Attention : Script is not a complete representation of the oral lecture !!! 
	Script is not yet completely finished !!!
 

Observation and  Measurement; general Ideas

The general case of everyday observation is not sufficient for scientific purposes.  In the non-scientific case the  observer (Obs)   is  observing  the H-UI-M-Relationsship as an object without additional aids, i.e. the process of measurement is 'empty'. In this case does the observer relays only on his subjective experience as far as this experience is conscious. Thhis subjective experience is extremely powerful, because it is based on a biological machinery which has evolved during about 3.8 billion years of biological life. But nevertheless this powerful machinery is a dedicated machinery which certain inherent properties which are influencing the construction of the inner model of the observer (cf. figure).

Figure: Basic epistemological structure of observer

As we know today by neurobiology and  neuropsychology is the brain (N) collecting all the different Stimuli (S) from inside of the Body (B) and is continuously  constructing an Internal World Model (IWM) which is the hypotheses of the brain, how all these different individual stimuli can be arranged in a way which represents their dynamics in a Presupposed External World (PEW), which is functioning as the outside Environment (E). In this perspective is the brain working as a function f mapping the representations S' of Stimuli S of the external world together with numerous other Internal States (IS) into an internal 3D-Model of an assumed dynamic context of the individual states IS and S'.  The presence of this model in the brain has the 'appearance' of 'Knowing of', i.e. the indiviual biological system, which is endowed with such a brain, has  something which we call 'intuitively' Consciousness. We are aware of something. Philosophers are naming the 'content' of this kind of consciousness Phenomena (P) (one of the most influencing writer for Phenomenology in Philosophy was E.Husserl; for a critical discussion see (Petitot et al, 1999)).

Thus when an Observer (Obs) does observe  H-I-M-objects (:= Human <-->User Interface : Machine-interactions) the observer does not perceive the 'object as such' (the 'Ding-an-sich of Kant (cf. Kant, 1781/1787)), but only the stimuli, which the body has registerd and the model of the object, which the brain has constructed. Because of the very long co-evolution of body and brain with the environments, in which the body occurs, is the internally constructed model --as everyday experience shows-- for many purposes 'reasonably good', but with a closer look to its output psychologists could find many kinds of distortions which are inherent in this model. The internal model of the brain is not a 1-to-1-mapping of the outside world, it is an ingenious abstraction of important properties to enable the organism to respond quickly and successfully for his purposes, but not to built  an completely 'objective' picture of the external world (To cite the relevant literature for this argument one had seriously  to cite the whole history of philosophy and all of modern science).

The question, how one can  observe the real world to overcome this 'built-in' bias in observing the real world has led --step by step-- to the evelopment of modern sciences.  To free the results of observations from the subjective internal states of a concrete observer one has to introduce procedures which are producing data which are independent from the person who is applying these procedures. These  observation procedures can include all kinds of instruments.   Using such kinds of measured data can improve the observations, but one has to keep in mind that the oberserver has to interprete these individual data as part of relations and dynamic sequences of events.  For to do such an interpretation the observer has to construct explicit models (MOD) which are more powerful than the data as such. Therefore is this kind of model construction an area of potential distortions with regard to the real world 'as such'.  The data D produced by such observation procedures are finally also phenomena P_D (the internally represented stimuli S' which are 'measured') in the consciousness of some human observer, but these phenomena P_D are a certain subset of the set of all subjective phenomena P, i.e. P_D C P.  The subset P_D  is characterized by the fact, that this subset P_D  is based on inter-subjective events D, which can be reproduced every time a person is applying the observation procedure.

As in the non-scientific case does the group of observers also in the scientific case need a language of observation L_Obs which allows to speak about all objects, properties and actions involved in this kind of observation.

Figure: Diferent Kinds of Observation + Measurement

Furthermore  one has to distinguish different kinds of data with regard to the object as such. Important basic types of data are:

• Behavioral Data (S-R)(Psychological, Black-Box): The observation process is only looking to the behavior of a system (H or M with UI).  The behavior is occuring as response R of a system to another system. For the receiving system the perceived events are stimuli S. Thus the response R_UI of the user interface UI of some machine M  can become the stimuli S_H of a human system H and vice versa the responses R_H of a human system H can become the stimuli S_UI of some user interafce UI of a machine M. Restricting the observation of a machine M to its observable input-output behvior(I-O) is also called Black-Box Method. In the social sciences is a behavior based approach (S-R) usually part of behavioral (experimental) psychology.
  

• Internal Data (IS)(Neuronal (N), Machine states (M)): If it is possible to include the internal states IS of a system into the observations, then does this provide an additional kind of data which in the case of human systems H are called physiological or --more specialized-- neurological (N) data, and in the case of a machine these are called machine states (M). The disciplines, which are dealing with internal states of human systems are Physiology and Neurology.
  

• NeuroPsychological (S-N-R), Black-and-White-Box (S-M-R): An important combination is the combination of human  behavioral (S-R) data and internal data (N) is called neuro-psychological data (S-N-R); in this case one correlates behavioral (S-R) data with neuronal (N) data. In case of a machine one has to combine input-output behavior (I-O) with internal machine states M resulting in (I-M-O). The last kind of data leas to a Blach-and-White-Box Approach.
  

• Behavioral-Phenomenological (S-P-R): another possible combination in the case of human systems is the correlaion of behavioral data (s-R) with phenomenological data (P) to a phenomenpsychological approach (S-P-R). 

• NeuroPsychologicalPhenomenal (S-N,P-R): recently based on the progress in  neurobiology and new trends in philosophy one can observe first examples of correlations of behavioral (S-R), neuronal (N), and phenomenal (P) data to enable a more generalized and integrated theory.
  

Observation as such must not necessarily be a measurement process. To turn an observation into a measurement process (MEAS) one has to compare some target objects (TO) with some reference objects (RO).  And in the scientific case the  target object (TO)  has to be compared to a reference object (RO) in a way which is independent  ('invariant') from a concrete observer.A measurement process invariant with regard to a concrete observer   is called objective or empirical.

Simple examples from physics of reference Objects --which are also called standards-- are 'the meter [m]', 'the kilogram [m]', 'the second [s]'. Measurement in physics means that there exists some kind of a procedure during which a target object will be compared with a reference object --e.g. one side of a table with a meter-- and the result of the measurement will be a number telling 'how much times' the target object coult be compared to the reference object (e.g. 5 m, 7 kg, 8 s). There cann also be fractions or multiples of reference objects for to simplify the tak of measurement.

Thus we can say measurement (MEAS) is a mapping from TOs and ROs to numbers NUM:

MEAS: RO x TO ---> NUM

An important distinction here is fundamental masurement and derived mesaurement. Fundamental measurement (FMEAS) occurs if one is applying reference objects directly to target objects (like in the case of measuring the length of an object with a meter). Derived measurent (DMEAS) occurs if one is taking fundamental measures and combines these with formally defined operations  into new values.

START