INTERROGATIONS #2 9/16/2004 General Perspectives II: Critical Realism 1. Fabian Pfeffer While I feel that critical realism generally fits sociological reasoning quite nicely, I understand from R. Pawson that its application rules out most forms of ‘variable sociology’. Is this correct; and if yes, is this necessarily so? Where else will we draw empirical evidence from? As a (future) Sociologist, I feel relatively comfortable with the central claims of critical realism. They seem to nicely parallel common sociological reasoning and theory constructing. What else is sociological investigation if not ‘digging deeper’, searching for underlying mechanisms and structures. The idea of ‘the stratified reality’ and the call to search the explanans on a lower level in order to elucidate the explanandum on a higher level are merely other words for what e.g. Coleman advocates by his macro-micro-models. (Besides: the opposite direction of explanation is also put forward). Anti-reductionist defenses of the independent existence of each ‘layer of reality’ also seem to be a sociologist’s motto. (I will leave the issue of experimentation and closed systems untouched here; even I expect this to be a crucial obstacle for the application of critical realism to social sciences). How sad on the other hand, to hear that the complexity of mechanisms at work cannot be traced by standard quantitative methods used in social scientific investigations. The mechanisms under discussion are “generative mechanisms [which are] not the same thing as a spurious cause or an intervening variable … A mechanism is not thus a single variable but an account of the constitution and behavior of those things that are responsible for the manifest regularity.” (Pawson: 130) How then can we discover the ‘underlying forces’ if not by searching for a lower-level variable? The solution proposed is to “isolate closed systems experimentally to see if they correspond to the conceptual systems we have modeled theoretically” (Pawson: 154). Let us suppose now, that we have elaborated a thorough theoretical concept, found single (other-level) variables which capture our concepts adequately, mapped these variables to the concept and finally examined the correspondence of the model with empirical reality by statistical analysis (our alternative to experiments). What is so wrong about it? What if we were so lucky (or so good) to match the ‘independently existing mechanisms’? Could the thoughtful use of variable- work be a kind of twisted ‘transcendental arguing’? If not, I have difficulties in seeing what else would count as empirical evidence. Neither Bhaskar’s reference to theoretical models in order to answer this question nor Pawson’s idea of network models “entrenching new models into the existing system” (isn’t this just ‘paradigmatism’?) convince me. [I don’t think that the implication of Pawson’s account of realism is that variables – in the straightforward sense of quantitative measurement – no longer has any place in social science. The associations among variables will always provide the raw material for the elaboration of the theoretical models of generative mechanism that explain how those associations are generated. In one of the examples given by Pawson, the quantitative relationship between variations in temperature and pressure in gas are crucial for any account of the relevant mechanisms. What the mechanisms in fact do is explain thisReading Interrogations #2. General Perspectives II: Critical Realism 2quantitative association. So, the research program requires postulating such underlying mechanisms. The question then is how do we validate such claims? Pawson, like many realists, invoke the experimental strategy as the best way of doing this. Statistical research of a properly designed sort can be viewed as a good second-best strategy in the context of open-systems where experiments are impossible. The notions of “controls” in a statistical model attempt to simulate an experiment. Of course this is always only partially successful, but I don’t think it has a logically different status from an experiment.] 2. GOKCEN COSKUNER 1. Bhaskar asserts that the concept of activity dependence of social structures enable social scientific work, rather than limit it. He argues that the hermeneutic dimension of social life is the necessary starting point for social science and that the self-referential character of sociology encourages a beneficial methodological reflexivity, which is less evident in the natural sciences. (Philosophy of Science, p. 134-135) How does being self-referential work in advantage of social scientific work? [I think the idea here is that it makes social scientists more acutely aware of how their relationship to the objects of study affects the observations they make of those objects. Our observations are always a form of intervention in which we are at least partially constituting the things we observe. The reflexive aspect of sociology makes sociologists aware of this to a greater extent, perhaps, than natural scientists.] 2. Benton and Craig provide a hierarchy of level for sciences. (philosophy of science, p 126) The way of the ordering of the sciences are justified in terms of the way the mechanisms characteristic of each level are explicable in terms of those of the next one below it. And according to realists reality is stratified where at the bottom level is the empirical level of observed events and at the top is the real world mechanisms, powers etc. which science seeks to discover. So from a realist point of view is there a relation between the two stratifications? [Interesting question. I think a key issue here is the problem of reductionism. If each science were reducible to the science at a “lower”/deeper level, then these two stratifications might collapse into one. I think the antireductionist argument means that each science retains its own specificity, and therefore each involves a set of specific explanatory problems of mechanisms/events/experiences.] 3. Matt Dimick Realism, Anti-Realism, and Critical Realism: What’s At Stake? What’s at stake between the rival views of realism, anti-realism, and critical realism? As between realism and anti-realism, one of the debates would seem to be whether the goal of science is to provide a “true” picture of the world. For realists the answer is