Confusion worse confounded

Commentary on Sheldrake

Journal of Consciousness Studies, 12:6, 64-6, 2005

This is the version originally submitted. It was slightly edited before publication.

Sheldrake’s two papers are so deeply confused that they should never have been published in JCS. Given that they are published I shall discuss just two worrying confusions; one from each paper.

In Part 1. Sheldrake consistently mixes up the sense of being stared at that derives from the normal senses of vision or hearing, with a putative sense that he claims can operate without normal sensory cues. He might have avoided confusion by giving a name to this proposed paranormal sense, and then made clear, throughout the paper, which he was referring to at different times, but he did not. The reader might try to understand the paper by assuming that when Sheldrake talks about “the sense of being stared at” he always means this proposed paranormal sense, but this is not so; he uses the same phrase to refer to the normal sense. For example, in the opening quotation from Conan Doyle the man looks up to meet the eyes of the person staring at him, implying that the starer was probably visible in peripheral vision.

This confusion permeates the paper. Detectives should not stare at someone’s back for the obvious reason that they might not be able to avert their gaze fast enough, or convincingly enough, if the person happens to turn round. Detection with binoculars might seem impossible by normal means, but binocular lenses can have highly reflective surfaces and may look like eyes from a distance. The surveys discussed seem to include some questions that refer only to a paranormal sense and some that might include normal sensing. This means that when Sheldrake discusses the evolutionary function of the sense of being stared at the reader cannot tell whether he means a normal or paranormal sense. Obviously there would be an evolutionary advantage in being able to detect another’s gaze and in fact we, and other species, have visual systems designed to be good at this. A pair of eyes is a salient stimulus. We can pick it out easily from complex scenes, attention is automatically drawn to it, and eye movements are made towards it without prior identification. We can also tell from very small differences whether someone is looking straight at us and focussing on us, or not.

Sheldrake asserts that research has neglected the sense of being stared at because people believe it is impossible. Unless he makes it clear that he is referring to a paranormal sense, this claim is ridiculous. He also refers to a taboo against psychic phenomena. As a former parapsychologist, I do not believe there is any such taboo. The fact is that evidence for paranormal phenomena is weak and usually unreplicable, and there is no plausible theory to explain such evidence as there is. Most scientists choose not to investigate paranormal claims, not because of a taboo, but because they have more promising and exciting things to spend their precious time and research resources on.

There follows a review of experiments some of which were done under conditions that would rule out the use of the normal senses, and some of which would not. Sheldrake admits to the flaws in some of these experiments but then goes on to use all of them – flawed or not – for his assessment of the overall pattern of results. He also gives detailed results of some highly flawed studies but then gives only a cursory description of experiments that would, if valid, be very impressive.

Finally, Sheldrake’s conclusion reveals the same confusion. Of course most people say they have sensed when they are being stared at. It is a normal, evolved, human talent. Whether there is also a paranormal ability to detect staring, as Sheldrake seems to believe, remains unknown, but we will not get closer to knowing the truth by reading this misleading and confused paper.


In part 2 the entire discussion is marred by a confusion between active v. passive theories and intromission v extramission theories. Sheldrake points out that intromission theories have tended to regard vision as passive while extramission theories have tended to regard it as active. This may be true as a historical fact but this is no reason to conflate two fundamentally different distinctions. All modern theories of vision are intromission theories; they assume that light enters the eye and that nothing leaves it. This fits with the physics of light, the structure of the eye, and the principles of sensory systems. Any theory that proposes, as Sheldrake does, that “An influence seems to pass from the observer to the observed” is a paranormal theory. The problem for normal theories of vision is to understand how incoming information eventually leads to visually guided behaviour and visual experience, and they vary from more active to more passive theories. There has recently been a revival of Gibson’s ecological approach to vision and renewed enthusiasm for so-called active and embodied theories of vision, but these do not involve anything leaving the eye and projecting out into the world; their main point is to emphasise how much active processing of the incoming information has to be achieved.

The most extreme of these theories is probably O’Regan and Noe’s sensorimotor theory of vision. This does away entirely with any notion of a picture-like representation in the visual system and replaces it with the idea that vision is a kind of action; seeing is doing. As they put it, seeing is mastering sensorimotor contingencies, or playing with the relationships between one’s own actions and the changing input. This is an active theory par excellence but it contains no hint of extramission. Sheldrake praises Gibson’s theory because it “leaves open the possibility of an interaction between the perceiver and the perceived.” But this is crazy – for what theory does not? Vision is and must be an ongoing interaction between the perceiver and perceived. By confusing two fundamentally different distinctions between theories of vision Sheldrake has created nothing but an unhelpful muddle.


O’Regan, J.K. and Noë, A. (2001) A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences, 24(5), 939-1011