The studies reported in Chapters 2 and 3 sought to determine which of two methods, exemplar training and name training, would most readily facilitate the derived transformation of functions in accordance with symmetry. Twenty one out of twenty four children, aged between four and five years old, failed to show derived object-action or action-object symmetry until they received explicit symmetry training. Thirteen of these children had received name training. Overall, the data are consistent with Relational Frame Theory, but not with Naming Theory. The studies reported in Chapters 4 and 5 were concerned with establishing specific patterns of relational responding when they were found to be absent in children aged between four and six years old. Problem-solving tasks were developed to test and train patterns of relational responding in accordance with the relational frames of more-than, less-than and opposite. Interventions suggested by Relational Frame Theory were successfully used with all subjects to establish the target relational responses as well as increasingly complex patterns o f relational responding. Generalisation tests demonstrated that the relational responding successfully generalised to novel stimuli and experimenters, and contingency reversals indicated that the trained and tested relational responding may usefully be considered a form of generalised operant behaviour. These findings once again lend positive support to Relational Frame Theory’ s approach t to derived relational responding, and to the functional analysis of human language and cognition. The study reported in Chapter 6 argued that the emergence of deictic relations such as "I and you," "here and there" and "now and then" is critical to the development of perspective-taking. A testing and training protocol was developed to analyse responding in accordance with I-YOU and HERE-THERE relations. Two case studies that employed this protocol were presented in which complex forms of generalised perspective-taking were established for two young children. The findings suggest that Relational Frame Theory, and behaviour analysis more generally, may have an important contribution to make to the study of perspective-taking. Finally, Chapter 7 synthesises the empirical work presented in the preceding chapters and addresses a number of theoretical issues that arise from this work.