In the first case there is held to be a change in the individual’s impairment. When the studies with methodological weaknesses were excluded, then 11 of the 44 people given phonological or orthographic information showed some generalisation to untreated items. Thus, around a quarter of participants in these studies improved on untreated as well as treated items. Findings from approaches involving ‘strategy’ and aimed at re-organising processes, such as orthographic self-cueing, were even more encouraging.

Thirteen of nineteen cases showed some GDC-0068 chemical structure generalisation. Such approaches are, however, suitable for only some individuals with particular strengths (e.g., in retrieving orthographic knowledge). Interestingly, in a case series intervention using written cues, sixteen of eighteen participants improved on written naming, and four of these showed transfer to untreated items (Deloche et al., 1997; see also Carlomagno et al., 2001). This mirrors Nickels’ review in suggesting around one quarter may demonstrate generalisation in word production. There are several experimentally controlled single case studies with participants with deficits in post-lexical processing where intervention resulted in improvement on both treated and untreated items (Fisher et al., 2009; Franklin et al., 2002; Robson et al., 1998) For example,

Fisher et al. (2009) worked with a man with ‘mild phonological encoding impairment’. He showed significant generalisation to untreated items from an intervention which involved Gefitinib attempting to name pictures with unrelated names or with shared phonology (magnet, mattress, macaroni). In contrast, Waldron et al. (2011) found no generalisation to untreated items, despite employing a previously successful intervention (Franklin et al.,

2002). The participants in Waldron’s study had a combination of lexical (stage 2) and post-lexical (stage 3) impairments. Raymer et al. (2012), in a study investigating errorless naming treatment and gestural facilitation of naming did not obtain generalisation to untrained items for the three participants with semantic anomia, but obtained some generalisation in naming for three of five participants with phonological anomia. Finally, studies using orthographic cueing aids demonstrate convincing generalisation to untreated BCKDHB items (Best et al., 1997; Bruce and Howard, 1987; Howard and Harding, 1998). We aimed to explore the effects of a cueing hierarchy, especially generalisation to untreated items, and to relate the outcome to level of breakdown in naming. Specifically, we ask: (i) Can a cueing therapy improve word production (i.e., retrieval of meaning and form and phonological encoding) in participants with aphasia? From previous studies we predicted: (a) those with a post-semantic deficit, stage 2, with relative strengths in semantic and phonological output processing and a specific deficit in retrieving lexical forms will show item specific changes in naming (following e.g.