Little is known about the processing of lexical stress during visual word recognition. However, previous research in Italian has suggested that the processing of lexical stress during disordered reading can be explained using dual-route models of reading. In order to find out more about the processing of lexical stress during disordered reading of English we tested an individual with phonological dyslexia, PD, using a visually presented grammatical classification task. Prior to testing PD, Arciuli and Cupples (2000) carried out an experiment using the same task with 29 non-brain-damaged participants. Participants were asked to classify 120 individually presented words as nouns or verbs. The stimuli varied systematically in terms of frequency of usage and stress regularity. Responses were significantly faster and more accurate to words with regular than irregular stress patterns. However, there was no interaction between stress regularity and frequency. Arciuli and Cupples then carried out a second experiment using the same items with a lexical decision task. Again, the error rate data showed a main effect of stress regularity but no interaction between stress regularity and frequency. We reasoned that if lexical stress is assigned non-lexically, then regularity effects might not be observed in someone with phonological dyslexia because the non-lexical route is not functioning normally. If, on the other hand, lexical stress is assigned lexically, someone with phonological dyslexia should show ‘normal’ effects. PD’s error rates did not show a statistically significant difference between the processing of regularly and irregularly stressed words, suggesting that the effects observed by Arciuli and Cupples (2000) in typical, skilled readers might well have reflected a non-lexical influence. However, detailed inspection of the results for individual control participants indicated that (a) not all of them showed a lexical stress effect, and (b) the size of PD’s stress regularity effect was within the normal range. The results of a third analysis compared PD’s results with individuals from the normal sample who were similar to PD in terms of overall error rate (making 10% errors or more). Overall, this subgroup of control participants showed a significant effect of stress regularity, as expected; and once again, PD showed a stress regularity effect within the normal range. The results obtained using these three different approaches demonstrate the different theoretical interpretations that are available concerning the effects of lexical stress during reading. The central question is whether impaired non-lexical processing leads to similar or different effects of lexical stress to those seen in non-brain-damaged reading. On the basis of the present results, the answer remains unclear.