Two Notions of Scientific Justification

Scientific claims can be assessed epistemically in either of two ways: according to scientific standards, or by means of philosophical arguments such as the no-miracle argument in favor of scientific realism. This paper investigates the basis of this duality of epistemic assessments. It is claimed that the duality rests on two different notions of epistemic justification that are well-known from the debate on internalism and externalism in general epistemology: a deontological and an alethic notion. By discussing the conditions for the scientific acceptability of empirical results, it is argued that intrascientific justification employs the deontological notion. Philosophical disputes such as those on scientific realism can by contrast be shown to rest on the alethic notion. The implications of these findings both for the nature of the respective epistemic projects and for their interrelation are explored. 1. Scientific and philosophical assessments of science The justification of scientific claims regularly displays a familiar, yet puzzling ambiguity. Scientific theoretical claims, for instance, can on the one hand be justified according to the scientific standards, e.g. by being confirmed by specific empirical evidence and by possessing theoretical virtues such as simplicity and explanatory power. On the other hand, in disputes on scientific realism, the epistemic status of scientific theories and standards of confirmation is discussed with reference to philosophical arguments such as the abductive no-miracle argument or Laudan’s pessimistic meta-induction (Boyd 1984, Laudan 1981). In a similar way, empirical results are acceptable as evidence in science if, for instance, expected errors are controlled or corrected and our theoretical understanding of the measurement process supports the reliability of the results. At the same time, philosophical investigations of the theory-dependence of observation consider under which conditions empirical results that rely on theoretical reasons are acceptable (for instance Brown 1989; Kosso 1989; Culp 1995; Adam 2004). One might well wonder which questions of justification remain to be settled by the philosophical discussions if theories or observations have already been assessed scientifically. It might seem that in both cases, the philosophical assessment in some way duplicates the epistemic assessment ∗ The original publication will be available at www.springerlink.com, http://www.springerlink.com/openurl.asp?genre=article&id=doi:10.1007/s11229-006-9052-x. M. Adam: Two notions of scientific justification 2 that is already an integral part of scientific practice. The questions therefore arise in which respects the two kinds of assessment – intrascientific and philosophical – differ, what their respective roles are, and how they relate. According to the view that appears to be widely accepted in the philosophy of science, the two assessments both make sense since they are concerned with two different levels of assessment. While the scientific assessment evaluates scientific factual claims on the basis of methodological standards, the philosophical enterprise is seen to focus on the justification of the methodological standards themselves. According to this view, the main epistemological concern of the philosophy of science is methodological, and thus does not primarily address the epistemic status of scientific factual claims. Imre Lakatos is a prominent advocate of such a view. According to him, the epistemological task of the philosophy of science is to build a theory of scientific rationality, i.e. a theory of the standards that define the scientific method. As he sees it, this theory should be tested against the “basic value judgments” of leading scientists. The philosophical theory is confirmed if it successfully reproduces the methodological decisions taken by the scientific élite. The philosophical aim is to explicate the standards of rationality implicit in the scientific method (Lakatos 1971). In Lakatos’ view, the proper objects of the philosophical accounts are thus methodological standards and methodological decisions, not factual scientific claims. Larry Laudan has a different, yet also methodological view of the role of the philosophical assessment. According to Laudan, the aim of the philosophy of science is to evaluate the effectiveness of the scientific methodological standards in furthering the epistemic aims of science. Contrary to Lakatos, the evaluation of these standards is not based on what the scientists themselves or the scientific community of their time took to be good scientific method, but depends on the standards’ instrumental value for what we take to be the scientific epistemic ends (Laudan 1984, 1987). Still, also in Laudan’s picture, the philosophical epistemic enterprise is primarily concerned with methodological standards, not with factual scientific claims. Also in the disputes on scientific realism, the philosophical project is often portrayed as being mainly methodological, even though the position of scientific realism as such is explicitly concerned with the likely truth of scientific theories and thus with the epistemological status of scientific claims (see section 5). Still, Arthur Fine, for instance, distinguishes two levels of reasoning within Richard Boyd’s M. Adam: Two notions of scientific justification 3 prominent conception of scientific realism. On the ground level, scientists are seen to use rules of abductive reasoning to draw inferences from observables to unobservables. On the philosophical meta-level, the question is addressed why the scientific methods – including the rules for scientific ground-level abductive reasoning – are empirically successful. The position of scientific realism is taken to be located on the meta-level (Fine 1984, 84). Yet, if the philosophical arguments are primarily concerned with the methodological level, only the ground-level reasoning seems to address properly the epistemic status of theoretical scientific claims. Fine therefore concludes that it is altogether sufficient if one’s theoretical beliefs are “tutored by ordinary relations of confirmation and evidential support, subject to the usual scientific canons” (Fine 1984, 98). Any additional philosophical assessment of scientific theoretical claims, as implied in scientific realism, would only add a “deskthumping, foot-stamping shout of ‘Really!’” (Fine 1984, 97). A division of labor between scientific and philosophical assessments along the lines of the distinction of levels thus appears to make incomprehensible how philosophical arguments can address the epistemic status of scientific factual claims. Against these positions, I argue in this paper that the division of labor between the intrascientific and the philosophical assessment cannot be analyzed by reference to the distinction between the justification of factual claims and the justification of methodological standards. Instead, I maintain that the philosophical assessment is often primarily concerned with the epistemic status of factual claims as well. Still, the philosophical assessment does not duplicate the intrascientific assessment, since the two assessments differ with respect to the employed notion of epistemic justification. While the scientific assessment makes use of a ‘deontological’ notion of justification, the philosophical evaluation uses an ‘alethic’ notion. This analysis leads to a new understanding of the relationship between philosophical and scientific epistemic evaluations. While they both mainly concern factual scientific claims, they divide the justificatory labor by differing in scope and range of reasons adduced, with the philosophical assessment being fairly general and being often based on rather broad assumptions on our epistemic situation, while the scientific assessment 1 Richard Boyd supports the impression that the concern of scientific realism is primarily methodological by claiming that the abductive no-miracle argument aims to explain why the scientific method regularly leads to the acceptance of empirically adequate theories (Boyd 1990). According to him, it is “the business of scientific epistemology to explain the reliability of [principles of scientific methodology]” (Boyd 1973, 3). M. Adam: Two notions of scientific justification 4 concerns particular claims and is based on specific scientific reasons. The conceptual relationship between the two notions of justification leads to an intricate interdependence of the two types of epistemic assessment. Since the distinction between the deontological and the alethic notion of justification features prominently in the discussions on internalism and externalism in general epistemology, I will first explicate the distinction by reference to this debate (section 2). Then, I will show by a number of examples that the different notions in fact underlie intrascientific viz. philosophical assessments (sections 3 to 5). Finally, I will defend the new analysis against the traditional distinction between assessments of factual claims and of methodological standards, and explore the relationship between science and the philosophy of science and their division of justificatory labor to which the new distinction of epistemic assessments gives rise (section 6). 2. Two notions of epistemic justification and the internalism-externalism debate in epistemology In general epistemology, one often finds two different notions for assessing the epistemic justification of beliefs from a third-person perspective. William P. Alston has provided a representative characterization of the two notions by distinguishing between a “deontological” and an “evaluative” concept of epistemic justification (Alston 1985 and 1988). According to the deontological concept of justification