Teaching Language to the Truly Naive Learner — An Analog Study Using a Tactual Vocoder

A theoretical model for the analysis of sets of unfamiliar language stimuli is presented and some implications for the programming of language learning are discussed. An experiment is described illustrating research questions raised by the model In the experiment highly unfamiliar word discriminations were taught to 20 subjects distributed in four groups. Each group learned to discriminate the same 12 words presented through a tactual vocoder. The experimental conditions were: (1) low and high criteria of performance before new words were cumulatively added to the known set, and (2) “easy” and “hard” sequences. The hard sequence required more difficult discriminations to be made during learning of the first six words. The results show that difficult tactual discriminations could be learned by most subjects in the time limits allowed. The best performance was shown by the Hard Sequence-Easy Criterion group. This finding is consistent with the expectation that the “Hard Sequence” would produce faster learning by requiring a focus on critical discriminations early in the training, and the likelihood that lower criterion facilitated faster introduction of the larger set and avoided over learning of false discriminations. On a location transfer test (the vibrators were moved from the arms to the legs), the group was correct 72.1 percent of the time, demonstrating a central integrative explanation for the learning going on. The study also showed that unfamiliar problems are difficult learning tasks, even for college level adults.