analogy task across testing and children in the Inference condition also performed better on
the inference task than the other groups. These findings suggest an immediate mapping
between input and initial inferences. Those findings are in accordance with other research
showing that children can effectively use the linguistic context to infer the meaning of the
unknown words (Au, 1990; Dickinson, 1984; Dollaghan, 1985). The role of contextual
support is also supported by Sternberg and PowelTs (1984) theory of word learning from
context, where children use certain aspects of the context where the novel word is presented
to infer the meaning of a word.
Additionally the above findings indicate that the context where a child encounters a word has
a significant impact on the types of semantic representations he can make. Therefore, these
data clearly illustrate that a constraints model is not sufficient to explain the development of
full lexical representations. The above findings are more in agreement with Nelson’s (1988,
1990) interactive functional model. According to the model, children learn new words by
interacting with adults linguistically and поп-linguistically in a variety of contexts. In such
situations the adult is a collaborator who supports inferences and provides feedback.
D. Some of the children ,s representations reflected a basic level category principle
The children used a basic level word either in order to provide an analogous or different item
from the target item or to provide a definition for the target word. This suggests that in those
cases the principle of a basic level categorisation (Rosch, 1978) was operating. Furthermore,
the Error analysis of the naming task revealed that the use of a basic level word was the most
frequent error, when they were asked to name the target items (naming task). It seems that
when the children acquire a target word, they choose to make relations with other words
which belong at the basic level of the categorisation than the superordinate or the
subordinate, as they were defined by Rosch (1978).
E. Children performed better on the immediate than the delayed post test
The measurement of word learning twice (immediate and delayed post test) provided some
indication that children’s performance was better during the immediate than the delayed post
test for some of the tasks (naming, multiple choice and contrast tasks). Different possible
explanations can be offered for the above finding. Thus, may be children’s memory probably
did not help them to remember the relevant aspects of the word’s meaning for succeeding on