Dehaene S, Cohen L
INSERM, Laboratoire de Sciences Cognitives et Psycholinguistique, EHESS & CNRS, Paris, France.
J Exp Psychol Hum Percept Perform. 1994 Oct;20(5):958-75. doi: 10.1037//0096-1523.20.5.958.
Do people have to count to determine visual numerosity, or is there a fast "subitizing" procedure dedicated to small sets of 1-3 items? Numerosity naming time and errors were measured in 5 simultanagnosic patients who suffered from severe difficulties in serial counting. Although these patients made close to 100% errors in quantifying sets comprising more than 3 items, they were excellent at quantifying sets of 1, 2, and sometimes 3 items. Their performances in visual search tasks suggested that they suffered from a deficit of serial visual exploration, due to a fundamental inability to use spatial tags to keep track of previously explored locations. The present data suggest that the patients' preserved subitizing abilities were based not on serial processing but rather on a parallel algorithm dedicated to small numerosities. Several ways in which this parallel subitizing algorithm might function are discussed.
人们必须通过数数来确定视觉上的数量吗？还是存在一种专门用于处理1至3个项目的小集合的快速“数感”程序呢？我们对5名患有严重连续计数困难的同时失认症患者的数量命名时间和错误情况进行了测量。尽管这些患者在量化包含超过3个项目的集合时出现了近100%的错误，但他们在量化1个、2个以及有时3个项目的集合时表现出色。他们在视觉搜索任务中的表现表明，由于根本无法使用空间标记来跟踪先前探索过的位置，他们存在连续视觉探索缺陷。目前的数据表明，患者保留的数感能力并非基于序列处理，而是基于一种专门用于小数量的并行算法。本文讨论了这种并行数感算法可能发挥作用的几种方式。
