Aphasia is a language disorder entailed by brain damage. Naming errors are essential in aphasia and so they can be used as indicators of the disorder. Naming, in this context, means confrontation naming where the subject of a neuropsychological experiment is asked to utter the name of a target presented to him or her. We have devised a simulation technique modelling the mental word processing of an aphasic by employing a succession of four neural networks with semantic, phonological, syllabic, and phonemic processing functions, respectively. The simulation model is built using a bounded set of Finnish words in their base form. Simulated naming errors are generated in the system by spreading activation in the networks. Although we consider naming errors, and thus human speech, in executing our simulation, we treat words and their component parts as textual units that can easily and unambiguously be input, processed, and output using computer programs.