The construction of units ordered language is an object of study of syntax and of
fundamental interest to research the study of behavior. Syntactic performances often
involve changes that take place in word order between different linguistic contexts. The aim
of this study was to investigate the emergence of the composition of sentences written from
the teaching of overlapping words with conditional control. It was still intended to examine
the maintenance of performance from the reapplication of the test with new sentences. A
software was developed for submission of programmed contingencies of teaching, testing
and recording of behavioral data. The study included six children ranging in age from eight
to ten years of a social institution. Four sentences, two in the active voice and passive voice
in two participants were taught the procedure of overlapping words. The participant's task
was to compose sentences written by choice, the computer screen, words presented in
unpredictable order. After teaching for overlapping sentences in active and passive voice,
they were conditionally related to the red and green respectively. Substitutability tests
checked the production of new sentences. After about forty-five days without contact with
the experimental contingencies, the emerging performance was again assessed. All
participants reached the criterion for accuracy at the stages of education. In tests of
substitutability, a participant has reached 87.5%, two had 75% and for others, the
percentage of correct answers was 62.5%, 50% and 37.5%. A second study was carried out
increasing the number of attempts to teach for three of the participants who had a poor
performance with conditional control. The results showed that two of the participants (LUC
and DAN) had 100% and 87.5% accuracy in the tests with new sentences. The participant
POL reached only 62.5% accuracy in testing. The results suggest that refinement of the
procedure for two of the participants produced sentences with the emergence of conditional
control, regarding the performance of those participants in the first study. The participants'
performance in re-testing is still debatable with maintenance only one participant. This
study shows that the teaching procedure for overlapping words can generate the
composition of sentences with few errors. The conditional control needs to be well
established and variables such as order of training needs to be investigated.
|
