Justification of final cause in action theory.

first: 2011-10-20 last: 2011-10-22

The concept and mathematical basis of causation in action theory with Feedback loops

The basic concept of a feedback loop control is the structure by which the actual value of a process indicator (R), for instance a temperature, is compared with a desired value (D). The difference between these two is used to control the process.

As shown in the diagrams in Action theory and final cause Fred Dretske indicates the interrelation but not the key element, being the substraction of desired and realized output as the "driving force" in a (negative) feedback loop system. 

Both Davidson and Dretske refer to the example of a thermostat, but they do not refer to this key of the basic concept. Dretske mentions the bi-metal strip that forms the heart of old thermostats. The disadvantage of this example is that it is an ingenious combination of a measuring, reference and switch function, but this makes it not ideal to understand the basic concept of a feedback loop system.

A more specific concept structure is the so called servo loop. These processes can also be of mechanical nature as used in factory mechanization and robotics. De basic structure is the same but the difference is that the desired value is the variable. By changing the variable the system realizes the desired value, also when environmental conditions (with in some limitations)change. This structure is shown in figure A1.2.



 \small figure A2.2: Mathematical base of feedback loop concept\normalsize

The key characteristic again is that the difference between the desired and the actual value is used to realize a result. The control algorithm supplies the reference for the force that realizes the required result.

The nature of the desired input and realised output can be the same as for instance in case of power steering. But more interesting in this context is that in general there will be an transformation in the nature of input and output. The input value can be any kind of electrical signal or even a digital code with the required output of a mechanical movement.

Feed Back Loops in neural systems

 Already in the 50th of last century McGulloch en Pitt showed that there exist feedback-loop configurations of neurons in the brain. [Pask 1961] In practice with this kind of neuron structure more feedback loop configurations work together where one configuration can even correct an other one.

The Vestibulo Oculair Reflex is a kind of prototype example. The eye muscles make corrections for the movement of the head in order to stabilize the projection on the retina. This control mechanism is handled by a group of neuron cells in the cerebellum. These cells combine the inputs from the organ of balance, the eye muscles and the retina and transmit these to a special group of cells. These cells control the action. In case the feedback is insufficient or too strong the shift on the retina is send to a third group of cells (the lower olivecernels), these send a correction signal to the Purkinjecells which adjust the feedback mechanism.

concept of adaptive system

When we extrapolate these structures to the processes in the brain in general it becomes clear that one can not speak of states in the brain as entities but that we should consider these as complex dynamic systems that can also be understood and explained by a final cause explanation.

Justification for explanations

 In todays philosophy of science, as Wesley C. Salmon points out, the Deductive-Nomological (DN) model of Hempel-Oppenheimer is generally considered as too strict. [Salmon 1972 :20-23 paragraph 'problems for the DN pattern of explanation]

Scientific understanding is, after all, a complicated affair; we should not be surprised to learn that it has many different aspects. Exposing underlying mechanisms and fitting phenomena into comprehensive pictures of the world seem to constitute two important aspects.' ...'these two types of understanding frequently overlap. When we find that the same mechanisms underlie different types of natural phenomena this ipso facto constitutes theoretical unification [Salmon 1972: 39]

Salmon refers to three main elements:
The first element to be considered, as also identified by Van Fraassen, is an analysis of the problem under consideration... (also [Juarrero 1999: 182] refers to this aspect)
To be followed by the two other elements being the clarification of the underlying mechanisms and the fitting of the phenomena in a consistent representation of the world.
These are the elements also applied in this website.

It should also be clear from above explanations that the specific property of the feedback-loop configuration can not be reduced to a lower physical level. This observation is compatible with the more generic metaphysics of Lynne Rudder Baker. [Baker2007 Chap. Commonsense causation]

With the knowledge of feedback processes we can pick up Davidson's example to explain how a mental desire can result in a bodily action. We start at the stage where the `` practical reasoning'' as mentioned above has decided to call for a taxi. The person with a paralyzed arm and carrying something with his other arm is now equipped with a robotic harnas-arm for the paralyzed arm and a voice recognition device that can start the servo system of the robotic arm by the command to raise the arm when the vocal `taxi' is detected. As Davidson states ``I take it as obvious that linguistic behavior is intentional" we can explain that the intention caused the action. Assuming the similarity of the earlier described neural systems is accepted, this would also applied to non paralyzed arms without robotic harnas.


 More complex systems

 The above mentioned example of raising the arm covers only a well defined relative low level action process, starting with an already available mental desire to raise an arm as typical for calling a taxi. The desire could be made well defined as an intention by the linguistic interface to avoid subconscious processes that might have plaid a role in the decision process to call for a taxi rather that to walk or use public transport. That kind of decision processes requires other type of analysis.

It is interesting to notice that Salmon, in above mentioned quote, moved from a justification of explanation to understanding. At the level of basic mechanisms final cause can be used as a justifiable explanation. In case of more complex systems criteria for the label justified explanation should be developed. Gradually one will move into understanding.


 Conclusion

By referring to the basic feedback loop structure the final cause can be used as a justifiable explanation. With its underlying mathematical relations this explanation can be as solid as the working cause explanations based on Newtons laws.