Critical realism


 * Critical realism
 * Reject ontological atomism
 * Logic of inquiry is retroduction
 * Acknowledge that the world is structured: the empirical (what is observed), actual (what could be but is not observed), real (what is not observed but consists of those structures)
 * Aim of scientific investigation is to explain observable phenomena with regards to underlying structures and mechanisms
 * Critical Realism links the deductivist approach to positivist ontology. The ontological difference between positivist and critical realist perspectives is that the latter distinguishes three ‘levels’ of reality where the former addresses only one. The positivist deals only in the empirical, what can be sensed. The critical realist distinguishes the empirical from the actual and in turn from the real. The level of the actual contains the events that produce, but are distinct from, empirical impressions. The level of the real refers to structured objects and their related powers, exercised through mechanisms that may not be directly observable (eg gravity), but do create tendencies in the course of actual events. In an open system, different mechanisms may conflict but are none the less always operative, leading to tendencies, as a leaf tends to fall to the ground despite being lifted by the wind. On this definition, objects at this level are real in the sense that they exist independently of our knowledge of them. They are structured in the sense that they cannot be reduced to the actual events or empirical impressions they produce. Within this ontological context, retroduction may be contrasted with induction and deduction as follows. The latter two relate empirical events and the laws or functional relations between them. The inductive method moves from observation of (a series of) particular events (eg “I have seen white swans on several occasions”) to a general claim (“all swans are white”). The deductive method moves from a general claim to the prediction of a particular event (“the next swan I see will be white”). Both methods operate at the empirical level. Retroduction is the inference from knowledge of phenomena at one level of reality, to non-empirical explanatory events or mechanisms at a deeper level (“the colour of a swan is determined by its genetic structure”). Indeed it is the ascription of causality rather than direct perception that identifies the hidden reality. These levels are not necessarily fixed. Physical science has proceeded from a theory that actual substances can be understood in terms of earth, water, wind, and fire through the full periodic table of the elements and on to quantum mechanics. At each stage the development of knowledge depended on the invention of suitable measuring instruments, to turn the actual into the empirical, and thereby confirm the validity of a new understanding of the real. Each generation of scientists has laboured to uncover the next deeper layer of reality beneath the current level of empirical observations. There appears to be no limit to this process. The feature of social reality that distinguishes it from the natural world is its relation to human agency. Social structures are the product of human choices, and in turn influence human choices: they are mutually dependent. Human choice (in the sense that someone could always have done otherwise) is not reducible to actual events and consequently is always part of an open system. It follows that social structures and the mechanisms by which they influence actual events cannot usually be assessed by the use of deductive methods under quasi-experimental conditions. On the other hand their existence can be inferred by retroduction based on underlying causality rather than empirical pattern