First Order Control

First Order Control is when a desired result is attempted by adjusting a scalar (first order) control. Such mechanisms, where first order change sucessfully produces desired results, are called first order mechanisms. First order control is contrasted to Higher Order Control.

While many simple mechanisms are completely first order, most complex mechanisms are only first order within a range of operation.

And example of such a control mechanism is a trackball used to control a mouse cursor: a particular motion of the trackball causes a corresponding motion of the mouse cursor, until the cursor reaches the edge of the display. (The rigorous will note that a trackball is truly two scalar controls.)

Failure of First Order Control: Range [obviously] if a control mechanism is manipulated in a first order fashion outside the range where it exhibits first order behavior, the results may be unpredictable. The failure to recognize that a mechanism is not first order or is operating outside its first order range is one of the primary causes of those mishaps described as being due to operator error. An example of this is the behavior of a car on rain-slickened asphalt: it is not true that the car cannot be controlled, but the rules for controlling the car have changed (see Higher Order Control), and if an inexperienced driver does not understand the changes a mishap may result.

Failure of First Order Control: Precision if a control mechanism is less fine-grained than observable distinctions in the result, then it is usually impossible to achieve specific results with any regularity. The attempt to control a mechanism below the granularity of its input generally results in wavering or wobble, or can have catastrophic results as with Negative feedback loops[?]. Note that this situation must arise any time a continuous system is controlled by a discretely valued input.

Failure of First Order Control: Slack closely related to "precision" and directly involved in most Negative feedback loops[?] is "slack" the difference in time or distance between the input to a system and the result of that input. Delay in time must be allowed for when adjusting the system: changes in the system in the future because of adjustments made in the past must be allowed for when making present adjustments. The simple solution ot this problem is to allow the system to "settle". Examples of difference in distance generally come from Static friction[?]: a control may be perfectly linear once static friction has been overcome, while being very non-linear from a standing start. A common example is pushing a heavy crate across a not very slick floor: a great deal of force may be applied without moving the crate at all; a fraction more and it shoots across the floor: after that less force is needed to keep it moving.

Social implications of First Order Control

Many common beliefs about social organization, such as "Increasing police decreases crime", or "corruption can be combatted by legislation" imply a first order mechanism. In order to understand various social disasters it is important to recognize the limitations of range, precision, and slack in social controls.

For instance to reduce the per capita murder rate in a society it may be effective to increase funding for police departments and homicide investigations. But to reduce the per capita murder rate to exactly zero could require a fundamental change in every aspect of the society's organization that could not be afforded and would have many other implications. The implication would be that it is better to simply accept a certain amount of murder as inevitable; but such a position may in turn be seen as non-viable politically, causing politicians to espouse actions or goals that are functionally impossible.

First order social mechanisms based on false assumptions are even more problematic.

As a hypothetical example Houses are being burglarized at night. It is falsely assumed that the crimes are being committed by adolescents. A first order solution is imposed: a curfew restricting adolescents to their family homes after dark. The police are directed to enforce this curfew, but forbidden to use deadly force. Many adolescents are arrested by the police; many adolescents are seen by the police but fail to be apprehended. The newness of the situation and the increase in police patrols causes a temporary decline in burglaries. This is heralded as "proof" that the curfews are working to decrease crime, and "proof" that adolescents were responsible for the burglaries.

But in time burglaries begin to rise again to the former levels, as the burglars become more comfortable with the new rules. Addressing this in first order fashion, the curfew is tightened. Eventually the police will be directed to use deadly force against the adolescents in order to collect those who had been successfully evading their patrols. The continuance of burglaries is used as "proof" that adolescents continue to violate curfew.

Fallacy of First Order Control a Vicious cycle[?] where a social mechanism is treated as a first order mechanism, but the way the mechanism is operated causes the opposite of the "desired" effect, sometimes but not always in conjunction with the temporary appearance of success.

As a specific example: Civil unrest by African Americans in the 1960s led to increased persecution (the control mechanism) of arbitrary African Americans by the Police, FBI, and other formal and informal groups. Increased persecution caused radicalization of previously ambivalent African Americans; some of these radicalized citizens contributed violent crimes and led civil unrest. The increase in violent crimes led to increased persecution; and so on.

On the surface the Fallacy of First Order Control seems undesirable, but it can offer Secondary Gain to some of the participants, so it is often intentionally provoked. Certain to benefit from fallacious control that involves violence are hawkish and fascist political parties, armament manufacturers, and security forces. Other forms of fallacious control exist and should not be discounted, as when increased taxes inspire increased tax evasion: this can lead to the impoverishment of honest tax-payers and success for the dishonest.

References

Change Paul Watzlawick[?] [find urls]