2018 Friday Session B 0930

Friday, November 2, 2018 | Session B, Conference Auditorium | 9:30am

The dynamics of spoken and visual word recognition in school-aged children
K. Hendrickson, B. McMurray

Successful language comprehension requires efficient word recognition. Both spoken and visual word recognition are described as a competition process by which children activate lexical candidates (the target and similar words; e.g., cat, can), which then compete for recognition (McClelland & Elman, 1986). Recent work suggests that this competition may develop through adolescence (Fernald, Perfos, & Marchman, 2006; Rigler et al., 2015), and competition in written words predicts reading outcomes (Roembke et al., in press). However, it is unclear how the developmental timecourse of competition relates in spoken vs. visual word recognition. Mechanisms of spoken language are available from birth, while visual word recognition must be directly taught later in development. Therefore, the mechanisms of lexical competition may change as children improve in decoding and reading.

Secondarily, it is unclear if the structure of the competitor sets differs for spoken and visual words. A crucial factor is the order of phonemes or letters. While for spoken words phonemes unfold over time, for visual words, order is defined spatially. Toscano, et al. (2013) showed adults activate spoken words with the same phonemes in reverse order (anadromes, e.g., pit and tip), however words that preserve order (cohorts e.g., pit and pin) are more active. For visual words, lexical activation is observed when order is disrupted (letter transpositions, e.g., casino–caniso; Perea & Lupker, 2004). However, the development of the ability to use spatial or temporal order is unknown.

We asked how competition differs for spoken vs. visual words and whether these processes change with development. We measured real-time word recognition with the Visual World Paradigm at 7-8-years (n=22), 12-13-years (n=18) and in adulthood (n=38). Participants heard or read a word and clicked the referent from a display containing the target, an anadrome or cohort, and two unrelated pictures. Eye movements were monitored over time. For target fixations, we fit logistic functions for each participants looking over time. A linear mixed effects model found that crossovers (a measure of timing) were later in earlier ages (7-8, 12-13, adult) for both spoken and visual words (p < .0001). There was no Age x Domain interaction suggesting the developmental timecourse of spoken and visual word recognition was similar.

To analyze competitor activation, we ran a model with the difference in fixations to the competitor (anadrome, cohort) and unrelated objects. Anadrome and cohort competitors showed evidence of activation across domains (p < .0001), though cohorts demonstrated significantly more fixations than anadromes for spoken but not visual words across age (p < .0001; see Figures).

These results suggest that the ability to recognize highly familiar words undergoes a protracted period of development. Indeed, children are still improving in their ability to recognize words (spoken and visual) well into the school-aged years. Further, there are different profiles of lexical competition for spoken vs visual words as young as 7-years, and these profiles appear to remain stable until adulthood. For spoken words, temporal order outweighs the amount of phonemic overlap, whereas for visual words, order and overlap are weighed equally.