T. D. M. Roberts (1997) Understanding Balance: a Scheme With Snags. Psycoloquy: 8(06) Posture Locomotion (5)

Volume: 8 (next, prev) Issue: 06 (next, prev) Article: 5 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 8(06): Understanding Balance: a Scheme With Snags

Reply to Mergner et al. on Posture-Locomotion

T. D. M. Roberts
11 Menzies Drive
FINTRY, Glasgow, G63 0YG
Scotland, UK



The "alternative scheme" outlined by Mergner, Bolba and Becker (1997) depends on the unexamined acceptance of some of the very concepts that the book under review sets out to show to be invalid. This reply draws attention to some of the detailed points at issue.


Balance, connectionism, gestalt, learning, locomotion, motor control, proprioception, recognition.
1. The review of my book, Understanding Balance (1995), by Mergner et al. (1997) presents an alternative picture of balancing behaviour which is not so much a "violation of assumptions" (for which the authors in fact apologise) as an almost total disregard of the evidence presented in the book under review. The authors claim to be addressing the problem of the "non-ideal character of vestibular measurements", and imply that this problem is not solved in the book, whereas a good deal of space in the book is devoted to setting out the difficulties involved in supposing that sense organs, of any kind, deliver the sort of quantitative signals that one might wish to use as a basis for any scheme of motor control. In contrast, the reviewers do not specify the nature of the signals that are supposed to generate the variables to be used in their alternative scheme.

2. When it comes to co-ordinate transformations, the authors do not make clear what is to be transformed. A subjectively perceived direction may be represented by a unit vector, but a tilt requires a torque vector, with a magnitude as well as a direction representing the tilt axis. A tilt of the trunk implies a horizontal component of displacement, relative to the trunk, of the origin of the frame of reference for the skull, so that a co-ordinate transformation from head to trunk involves linear as well as rotational motion. This makes the computation process very complicated. It is not a simple matter of matrix multiplication such as might serve when all the motions are about a common vertical axis, as in the experiments reported in the earlier papers from Mergner's group.

3. In their scheme, the trunk is regarded as a rigid structure, articulated only at the atlanto-occipital joint and at the ankle. This scheme takes no account of the fact that, in addition, each vertebra has to be balanced on the one below, and that quite different problems have to be faced for balancing the pelvis on the legs. A different set of co-ordinate transformations is required for each joint. There is no indication in the proposed scheme of how motor control is to be handled for muscles spanning more than one joint. Indeed, balancing is supposed to be achieved entirely by adjusting the torque at the ankle. Force-plate studies (Roberts & Stenhouse, 1976), reported in the book, indicate that this is not how balance is maintained in practice.

4. In the latter part of the review, the authors appear to favour "psychophysical" explanations based on perceptual processes. In this they may be approaching a convergence with the conclusions set out in the book under review.

5. Along the way, however, they display many of the misconceptions highlighted in the book. For example, their use of "Centre of Mass" to mean "Centre of Gravity" is unfortunate, though quite common. The points indicated by the definitions of these two terms coincide only for objects that are radially symmetrical in all directions, both in shape and in density-distribution. They do not coincide for such a structure as the body of an animal or of a person.

6. The notion of a "gravitoinertial vector experienced by the body" involves further common misconceptions. Gravitational forces on the body are not perceived as such, and the so-called "inertial forces" are only pseudoforces that do not obey Newton's Laws, so it is not at all clear what sort of thing is intended by a "gravitoinertial vector".

7. This means that the concept of "nulling the angle between two perceived directions" breaks down right at the outset. The two directions referred to are not available except as mental constructs, and these have no quantifiable relation to such neural signals as may be generated by the vestibular and proprioceptive sense organs.

8. This article presents not so much a framework for a computational algorithm, which it may appear to be at first sight, as a metaphor for a possible segmentation of the associated mental processes

9. On the other hand, it is useful to have one's attention drawn to the various illusions of relative movement that have been uncovered in the series of experiments reported earlier by Mergner's group, involving separate imposed rotations of head, trunk, feet, and visual targets, in various combinations, about a common vertical axis.

10. The explanations offered for these illusions depends, however, on the supposed existence of a central mechanism for detecting, and compensating for, defects in the labyrinth signals. It is not at all clear how such a detection could be achieved.

11. An alternative explanation could be set up using, as a basis, the fact that an imposed neck torque is likely to generate a quite different afferent pattern from that accompanying a neck movement initiated by the subject either voluntarily or reflexly (Roberts, 1963). The relevant proprioceptors in the neck are muscle spindles, with gamma innervation, not joint receptors of the kinds found elsewhere. Their discharges are therefore likely to be affected differently by the different gamma activity associated with forced, as opposed to active, movements. Furthermore, in imposed movements, muscles that would be inhibited during an active movement are being forcibly stretched, and muscles which would contract to produce an active movement are being relieved of tension and may go slack. These differences between active and forced movements should be taken into account.

12. Concealment of known complexities allows the authors to set up a deceptively simple scheme which, while seductively attractive to the unwary, lacks direct application to the real state of affairs and may therefore be regarded as irrelevant to the study of balancing behaviour, except in so far as it may prompt the formulation of new experimental tests. It is important to remember that, whatever one's opinion of the interpretations put forward by others, one should always preserve a proper respect for their actual observations.


Mergner, T., Bolba, B. and Becker, W. (1997) A Modified Version of the Centre-of-Mass Balancing Hypothesis. PSYCOLOQUY 8(2) posture-locomotion.4.mergner.

Roberts, T.D.M. (1963) Rhythmic excitation of a stretch reflex, revealing (a) hysteresis and (b) a difference between the responses to pulling and to stretching. Quart. J. Exp. Physiol., 48, 328-345.

Roberts, T.D.M. and Stenhouse, G. (1976) The nature of postural sway. Agressologie, 17A, 11-14.

Roberts, T.D.M. (1995) Understanding Balance: The Mechanics of Posture and Locomotion. London, Chapman & Hall.

Roberts, T.D.M. (1996) Precis of: Understanding Balance: The Mechanics of Posture and Locomotion. PSYCOLOQUY 7(2) posture-locomotion.1.roberts.

Volume: 8 (next, prev) Issue: 06 (next, prev) Article: 5 (next prev first) Alternate versions: ASCII Summary