Errorless learning has had positive reports across both clinical and non-clinical domains. Within aphasia therapy studies, decreasing cue methods have been proposed as the optimal combination of low error and high effort. This study aimed to evaluate a modification to the decreasing cue method in which the set size of target nouns and verbs for training was extended dependant on participant progress in naming therapy. Four participants with word retrieval symptoms as part of their aphasia took part in a case-series study in which a decreasing cue hierarchy was applied to a dynamic list of noun and verb targets. An assessment probe at the start of each session determined how many new items were to be taken on in therapy, dependant on participant responsiveness to that point. All participants made significant improvements in naming accuracy for treated items from baseline and compared to control items. Two participants continued to take on new items which reflected consistent trajectories of improving naming accuracy, while two participants reached saturation points at which the frequency of errors in naming accuracy limited their set sizes. The varying trajectories of the participants were evaluated in relation to their baseline language and cognitive skills. Clinical implications from the findings and themes for further research were also considered. © 2012 Copyright Psychology Press.