It has been hypothesised that different word-finding impairments in people with aphasia, will be best remediated by different treatments (e.g., Hillis & Caramazza, 1994). To date, however, no clear or predictable relationship has emerged between the type of word-finding problem and the most appropriate treatment task (Best & Nickels, 2000). As therapy is a time consuming and costly process, it is clearly desirable for both the client and the clinician to be able to match the nature of the word-finding problem to the most appropriate treatment, as quickly as possible. This paper reports preliminary results from the first phase of a larger project, which aims to identify whether the response to using a task once, “facilitation”, can predict the response to the same task used repeatedly over time, “therapy”. The first phase, reported here examines the effects of two different tasks used once in a facilitation paradigm on subsequent picture naming of people with aphasia. This project aims to address three questions: • Do tasks that focus on word meaning (semantics) and those focusing on word sound (phonology) produce equal facilitation effects or do they vary in efficacy across and/or within individuals? • Do individuals with different levels of impairment in spoken word production (e.g., semantic or phonological) respond differently to facilitation? • Is there an interaction between 1 & 2 (i.e., are some tasks more effective than others for individuals with a particular level of impairment)? To date, five individuals have been tested. Only one individual showed a significant benefit from a facilitation task, on subsequent picture naming at least 10 minutes later. This individual, benefited from phonological facilitation (repetition in the presence of a picture) but failed to benefit from semantic facilitation (feature verification, e.g., Does it bark?). No other individual showed any benefit from the facilitation tasks. Why have the results of facilitation been so limited? For phonological facilitation it is possible that retesting happens too late: some previous studies have found that there were only short term benefits from facilitation, with no lasting effects observable 10 minutes later (the time at which we retest). For semantic facilitation it is possible that the fact we do not include the phonological form of the word in the question influences efficacy: LeDorze et al. (1994) found facilitation effects from a semantic task ONLY when the phonological form was provided. Further investigation is currently underway, to test these hypotheses, with the individuals tested to date, and additional aphasic participants.