The facilitations used were either semantic in nature (sentence completion) or morpholexical (the first letter or first syllable or anagram of the target word). The global results indicated specific effects of the training of written naming: drilled items improved from .46 to 65, but non-drilled items also significantly improved from .50 to 58. There was no significant effect of response modality: improvements were significant in both modalities and thus generalized from written training to untrained oral naming.