Gender bias in morphological inferences

The present study explores how language learners apply gender stereotypes in learning a novel language with grammatical gender. Adult, English-speaking participants were exposed to picture–sound pairs from a miniature language. Each picture was of a matched gendered professional (e.g., male tennis player, female tennis player) with a nonsense form [CVCV-go/gu]. Participants were exposed to 32 picture–sound pairs (16 male, and 16 female, all matched) five times in a randomized order. Following training, participants were given a two-alternative forced-choice test with novel picture–word pairs. Participants were presented with a novel picture paired with two words (e.g., [befegu vs. befebo]) and were asked to choose which word most likely portrayed the meaning conveyed by the picture. These novel items contained gender-matched professions (e.g., male and female chemist), neutral items (office supplies), stereotypically female items (makeup), and stereotypically male items (tools). Participants assigned the appropriate gender to the novel professions, and assigned gender in line with the stereotyped objects at a rate significantly greater than chance (but not for neutral items). These results support the hypothesis that learning a language with a binary grammatical gender might be influenced by gender stereotypes.

Frequency of Occurrence and Identification of Ambiguous Perceptual Forms

Differential frequency of stimulus presentation was manipulated in a training session during which 144 Ss identified nonsense form prototypes. Knowledge of results was given. In a test session Ss attempted to identify variations of the eight prototypes seen during training. The test task was to decide as to which of two prototypes a given variation was most similar, under some conditions a neither response being allowed. The results showed that the manipulation of differential frequency during a training session can lead to low-frequency responding to one type of test stimulus and high-frequency responding to another. These results were interpreted as supporting a perceptual set rather than a response bias explanation of the influence of frequency of stimulus occurrence.