Three Senses of 'Explanation'

Three Senses of ‘Explanation’ Jonathan Waskan (waskan@illinois.edu) Ian Harmon (iharmon2@illinois.edu) Andrew Higgins (higgins9@illinois.edu) Joseph Spino (spino2@illinois.edu) Department of Philosophy, 810 S. Wright Street Urbana, IL 61801 USA Abstract ‘Explanation’ appears to be ambiguous between a representational-artifact, an objective, and a doxastic sense. That the distinctions between the three are still poorly understood we regard as an impediment to progress in the philosophy of science and as a source of the field’s resistance to greater integration with experimental psychology. We begin to elucidate the overlapping contours of the three sense of ‘explanation’ using a variation on Powell & Horne’s Semantic Integration paradigm, showing that both laypeople and scientists regard doxastic explanations as constitutive of representational-artifact, but not of objective, explanations and accuracy as closely connected to objective, but not representational-artifact, explanations. Keywords: Explanation; understanding; philosophy of science Investigating Conceptions of Explanation In the early 20 th century, many positivistic philosophers claimed that science rightly concerns itself only with questions of ‘what?’ not ‘why?’ because, as Stace (1935) put it, “‘why?’ does not proceed from the intellect, but from the emotions.” The general sentiment was that answers to why-questions engender mere feelings of understanding, empathy, or familiarity. Nevertheless, by the mid-20 th century, there was general agreement that science has the function of answering both sorts of questions, that it not only describes but also explains (Hempel & Oppenheim 1948). This shift in attitude had much to do with the pioneering work of Hempel, who regarded explanations as representational artifacts – that is, sets of statements that answer ‘why?’ questions. On one common manner of speaking (e.g., “There is an explanation for the odd trait on page 25”) the noun ‘explanation’ does seem to refer to a set of representational-artifacts; let us call them explanations r-a . Hempel (1965) countered worries about the affect-inducing character of explanations r-a by claiming that they are doubly dissociable from psychological states, and he embraced the apparent corollary that the philosophical study of explanations is, like formal logic, autonomous from scientific psychology. Accuracy, on the other hand, he regarded as essential for explanation. Many subsequent philosophers would agree with Hempel’s anti-psychologism and his claims about accuracy. One holdout is Achinstein (1984), who contends that some statements E constitute an explanation r-a for q only if an explainer could cite E and thereby render q intelligible to an audience. On his view, E need not have actually been used to render q intelligible; it suffices that it could be used in this way. Nor, for Achinstein, need E be accurate. Others made a cleaner break with positivism. Salmon (1984), for instance, claimed that explanations are sets of objective happenings, what we will call explanations o . Here too we agree that there is a common manner of speaking (e.g., “The explanation for combustion is oxidation”) on which ‘explanation’ appears to refer to a set of objective happenings. Similar to Hempel, Salmon argued that explanations o have nothing to do with what anyone feels or thinks and, as such, are non-psychological. As for accuracy, since explanations o are not representations, they cannot be right or wrong, “they just are” (Craver 2007). Meanwhile, as early as Craik (1943), psychologists have undertaken their own investigation of explanations, which they construe as mental states, or more particularly as having a belief about what might have produced the target phenomenon (see Waskan et al. 2013). There does appear to be a common manner of speaking (e.g., “Lavoisier had an explanation for combustion” or “The preverbal infant had an explanation for the meowing coming from the closet”) on which ‘explanation’ refers to a doxastic state; let us call it explanation d . The relevant beliefs are those in virtue of which one understands how or why, at least possibly, the phenomenon came about or, more succinctly, in virtue of which one finds the phenomenon intelligible. Psychologists often study the explanations d of children and science-naive laypeople, whom they often regard as misrepresenting the state of the world, the tendency thus being to disregard accuracy as an important feature of explanation d . It may be that all sides are correct in that the noun ‘explanation’ is actually ambiguous between at least these three senses. We are not the first to propose this. Craver (2014), for instance, claims that ‘explanation’ can refer to “a representation or text…a cognitive act, and…an objective structure” (also see Waskan 2006). Salmon (1998) likewise shows sensitivity to “The radical ambiguities of ‘explanation,’” which, he claims, “create almost endless opportunities for obfuscation and confusion.” Though ‘explanation’ seems to designate either artifacts, objective happenings, or mental states, the three senses of the term are closely intertwined. For instance, what explanations d we have may depend upon what explanations r- a we have read (e.g., phlogiston or oxidation theory). And what we believe to be the explanation o for an occurrence (e.g., combustion) may depend upon what theories we take