The counter model (R. Ratcliff & G. McKoon, 1997) was designed to explain the normal processes of word identification and how they are influenced by a prior encounter with a word. The model accounts for the findings of word identification experiments in which words are flashed briefly. A crucial finding is that prior encounters with words typically lead to biases such that a previously encountered word is more likely to be given as a response. However, for low-frequency words, a prior encounter can improve overall performance (J. S. Bowers, 1999; E. M. Wagenmakers, R. Zeelenberg, & J. G. W. Raaijmakers, 2000). The authors show how the model can explain this result. Also, J. S. Bowers (1999) has claimed that some earlier data concerning dissimilar alternatives in forced-choice experiments that support the counter model are spurious, but the authors show that his claims are incorrect. In sum, the authors argue for a theoretical approach that offers a detailed description of the cognitive processes of word identification and explains performance across tasks, measures, and independent variables.