size–weight illusion

Lifting weights was a favourite activity of experimental psychology in its early days, but now, although these experiments involve much of interest, they are unfashionable and are seldom even mentioned in textbooks. This may change, however, as manipulative skills in space at zero g give rise to new problems in relation to perception (see space psychology). One of the most theoretically significant of perceptual illusions, which also has practical importance, is the size–weight illusion: when objects are lifted by the hand, a larger object feels and is judged to be lighter than a smaller object of the same scale weight. This makes a nice kitchen experiment, which can easily be carried out, with a pair of tins of different size, partly filled with sugar or salt (or lead or sand) to have the same scale weight. When lifted a few inches from the table, the larger tin will feel lighter than the smaller tin — though they both have the same scale weight. Ideally, the tins should be lifted by identical handles or wire loops, so that there are no significant touch signals from the different sizes of the tins — for the whole point is that sensory signals of weight are the same for both tins, though the smaller feels heavier. This illusion is interesting because it shows very clearly that the sensation of the weight — which one might think is directly and simply given from forces on the hand and arm — is markedly affected by the assumption of how heavy it 'should' be. Further, it turns out that it is the surprise that the larger weight is not the heavier (as it would usually be) that produces the illusion. (Weight illusion also arises from the colours and shades of objects — darker objects being accepted as probably heavier than lighter ones.)

Even the 'purest' sensory signals must be 'read' or interpreted according to general knowledge or assumptions in order to provide information, sometimes fallible, about the objects handled. Thus, sensed weight is not given simply by skin and muscle receptors, but also from our knowledge that larger objects are usually heavier than smaller objects. Anticipated weight is important when lifting objects, since the muscle force must be set to give a reliable smooth lift and avoid injury. That we do not always correctly anticipate the force required is obvious from the familiar 'empty suitcase' effect: that is, when we pick up an empty suitcase assumed to be full, it flies up in the air! For then our anticipation of its weight was wrong — which shows that we do anticipate weight before we lift things.

The size–weight illusion is a large (up to about 30 per cent), stable, and repeatable effect, which offers an opportunity to challenge an assumption of psychophysics — that our ability to discriminate between small differences depends, simply, on relative stimulus intensities. It is possible to test whether the Weber–Fechner law depends (as is generally assumed) simply on the weight stimuli, or whether it is related also to our knowledge and assumptions of what objects are like. We can test this by finding out whether it is more difficult to distinguish differences between weights of small objects than it is in the case of larger ones. It turns out that there is an effect of size, and so of apparent weight, but it is not merely a matter of an increase of the Weber fraction in proportion to apparent weight. Rather, weight discrimination is best — that is, the Weber fraction is smallest — when the density of the lifted weights is about 1, which is roughly the average density of objects. Both lower and higher densities give impaired weight discrimination.

Why should this be so? A possible reason is that the neural signals for weight are compared with expectations, and that the signalled and anticipated weights are nulled — as in delicate measuring instruments such as Wheatstone bridges. Cancelling of received input against anticipated values is useful as a way of gaining sensitivity over a wide range of input values, even when the components (whether electronic or those of the nervous system) have only small dynamic ranges. A nulling arrangement gives high stability even if the components are liable to 'drift' or fatigue, as neurons are. That our weight-sensing system is labile is obvious — for after carrying a heavy object even for a short time the arm, and even the whole body, feels light when it is released. And the arm may float up, almost out of control. This is the basis of the party 'levitation' trick: after someone's shoulders have been pushed down in the dark, that person may have the impression of rising towards the ceiling, for such is the lability of the nervous system.

(Published 1987)

— Richard L. Gregory

Bibliography
• Ross, H. E. (1969). 'When is a weight not illusory?' Quarterly Journal of Experimental Psychology, 21.
• — —  and Gregory, R. L. (1964). 'Is the Weber fraction a function of physical or perceived input?' Quarterly Journal of Experimental Psychology, 16/2.
• — —   — —  (1970). 'Weight illusions and weight discrimination: a revised hypothesis'. Quarterly Journal of Experimental Psychology, 22.

---