Do activated letters influence lexical selection in written word production?

Abstract

Background: Written communication has become increasingly important since the advent of digital media such as emailing and texting. Research into the nature of interaction between the cognitive processes underlying production have focused on spoken language. Less is known about the processes underlying written word production.

Aims: We investigated the presence of feedback from letter activation to lexical selection in written word production. The presence of feedback would lead to greater-than-chance orthographic overlap between targets and semantically-related word errors in acquired dysgraphia, as the semantic neighbours of a target word sharing orthographic structure would receive extra support during lexical selection.

Methods & Procedures: The orthographic overlap of written semantically-related word errors from an individual with acquired dysgraphia was compared to chance distributions of overlap generated using a word association database. We used a Monte Carlo procedure to select random “pseudo-errors” from the top semantic associates for each target word. 10,000 hypothetical datasets were generated for two distinct hypotheses regarding the likelihood of selecting particular semantically-related word errors.

Outcomes & Results: The orthographic overlap between target words and semantically-related word errors for the individual with dysgraphia was greater than any of the hypothetical datasets, indicating that the target-error pairs shared more orthographic structure than expected by chance (p < .0001).

Conclusions: We obtained evidence that feedback from letter activation influenced lexical selection in a case of impairment. The presence of feedback from letter activation to lexical selection may be useful in developing strategies leading to improved language production in individuals with dysgraphia and aphasia.

Keywords:

• Aphasia
• Dysgraphia
• Writing
• Language processes
• Lexical selection

Acknowledgments

The authors gratefully acknowledge RMI for his hard work and dedication to our extensive testing. We also thank Christine Asaro, Tori Miner, Alexandra Muratore, and Daniela Velez for help with stimulus creation, testing, and data analysis throughout this project.