Constraints are the solution, not the problem

Reichle and Reingold (2013) presented the hypothesis that parafoveal preview is a requirement for the average fixation during reading in order for lexical processing to control eye movements. Taking 240 ms as the customary fixation duration in reading a single word, they arrive at their conclusion by decomposing these 240 ms into stages of a strictly serial-processing model. This subtractive accounting finds the visual system coming up short so long as lexical processing is strictly foveal, that is, so long as lexical processing addresses only the word directly in front of the fovea during a fixation. Neurophysiological estimates of latencies in retinal transmission to the brain and in visual encoding leave little more than 60 ms for lexical processing and 90 ms for programming the next saccade. Reichle and Reingold's solution to this processing bottleneck is to propose that lexical processing of the next word usually begins parafoveally, that is, during the fixation of the current word. We question the generality of the solution that Reichle and Reingold (2013) offer. Rather the observation of comparatively short fixation durations may tie in with a wider range of findings on remarkably fast response times in which cognitive, neural, and physiological events unfolding at various time scales conspire to poise the human organism for a variety of stimuli and dramatically reduce the time needed to arrive at sensible response from stimulus onset (Wallot and Van Orden, 2012). We argue that the solution to the problem can be found by investigating the contextual constraints organizing cognitive activity.