Fig. 19.

Simulations of Swingley and Aslin (2007) where (i) novel words are not part of the lexicon (no learning has occurred) and where there is no intrinsic salience associated with either object on display. Presentation of novel words that are neighbours of the familiar picture’s label (weak mispronunciation) activate that label more than when a non-neighbour word is presented (strong mispronunciation), leading to a decrease in looking time to the novel object. (ii) A novelty-based salience leads to longer looking times towards the novel object when named, providing a qualitative match to Swingley and Aslin’s results. (iii) When novel words are introduced into the lexicon as low-frequency entries, both images can be associated with a lexical item and the Luce choice rule can be applied. Target looking patterns closely match the experimental results.