The process-dissociation procedure was used to estimate the influence of spatial and form-based processing in the Simon task. Subjects made manual (left/right) responses to the direction of arrows (> or <) presented to the left or right of fixation. Manipulating the proportion of incongruent trials (e.g., a right-pointing arrow presented to the left of fixation) affected both the size and direction of the Simon effect. To account for this pattern of data, we compared process estimates based on three possible relationships between spatial and form-based processing: independence, redundancy, and exclusivity. The independence model provided the best account of the data. Most telling was that independent form-based estimates were superior at predicting observed performance on arrows presented at fixation and did so consistently across conditions (r's > .80). The results provide evidence that the form ("what") and spatial location ("where") of a single stimulus can have functionally independent effects on performance. They also indicate the existence of two kinds of automaticity--an associative ("implicit learning") component that reflects prior S-R mappings and a nonassociative component that reflects the correspondence between stimulus and response codes.