Abuzaid,

Allow me to reflect on Psillos' definition of determinism. I think he is right to distinguish ontological and epistemological determinism, with predictability clearly belonging to the latter. In the modern western tradition, epistemology was given priority: true statements about the world are constructed (inferred) by the mind from sensory data, and so the world becomes a function of the powers of the mind. I believe there is need for a better justification for this than has so far been offered. In fact, outside the modern West (Aristotle or feudal Europe, for example), the world was believed to be more than can be represented in thought. A little example: St. Augustine in his Confessions said that he very well knew what time is, but could not answer anyone who might ask him what it is.

When it comes to the ontological or metaphysical defintion, I suspect Psillos might be skating on thin ice when he claims that a non-epistemological definition amounts to a claim for universal causality. In historical terms, in the late feudal west (16th century) neo-platonic mysticism became the fashion, and the cosmos was understood to be coherent. This is often cited as a foundation of modern European science, for it claimed, roughly, that all things have some relevance for each other and so natural knowledge is not limited to concrete particulars. However, this cosmic coherence can mean several things, only one of which is universal lawfulness. Folks like Newton and Laplace realized not only that generalization based on experience is possible (hardly a new insight), but that this generalization justified a reification of natural law as an ontologically independent universal force. In recent years, this reification of an epistemological artifact has made people nervous, and so there has been a shift to a singular causality that explains a development by a description of the inner caual mechanism at work in particular situations (this nicely brings the physical and historical sciences into greater accord, although it also has problems).

My point is that there may be several kinds of coherence other than on natural law or on universal causality. As for universal lawfulness, there has also been a coherence based on rationality (Kant) or objective idealism (Hegel), and my own preference is for a coherence based on mutual enabling, that one thing enables another. The point is that there is no a priori reason why coherence must imply natural law, for there are other options and the current trend seems to be away from it. Given the choices, whatever one adopts must be justified. As for universal causality, that seems particularly problematic. It seems quite Eurocentric and ideological, but here I'd only like to argue that it is also an effect of epistemological primacy. In brief, if the world is what we infer from sensations, then the world is reduced to only observable (empirical, local, intrinsic) properties --  properties that can enter a causal relation with our sensory apparatus.  Causality (and deductive logic itself) is an artifact of a closed conceptual or physical frame: the greater the closure, the more will the outcome of a process be unequivocally determined by the intrinsic properties within it. Absolute determinism implies complete closure, which today is considered only a hypothetical limiting case. The modern Western notion of ontology is therefore based on a hypothetical limiting case, not the way things actually are. The weakness of cauality is clearly shown by there being no agreement today over just what causality is, and the reason seems to be that it is epiphenomenonal.

I also wonder about Psillos's implication that a denial of determinism is indeterminism. This issue comes up naturally in efforts to define "constraint". For example, in evolutionary biology, phylogenetic constraint is defined in two ways: a loss of degrees of freedom (probability 0) or reduction of the probability of possible states (probability between 0 and 1). For example, the highway speed limit does not limit my degrees of freedom, but only makes my going excessively fast less probable. The word indeterminism might imply no relation, but there are relations other than those of unequivocal determinism. At the end Psillos seems to be in agreement with some of my points, but he is not emboldened to define "determinism" as just a hypothetical mental construct with limited ontological relevance.   

Not sure if the "two versions" of determinism are yours or belong to Psillos. In any case, the above discussion might suggest that weak determinism/strong determinism is a false dichotomy, simply a reflection of the Self-Other dichotomy pecular to the modern West, which in turn is ideological. The Laplacian strong strong determinism is epistemic and anthropocentric Self-centered; the weak determinism is ontic, Other-centered and is objectivist, positing in a very contradictory fashion that universal closure is the basis of universal coherence!

Sorry to go on at such length over Psillos, and I turn finally to your "pre-determination". You seem to reject both notions of determinism, but I'm not clear what you are offering in its place. Is it a probabilistic causation? Your attitude toward it seems ambivalent, which seems to leave a vacuum, and maybe the whole thing is a fantasy to begin with. The relation of all events seems to be probabilistic in principle, and perhaps we should start with that instead of determinism. Usually discussions of probabilism are epistemic, such as the reliability of prediction (standard deviation), but the issue is, I believe, to explain probabilistic behavior. One approach, the n-body problem, might be that all properties are approximations, such as a reduction of a continuum to integers, or assuming that the mass of a body is located at its center, etc., and these approximations come back to bite us in the butt, or even give rise to wildly divergent trajectories (chaos). But this looks at it epistemologically, not ontologically. In terms of the latter, if properties have specific values, and the outcome is due to them, then it would seem to be unequivocally determinant -- a conclusion neither you or I accept.

This seems why there is a broad sense that there's more to things than just observables. For example, William James' "fringes", Henri Bergson's vital force, a functional property held by the components in an emergent system, Arthur Koestler's holon, David Bohm's implicate order, although all of these are now seen as inadequate. However, my sense is that there's broad agreement that what accounts for both emergent and probabilistic behavior is the presence of unobservable real properties. If so, then whatever these properties might be, they are obviously incompatible with laws.

Laws seem to apply only to a relation of closed systems, where closure is an effect of systems entering into a relation in which one frames the other and gives specific values to its observable properties (this not so odd: a RF wave has specific values such as strength at any particular point, but as a whole its property of strength has no specific value because it is not local). This relation (which I prefer to analyze as "superimposed" processes) are conventionally described as causal. An example of such a causal relation is observation, where the relation of observer and object gives rise to specific values for the properties of that object. Laws therefore seem a generalization of observations, where the observation constructs the data of observation, the values for the properties of the object which are the basis of law. 

I've already run on too long and have not directly addressed your concern for levels. Your assumption that higher levels are more complex strikes me as either ambivalent or a tautology (if I put more eggs into the Easter basket, the basket of eggs has more constituents than it did before). Another issue: just what does your "level" refer to? Cosmos, biosphere, noosphere; quantum mechanics, physics, chemistry; animal, vegetable, mineral; dogs, cats, rats; the rich, middle class, and poor? These examples may imply different notions of "level", and I suspect the place to start is why one notion is more important or more universal or more useful than another.  The word "structure" is also ambivalent. In the physical sciences, it is a closed conception of things that are related in some way, but in the social sciences, it tends to refer just to the relationships, not the related nodes. And the word process is a problem because no one has been able to define it. Except for occasional hints, process generally refers to a temporal sequence of the static states of being of an identity which differs only in terms of accidental properties. This notion srikes me as a philosophical disastor for several reasons (time, entity, essential/accidenal properties).  I agree that we need to define inter-level relations in a way that is explanatory. You seem to hint that the path to follow is analytic philosophy (start with a set of unequivocal and logical concepts such as process, structure and complexity), but I'm not so sure I'd embrace the Cartesian cogito ergo sum, but instead look to action (not in Parsonian sense), where mind and world are not separate entities, but aspect of one process of change and in which we directly experience the unobservable possibilities of things.     

Haines Brown