Reading processes: on the recognition of single words in eccentric vision

In normal reading, the eyes show fixational pauses separated by quick jumps (saccades) along the lines. During the pauses, foveal and parafoveal images give rise to the recognition of words, and the successive images are centrally integrated into a meaningful whole. In these recognition processes, visual as well as language factors are important. Analysis of the processes is a main theme of our research on normal reading and we concentrate on the visual factors. Our approach is in line with that of several authors around the turn of the century (Pillsbury, 1897; Erdmann and Dodge, 1898; Messmer, 1904; Huey, 1908; Korte, 1923). One focus of interest is the functional visual field, i.e. the visual field from which some particular information is picked up during a single fixational pause. Our technique of investigation is to determine in the para foveal field the limits of eccentricity at which language material can still be recognised in brief presentions. The functional visual field is not an invariant of the visual system, but it depends decisively on the stimulus material used and on the type of recognition required. For example, if isolated typewriter letters have to be recognised at normal reading distances, the functional visual field extencs some 15 typing spaces L(eft) and 15 spaces R(ight) of the point of fixation, if a criterion of 80% correct responses is applied. For letters embedded between two letters Ix/, we find the field much reduced, viz. to about 5 spaces Land 5 R. Apparently, the visual processing of letters involves adverse interaction effects. These effects are asymmetrical in a sense that the interference is stronger if only a single /x/ is at the peripheral side of the unknown letter than if it is at the foveal side (Woodworth and Schlosberg, 1954; Mackworth, 1965; Bouma, 1970). As concerns the recognition processes proper, an analysis has been carried out on the relative importance of certain letter properties for the recognition of isolated letters (Bouma, 1971). Because of the interaction effects, it is not a simple matter to extrapolate this analysis towards letters presented in groups. The possible nature of some interaction effects is the subject of separate contributions in this issue (Beerens and Bouma, 1970 a,b). Owing to interaction effects, letters embedded in lettercombinations have lower recognition scores and correspondingly narrower functional fields than have initial and final letters. Moreover, because of the asymmetry, L initial letters and R final letters (outward letters) have wider functional fields than L final letters and R initial letters. The question which will concern us here is whether these effects do operate in the recognition of single words. If this is so, outward letters would be expected to playa prominent role in word recognition. We shall also look into the contributions of second and pre final letter positions and into word length and word contour. We report here on an exploratory experiment, in which subjects read single words, presented briefly in eccentric (parafoveal) vision. The stimulus words were of various lengths and of two categories of frequency of occurrence in the printed language