The role of morphemes during reading development – MORPHEME

This ANR DFG project brings together experts from cognitive psychology, education and computational modelling to investigate the development of reading skills using a mixed cross-sectional and longitudinal cohort design, with pre-school and primary school children, combined with a cross-linguistic approach. Past research on reading acquisition has focused heavily on the mapping between letter strings and their pronunciations, but the role of units that carry meaning (“morphemes”) has often been ignored. This is a significant problem, because approximately 90% of all words comprise multiple morphemes, yet there is very little understanding of how children might use such morphological structure when learning to read. Thus, the focus of our research will be on words comprising multiple morphemes (‘walked’, ‘walker’, ‘walking’).

Previous research from adults, including by members of our own team, suggests that skilled readers rapidly “decompose” derived words into their morphemic subunits (‘teach’ + ‘er’). However, despite extensive research in adults, we still know little about how and when morphological knowledge becomes incorporated into the reading system during reading acquisition. Virtually all literature is based on studies with adults, and nearly all have looked at skilled processing, not learning. This is unfortunate, as to understand properly how morphological knowledge is represented and used during skilled word recognition, we need to understand its acquisition and developmental time course, as children learn to read. Moreover, since little is known about when the reading of morphemes becomes more intuitive and automatized, literacy training programs fail to address this important skill appropriately.

We propose to address this research gap by exploring the acquisition of morphological knowledge based on an original mixed cross-sectional and 3-year longitudinal cohort design, using a large test-battery that includes experimental paradigms such as auditory lexical decision, eye-tracking, and masked priming combined with visual lexical decision and rapid naming tasks. We test children in three different age groups, which we will follow over the course of three years.

The strength of this project is not only that it brings together experts from different fields to work on the exploration of how morphological structure in combination with other factors (e.g., phonological awareness and vocabulary growth) can influence reading development in young children, but it also involves a crucial cross-linguistic comparison of the process of reading acquisition. We aim to compare the development of morphological processing in German and French, two languages that differ fundamentally in terms of word length, frequency and morphological productivity. This cross-linguistic comparison will provide important insights into the developing German and French reading systems.

An additional strength of our project is the use of computational modeling to constrain interpretations of the obtained data, and to provide a theoretical framework that will help specify the key mechanisms driving potential effects of morphological structure on reading acquisition. The Connectionist Dual Process model of reading aloud (CDP++; Perry, et al., 2010) will be used to examine the extent to which the grapho-syllabic parsing mechanisms of this model can account for observed morphological effects. We will also implement and test a novel connectionist model of the acquisition of morpho-orthographic representations during learning to read. The expected results are not only crucial for understanding the fundamental processes involved in normal reading development, but also have the potential to be of great value in motivating changes to remediation programs for children with language and reading deficits.