TWO STUDIES OF THE DEPTH HYPOTHESIS
Patricia Wright · British Journal of Psychology · 1969
Two experiments are reported which examine whether the memory load imposed by sentences of different syntactic forms is better predicted by measures derived from transformational generative grammar or from the phrase structure analysis of sentences (Yngve's measure of sentence depth). The first experiment shows that if an adverbial phrase is introduced into an active sentence, so that the phrase structure of the active sentence resembles that of a passive negative sentence, the memory requirements are also those of a passive negative. This supported the depth hypothesis. In the second experiment a subordinate noun clause was introduced into the sentence so that when an active sentence was made passive its phrase structure was less complex than the corresponding active. Here it was found that transformational complexity affected memory requirements but phrase structure complexity did not. It was concluded that both syntactic measures of complexity are reflected in performance, but, with content factors closely controlled, transformational complexity had the greater effect.