Evaluated hypothesis that cognitive antecedents of word recognition are domain-specific and unrelated to higher-order domain-general cognitive abilities in a longitudinal study of Hebrew-speaking children. Found that kindergarten domain-specific measures accounted for 33 percent of variance in Grade 1 word recognition, even after controlling for domain-general and higher-order language tasks, whereas domain-general measures explained only 5 percent of the variance. Reading comprehension was predicted by print-specific skills and domain-general abilities. (Author/KB)