Three experiments, and a replication of each, investigated the noncons
cious prerecognition visual processing given left visual field (LVF) a
nd right visual field (RVF) letter inputs. Each input was a vertically
arrayed pair of letters in which three variables were manipulated: (1
) the same letter twice vs one each of two letters, (2) same vs 180-d
egrees difference in orientations within each letter pair, and (3) nor
mal vs mirror-image letter form. The procedure presented all pairs of
letter combinations in pairs of pre- and postmasked 10-msec flashes; t
he subject's task was to report which flash of each pair appeared to l
ast longer. When letter pairs differed on all three variables, RVF pre
sentations of mirror-image letters were judged to be longer than equal
presentations of normal letters; the reverse occurred for LVF present
ations. When one normal and mirror-image letter were presented, RVF pr
esentations of mismatched orientations were judged to be longer than m
atched orientations, and the reverse was true for LVF presentations. W
hen pairs of two normal letters were presented, no processing differen
ce between LVF and RVF presentations was observed. A fourth experiment
tested presence/absence detection of the letter pairs under the input
conditions of the main experiments and showed those conditions to pro
duce chance-level presence/absence detection. These results suggest th
at each hemisphere can perform its own prerecognition operations and t
hat neither hemisphere is necessarily specialized for any particular p
rerecognition visual operation.