A case study by Brennen, David, Fluchaire, and Pellat (1996) reported the case of a patient who could occasionally name celebrities with very low concurrent levels of semantic access, which is difficult to reconcile with current models of face identification. Hodges and Greene (1998) attribute Brennen et al.'s case study to artifactual explanations and provide new data, which, they claim, is evidence against the "theory of naming without semantics". This reply demonstrates that Hodges and Greene's arguments are unconvincing and that aspects of Hodges and Greene's data in fact provide support for Brennen et al.'s conclusions. It is also argued that in cases of dementia direct naming has been reported for other stimuli domains as well.