2
A. LOCKER
systems ongmates in the human mind and is inevitably determined by
decisions made previously in order to specify the purpose of a system,
it may be concluded that between such ambiguously conceived systems on
the one hand and models, as their still more or less vaguely conceived
representatives on the other, no sharp distinction can be made-at least not
in the initial phase of the process of cognition. Only when systems have
been "concretized" -by ascribing certain properties to them-are we able to
infer the existence of certain properties as existing in other objects or systems
as well; for various reasons, e.g. convenience, these are called models.
The relation between original (i.e. system) and model with which we
are faced here may ultimately be formalized and this puts us into the
position of making statements about properties which are characteristic of
models in general. It is the merit of STACHOWIAK [12] to have thoroughly
analyzed the main traits occurring here and to have emphasized-besides
having introduced a suitable terminology-the distinction between the two
possible main types of adjustment of the model to its relevant system or
original. If the adjustment of the model to the original is based on structural
features alone, then in its maximum degree of perfection the adjustment
may be termed isomorphism; in the other instance where this adjustment
is based on qualitative characteristics alone, its maximum degree of perfection leads to isohylism. However, between the two extreme instances,
so to speak, an intermediate one may be conceived if, for a certain prescribed
structural adjustment, a well-defined qualitative adjustment (or even an
isohylic adjustment) is assumed; this clearly represents a case of analogy
and the model is thus becoming an analog model. It is precisely this range
to which all the models science-including quantitative biology-is concerned with, belong, e.g. simulation models programmed on computers,
electronic circuits, hydrodynamic analogs and any other physical realization.
The term "partially analog model" has also been proposed [12] in order to
characterize models of this range, mapping time-dependent processes.
These types of adjustment between models and relevant originals hold
true, irrespective of the role models play in the process of cognition; in
this process models may induce further cognition or they may merely
exemplify by demonstration what is already known.
To characterize models appropriately, we have to take into consideration
the following three characteristics which reveal the functional nature of
models [12] or their "pragmatic" significance: (1) The mapping function
deserves prime attention since here a principal feature may be conveniently
formalized. Mapping depends on the degree of adjustment previously
achieved between the model and the original. (2) The reducing function
implies-which is what makes a model so useful-that certain traits on
which a pre-decision has been made will be retained in the relationship
between model and original-whereas other traits will have to be abandoned.
A. LOCKER
systems ongmates in the human mind and is inevitably determined by
decisions made previously in order to specify the purpose of a system,
it may be concluded that between such ambiguously conceived systems on
the one hand and models, as their still more or less vaguely conceived
representatives on the other, no sharp distinction can be made-at least not
in the initial phase of the process of cognition. Only when systems have
been "concretized" -by ascribing certain properties to them-are we able to
infer the existence of certain properties as existing in other objects or systems
as well; for various reasons, e.g. convenience, these are called models.
The relation between original (i.e. system) and model with which we
are faced here may ultimately be formalized and this puts us into the
position of making statements about properties which are characteristic of
models in general. It is the merit of STACHOWIAK [12] to have thoroughly
analyzed the main traits occurring here and to have emphasized-besides
having introduced a suitable terminology-the distinction between the two
possible main types of adjustment of the model to its relevant system or
original. If the adjustment of the model to the original is based on structural
features alone, then in its maximum degree of perfection the adjustment
may be termed isomorphism; in the other instance where this adjustment
is based on qualitative characteristics alone, its maximum degree of perfection leads to isohylism. However, between the two extreme instances,
so to speak, an intermediate one may be conceived if, for a certain prescribed
structural adjustment, a well-defined qualitative adjustment (or even an
isohylic adjustment) is assumed; this clearly represents a case of analogy
and the model is thus becoming an analog model. It is precisely this range
to which all the models science-including quantitative biology-is concerned with, belong, e.g. simulation models programmed on computers,
electronic circuits, hydrodynamic analogs and any other physical realization.
The term "partially analog model" has also been proposed [12] in order to
characterize models of this range, mapping time-dependent processes.
These types of adjustment between models and relevant originals hold
true, irrespective of the role models play in the process of cognition; in
this process models may induce further cognition or they may merely
exemplify by demonstration what is already known.
To characterize models appropriately, we have to take into consideration
the following three characteristics which reveal the functional nature of
models [12] or their "pragmatic" significance: (1) The mapping function
deserves prime attention since here a principal feature may be conveniently
formalized. Mapping depends on the degree of adjustment previously
achieved between the model and the original. (2) The reducing function
implies-which is what makes a model so useful-that certain traits on
which a pre-decision has been made will be retained in the relationship
between model and original-whereas other traits will have to be abandoned.