ABSTRACT
There is a position in each word, the “convenient viewing position,” where the eye should first fixate in order for the word to be recognised most quickly. For every letter of deviation from this “ideal” initial fixation position, there is a penalty of about 20msec in recognition time.
This phenomenon, as well as the detailed underlying eye-movement behaviour that generates it, can be explained satisfactorily in terms of the combined action of three types of factors: sensory (acuity, lateral masking), lexical (word structure), and oculomotor. It seems that the time loss arising when the eye initially fixates in an “inconvenient” position arises predominantly, not from increased fixation durations, but from a higher probability of making additional fixations in the word. When the initial fixation is near the convenient viewing position, only a single fixation need be made. This fixation will be of comparatively long duration, and its duration may be affected by linguistic variables. When the initial fixation is far from the convenient viewing position: two fixations are made in a word; processing is distributed over both; each of these is of short duration; and their summed duration is about 75msec longer than if a single fixation had been made. In the two-fixation case the duration of the first of the two fixations is determined mainly by oculomotor and not by lexical processing.
