Epistemological Significance of Models in Science
7
much as it depends on the previous internal state as well as on the instantaneous stimulus. The fundamental difference between the two types of
automata lies in the fact that the automaton with a finite memory has an
explicitly expressed dependence of its response on the previous stimuli and
responses; in the automaton with a finite number of states, the response is
expressed implicith-by means of the internal states [6].
It is certainh" not without general (epistemological) significance that
models, systems and automata have several features in common with
respect to (1) the mathematics inherent and (2) to their behavior-if we
define behavior as the relationship between stimuli and responses or
activity and reactivity to environment. Besides deterministic and stochastic
automata-with completel\' different onsets-additional classes of automata, nameh" prospective and retrospective ones, may be mentioned as
examples. With the exception of some still problematic attempts at continuous automata, the theory of automata is at present entirely based upon
time-discreteness. Three particular forms of automata are of special interest
for biology, i.e. self-organizing, growing and self-reproducing automata.
Self-Organizing Automata
Self-organization means that the structural order increases. A presupposition is, of course, that the system (automaton) is open to its environment, a fact which allows it to adapt to prevailing situations and to attain
an improved organization. One may also state that in those automata the
variety of responses produced may under certain circumstances change and,
in general, must diminish [6]. Practically the same holds true for the learning
automaton which improves its efficiency with respect to a certain task that
is to be fulfilled. An automaton is also considered to be learning if it changes
in such a manner as to require different quantities for its description at
different times.
Self-Reproducing Automata
In considering automata with an infinite number of states, we are
confronted with the TCRINC machine named after the famous British
mathematician, A. M. TCRIK'G. It consists of an indefinitely long tape, divided
into squares, along which a reading head moves; the latter is able to write
or erase symbols in the squares and moves only one square per time interval.
The movement is governed bv standard code instructions; finally, there
is also provision for recording the current state of the machine. However,
this device represents only a basic one, called the TURING automaton by
VON NEmIANN, so as to distinguish it from the TURING machine proper
where the device is provided with a tape of infinite length in both dire<;-
7
much as it depends on the previous internal state as well as on the instantaneous stimulus. The fundamental difference between the two types of
automata lies in the fact that the automaton with a finite memory has an
explicitly expressed dependence of its response on the previous stimuli and
responses; in the automaton with a finite number of states, the response is
expressed implicith-by means of the internal states [6].
It is certainh" not without general (epistemological) significance that
models, systems and automata have several features in common with
respect to (1) the mathematics inherent and (2) to their behavior-if we
define behavior as the relationship between stimuli and responses or
activity and reactivity to environment. Besides deterministic and stochastic
automata-with completel\' different onsets-additional classes of automata, nameh" prospective and retrospective ones, may be mentioned as
examples. With the exception of some still problematic attempts at continuous automata, the theory of automata is at present entirely based upon
time-discreteness. Three particular forms of automata are of special interest
for biology, i.e. self-organizing, growing and self-reproducing automata.
Self-Organizing Automata
Self-organization means that the structural order increases. A presupposition is, of course, that the system (automaton) is open to its environment, a fact which allows it to adapt to prevailing situations and to attain
an improved organization. One may also state that in those automata the
variety of responses produced may under certain circumstances change and,
in general, must diminish [6]. Practically the same holds true for the learning
automaton which improves its efficiency with respect to a certain task that
is to be fulfilled. An automaton is also considered to be learning if it changes
in such a manner as to require different quantities for its description at
different times.
Self-Reproducing Automata
In considering automata with an infinite number of states, we are
confronted with the TCRINC machine named after the famous British
mathematician, A. M. TCRIK'G. It consists of an indefinitely long tape, divided
into squares, along which a reading head moves; the latter is able to write
or erase symbols in the squares and moves only one square per time interval.
The movement is governed bv standard code instructions; finally, there
is also provision for recording the current state of the machine. However,
this device represents only a basic one, called the TURING automaton by
VON NEmIANN, so as to distinguish it from the TURING machine proper
where the device is provided with a tape of infinite length in both dire<;-