Source
Turvey (1977), “Preliminaries to a theory of action with reference to vision”, LIT-904, read in NOTE-826 in its preprint form, Haskins Status Report SR-41 (1975), pp. 1–49: the argument against the keyboard executive (pp. 5–6), coordinative structures (pp. 7–9), executive ignorance (pp. 14–15), tuning and pretuning (pp. 18–30), and vision and tuning (pp. 32–38).
What was actually shown
The argument. Innervation and movement are equivocally related: for a limb segment, I·d²α/dt² = F(E, α, dα/dt) + G(α), so the same impulses give different movements in different states and loads. The controlling system has far fewer degrees of freedom than the controlled one. So the executive cannot command muscles. It sets “the modes of interaction” of lower units, “reflexes and functional combinations of reflexes”, which are “relatively autonomous”. What it specifies is not a state but an equivalence class, a “ballpark” from which the lower level brings the state in by itself.
The evidence at the spinal level (reviewed, not new).
- Tendon and H-reflexes probed between command and movement are enhanced, specifically in the muscles about to act and nonspecifically elsewhere, beginning about 60 msec before agonist activity (Gurfinkel et al.; Gottlieb et al.; Kots).
- Reciprocal inhibition of the antagonist is enhanced in the last 30 msec without any change in the antagonist’s own excitability (Kots and Zhukov).
- The reorganization goes from diffuse to localized as movement nears.
- A joint behaves like a spring, F = −S₀(l − λ₀): its equilibrium is set by parameters fixed beforehand and is independent of initial conditions, even when loads are removed unpredictably before the movement (Asatryan and Fel’dman; Fel’dman).
- In one data set the reflex changes do not track the EMG, so tuning and activation appear to be separately controlled (Gottlieb et al.).
What could have come out otherwise: the probes could have shown no change before the first EMG sign of movement, which would have made the pretuning only the start of activation. They did not.
What this does not say
- Not shown above the spinal level. That every coordinative structure, at every level, is a “biasable, self-regulating vibratory system” tuned the same way is extrapolated “on the assumption that the regulatory principles for large systems follow very much the pattern of small systems” (p. 30).
- Not shown for environment-related acts. Turvey says there were “no experiments on segmental pretuning for acts that depend on the detection of environmental properties” (p. 22). That optical properties set tunings “without the intervention of executive procedures” is argued from the cat’s leap and the mantis’s strike, not measured.
- Not a model. The equation m·x″ + k·x′ + s·x = 0 is “given only to illustrate a principle” (fn. 7).
- Not the source of the term. Turvey borrows “coordinative structure” from Easton (1972a) and widens it. Nor is reciprocal compensation, later taken as its signature (LIT-902), part of this account.
- Not about more than one body. The units are reflexes and their combinations within one organism.
Connections
- THEORY-069. The Haken–Kelso–Bunz account of bimanual coordination is the later dynamical form of the same programme; this account precedes it and models units as tuned springs rather than coupled oscillators in relative phase.
- CLAIM-175. If “coordinative structure” is used for the dancing couple, the defensible content is this account’s: two relatively autonomous units, each tuned rather than commanded, with the details left to them. Extending it across bodies is the record’s step, not Turvey’s.
- Bernstein (LIT-907, unread) is the source of the degrees-of-freedom problem the account answers.