You wonder if maybe you misunderstood what I said here, for example:
Because of this necessary correspondence, to allege that the perceptual signal is controlled according to the *reference signal, but to deny that the perceived observed value to which it corresponds by definition is controlled according to the observed reference value for it, is to miss the point entirely.
This is not a correspondence to ‘Real Reality’. It is a correspondence of one perception to another. It is a correspondence of a measurement by the experimenter (a perception) to a quantity (another perception) in the model or in a simulation which is by definition proportional to that measurement.
Claims about Real Reality come in only because that is the nature of the claims that any science makes. The point of doing science is to make a description of Real Reality. We know that these descriptions (a.k.a. scientific theories) are provisional, incomplete, liable to be superseded in some domains (Newtonian physics) or in their entirety (phlogiston theory of heat, luminiferous theory of electromagnetic propagation). Every theory of any science is an avowedly defective claim about Real Reality. You could say that the point of doing science is to reduce the portion of time that our senses deceive us, though that is really the point of view of applied science and engineering.
But what I wrote said nothing about such claims about Reality. What I wrote is only about the correspondence of one set of perceptions (‘external’ quantities that an experimenter measures as observations of behavior) and another set of perceptions (‘internal’ quantities that a modeler implements in a simulation of that behavior). This is a necessary correspondence because the quantities employed in a simulation that is built according to the PCT model are defined so that the ‘internal’ quantities are proportional to the ‘external’ quantities measured in experiments.
When the simulation runs successfully, we say (according to the PCT model) that p, the ‘internal’ quantity identified as the perceptual signal, is the controlled variable. In the simulation, p is controlled so that its value stays near the internally maintained reference signal r. But p is defined as a function of the input quantity q.i; q.i is a measure of the environmental effect on the subject’s sensory organ(s) from an aspect of the environment as perceived by the experimenter and as determined by the TCV; and r is defined as a linear function of the reference value that the experimenter measures of the experimenter’s perception of that aspect of the environment which the experimenter specifies as q.i.
As a successful simulation numerically replicates the measured behavior of the subject, the interpretation of the model is that p represents a Really Real rate of firing in the subject’s Really Real brain, controlled relative to a Really Real rate of firing in the subject’s brain called r; that this rate of firing p is a linear function of the input quantity q.i that is a measure of a Really Real aspect of the environment, and that the rate of firing r is a linear function of the measured reference value for q.i. The interpretation of the model, just like the interpretation of any theory of any science, is that q.i and its reference value are measures of something Really Real in the environment and that the firing rates p and r postulated by the theory are proportional to actual firing rates that are really happening in the Really Real brain of the subject. The firing rates p and r are theoretical entities postulated by the theory, but the interpretation of the theory is that they correspond to Really Real realities.
To justify this interpretation, we rely on the findings of prior sciences including physics, chemistry, and physiology. These sciences also correlate perceptions of different kinds with each other in systematic ways, construct complex perceptions (theories, models, simulations, hypotheses, tests, methodologies, etc.) and then interpret their measurements, models, and so forth to be representations of Real Reality. Every science then involves more experiments to test the proposed correspondence of perceptions to Real Reality.
Because simulations built according to the PCT model control input numbers that disturbance numbers would cause to deviate from reference value numbers if there were no control, using the same numbers that the experimenter measures and (unproveably) perceives to be in the Really Real environment (disturbance, reference value, and input), the interpretation of the model is that the structure of the model, and the subject-internal variables that it postulates (p, r, e, weights, combinings in diverse input functions and output functions, etc.), all correspond to structures that could be perceived in the Really Real subject if and when we have the means to make direct in vivo operations. Such observations would still be no more than perceptions, of course. The weight of science–PCT and all prior sciences on whose shoulders it stands–would strengthen the interpretation that these perceptions represent Really Real Reality. But science proves nothing conclusively. Proof is possible only for a closed, essentially tautological system such as mathematics or logic.
PCT works with correlations of experimental perceptions with imagined perceptions postulated by the theory. A simulation quantifies these perceptions, and demonstrates that a simulation constructed according to the theoretical model or PCT ‘behaves’ in the same way as observed subjects do. (I put ‘behaves’ in scare quotes because the behavior of living subjects is abstracted to numerical quantities.) This close correspondence justifies the interpretation that the internal structure of the simulation represents really real structure inside the subject. The body of such work justifies the interpretation of PCT that it describes how really real organisms function in the really real world.
One could refuse to indulge in such interpretations of PCT. But then what’s the point of doing science?
The neoscholasticism arguments about the CV do not depend upon the interpretation that theory describes reality. The observational perception q.i co-varies directly with the imagined perception p because the latter is defined as a linear function of the former. In the model and in a simulation to say that p is controlled but q.i is not is simply wrong. The control loop keeps q.i at or near the reference value that was measured for it in the environment simultaneously as it keeps p at or near r, and this is necessarily so because p is defined as a linear function of q.i and r is defined as a linear function of the reference value for q.i that was determined by the TCV. These are all correlations of one perception with another. To say that q.i is controlled if p is controlled makes no claim that q.i is a Really Real Reality in the Really Real environment. That claim is an interpretation of PCT.
The fact that we routinely make that claim of our perceptions is what makes all our discussions of this sort laughable. Are quarks Really Real? Strings? Dark Matter? Dark Energy? Physics and Cosmology are interpreted to say that the theoretical entities that they posit must be Real. But that’s an interpretation of their correlations of observed and imagined perceptions. Such is science.