In the mechanistic world view, there is no place for purpose or goal-directedness. All mechanical processes are determined by their cause, which lies in the past. A goal, on the other hand, is something that determines a process, yet lies in the future. To a Newtonian scientist, the idea that an as yet non-existent, future state could influence the present, seems wholly unscientific, not to say mystical.
The thesis that natural processes are determined by their future purpose is called teleology. It is closely associated with vitalism, the belief that life is animated by a vital force outside the material realm. Our mind is not an aimless mechanism; it is constantly planning ahead, solving problems, trying to achieve goals. How can we understand such goal-directedness without recourse to the doctrine of teleology?
Probably the most important innovation of cybernetics is its explanation of goal-directedness. An autonomous system, such as an organism, or a person, can be characterized by the fact that it pursues its own goals, resisting obstructions from the environment that would make it deviate from its preferred state of affairs. Thus, goal-directedness implies regulation of--or control over--perturbations.
A room in which the temperature is controlled by a thermostat is the classic simple example. The setting of the thermostat determines the preferred temperature or goal state. Perturbations may be caused by changes in the outside temperature, drafts, opening of windows or doors, etc. The task of the thermostat is to minimize the effects of such perturbations, and thus to keep the temperature as much as possible constant with respect to the target temperature.
On the most fundamental level, the goal of an autonomous or autopoietic system is survival, that is, maintenance of its essential organization. This goal has been built into all living systems by natural selection: those that were not focused on survival have simply been eliminated. In addition to this primary goal, the system will have various subsidiary goals, such as keeping warm or finding food, that indirectly contribute to its survival. Artificial systems, such as thermostats and automatic pilots, are not autonomous: their primary goals are constructed in them by their designers. They are allopoietic: their function is to produce something other ("allo") than themselves.
Goal-directedness can be understood most simply as suppression of deviations from an invariant goal state. In that respect, a goal is similar to a stable equilibrium, to which the system returns after any perturbation. Both goal-directedness and stability are characterized by equifinality: different initial states lead to the same final state, implying the destruction of variety. What distinguishes them is that a stable system automatically returns to its equilibrium state, without performing any work or effort. But a goal-directed system must actively intervene to achieve and maintain its goal, which would not be an equilibrium otherwise.
Control may appear essentially conservative, resisting all departures from a preferred state. But the net effect can be very dynamic or progressive, depending on the complexity of the goal. For example, if the goal is defined as the distance relative to a moving target, or the rate of increase of some quantity, then suppressing deviation from the goal implies constant change. A simple example is a heat-seeking missile attempting to reach a fast moving enemy plane.
A system's "goal" can also be a subset of acceptable states, similar to an attractor. The dimensions defining these states are called the essential variables, and they must be kept within a limited range compatible with the survival of the system. For example, a person's body temperature must be kept within a range of approximately 35-40 degrees C. Even more generally, the goal can be seen as a gradient, or "fitness" function, defined on state space, which defines the degree of "value" or "preference" of one state relative to another one. In the latter case, the problem of control becomes one of on-going optimization or maximization of fitness.
Heylighen F. & Joslyn C. (2001): "Cybernetics and Second Order Cybernetics", in: R.A. Meyers (ed.), Encyclopedia of Physical Science & Technology , Vol. 4 (3rd ed.), (Academic Press, New York), p. 155-170