Comment: Implication Analysis as Abductive Inference

It is a pleasure to thank the editor of Sociological Methodology for the kind invitation to comment on the interesting paper on “Implication Analysis: A Pragmatic Proposal for Linking Theory and Data in the Social Science by Lieberson and Horwich (LH; this volume: p. 1). In our opinion, LH ask all the right questions about missing ingredients in the process of weighing evidence in the generation and testing of sociological theories. However, their paper falls substantially short of what it might have achieved in responding to the issues raised. This criticism applies not only to LH, but also to Ni Bhrolchain and Dyson (2007), whose checklist for assessing evidence in demography is highlighted by LH in their outline of how “Implication Analysis” should develop. In our view, the root of the problem lies in the overemphasis through much of the social science literature on a dichotomy in the logic of inference—namely, induction and deduction. Neither of these forms of inference, alone or in combination, provides an adequate basis for the weighing and integration of evidence of disparate types and varying quality that LH would like to see. Missing from the discussion in LH, and throughout the sociological literature, is a treatment of abduction, or “inference to the best explanation.” This form of inference has roots in the 17th century (Hobbs 2006), but was explicitly formalized by Charles Sanders Pierce in the 1870s and further developed by him (Pierce 1955). Abduction represents the reasoning that goes on in much of science— physical, biological, and social—and in everyday life, although it is rarely identified as the mode of inference that is actually being used.