An important limitation in animal models of cognitive learning has been the difficultly in specifying how animals code (or represent) events. Coding processes are particularly important in the development of stimulus classes. Recent evidence suggests that pigeons (and other animals) can form stimulus classes from originally unrelated stimuli. The characteristic of members of such stimulus classes is that they are able to substitute for one another in much the same way that arbitrarily-designated spoken and written words can be substituted for objects and actions in human language. The proposed research with pigeons will use delayed-matching procedures in which pigeons have been shown to develop single-code-default strategies (one of two samples is coded - the other comparison is chosen by default). Evidence for single-code-default strategies restricts the class of possible codes and allows one to gain access to the underlying codes. The immediate goals of the present research are (a) to examine the coding processes used by pigeons, (b) to determine the conditions under which these coding processes are used, and (c) to identify the nature of the underlying codes. In the first section of this proposal we will examine commonalities in the coding of present/absent samples. In the second section we will apply similar strategies to the analysis of duration sample coding. Finally, we will examine the similar coding of samples that are related only through their association with a common comparison. The long-term goal of this research is to be able to specify the bases of stimulus class formation by animals so one can determine which components might be missing from humans who fail to acquire the stimulus classes needed for language development.
