Reply to R. A. Sayer

Abstract

Listen

Translate article

Sayer concedes several points in his second reply, but he still finds Conceptions of Space in Social Thought: A Geographic Perspective to contain errors, specifically five. The first (his point four) refers to my analysis of generalizations. The second (his point five) concerns explanation. The third (his point six) concerns closed and open systems. The fourth (his point seven) concerns science. The fifth (his point eight) concerns hermeneutics. Then there follows a discussion about meaning which seems to be part of the hermeneutics argument. As I see it, the first four (his points four through seven) can be addressed as one issue--that of science. The fifth point, hermeneutics (which is his eighth point) can be considered an application of science. I will consider first the issue of science and then the problem of hermeneutics. My remarks about science which are most pertinent to Sayer's criticisms are found on pages 35 and 41-2 of my book. I do not say empirical regularities are all there is to science. Yet this is what Sayer seems to believe I said. He sets up two types of explanations (which are not clearly differentiated) and claims my conception of science is about only one of them--the wrong one, i.e., the one linked to something like empirical regularities. Supposing he has read my statements about science, how can his remarks be interpreted? One possibility is that his own conception of science and explanation differs so radically from the one I outlined above that he would believe any provision for statistical testing and verification irrevocably dilutes the search for, and proof of, true explanations. In other words, he cannot entertain a science incorporating something like both of the explanatory types he outlines. Take his example of the nation state. He states 'to exist as a nation state certain social relations must be established on the basis of some degree of understanding (not necessarily a clear or very correct one) of the meaning of its concept'. Though I do not pretend to know precisely what is meant by this, his subsequent discussion suggests that trying to test it, or trying to see somehow if it corresponds to evidence, is taking the wrong approach. In my book I have discussed how testing is frought with philosophical and empirical problems, and how it alone does not establish cause and effect or give us true explanations. But I maintained that something along these lines must be done and is done in the practice of science. If Sayer is arguing that correspondence to evidence is not necessary at all, then he is indeed presenting a radically different picture of science and explanation than I have presented. It is also one that may be worth developing, although I cannot imagine how it could avoid being completely idealistic and solipsistic. But Sayer may not be saying that testing is unnecessary. Rather he may be saying that it is impossible because among other things: human behaviour is composed of non-repeatable or unique events; the causes of behaviour are so varied; the systems are so open; and there is an interdependent relationship between the language of observation and the theory to be tested. Here I would agree that human behaviour appears to present such conditions in abundance, (whether more than non-human action is questionable) and that it does create problems of testing, validating, explaining, and so on. But I would hasten to remind Sayer and the reader that after a long history of wrestling with these problems (often within the contexts