Impact of dialect use on a basic component of learning to read

Abstract

Can some black-white differences in reading achievement be traced to differences in language background? Many African American children speak a dialect that differs from the mainstream dialect emphasized in school. We examined how use of alternative dialects affects decoding, an important component of early reading and marker of reading development. Behavioral data show that use of the alternative pronunciations of words in different dialects affects reading aloud in developing readers, with larger effects for children who use more African American English (AAE). Mechanisms underlying this effect were explored with a computational model, investigating factors affecting reading acquisition. The results indicate that the achievement gap may be due in part to differences in task complexity: children whose home and school dialects differ are at greater risk for reading difficulties because tasks such as learning to decode are more complex for them.

Keywords: African American English; achievement gap; dialect; reading.

Figure 1

Mean naming latencies for Contrastive and Non-contrastive words for each child, as a function of the number of AAE features children produced in the Charity et al. (2004) sentence repetition task.

Figure 2

Left: Rules used to create the AAE corpus (Craig et al., 2003). Right: Architecture of the computational model. Each layer consists of a set of units. Arrows indicate weighted connections between units; all units at one layer (e.g., orthography) have connections to all units at the connected layer (e.g., hidden). Spellings are represented as patterns of activation over the orthographic units; pronunciations are represented in an analogous manner on the phonological layer. Cleanup units and interconnections between phonological units create attractor dynamics such that the model settles into a phonological code over time. Hidden units allow the model to represent the complex contingencies between orthography and phonology that exist in English. Functions of the context units are discussed in connection with Simulation 2.

Figure 3

Left graph: Performance of the model over the course of training in the three training conditions as well as performance for Contrastive and Non-contrastive words in the Mismatch condition. Right graph: Cross-entropy Error for Contrastive and Non-contrastive words in the MAE Match and Mismatch conditions after 1000 epochs of training.

Figure 4

Model performance on the reading task in the Bidalectal conditions compared to the MAE Match condition from Simulation 1.