The Objectivity Question

Thus far we have dealt with the former question dividing realism and antirealism, whether unobservables entities can be known. This question arises from the limits of our senses and cognitive apparatuses.

The objectivity question arises for knowledge in general, not just for science. Illustrious antirealist doctrines are Protagoras’ and the skeptics' relativism; Kant’s critical idealism; Nietzsche’s doctrine that there are no facts, only interpretations (1886: § 14, 1901: § 540); the Sapir-Whorf thesis that language shapes the world (Whorf 1956: 213); post-modern gnoseology (Vattimo, Rovatti 1983; Ferraris 2012). Special arguments have been used to deny the objectivity of scientific knowledge in particular. Foucault (1966) claimed that man is built by human sciences, and Latour (1998) denied that Rameses II was killed by the bacilli of tuberculosis, because they were discovered only in 1882; in the late 1950s and 1960s Hanson, Kuhn and Feyerabend, the “new philosophers of science”, held that meanings and experience are thoroughly theory-laden; Putnam (1978b, 1981) rejected “metaphysical realism”; Goodman (1978) claimed that we make the world, or actually, many worlds.

9.1 Perspectival Realism

Today “perspectivist” philosophers stress that science is always carried out within a “perspective” characterized by a priori factors like the above mentioned ones. Nickles considers “an illusion” the idea that we can gradually eliminate every human perspectival element and finally reach a completely objective science (Nickles forthcoming^a). However, perspectivists need not be subjectivist, relativist, or antirealist: one can acknowledge the perspectival character of knowledge while recognizing that objective reality plays a role at least as important as subjectivity. In fact, the two are not incompatible nor they limit each other, precisely as each argument of a function determines exactly the values without preventing the other to do the same. Of course, just looking at the values we cannot distinguish the two arguments; but in knowledge we are given both the value and the subjective factor, and this gives us the objective factor (Alai 1994: §§ 2, 3.6, 3.7). No wonder that subsequent research in the history and philosophy of science has shown that the radical relativism of the “new philosophers of science” was at best exaggerated, and since 1994 Putnam has retracted his antirealism (Putnam 1994: 489-494, 502-506, 516-517; Putnam 2012: chs. 1-4).

Thus it is possible to embrace forms of perspectival realism (roughly what Putnam called “sophisticated metaphysical realism”), as also urged by Votsis (2012). Basically, a perspective may determine either (a) the particular aspects of reality that are selected for representation, or (b) how those aspects are represented, or (c) both (Giere 2006: 14, Votsis 2012: 90). (a) is perfectly compatible with realism, and if (b) can be shown to boil down to (a) then we have a realist answer to the objectivity question. Moreover, although the knowledge question and the objectivity question are in principle independent, many take a realist stand on both (e.g., Devitt 1984: 22; Sankey 2008: 12-18).

Perspectival realism, however should be distinguished from doctrines which are misleadingly called ‘realism’, like Kant's “empirical realism” or Putnam’s “internal realism” (Alai 1989, 1990), for they keep objectivity only by weakening it until it becomes compatible with subjectivism and antirealism (Agazzi 2014: 51-57).

9.2 Agazzi and Dilworth

A truly realist perspectivism, instead, was put forth by Agazzi in (1969). He called it “Gestalt view”, a term also used for the “new philosophies of science” (Suppe 1977), but he showed that the a priori features of science highlighted by them could be taken into account without jeopardizing realism. This view has been presented and discussed in depth again in (Agazzi 2014). In a nutshell, each scientific dis­cipline ‘clippes out’ its objects by particular empirical operations, which charac­terize its specific point of view, or “Gestalt” (Agazzi 1979: 42-44; 2014: ch. 2, pp. 83, 97, passim). Hence, “one and the same ‘thing’ can become the object of a new and different science every time a new specific point of view... is taken on it” (ibi: 84).

Thus we always work from within some particular perspective and on already structured materials; scientific objects are abstract and constructed, but the interac­tive operations by which they are “clipped out” ensure their reference to independent reality and an objective criterion of truth for claims about them. The intervention of the human subject simply brings to light different aspects of reality. “Under different conditions [and through different operations] reality would manifest itself under different aspects... but these too would be real” (ibi: 229). Therefore “(a) science attempts to represent a reality independent of science itself...; (b) what science states is an adequate representation of this reality ‘as it is'” (ibi: 263).

A similar conception was proposed by Craig Dilworth in (1981). It was called ‘perspectivism’, it borrowed some of Agazzi's key insights, and in reaction to the new philosophers of science it provided criteria for evaluating scientific progress (1981: 84-88).

9.3 Giere

New discussions on perspectivism have been spurred by Sosa (1991) and Giere, R. (2006), and the issue 84 (2012) of Philosophica collects various essays on them. Giere’s “perspectival realism” is meant to provide “a genuine alternative to both objectivist realism and social constructivism...” (2006: 14-15). On this view perspectives play a key role both in scientific observation and in theorization. In each of them perspectives affect scientific outputs in two ways: first, both the human visual system and instruments are sensitive only to some kinds of input. Second, “the output is a function of both the input and the internal constitution of the instrument” (ibi: 14).

As noticed by Votsis (2012), the first way has no bearings on the realism/antirealism discussion, it simply entails that different theories have access to different aspects of reality (like for Agazzi). From the second way, however, Giere infers that science does not describe objective features of the world, both as con­cerns everyday macroscopic objects and theoretical unobservable entities: “truth claims are always relative to a perspective” (2006: 81). “So even the claim that the sky is blue is not an absolutely objective truth. Rather, the sky appears blue to normal human trichromats” (ibi: 123). We can only say: “According to this highly confirmed theory (or reliable instrument), the world seems to be roughly such and such, [but not] ‘This theory (or instrument) provides us with a complete and literally correct picture of the world itself” (ibi: 6). Yet, Giere rejects ontological con­structivism, granting that the facts under investigation are objective (ibi: 81-82). Thus, his position resembles Kant’s, holding that we cannot observe things in and of themselves, but things-from-the-perspective of human-visual-apparatus, or things-as-represented-by such-and-such-instrument (ibi: 43, 56).

However, he faces a dilemma: either the output of each perspective is just an alternative projection of the same content (like different maps of the same territory), but then they are mutually compatible, without any threat to objectivity; or one may be correct and other incorrect, but then this can be decided by empirical evidence, and if it cannot we get just another instance of the general underdetermination problem (Votsis 2012: 101-ff).

9.4 Some Latest Proposals

A more distinctly realist position is taken by Sosa (1991), followed by Massimi (2012). According to them we can know non-perspectival facts, but we know them perspectively, because the justification of our beliefs is perspectival: it “always takes place within an epistemic perspective, including not only first-order beliefs about body x, but also beliefs... about our perceptual system, cognitive faculties, mea­surement devices, and their reliability as sources of beliefs (Massimi 2012: 40-41; see Sosa: 145, 210, passim). The result is a sort of “perspectival coherentism”, which makes justification and knowledge context-relative: “J. J. Thomson was justified to believe [that cathode rays were not electrons] from his own epistemic perspective as much as we are justified in not believing [that] from our own per­spective” (Massimi 2012: 47). But this conception “does not open the door to epistemic relativism of Rortian type or to Kuhnian incommensurability, because it is objective facts that make those beliefs true or false” (ibi: 48). So, it is realist about truth, even if antirealist about knowledge. In (2016) Massimi further tries to show how truth itself can be ‘perspectival’ while remaining correspondence to objective states of affairs.

Teller’s “panperspectival realism” maintains that “the world is too complicated for us to succeed in attaching specific referents to our terms” and “to get things exactly right”. However, “our representations tell us about an independent world without securing reference by showing that the world is very like the way it is represented in a range of different, often complementary modeling schemes”.

In fact, Teller’s perspectivism provides a further argument for realism on the knowledge question: since perception is as perspectival as theorization, and the former provides knowledge, so does the latter (ibi: 7). Thus Teller agrees with selective realisms that best theories are true, although not exactly (i.e., not com­pletely) true, and he agrees with structural realism that only structures are objective, while entities are features of our representations.

Interestingly, the discussions on knowledge and those on objectivity converge to the conclusion that both can be achieved, but only partially. In fact, this kind of modest realism is compatible even with another form of perspectivism, Nickles’ “nonrealism”. According to Nickles (this volume, forthcoming^, we don’t have sufficient evidence to believe that our best theories are true or nearly true, in the sense of providing us with a complete and final understanding of “what is really going on”. A modest realist may agree, granting that probably no accepted theory is finally and completely true, so we are still far from “the whole truth” (if there is one), and perhaps we'll never get there. Yet, such a realist may hold that some partial truths about unobservables have been found by science in the past and they still keep accumulating, probably at an increasing rate.

Acknowledgements I thank EvandroAgazzi and Vincenzo Fano for valuable suggestions.