awareness, it is the ability to treat language as an object. A similar argument has been proposed about language acquisition, in that as active participants in their linguistic environment, children must be sensitive to, hypothesize about and test language itself (Karmiloff-Smith, 1979; Macrae (Elliot), 1978). Our task was to devise situations or contexts that would allow children to use their language, their cognitive and their linguistic awareness skills. We hypothesized that it would be possible to do this.

Another hypothesis we propose is that the behavioural manifestations of linguistic awareness will develop gradually and will not appear in an all-or-none fashion. Because both language acquisition and cognitive development are no longer viewed as all -or -none processes (Bruner, 1975; Gelman, 1978; Howe and Hillman, 1973; Trabasso, 1975), we supposed that the expression of linguistic awareness would similarly follow suit. Howe and Hillman (1973) in their research on the acquisition of semantic restrictions in children, found that although stimulus sentence types were relatively homogeneous, children's responses were neither 100 nor 0 percent correct within a sentence type. This led Howe and Hillman (1973) to state:

"It must be recognized that while children may be acquiring one kind of restriction more rapidly than a second, neither is learned in an all -or -none fashion, and that the development of two or more may be occurring simultaneously." (page 138)

We have already discussed the studies that suggest the same hypothesis is valid for cognitive development.

The ultimate aim is to arrive at a working definition of linguistic awareness that will take into account the various theoretical and empirically- backed viewpoints.