摘要: Abstract Beginning with a consideration of visual confrontation naming, models requiring an intervening construct between semantic input and phonological access are distinguished from those which permit direct activation phonology upon identification picture. Discrete stage also non-discrete models. In considering how various types aphasic naming errors may be reconciled different it was concluded that substitutions readily accommodated by variety models, but the type characteristic aphasics most easily dealt spreading A discussion lexical in free discourse emphasized not only absence any stages related to processing, character activity leads running speech, contrast canonical representations picture description.