The paper aims to account for linguistic and processing factors responsible for the incidence of spelling errors in Hebrew. The theoretical goal is to disentangle a complex interaction between morphology, phonology, and orthography in production of written words. We focused on a specific spelling error in Hebrew: an overt representation of the word-internal segment/i/by the letter Y ([Hebrew character]). This Y-insertion goes against the prescriptive spelling rules (cf. substandard MYRPST [Hebrew characters] vs conventional MRPST [Hebrew characters],/[Hebrew characters]/'balcony') and yet in our data it affects 25% of nouns with an appropriate phonological environment. Corpus analyses of unedited texts further revealed that errors proliferated in lower-frequency words, but their occurrence was much less likely if it would disrupt a morphological unit. These results point to morphology and statistical patterns of language use in Hebrew as major mechanisms driving orthographic learning: the paper discusses repercussions of our findings for theories of reading.