7. Epistemic Retreat

This chapter is philosophically epistemological, providing a kind of constructive epistemology in the form of some architectural principles in relation to the representation of knowledge in networked storage systems and its processing by distributed processing systems involving both natural and artificial processing elements.

The chapter is therefore a conflation of aspects of philosophical epistemology with abstract architectural design for knowledge processing systems. This necessitates (or flows from) some novelty in my conception of both these enterprises, so I will begin with some discussion of the innovations involved.

Epistemology may most briefly be characterized as the theory of knowledge. It has typically been anthropomorphic, i.e. concerned specifically with human knowers, and sometimes linguistic, concerned with the specifics of the meaning of the concept of knowledge.

Here I am interested in knowledge in information systems, not in human brains, and seek to avoid giving any deep scrutiny to the meaning of the word know. The conduct of epistemology without attachment to the concept of knowledge is analogous to the preference in science for avoiding vague and relative concepts such as ``hot'' and ``cold'', in favor of objective and precise properties such as ``temperature in degrees Kelvin''. Instead of claiming knowledge, we will be aiming to provide more objective descriptions of grounds for the truth of propositions, or of evidence showing the reliability and fidelity of abstract models of the real world. I do not prescribe particular measures, but provide instead a context in which a plurality of comparative evaluations can be accommodated.

The first comparison is the one provided by Hume's fork, and this has a major impact on the epistemology and the knowledge architecture. Because of the great precision with which ``relations between ideas'' can be expressed, such propositions we regard as assert able, and in connection with such assertions ``epistemic retreat'' involves a form of assertion in which the grounds for asserting truth are made explicit. On the other side of Hume's fork, in relation to ``matters of fact'', epistemic retreat in the first instance reflects the imprecision in our knowledge of Plato's fleeting world of appearances. Scientific knowledge is regarded as embodied in abstract models of the real world, which are applied by deduction (this is our version of a nomological-deductive scientific method).

A principle effect of the status of Hume's fork in this epistemology is that our formal knowledge base asserts only logical or analytic truths. In it, scientific theories are represented through abstract models, which will in general have concrete interpretations, but which have the logical status of definitions rather than assertions.

Epistemology has been concerned with the refutation of scepticism. It here embraces an open scepticism (but not a dogmatic negative scepticism), and is therefore oriented not with the refutation of scepticism but with the establishment of a viable constructively sceptical system.

The first step in the moderation of a pyrrhonean scepticism is the recognition that, though no proposition is absolutely certainly true, some appear to be more certain than others, and among the more doubtful there are also degrees of doubt. Can we know with absolute certainty that some proposition is more certain than another? No, but we can form opinions about relative certainty which are themselves more solid than our opinions about truth (again, abstaining from absolutes). Among these comparative assurance judgements are the relations between a proposition and the evidence we have for it. It will generally be the case that our confidence in the evidence on which we base an opinion will by stronger than that in the proposition we infer from the evidence. Similarly, it will be the case that a statement effectively describing the evidential support for a theory will can be more confidently asserted than the theory itself.

This is the principal way in which Metaphysical Positivism interprets the positivistic principle that science should not go beyond presentation of observational data. It is not taken to impede the formulation of general scientific theories which go beyond the content of any possible body of experimental evidence. It is taken instead to impede the assertion of such empirical generalizations as truths. The positive scientist instead compiles various bodies of evidence which provide a basis for decisions about when the scientific generalization might be applied.

The experimental data obtained will provide information about the fidelity and accuracy of the theory as modelling the real world in a variety of circumstances, so that someone contemplating use of the theory will be in a position to form an opinion about whether the theory will provide a sufficiently reliable and accurate model for his purposes.

The mitigated scepticism of positivistic philosophy, following David Hume, accepts a priori truths of reason as certain, but expects of positive science that it does not go beyond what is entailed by the experimental or observational evidence. Since most scientific theories involve empirical generalizations which go beyond any finite collection of particular observations, positive science would seem by this doctrine to be eviscerated.

Metaphysical positivism recognizes two principles which constitute a kind of foundationalism. The first is connected with the sceptical doctrine that all we know is that ``appearances appear''. However, what we count as an ``appearance'' is not confined to sensory impressions. Any impression which we may form on the truth of some proposition, whatever its source, is counted as an appearance. Such appearances we accept as what they are, and the body of science is considered to constituted just an organized presentation of a great deal of such material. The enterprise is not solipsistic, it is collaborative, and we therefore recognize as significant the source of the impression, the identity of the person or entity who was the subject of the impression. Because of the broadening of the notion of appearance, propositions expressing such appearances are called ``opinions'' and are to be tagged or digitally signed by the entity whose opinion they are. The kind of entity which has opinions is called an ``authority''.

Because of the collaborative nature of the enterprise, an opinion will normally be formed by an authority on the basis of (as entailed or otherwise supported by) some collection of opinions of other authorities. It is therefore normal for an opinion to be expressed, on the assumption that various authorities can be trusted, or more specifically, on the assumption that all their previous opinions are true. The collection of authorities on the basis of whose opinions a further opinion is formed, together with the authority forms the new opinion, give a measure of the risk associated with the opinion which which I call a degree of assurance. These degrees of assurance are partially ordered. The more authorities whose opinions are involved, the lower the level of trust. The partial ordering becomes a lattice when we allow that an opinion may be endorsed by several authorities on the basis of distinct collections of other opinions. Adding more independent opinions increases the degree of assurance.

Roger Bishop Jones 2012-09-23