Eckhard Hess noticed it, by his own account, while looking at pictures of animals in dim light and being told by his wife that his pupils had become very large. The observation turned into a research program at the University of Chicago through the 1960s, and pupillometry has been a serious tool ever since — though not quite the tool Hess thought it was.
The pupil is controlled by two opposed muscles under autonomic control. The sphincter constricts under parasympathetic activity; the dilator expands under sympathetic activity. Neither has any voluntary motor pathway. You cannot dilate your pupils by deciding to, and no amount of training will give you the ability. This makes pupil size the least manipulable observable channel in this entire part of the guide.
What it responds to is well established. Light, obviously and overwhelmingly, which is the practical constraint on using it for anything. Beyond that: sympathetic arousal, and cognitive load. Daniel Kahneman's early work with Jackson Beatty in the 1960s established the pupil as a reliable index of mental effort — the pupil dilates as a task gets harder, tracks the moment-to-moment demand within a task, and contracts when the effort is released. Kahneman later described using pupil size to know exactly when a subject had given up on a problem. That line of work is solid, has replicated, and underpins the use of pupillometry in cognitive psychology and human factors research today.
Hess's stronger claims are the ones that have not fared as well. He reported that pupils dilate to images the viewer finds attractive and constrict to images they find distasteful — including a much-cited claim about differing responses to images by sex and sexual orientation — and he suggested pupil size functions as an interpersonal attraction signal, noting that belladonna, used cosmetically for centuries, is named for exactly this. A 2021 Monash-led project attempted to replicate five of Hess's studies and found partial support at best; the constriction-to-disliked-stimuli claim in particular did not hold up well, and several effects were much smaller than reported.
So the honest position is a split verdict. The channel is real, involuntary, and measures something. What it measures is arousal and cognitive effort, undifferentiated by valence. The folk interpretation — dilated pupils mean attraction — over-reads a signal that cannot distinguish attraction from fear from difficult arithmetic.
Which turns out to be the recurring structure of this entire part. A genuine autonomic signal, a real measurement, and an interpretive leap from arousal to a specific cause that the data does not support.
The practical use is therefore the same as the pacifying behaviors two chapters ago: time-locking. Under controlled illumination, dilation that occurs precisely when a particular photograph, name or number is presented tells you that item produced arousal or effort in this person. Card players have watched for it for a century, which is one of several reasons sunglasses appeared at poker tables. Consumer researchers use it because it does not depend on what people say.
And the same limit applies. It tells you that something happened. It does not tell you what, and the temptation to fill that in from context is the failure mode that the rest of Part 2 documents.
See Defense: separating arousal from meaning