Some thoughts on the Modelling Relation

Here, on this list, and I think elsewhere as well, we are bugged down with 'RR arguing with terms of the reductionst sciences' mostly using those terms. We never get to even 'think' by his thoughts on - what I call - wholeness aspects. Model biology, model physics, modelled with mathematical aspects, model-qualia of models, restricted views of topically boudaried thinking is going on. Of course he spent an overwhelming time of his efforts in such argumentation, but "Rosenism" in my views (I hope) is "the other side". Beyond the MR (in both senses).

It seems to me that there is no a priori limit on the modelling relation as far as the size (i.e., the number of elements which comprise S) of the natural system is concerned. At the one extreme, we could define an S to consist of zero elements (zero percepts and hence zero relational properties between percepts). In that case, the only (and largest) model of S would be what we might call the "empty" or "null" model. In this case, the ambience is essentially split into no system and all environment.

At the other extreme, we might define S such that any percept, and any possible relation between percepts, in the natural world are members of S. (And, if desired, we could expand S to include non-sensory sources of information other than percepts, e.g., mystical sources.) In that case, S might consist of an indefinitely large or perhaps, infinitely large, number of elements. In that case we have "split" the natural world into all system and no environment. I don't know if this might satisfy JohnM's idea of "wholeness aspects". Since we are talking about epistemology - how we comprehend the world - I'm not sure I can imagine what it would mean to say we comprehend something which we cannot even enumerate, such as a countable infinity of elements, much less an uncountable infinity of such. But we can leave that door open; in either case, I am personally not sure how one would proceed to create models which incorporate all those elements of S and their relations, without drawing further partitions and creating sub-models of smaller systems, which would seem to defeat the original purpose of defining S as all-inclusive.

As to going "beyond the MR", I'm not sure what that would mean. At root, the modelling relation occupies a central position in the Rosennean epistemological foundation. The reason that the arguments about largest models, anayltic vs. synthetic models, and so on carry any epistemological weight rests precisely on the presumption that the modelling relation is in fact "the habitat of all epistemology", to use Rosen's phrase from Essays [p.324]. Otherwise, none of the formal conclusions have any force of argument - they become mere mathematical excursions.

So it is that some criticisms of Rosen are based on equivocations on the status of the modelling relation. To regard the MR as merely pragmatic, or only meaningful as a formal oddity, or somesuch. Those kinds of equivocations fail to appreciate the central epistemological role of the MR in the Rosennean paradigm. Either one agrees that the modelling relation is "the habitat of all epistemology" or one does not. In the former case, the aforementioned equivocations are void, and in the latter case where one disagrees with the centrality of the MR, then none of the conclusions about largest models, analytic vs. synthetic models, simple vs. complex systems, etc. is epistemologically meaningful. In the latter case, one is no longer participating in the Rosennean paradigm, and it becomes prudent to ask of such a person: if not the modelling relation, then what exactly is their particular epistemological foundation and framework?

As I see it, the Rosennean paradigm consists of very few pieces. Roughly, and in no particular order: 1) the distinction between self and other, 2) the modelling relation as the habitat of all epistemology, 3) the 2-part assertion about causal entailment relations and their knowability, (i.e., "Natural Law"), 3) a lack of a priori conditions (e.g., computability) on models. And, in some sense, the modelling relation and Natural Law are two sides of the same coin. From these general foundations, the content of the Rosennean paradigm (which I feel is sometimes mistaken for the paradigm per se) arises as logical consequences. So in the "Note to the Reader" in Life Itself Rosen states:

So, once again, the edifice which is the content of the paradigm is built out of logic, all resting on the foundation of the Rosennean paradigm. Remove or alter that paradigm, and the entire edifice is no longer there. This also why the content of the Rosennean paradigm is not a la carte, any more than theorems of a particular system of math are a la carte; theorems arise from a particular set of premises or axioms, and to selectively choose some theorems as true and others as not true is to tacitly espouse some other particular set of premises or axioms. In either case, it is not a matter of orthodoxy, it is simply a matter of logical consistency.

Going back to the question of going "beyond the MR", maybe we could speculate thusly: is the modelling relation an instance of something more generic? That is, is the modelling relation as defined valid but merely circumscribes a subset of the entire habitat of epistemology, just as computable models are valid but circumscribe only simple systems? This question thus really requires answering: is Natural Law a specific instance of something more generic?