The Development of Transitivity of Length: A Comment on Braine's Reply

The validity of various methods of diagnosing transitivity of length in children has been the theme in a dialogue between Braine (1959, 1964) and Smedslund (1963). A series of methodological fallacies have been unearthed in the course of this discussion. It has been the position of Braine that children normally become capable of making transitive inferences of the types A > B. B > C: D : A > C and A means 'longer than and B. D .A > C and A B. D .A > C. Braine has replied that this study did not control for the possibility that the subject merely selected the measured stick without paying attention to the outcome of the measurement. He then did another experiment on pseudo-measurement (Braine, 1964), in which the measured stick was correct on some trials and incorrect on others. He argued that if preference for the measured stick were the only process involved, subjects should score high on the items where the measured stick was correct and low on the items where it was incorrect. The total score should not exceed what could be expected on the basis of visual discrimination alone. On the other hand, if the subjects consistently, or fairly consistently, applied both types of non-transitive hypotheses, the scores should be high on both types of items and the total score