Primer on Natural Resource Science

Chapter 3. Induction, Deduction, and Retroduction

CHAPTER 3 Induction, Deduction, and Retroduction Many, if not most, . . . graduate students do not even understand the differences between induction and deduction. —H. Charles Romesburg (1981:311) There is a tradition of opposition between adherents of induction and of deduction. In my view it would be just as sensible for the two ends of a worm to quarrel. —Alfred North Whitehead (cited in Peirce 2000:1340) [N]o body of observations can prove the hypothesis true, because [inductive] hypotheses make predictions that go beyond any body of evidence that we may have accumulated at any point in time. —Harold I. Brown (2000:196; emphasis in original) Nobody argues with the assertion that accumulation of reliable knowledge is the prime goal of science. Yet there is considerable difference of opinion among philosophers and practicing scientists on how to judge the reliability of knowledge gathered under different methods of reasoning (induction, deduction, and retroduction). Romesburg’s (1981) article on gaining reliable knowledge in- fluenced the manner in which wildlife scientists perceive knowledge. He pointed out that induction (generalizing from specific results) is the workhorse of wildlife science. Indeed, induction is policy in our journals as illustrated by the “Management Implications” section. This section operates, to some degree, under the premise that the results of a specific study will hold in a more general spatial and temporal context than where and when the study took place. Romesburg also pointed out that H-D experimentation is underused in wildlife science. H-D experimentation involves positing a research hypothesis , deducing outcomes that will hold under the hypothesis, and testing to determine whether the outcomes appear in experiment. Romesburg (1981:295) pointed out that untested hypotheses have gained “credence and the status of laws through rhetoric, taste, authority , and verbal repetition.” “The H-D method,” he argued, “is a way of raising the reliability of . . . speculation and, hence, the overall reliability of our knowledge. It is not a cure-all.” Like all human enterprises, natural resource science is a culture whose members perceive proprieties and improprieties in conduct. The philosophers of science, who are grounded in formal logic and who tend to see physics as science (with deference to Charles Dar- 24 Perspectives win), have laid down the proprieties. One is the notion that, philosophically , induction holds no claim whatsoever on truth (Howson 2000). Another is the notion that the H-D method, though imperfect, at least represents a logically supported approach to truth seeking. Accordingly, wildlife scientists have extolled H-D experimentation (Guthery, Lusk, et al. 2001; Guthery, Brennan, et al. 2005; Garton et al. 2005) since Romesburg’s (1981) article. We now realize there may be a major disconnect between the thinking of philosophers and the realities of natural resource science. This chapter introduces induction, deduction, and retroduction (after-the-fact explanations) as methods of reasoning about the truth of propositions that describe or explain nature. I provide definitions , examples, and philosophical and practical considerations regarding the methods of reasoning. I also explain and provide examples of the H-D method and point out that it may be flawed when applied to research problems in field ecology. Induction Definition Induction is reasoning from particular facts or individual cases to a general conclusion or principle. Also, the general conclusion is called an induction. Aldo Leopold (1933) observed that animals of low mobility requiring ⱖ2 cover types tended to be associated with edges between required cover types. From this observation, he induced the principle of edge: the density of animals on an area is proportional to the sum of the edges of required cover types on the area. Examples Inductions appear frequently in the literature of natural resource science. They can be identified whenever authors extrapolate the results of a particular study to a general situation, as in these examples: * “The results of this study indicate that [construction of earthen dikes across wetlands] provides predator access and decreases avian productivity by a factor of 1.5–3⫻ that of undisturbed wetlands. Thus, the widespread construction of [earthen dikes] in small prairie wetlands will be detrimental to the waterfowl resource” (Peterson and Cooper 1987:246). * “My results suggest that lead shot is available to feeding waterfowl for many years, and that exposure of waterfowl to 25 Induction, Deduction, Retroduction lead poisoning will likely occur for ⬎3 years after the lead shot is curtailed” (Flint 1998:1099). * Concerning three-toed woodpeckers...