The role of redundancy in the object-line effect

Abstract

Subjects' identification of briefly presented target lines can sometimes be facilitated by including noninformative lines of context in the display (the object-lineeffect). Previous explanations for this effect have claimed that stimulus properties such as three-dimensionality, connectedness, structural relevance, line-masking, or fixation-point detail, allow line-in-context stimuli to be processed more efficiently than single-line stimuli. Instead of postulating special processing consequences for these stimulus properties, we propose two general preconditions for the object-line effect. Three experiments demonstrate that the effect occurs when lines-incontext have perceptual attributes that are correlated with the target lines and when the linesin-context are perceptually more dissimilar from one another than the single target lines.

Perceptual versus postperceptual mediation of visual context effects: Evidence from the letter-superiority effect

“…Although not all of the findings reported by researchers of the OSE and OLE can be explained solely in terms of structural redundancy within the experimental stimulus set (see Enns & GHani, 1988;Lanze et al, 1985), the findings reported by Enns and Prinzmetal (1984) clearly demonstrate the potential relevance of this factor to the study of context effects in object perception.…”

“…The re-VISUAL CONTEXT EFFECTS 167 suIts of this experiment indicated that, across stimulus sets, line-in-context identification accuracy was positively correlated with the magnitude of the structural redundancy. In fact, contexts (i.e., triangle or arrow) facilitated lineidentification performance only under conditions of abovezero structural redundancy within the stimulus set.Although not all of the findings reported by researchers of the OSE and OLE can be explained solely in terms of structural redundancy within the experimental stimulus set (see Enns & GHani, 1988;Lanze et al, 1985), the findings reported by Enns and Prinzmetal (1984) clearly demonstrate the potential relevance of this factor to the study of context effects in object perception.2. Preexperimental knowledge ofstructural redundancies.…”

“…This is the case because identifying the presented object may enable subjects to infer the presentation of its parts, but not their color. This argument applies just as well to any structural redundancies that may be learned during the experimental sessions via the availability of diagnostic emergent features such as the ones proposed by Enns and Prinzmetal (1984).…”

“…This was assumed because given any of N contexts, C.. C 2 , ••• Ci, ... Cn, or the null context, Co (i.e., line alone), the alternative target lines, for example, A and B, had equal probability of occurrence, that is, p(Alcj) = p(BICi). However, as correctly pointed out by Enns and Prinzmetal (1984), target lines may interact with context lines to produce higher order local or global forms such as angles or bounded regions, which in tum may be correlated with target lines resulting in greater than zero redundancy. If, for example, the interaction between Line A and each one of the contexts, [Ci&A], always produces an angle in the shape of an arrow, whereas the interaction of Line B with the same contexts, [Ci&B], always produces a bounded region in the shape of a triangle, then there is perfect redundancy (i.e., a correlation of 1) between the higher order forms arrow and triangle, and the Target Lines A and B.…”

“…This is the case because identifying the presented object may enable subjects to infer the presentation of its parts, but not their color. This argument applies just as well to any structural redundancies that may be learned during the experimental sessions via the availability of diagnostic emergent features such as the ones proposed by Enns and Prinzmetal (1984).The suggested paradigm allows for several possible controls necessary for disentangling familiarity and figural goodness. Whenever possible, objects and nonobjects would be equated with respect to figural goodness.…”