The Organization and Activation of Orthographic Knowledge in Reading Aloud

This experiment is based on the following paper:

Glushko, R. (1979). The organization and activation of orthographic knowledge in reading aloud. Journal of Experimental Psychology, 5, 674-691. Copyright (1979) by the American Psychological Association. Reprinted with permission.


Early reading models had two parallel pathways for processing text. The phonologic route used spelling-to-sound rules of the language to determine pronunciations of unfamiliar words. Since the rules which were applied were fixed, words with irregular pronunciations (e.g., yacht) would be “regularized” when encoded in the phonologic route. The direct route determined pronunciation by accessing lexical memory. The pronunciation of irregular words would be processed accurately in this pathway so long as the word was known. Pseudowords (or pronounceable nonwords; e.g., mave) always would be processed in the phonologic route. Each route was considered independent of the other. Text was thought to be processed in both pathways in parallel, but the direct route was considered to be faster and so took precedence.

Glushko wondered if two routes were really necessary. He hypothesized that the direct route could process pseudowords by analogy, building a pronunciation by examining how other known words with a similar orthographic structure were pronounced. If the direct route could make such inferences, Glushko reasoned that pseudowords that had orthographic neighbors which only had regular pronunciations would be faster to pronounce than pseudowords that had irregular orthographic neighborhoods because extra time might be needed to resolve the conflicts. The dual-route model would show no difference in pronunciation latencies, since both pseudowords would be processed in the same phonologic pathway.

To test this idea, he compared pronunciation latencies for words with both regular and irregular pronunciations to pseudowords with regular and irregular orthographic neighborhoods. He predicted that latencies would be longer for the irregular words and pseudowords.


The full text describing the experiment is available here.

The experiments are available below. (72.2 KB) (130 KB)