This thesis is concerned with the explanation for young children's failure on Piagetian concrete-operations tests. The issue is of considerable importance because it relates directly to the process by which a child becomes capable of concrete-operational thinking. We investigate a claim that, contrary to Piaget's view, young children may fail on the tests because they are misled about the reference of the critical questions by the non-verbal communicative contexts. It is suggested that young children could be in the habit of giving only non-literal interpretations to test instructions. Nonliteral interpretations are constructed primarily on the basis of what the context of utterance suggests. Where these interpretations do not correspond with the intended literal interpretation of experiment¬ er's meaning, the children will be misled about the reference of the test instructions and will give incorrect responses based on the incorrect interpretations. We try to account for young children giving only non-literal interpretations in terms of the evidence that children depend on their sensitivity to social situations for the development of language, and for responding to linguistic expressions which are only partially grasped. We review studies by McGarrigle and Donaldson (1975) and McGarrigle, Grieve and Hughes (1978) who proposed that situational clues cause young children to give incorrect interpretations and incorrect responses in the Conservation and Class Inclusion tests respectively. We report experiments carried out to further investigate the new account. These experiments involved a test (Tower test) which was new but which was similar to one of the Piagetian tests (Conservation test) in design. The young children studied "behaved in the manner predicted "by the new account of their failure, giving considerable weight to what the situation of the test suggested about the experimenter's intention for the critical question asked. This resulted in responses which we termed Yes-responses. The responses given where less weight was assigned to the situational suggestion were termed No-responses. The results of a later experiment showed that the same children who give the supposedly more context-dependent non-conserving responses on the Conservation test are the ones who give Yes-responses on the new test. Similarly, the children who give conserving responses on the Conservation test are more likely to give No-responses on the new test. However, our final experiment led to the important discovery that young children do not assign preferential weight to situational suggestions where the expression to be interpreted has more determinate reference than the ones encountered in the Tower test or the conservat¬ ion question. This result indicated that the key to the children's difficulty might reside in the particular expressions used in these tests - the deictic term 'them' (Tower test) and the quantitative term 'more' (Conservation test) which behaves in a manner similar to deictics. In the conclusions, we propose an account of the developmental transition from non-conserving to conserving responses in terms of our speculations about the change from Yes-responses on the Tower test to No-responses. This account centres on the indeterminacy surrounding the key expressions in these tests and explains the developmental change by appealing both to a decreasing reliance on non-verbal clues and to a restriction of the indeterminacy of these terms imposed on them by the five-year old child.