Butler B E, Mewhort D J, Browse R A
Department of Psychology, Queen's University, Kingston, Ontario, Canada.
Percept Psychophys. 1991 Jan;49(1):91-9. doi: 10.3758/bf03211620.
Feature-integration theory postulates that a lapse of attention will allow letter features to change position and to recombine as illusory conjunctions (Treisman & Paterson, 1984). To study such errors, we used a set of uppercase letters known to yield illusory conjunctions in each of three tasks. The first, a bar-probe task, showed whole-character mislocations but not errors based on feature migration and recombination. The second, a two-alternative forced-choice detection task, allowed subjects to focus on the presence or absence of subletter features and showed illusory conjunctions based on feature migration and recombination. The third was also a two-alternative forced-choice detection task, but we manipulated the subjects' knowledge of the shape of the stimuli: In the case-certain condition, the stimuli were always in uppercase, but in the case-uncertain condition, the stimuli could appear in either upper- or lowercase. Subjects in the case-certain condition produced illusory conjunctions based on feature recombination, whereas subjects in the case-uncertain condition did not. The results suggest that when subjects can view the stimuli as feature groups, letter features regroup as illusory conjunctions; when subjects encode the stimuli as letters, whole items may be mislocated, but subletter features are not. Thus, illusory conjunctions reflect the subject's processing strategy, rather than the architecture of the visual system.
特征整合理论假定,注意力不集中会使字母特征改变位置并重新组合形成错觉性结合(特雷斯曼和帕特森,1984)。为了研究此类错误,我们在三项任务中使用了一组已知会产生错觉性结合的大写字母。第一项任务是条形探测任务,结果显示存在整个字符的位置错误,但没有基于特征迁移和重组的错误。第二项任务是二选一的强制选择检测任务,该任务让受试者专注于子字母特征的有无,并显示出基于特征迁移和重组的错觉性结合。第三项任务同样是二选一的强制选择检测任务,但我们操控了受试者对刺激形状的认知:在大小写确定条件下,刺激始终为大写字母,但在大小写不确定条件下,刺激可能以大写或小写形式出现。大小写确定条件下的受试者产生了基于特征重组的错觉性结合,而大小写不确定条件下的受试者则没有。结果表明,当受试者能够将刺激视为特征组时,字母特征会重新组合形成错觉性结合;当受试者将刺激编码为字母时,整个项目可能会出现位置错误,但子字母特征不会。因此,错觉性结合反映的是受试者的加工策略,而非视觉系统的结构。