00100		CHAPTER 9
00200	
00300	   MODEL EVALUATION
00400	
00500		An  evaluation procedure for a model involves the disarmingly simple
00600	question - `is it a good model?' The ordinary language term `good'
00700	means praiseworthy. But what is a model good at or good for in order to
00800	be praiseworthy? It depends on how well it serves the purposes for which it was constructed.
00900	Our primary aim in constructing this model was to explore and test a theory having explanatory verisimilitude.
00910	To satisfy this aim the model must meet norms of internal consistency
00920	and norms of external correspondence with observed phenomena. A secondary
00930	aim involved pragmatic norms.
01200		A model in the form of an algorithm consists of a structure of
01300	mechanisms whose inner workings are sufficient to generate the outward
01400	behavior under consideration. The theory embodied in the model is
01500	revealed by the set of statements which describes how the structure
01600	reacts under various circumstances.
01700		Theories have many functions. They can be summarized as follows
01800	[from Bunge?]
01900		(1)To systematize knowledge.
02000		(2)To explain facts by showing how they are the entailed
02100	              consequences of the systematizing hypotheses.
02200		(3)To increase knowledge by deriving new facts.
02300		(4)To enhance the testability of hypotheses by connecting them to observations.
02400		(5)To guide research by:
02500	              (a)posing fruitful problems
02600	              (b)suggesting new data to gather
02700	              (c)opening new lines of investigation
02800		(6)To map a portion of reality.
02900		It is a tall order for a theory to fulfill all of these
03000	functions. In undeveloped fields we should be happy with even one
03100	of them. Models can be assigned  these functions when they are
03200	theoretical, rather than replicative, in type. Our model was intended
03300	primarily to serve functions (2) and (4), testable explanation.
03400		What constitutes a satisfactory explanation has been treated in
03500	section 00.0. The `fit' with phenomena as indicated by some measurement
03600	or empirical test indictes truth, verisimilitude, or grains of truth.
03700	Our tests and measures were described in section 000.0. Acceptability of
03800	a model sometimes depends not so much on truthlikehood, an elusive state,
03900	as on whether a majority of the relevant expert community believes
04000	the theory or model to approximate truth to some unknown and unknowable
04100	degree. Truth or falsity cannot be proven with certainty but their
04200	presence can be assayed bt critical assesment and deliberation. A theory
04300	or model may bring cognitive or pragmatic comfort, not because it is TRUTH
04400	but because it represents an improvement or its explanatory rivals. Cognitive
04500	comfort is a type of intellectual satisfaction while pragmatic comfort
04600	accrues from applications to  problems in order to make things work the
04700	way humans want them to work.
04800		It would be a bonus if our model could satisfy those interested
04900	in function(3) listed above, making possible new knowledge through prediction.
05000	This novelty could arise in two ways. First the model might demonstrate
05100	a property of the paranoid mode hitherto unobserved clinically.  
05200	In principle this could come about because the I/O behavior of the model
05300	is a consequence of a large number of interacting hypotheses and assumptions
05400	chosen initially chosen to explain frequently observed phenomena. When the
05500	elements of such a complex conjunction interact with input they generate consequences
05600	in addition to those they were designed to explain. Whether any of these
05700	consequences are significant or characteristic of the paranoid mode remains
05800	a subject for future study.
05900		A second source of novelty would lie in the behavior of the model
06000	in some new situation. Since it is designed to simulate communicative
06100	behavior in an interview situation, the `new' circumstance would have to
06200	involve some new type of linguistic interaction which the model is capable
06300	of responding to. From its behavior one might then predict how paranoid
06400	patients would behave under similar circumstances. the requiste
06500	empirical tests and measures would show the degree of correspondence
06600	between patient and model behaviors.
06700		This possibility is of importance in considering therapies for
06800	patients caught in the quandaries of the paranoid mode. Language-based
06900	or semantic techniques do not seem very effective in the psychoses
07000	but they are useful in states of lesser severity. A wide range of
07100	new semantic techniques, including extremes, could be tried first on
07200	the model without hurting patients through blind experimentation.
07300		While we have used the model to explore a theory and to
07400	study psychiatric judgements, its potential use as a training device
07500	has not escaped out attention. Mental health professinals in their
07600	training need `disposable' patients to practice on. A model has the advantage
07700	of not being harmable and providing an opportunity to measure performance.
07800		A theoretical model is evaluated relative to rival explanations.
07900	Our model stands as a contender for the preferred psychological explanation
08000	of paranoid processes. The expert forum will decide its ultimate status.
08100	A theoretical model is partial, perspectival and has a short half-life.
08200	Hopefully it lives long enough to provide a first approximation from
08300	which better approximations can grow.