The main aim of this thesis was to investigate further the conditions under which
stimulus naming promotes responding in terms of equivalence relations and in particular the role played by the instructions used during name training. In addition, thirteen children (aged 4-5 years) participated in two studies carried out to investigate other factors that may facilitate conditional discrimination learning in children, and which formed the basis for the development of a novel matching-to-sample procedure.
This procedure was designed to overcome general problems when carrying out learning studies with children, such as motivation and boredom, besides problems specific to matching-to-sample tasks, e.g. discrimination between stimuli. The studies also provided useful information regarding the type of instructions suitable both for matching tasks with visual stimuli and for teaching stimuli names. The main changes to standard conditional discrimination procedures were the introduction of 3-D stimuli, the incorporation of an identity
matching task with the stimuli serving as samples on every trial and the inclusion of a blocked trial presentation order, similar to one used by Saunders and Spradlin (1989).
The procedure was then implemented with an second subject pool of 20 children, aged 4-5 years, who all carried out conditional discrimination training. Fourteen of these children learnt the baseline task and passed a subsequent symmetry test. Five children also carried out equivalence tests, which three passed. Three children failed the baseline task and three others learnt it but failed the symmetry test. These six, plus one child who failed the equivalence test, were then taught to name the stimuli.
Comparisons were made between the effects of naming-instructions that included a relational term ("That is X") and naming-instructions that were non-relational ("Say X"). In addition, speed of naming acquisition was compared across instructions. Comparisons were also made between the effects of names taught prior to the acquisition of the baseline task and those taught after the baseline task had been learnt.
Despite the small number of subjects some interesting trends emerged. Overall, the relational instruction led to more rapid acquisition of stimulus names. The acquisition of sample names resulted in immediate acquisition of the baseline task. Common names learnt through the relational instruction led to success on the symmetry test. Only one child learnt names through the non-relational instruction and he failed the symmetry test until the relational instruction was
introduced.
The findings may be of potential significance for understanding how naming and other aspects of language inter-relate with the emergence of equivalence in early childhood.