UC San Diego

There are two potential routes to identifying written words; meanings can be “looked up” directly based on written forms, or written forms can be recoded into phonological (sound-based) codes and used to access the meanings of words. The direct route is assumed to be faster, yet skilled readers still generate phonological codes. In this dissertation I investigated how rapidly skilled readers generate phonological codes and whether they use them to identify the meanings of words. Measuring the time course is important because only rapidly generated codes could actually be used to identify words.

In Study 1, I conducted survival analyses of eye movement data to determine how early skilled readers generate phonological codes. Results suggest that readers rapidly generate phonological codes, and that earlier code generation is associated with faster word identification—suggesting that skilled readers use phonological codes to identify the meanings of words.

In Study 2, I investigated which underlying language skills were predictive of early phonological code use. The rapidity with which readers generated phonological codes was not related to general reading skill, but instead to phonemic decoding ability specifically. Furthermore, the rapidity with which a given reader generated phonological codes was more predictive of word identification speed among highly skilled phonemic decoders and readers with lower general reading skill, suggesting that readers with different language skills might adjust their reliance on the different routes to meaning as their skills allow.

Finally, in Study 3, I investigated phonological coding in profoundly deaf readers. Results suggest that skilled deaf readers can generate phonological codes and may use them to identify the meanings of words. As with hearing readers, there was a high degree of subject variability in the time course, yet all of the subjects were very efficient readers, suggesting that readers may adopt different strategies for or rely more heavily on different routes to meaning.

Together, these studies demonstrate that the cognitive system is flexible and adaptive, and the processes associated with word identification can be adjusted to a given reader’s individual set of language skills to maximize the efficiency of word recognition during reading.