Citation/Abstract

The declarative/procedural model posits that expressive and receptive language depend on two memory systems that underlie the mental lexicon and the mental grammar, two aspects of language. The lexicon, the store of words and their meanings, is hypothesized to depend on the declarative memory system that underlies the learning, storage and use of information. The grammar is posited to depend on the procedural memory system which underlies the learning and use of "non-conscious" skills and habits, such as learning motor sequences or riding a bicycle. Study of this distinction has used an English production paradigm that compares the production of past tense forms that can be composed by applying a grammatical rule, the regular past tense forms (e.g. walk + ed); and the production of idiosyncratic irregular past tense forms (e.g. dug ), that are retrieved from the lexicon. However, there is evidence in the literature that some fully composed regular forms (e.g. walked ) might also be retrieved from the lexicon as well.

Women have better declarative memory, particularly for verbal information, relative to men. Tests of verbal fluency and verbal declarative memory are sensitive to circulating levels of sex hormones. We hypothesized that the female superiority at declarative memory may result in women retrieving regular forms from the lexicon rather than composing them with the grammatical rule. Using frequency effects, a marker of retrieval from memory, we found sex differences in the mechanism of production of regular forms, with women retrieving them from the lexicon and men composing them. We hypothesized that increased estrogen levels would enhance declarative memory and increase the retrieval of regular forms from the lexicon. We examined the influence of hormone therapy on past tense production. There were no effects of treatment on frequency effects or production of regular or irregular verbs. We examined the effects of estradiol, progesterone and testosterone across the menstrual cycle. We did not find effects of estrogen or testosterone, but we found that progesterone had a negative effect on frequency effects, particularly for regular verbs.The data did not support the hypothesis that estrogen levels enhance declarative memory nor did it increase the retrieval of past tense forms from the lexicon. A number of confounding effects in the hormone studies suggest that the effect of sex hormones on past tense production should be further investigated.