“…In Foygel and Dell (2000), the single-route SP model was used to simulate repetition, without explicitly modeling the auditory input, by assuming that perfect auditory recognition delivers a boost directly to lexical units, essentially just the second step of naming. Later, to account for patients with poor naming but spared repetition abilities, a direct input-to-output phonology route was added to the model (Hanley, Dell, Kay, & Baron, 2004). This dual-route model grafts the “nonlexical” route on to SP, leaving the architecture and simulations of naming unchanged; the two routes are used only during repetition.…”