Visual illusions have been studied extensively, but their time course has not. Here we show, in a sample of more than 550 people, that unrestricted presentation times—as opposed to presentations lasting only a single second—weaken the Ebbinghaus illusion, strengthen lightness contrast with double increments, and do not alter lightness contrast with double decrements. When presentation time is unrestricted, these illusions are affected in the same way (decrease, increase, no change) by how long observers look at them. Our results imply that differences in illusion magnitude between individuals or groups are confounded with differences in inspection time, no matter whether stimuli are evaluated in matching, adjustment, or untimed comparison tasks. We offer an explanation for why these three illusions progress differently, and we spell out how our findings challenge theories of lightness, theories of global-local processing, and the interpretation of all research that has investigated visual illusions, or used them as tools, without considering inspection time.

视觉错觉已经被广泛研究，但它们的时间进程还没有。在这里，我们在超过 550 人的样本中展示了不受限制的演示时间——与仅持续一秒的演示相反——削弱了艾宾浩斯错觉，通过双倍增加加强了亮度对比，并且不会通过双倍减少来改变亮度对比。当展示时间不受限制时，这些错觉会以相同的方式（减少、增加、不变）受到观察者注视它们的时间的影响。我们的结果意味着无论是在匹配、调整还是不定时的比较任务中评估刺激，个体或群体之间错觉强度的差异都会与检查时间的差异混淆。我们解释了为什么这三种幻觉的进展不同，