The Universal and Language-specific Features of Skilled Reading in a Brāhmī-derived Writing System: Case of Hindi/Devanāgari

Research on visual word recognition in many varying orthographies has helped to understand the universal and language-specific characteristics of reading. Hindi and similar orthographies are potent candidates to take this research forward because they have some unique structural properties. Two hypotheses have been influential in this regard. The first hypothesis is the orthographic depth hypothesis that states that the difference in the effect of a lexical variable on lexical decision and naming latencies will be larger in shallower than deeper orthography. The second hypothesis states that for readers of deeper orthography bigger grain size units will be more relevant than smaller ones. Both these hypotheses were tested in Hindi. Results from the first experiment supported both the above hypotheses, by showing that the effect of a lexical variable, familiarity, was stronger in lexical decision latencies than in naming latencies; it supported the other hypothesis by showing that the grain size of letter but not akshar is more relevant for reading Hindi. The findings from this study were further tested in two more experiments that used either a different experimental design, or paradigm. We found weak support from one experiment but null effect in the other. Implications of the results from the last two experiments on the finding of the first experiment are discussed