Abstract
Today the expression scientific realism designates a broad (perhaps too broad) range of epistemological conceptions against which stands an even broader range of ‘anti-realist’ conceptions. We shall not survey these positions, nor enter the critical discussion concerning their merits and shortcomings. Our goal is much more practical. Since it is clear from the whole of the foregoing discourse that we advocate a form of scientific realism (because we have maintained that “scientific objects are real”), we want to make more precise and explicit which form of realism we advocate and at the same time give (or better recapitulate and restate) the fundamental reasons we have found for maintaining realism.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
A more detailed development of the historical considerations presented in this section can be found in Agazzi (2001). Let us stress that, owing to the specific interest we have in characterising our conception of scientific realism, we are not going to discuss other valuable conceptions of scientific realism that have been proposed rather recently, of which Niiniluoto (1999) and Psillos (1999) are significant examples. Therefore, our attention will be rather concentrated on the discuyssion of the more significant anti-realist positions.
- 2.
For example, those who restrict existence to observable entities are making an ontological claim (Mach, and perhaps to a certain extent, van Fraassen); others consider realism simply as the contrary of idealism (Popper), though in defending their position they may introduce arguments of an epistemological or semantic nature, as we shall see in the sequel. In particular, it must be noted that even within the logical-empiricist tradition realism was not rejected, but simply accommodated to very restrictive conditions. An analysis of this issue, and of the evolution of the ‘reality principle’ is offered in Tarozzi (1988).
- 3.
Historical accuracy would oblige one to note that already at the beginning of the nineteenth century positivism had adopted an anti-realist conception of modern science, to which was already given fuel by Newton’s (and everybody else’s) inability to indicate the mechanism of gravity. Therefore one can recognise that it is fundamentally for this reason that anti-realism first gained a foothold; but it is also undeniable that the generalised decline of the realist conception of science was determined by the deep ‘foundational’ crisis in the exact sciences that occurred at the end of that century.
- 4.
For more details and relevant quotations we refer to Sect. 1.4 of this book, and to Agazzi (1994).
- 5.
These general methodological principles are clearly summarised, for instance, in the Scholium Generale of Newton’s Principia (Newton 1687), and, in a more elaborate form, in Question 31 of the third book of his Opticks (Newton 1704).
- 6.
This doctrine is presented in the first edition of the Critique of Pure Reason, and taken up again in the Prolegomena. It tends, first, to free the ‘idealist’ from the charge of denying the existence of external things:
The term “idealist” is not, therefore, to be understood as applying to those who deny the existence of external objects of the senses, but only to those who do not admit that their existence is known through immediate perception, and who therefore conclude that we can never, by way of any possible experience, be completely certain as to their reality (A368–369).
The reason we can never be completely certain of the existence of such objects is that it should be inferred as a causal explanation of our “inner perception,” and “the inference from a given effect to a determinate cause is always uncertain, since the effect may be due to more than one cause” (ibid.).
However, while the general idealist finds himself in this position, the transcendental idealist is—according to Kant—much better off, since he subscribes to “the doctrine that appearances are to be regarded as being, one and all, representations only, not things in themselves, and that time and space are therefore only sensible forms of our intuition, not determinations given as existing by themselves, nor conditions of objects viewed as things in themselves” (A369). The important consequence is that such a ‘transcendental idealist’ can also be a realist: not, of course, a ‘transcendental realist’ who “interprets outer appearances (their reality being taken as granted) as things-in-themselves, which exist independently of us and of our sensibility” (A369). He can only be an empirical realist, that is, someone “who may admit the existence of matter [we should say of the ‘external world’] without going outside his mere self-consciousness, or assuming anything more than the certainty of his representations” (A370). From all this the desired conclusion follows:
The transcendental idealist is, therefore, an empirical realist, and allows to matter, as appearance, a reality which does not permit of being inferred, but is immediately perceived (A371).
However, the Achilles’ heel is here that the existence of matter is immediately secured just as appearance. There is no need to ‘infer’ it only because it is nothing but one of two different kinds of representation (the representations affecting the so-called “outer sense,” which is itself a structure of the thinking subject):
External objects (bodies), however, are mere appearances, and are therefore nothing but a species of my representations, the objects of which are something only through these representations. Apart from them they are nothing. Thus external things exist as well as I myself, and both, indeed, upon the immediate witness of my self-consciousness. The only difference is that the representation of myself, as the thinking subject, belongs to inner sense only, while the representations which mark extended beings belong also to outer sense (A370–371).
One must admit that the consolation is rather meagre. If all of what we can be certain is the existence of our representations, of appearances, it is not the fact that we can classify certain of them under the special rubric of belonging to our ‘outer sense’ that may entitle us to be realists in any serious sense, that is, the sense that reality does not reduce to representation. Kant himself was certainly dissatisfied with this doctrine, which occupies several pages in the first edition but is completely dropped in the second. Here it is replaced (but in a different place and context) by the “refutation of idealism” (B274–279) which, however, is no less cumbersome and unconvincing. The basic reason for this frustration is that this doctrine depends on that epistemological dualism of which we have spoken several times (indeed, the pages presenting this doctrine are among the most expressive documentation of Kant’s adhering to this gratuitous presupposition). Therefore, it is only by abandoning the unjustified presupposition that what we know are just our representations that one can be a realist in a non-vacuous sense.
We have devoted special attention to this Kantian doctrine because it has been recovered, almost literally, in Putnam’s doctrine of ‘internal realism.’ Therefore, it is not only relevant to the contemporary debate concerning realism, but also shows features which are also those of certain positions in this debate as well.
- 7.
See, for example, Smart (1968), Putnam (1975, Vol. 1, p. 75), Musgrave (1988), Niiniluoto (1999, p. 197).
- 8.
For example, even J. C. Maxwell, in the last pages of his Treatise on Electricity and Magnetism, indicated as a task for future generations that of finding a mechanical description of the electromagnetic field, for which he had offered his famous equations. For details on this issue see, for instance, the large Introduction to my Italian translation of Maxwell’s work (Maxwell 1972), as well as my article Agazzi (1975).
- 9.
Mach’s work on the historical development of mechanics, which substantiates his views on this issue, was actually published in 1883 (see Mach 1883).
- 10.
van Fraassen (1980), p. 8.
- 11.
Op. cit., p. 12.
- 12.
Actually in van Fraassen (2008) the statement view is abandoned in favor of a representational view that the author qualifies as ‘empiricist structuralism.’ In the Introduction of this new book (p. 3) however, he explicitly confirms the continuity with the position defended in his book of 1980, in particular as far as the issue of realism is concerned, and he does not offer new arguments for advocating his anti-realist position (though, under a closer scrutiny, one could say that anti-realism is more definitively affirmed here). This is why we are going to concentrate our attention on the 1980 work. By these remarks I do not want to underestimate the rich harvest of novelties and interesting reflections that are offered in the 2008 work, and even the presence of significant points of contact with views that I had been defending for many years (and are also presented in the present work). For example, the non-‘epistemologically dualist’ conception of phenomena he defends, the contribution of experiments to theory construction, the ‘pragmatic’ status of scientific concepts, representations and theories, the ‘indexical’ component of scientific statements, representations and theories, and the role of intentionality—though on each of these points the affinities are also accompanied by significant differences that will be pointed out when they are met with in the course of this work. The reason for such differences is constituted in general by the strict radical empiricist orthodoxy to which van Fraassen intentionally sticks, as opposed to the more open attitude I adopt towards the role of reason.
- 13.
Sellars (1963), p. 97.
- 14.
See what we have said in Sect. 4.6 regarding the appropriate use of the adjective “true.”
- 15.
Remarks similar to those made here are expressed in Ellis (1985), where it is stressed, in particular, that in the case of theories characterised by the aim of providing causal explanations of phenomena, it is irrational not to admit the existence of the postulated causes.
- 16.
Fine (1986), p. 127.
- 17.
For example, the activity of fishing is characterised by the intrinsic goal of trying to catch fish, but this does not mean that the primary intention of everyone who goes fishing is to catch fish. Professional fishermen catch fish primarily for the sake of their livelihood, while anglers do it for amusement. However, it remains true that these persons actually fish only to the extent that their immediate purpose coincides with the defining goal of fishing. Let us also note that the pursuit of this goal remains unaffected by whether the person performing this activity actually reaches the goal. (One is fishing independently of whether one succeeds in catching any fish, or even whether there are any fish to catch, provided one intends to catch fish.)
- 18.
Several authors (e.g. Laudan 1977, van Fraassen 1980, Putnam 1981, Popper 1983) have pointed out the difference between the intrinsic goal or aim of science and the contingent intentions of those who are concerned with science, and have admitted that science has a goal. However, they do not agree as to the specification of this goal. For Laudan, the aim of science is problem-solving in a very broad sense; for van Fraassen, it is the construction of empirically adequate theories; for Putnam, it is producing a rationally acceptable representation of the world; for Popper, it is providing satisfactory explanations of what strikes us as being in need of explanation, and so on. However, I believe that all these goals implicitly presuppose, or entail, the pursuit of truth. This is even explicitly admitted by Popper (1983, p. 132). van Fraassen and Putnam do not exclude truth, but rather restrict the sense in which it may apply to science. As to Laudan, it seems undeniable—unless we limit ourselves to a purely genetic or psychological account (i.e. to acknowledging that science arises out of the spontaneous curiosity of human beings)—that what is aimed at in science is not some vague ‘solution,’ but the true solution of problems. Therefore, we believe that Laudan’s proposal is not at variance with, but rather provides a useful genetic complement to the characterisation of science as a truth-seeking human activity. These are the reasons for which we feel right in suggesting the search for truth to be the aim of science. But, of course, the soundness of this claim depends on several clarifications regarding the notion of truth, many of which we have already proposed, and some of which we shall present later.
A last remark: by making the search for truth the characteristic mark of science we do not want to intend that whatever truth-seeking activity is a science. We want truth be not only ascertained, but also understood (according to that synergy of empiricity and logos that has been included in the definition of science since antiquity). For this reason we shall use the expression “full truth” in the rest of the present subsection.
- 19.
Rescher portrays this situation well:
We realise that there is a decisive difference between what science accomplishes and what it endeavours to do. The posture of scientific realism—at any rate of a duly qualified sort—is nevertheless built into the very goal-structure of science (Rescher 1982, p. 249).
- 20.
This methodological rule of ‘non-exclusiveness’ applies in particular to the instrumentalist view of science. Instrumentalism is right in what it affirms, namely, that scientific theories must (in a broad sense) be ‘reliable’ (be able to account for data, make predictions, permit useful applications and so on). This does not imply, however, that other goals should be denied, that is, be considered as illusory or misleading. In particular, the search for truth and for a true description of reality cannot be dismissed as illegitimate unless sound reasons for such claims are convincingly advanced. Therefore, the realist can accept the instrumentalist requests, but in addition she maintains that other goals can legitimately be pursued by scientific inquiry.
- 21.
These positions are sometimes labelled “referential realism” and “truth realism” respectively (e.g., in Harré 1986, pp. 65ff.), but what we have in mind perhaps does not fully coincide with these characterisations. Therefore we prefer to avoid using any kind of terminological classification.
- 22.
Since this task is rather complex, and sometimes demands reference to certain issues which have already been tackled in other parts of this study, we shall once again take up some of these points in the present discussion, rather than simply refer back to them. This will sometimes involve a few repetitions, but this might not be too high a price if compared with the (practical) advantage of making this chapter more self-contained, and the (conceptual) advantage of stressing more explicitly the links between these former positions and the general question of realism.
- 23.
Bunge (1989), p. 130.
- 24.
Mach (1872).
- 25.
See, for instance, Popper (1972), p. 38; (1983), p. 80.
- 26.
Popper (1983), pp. 82–83. The strongest of such not irrefutable arguments has been sketched by Popper e.g. in his Objective Knowledge. They are: the agreement of realism with common sense and with a general scientific mentality, the descriptive and referential characteristic of language, the avoidance of postulating the absurd idea that we create that which we perceive and know (Popper 1973, pp. 39–42).
- 27.
For a perhaps too severe, but essentially correct, appraisal of Popper’s doctrine on this point, see Keuth (1978); also, Popper’s realism is qualified—not illegitimately—as fiduciary in Harré (1986). In Buzzoni (1982) it is shown that the lack of an operational criterion frustrates Popper’s efforts to provide the foundation of a consistent realism.
- 28.
van Fraassen (1980, p. 8). In such a way, van Fraassen expresses the methodologically sound requirement of avoiding that which Harré calls “the fallacy of high redefinition,” that is, “the move by which some established metascientific concept… which has a well-understood use in scientific discourse, is redefined in such a way that there are no conditions under which it could reasonably be applied” (Harré 1986, p. 38).
- 29.
Putnam (1981), p. 49.
- 30.
The ‘linguistic turn’ may be considered as the expansion of the thesis of the ‘impossibility of transcending language’ according to which any investigation that apparently regards a certain subject-matter cannot avoid being an investigation of the discourse or discourses in which such a subject-matter is described or treated. Therefore, any reality whatever is always given in a language and, to express the point using a famous claim of Wittgenstein, “the limits of my language are the limits of the world.” This thesis is very similar to the fundamental thesis of the ‘impossibility of transcending thought’ defended by idealists, according to which it is impossible to affirm something ‘external’ to thought, since by this very affirmation one would include this something in thought. We have discussed this point in Agazzi (1989), our discussion justifying the use of the expression “linguistic idealism” for qualifying the linguistic turn; and we have noted that this position shares with idealism not only the correct claim that reality cannot be ‘separated’ from language (or thought), but also the mistaken idea that reality can be ‘reduced’ to language (or thought). An inability to separate does not imply the lack of a distinction. This is why the anti-realism based on this reduction or ‘identification’ must be held to be as naive as the anti-realism based on a similar ‘identification’ made by idealists. Several parts of the quoted paper are reproduced in the present section.
- 31.
Here again we refer to the analyses of these positions provided in foregoing sections, from which it has also appeared how these different claims are not equivalent, nor really logically entailed in the way sketched here. Moreover, one could also add that arguments of a sociological nature have joined the linguistic arguments just mentioned, especially after the publication of Kuhn (1962). We shall consider these additional elements later.
- 32.
See Sect. 4.1 for details.
- 33.
As a matter of fact, the philosophy of science inspired by logical empiricism, and continued within the analytical tradition, has been characterised by a strong ‘syntactic’ approach that, in part, reflects the spirit of the ‘formalistic’ view that became dominant in mathematics and, in addition, that allowed for the application of the sophisticated tools of mathematical logic (with the skill and sense of pride entailed by the ability to master such complicated techniques). Although some interesting results were obtained using such an approach, it seems undeniable that it proved much less fruitful than its complicated machinery seemed to promise, and that it even distracted attention from philosophically central issues, in favour of rather artificial questions. This is why, in the present work, we did not follow this syntactic approach.
We are not alone in this evaluation. van Fraassen, e.g., declares: “Perhaps the worst consequence of the syntactic approach was the way it focused attention on philosophically irrelevant technical questions. It is hard not to conclude that those discussions of axiomatisability in restricted vocabularies, ‘theoretical terms,’ Craig’s theorem, ‘reduction sentences,’ ‘empirical language,’ Ramsey and Carnap sentences, were one and all off the mark—solutions to purely self-generated problems and philosophically irrelevant. The main lesson of twentieth-century philosophy of science may well be this: no concept which is essentially language-dependent has any philosophical importance at all” (van Fraassen 1980, p. 56). We essentially agree with this judgment, and especially with its conclusion, though we think that a more balanced evaluation should be made of the formalistic approach taken as a whole (see Agazzi 1990). In any case, we want to stress that the mentioned work of van Fraassen (1980) really meant a decisive overcoming of the strongest limitations of the logical-empiricist paradigm, by introducing (with his ‘constructive empiricism’ a significant appreciation of the semantic dimension. (Of course, the introduction of a semantic approach in the analysis of the structure of scientific theories had already known a significant start at the end of the 1960s and beginning of the 1970s of the twentieth century, in the works of Suppes and Sneed, later developed by Stegmüller and his collaborators of the so-called ‘structuralist school,’ of which we have already had the opportunity of speaking in Sect. 3.2 of this work. One can also mention Agazzi (1976) as a contribution in this direction.) Yet van Fraassen still remained partially prisoner of the linguistic approach and especially of radical empiricism. An overcoming of the first limitation is attained in his other fundamental work (van Fraassen 2008), where the notion of ‘scientific representation’ entails abandoning the previous approach to philosophy of science in terms of theories and their truth (where truth was replaced by the notion of ‘empirical adequacy’), and the new perspective of an ‘empiricist structuralism’ is presented, in which theories are conceived as systems of models and representations to which the predicate ‘true’ does not apply. The rather curious consequence is that, whereas van Fraassen’s first view attributed to science at least a rough empirical realism, the second view is much more radically anti-realist, since there is no possibility of finding, in the different levels of models proposed by the author, a ‘bottom level’ of which it could be said that it ‘represents’ the external world.
- 34.
What we have said explains why, for example, a phenomenalist could also accept our thesis and say that scientific discourse is not a simple language game, that it must have referents, but that these are only phenomena, and not things as they really are, or things in themselves. The essence of the views of van Fraassen (1980) and Putnam, for example, could be included in this line of thought. This means that they actually share some of the features of realists, in the sense explained here. The fact that they do not accept this qualification (at least in full) clearly depends on certain additional requirements they impute to realism, to which they are not prepared to subscribe. The examination of these requirements will occupy us soon, but what we have seen is already significant because it explains why the most recent debates concerning realism have increasingly abandoned the orbit of philosophy of language in which this discussion was couched for many years, and take into account more traditional non-linguistic factors, such as scientific progress, the success of science, the aims of theories, and so on.
- 35.
As expressed by Shapere, whose statement we fully endorse: “According to a widely cited slogan, the philosopher of science must pay attention to what scientists ‘do’ rather than to what they say. I believe, however, that we must attend to both, though of course with a great deal of critical awareness” (Shapere 1984, pp. XXXVI–XXXVII).
- 36.
See the introductory considerations of Agazzi (1976) for more details.
- 37.
Here again we speak of true or false theories, following a common way of speaking, without recalling what we have already said in Sect. 5.2.1 about the legitimacy of this extendend use of the notion of truth.
- 38.
This is also a central point in Dilworth (2008).
- 39.
This point has already been presented in greater detail in Sect. 3.3, and corresponds to a ‘minimal’ condition of stability, which is sufficient for our purposes. However we would at least like to mention a more elaborate doctrine developed in several papers by Shapere which convincingly accounts for the historical change of meaning (and even of reference) of scientific concepts, without entailing incommensurability, provided one considers the actual reasons that have determined the piecemeal change in question:
The idea of ‘chain-of-reasoning-connections’ disposes of the problems of “incommensurability” which have been the source of so many relativistic and sceptical views of science during the last two decades.… Later ideas in science are often rational descendants of earlier ones, even if they abandon a great deal, or even all, of what was in those earlier ideas. (Shapere 1984, pp. XXXVII–XXXIII)
The requirement of the stability of reference has been notoriously investigated in the theories of ‘direct reference’ and ‘causal theories of reference’ elaborated by Putnam (1975) and Kripke (1980). Since they are well known and, moreover, are of a general character and less related to the problem of ‘scientific’ realism, we believe we can dispense with presenting them here.
- 40.
For a good analysis of this topic from the point of view of the philosophy of language, see Smith (1981), pp. 106ff.
- 41.
It may be useful to note that there is a difference between our position and that presented by Dilworth (2008): he suggests that sameness of operations is only a criterion for sameness of referent. Sameness of referent—according to him—is determined by the intentions of the persons applying the conceptual scheme or theory in question. This not negligible difference depends on the fact that Dilworth strictly links reference with the personal intention of the subject using a term, and does not consider it (as we do) to be a semantic constituent of the meaning of a concept. This is why his “perspectivist” conception of science, though having many points of contact with our view, has certain points of difference as well. The main reason for this is probably the fact that we are investigating scientific objectivity and, therefore, leave out of consideration the subjective factors. In particular, Dilworth can easily admit that incompatible attributes can be assigned to the same referent (because two different subjects can ‘refer’ to the same object though attributing it incompatible properties), whereas for us the referent is intersubjectively identified within a certain linguistic community by means of a certain term, and for this reason cannot be characterised by incompatible attributes.
- 42.
Such a very obvious fact has been stressed by those authors who have been aware of the importance of not reducing the whole of meaning either to sense or to reference. Let us simply quote two examples. Harré says:
We must routinely distinguish the business of establishing that something exists from the ever open possibilities for further research into what it is that exists. A referent can persist as the focus of empirical research and as the subject of predication even through recategorisations of a rather drastic sort. We can maintain our focus on an existent while our researches into its nature lead us to abandon every statement we once thought true of it, except that it exists and that its nature is such as to secure it a place in some referential grid (Harré 1986, p. 66).
And Putnam, in his paper ‘Explanation and Reference’ (in his 1975, II, pp. 196–214) attacks the alleged dependence of reference on sense and context (“concepts which are not strictly true of anything may yet refer to something; and concepts in different theories may refer to the same thing”) and, moreover, recognises that, though the intension of a concept might not be well determined, elements having referential relevance may occur in it:
I said before that different speakers use the word “electricity” without there being a discernible ‘intension’ that they all share. If an ‘intension’ is anything like a necessary and sufficient condition, then I think that this is right. But it does not follow that there are no ideas about electricity which are in some way linguistically associated with the word. Just as the idea that tigers are striped is linguistically associated with the word “tiger,” so it seems to me that some idea that “electricity” (i.e. electric charge or charges) is capable of flow or motion is linguistically associated with “electricity” (II, p. 200).
It is interesting to note that, in the same paper, Putnam strongly attacks the philosophy of science of logical empiricism, charging it with “idealism.”
- 43.
If we wanted to be extremely scrupulous we should say that a discourse in itself does not ‘intend’ anything, but only the speaker who uses this discourse intends to say something. This is in a way correct, but we have already explained when we spoke of the possible differences between sentences, propositions and statements, and more extensively when we spoke of judgments and of apophantic logos, that we tacitly (and obviously) understand that the discourses we consider are stated by persons. But, on the other hand, we can also tacitly presuppose that there is concretely a broad consensus regarding the ‘intensions’ of the linguistic expressions belonging to a certain natural or disciplinary language, a consensus that dispenses one with the need of knowing the subjective individual intentionality of every single speaker. This is what actually happens when we read a book, a paper, or consult a dictionary, and even more strictly when we have to do with a disciplinary language (which is, after all, the subject matter of our investigation). Let us simply refer to the general discussion we have devoted to intersubjectivity.
- 44.
The reader can easily see that the content of this section is to some extent a recapitulation of certain results presented in greater detail especially in Sects. 4.2 and 4.3. See also Agazzi (1997b). We believe that these repetitions are justified due to their making this chapter on scientific realism self-contained.
- 45.
This kind of realism can be accepted even by those who feel inclined towards a certain ‘phenomenalism,’ to employ an expression we have already met. In fact, if such a phenomenalism does not want to be confused with the incorrect thesis that we only know phenomena but not reality, it must identify phenomena with what we previously called the aspects of reality that are considered as objects by a particular science. In this sense every scientist can or must be a phenomenalist as far as he is aware that only certain features of reality are treated by his science but, once this correct admission is made, he must also recognise that these phenomena are part of reality, and thus he must confess to being a realist. For this reason it seems advisable not to speak of phenomenalism, since such a word is frequently used to designate the doctrine that we can only know certain appearances and not reality as it actually is.
- 46.
This criticism also applies to those scholars, such as Ian Hacking, who are ready to admit a ‘realism of entities’ but not a ‘realism of theories’ (see, e.g., Hacking 1983), because theories are nothing other than the way of expressing the properties of the entities they admit, i.e., of saying ‘what these entities are,’ and without this one would not be able to say what the entities are whose existence one admits. To be fair, we must recognise that Hacking is pointing (more implicitly than explicitly) to that requirement of operationality which, as we have seen, is fundamental for the referentiality discourse and for the claims of existence. In fact he maintains that even unobservable entities can be credited with existence if they can be used (to put it briefly) in the building and functioning of some scientific instrument or concrete process, because this fact shows that they are endowed with some causal ability, which indicates that they are ‘real.’ We note that this indicates that in scientific discourse and practice they have become so familiar that they can be taken for granted and ‘used,’ and this (to use our terminology) amounts to their being considered as things. This does not eliminate the fact, however, that they have been made identifiable through a long process of determining their properties. For instance, the electron (to use Hacking’s example) was submitted to experimental manipulations in order to establish its charge, to determine that it can be diffracted, that scattered electrons conserve their total energy, that they can be filtered, etc., and in such a way they could result (thanks to considerable theoretical work) in that kind of entity that could be concretely ‘sprayed’ in the Fairbank attempt to detect free quarks. Their ‘reality’ as entities cannot be separated from a realist interpretation of the theory determining what kind of entities they are.
- 47.
For the example of phlogiston see the discussion in Smith (1981), pp. 112ff. Also Poincaré once made a remark regarding the alleged elimination of the referents of old theories: “Barely 15 years ago, was there anything more ridiculous, more quaintly old-fashioned, than the fluids of Coulomb? And yet, here they are reappearing under the name of electrons” (Poincaré 1902, p. 164).
- 48.
We shall provide more details on this point when we speak of scientific progress.
- 49.
Tarski (1944), p. 363.
- 50.
Here is a paradigmatic example. The opening lines of Fine (1984) present a strong anti-realist declamation:
Realism is dead. Its death was announced by the neopositivists.… Its death was hastened by the debates over the interpretation of quantum theory …. Its death was certified, finally, as the last two generations of scientists turned their backs on realism and have managed, nevertheless, to do science successfully without it” (op. cit., p. 83).
However, when one considers what Fine’s own position is—which he calls the “natural ontological attitude,” which he claims to be equidistant from realism and instrumentalism—one finds that it consists in a so-called “homely line of argument” thanks to which “it is possible to accept the evidence of one’s senses and to accept, in the same way, the confirmed results of science” (p. 95). The consequence is that even the most classical theoretical constructs of scientific theories are to be taken as existing:
I have similar confidence in the system of ‘check, double-check, triple-check’ of scientific investigation, as well as the other safeguards built into the institutions of science. So, if the scientists tell me that there are molecules, and atoms, and y/J particles, and who knows maybe even quarks, then so be it. I trust them and, thus, must accept that there really are such things, with their attendant properties and relations (op. cit., p. 95).
How is one to distinguish this position from the usual realist positions (and indeed the most committed ones) is hard to say. The difference may be seen in the kind of argument adopted (quite a weak one that relies on a fiduciary trust in the scientific community), and in the almost unintelligible affirmation, according to which true realists “add onto this core position” something like “a desk-thumping, foot-stamping shout of ‘Really!’” (p. 97). This is certainly no argument against realism, but rather the expression of an emotional defence against being considered a realist. By our remarks, however, we do not exclude that, by a deeper scrutiny of Fine’s position, one might perhaps conclude that he is ready to accept a form of ‘empirical realism,’ but we are not interested in this additional point.
- 51.
An example is Popper. We have already noted that he recognises himself to be unable to provide conclusive arguments in favour of realism. However, he says:
This robust and mainly implicit realism which permeates the Logic of Scientific Discovery is one of its aspects in which I take some pride. It is also one of its aspects which links it with this Postscript, each volume of which attacks one or another of the subjectivist, or idealist, approaches to knowledge (Popper 1983, p. 81).
Clearly, philosophical positions are not something of which one is to be or not to be proud, and Popper is here using the language of someone who has fought for years a long battle for a noble cause rather than the language appropriate to an ‘objective’ philosophical discussion.
- 52.
This fact is also discussed in the first chapter of Dilworth (2007). As examples of misrepresentations of realism that work as presupposition for the defense of certain forms of antirealism we can mention those of van Fraassen and Putnam already considered in Sect. 5.2.3.
- 53.
Let us give a few examples of these ‘dualistic’ portrayals of realism. Ellis says:
We can investigate nature and develop a theoretical understanding of the world, but we cannot compare what we think we know with the truth to see how well we are doing. (Ellis 1985, p. 69).
Also Fine, though beginning his presentation of realism in a seemingly neutral way (1986, p. 150), imperceptibly shifts to a neatly dualistic misrepresentation of it:
The problem is one of access. The correspondence relation would map true statements (let us say) to states of affairs (let us say). But if we want to compare a statement with its corresponding state of affairs, how do we proceed? How do we get at a state of affairs when that is to be understood, realist-style, as a feature of the World?… The difficulty is that whatever we observe, or, more generously, whatever we causally interact with, is certainly not independent of us. This is the problem of reciprocity (Fine 1986, p. 151).
- 54.
Indeed, as we have shown in our historical reconstruction (Sect. 5.1), modern science was realist (with few exceptions) until the end of the nineteenth century, and anti-realist positions emerged when natural science began to be typically concerned with unobservables. For this reason we have maintained that the most suitable way of characterising the issue of scientific realism, as distinct from general philosophical realism, is that of considering the reality of unobservables as its central problem.
- 55.
For van Fraassen, this actual and concrete observability by unaided human senses is a defining characteristic of what is real, so that, for example, even certain entities which are theoretical according to the usual standards in philosophy of science (e.g., stars or galaxies which cannot be directly observed de facto), may be considered real for him, since they could in principle be approached and seen directly, let us say, by astronauts in a spaceship. (For a characterisation of this position see, for example, van Fraassen 1980, pp. 13–19.) In order to be completely fair one should recognise that van Fraassen’s position is more sophisticated, in the sense that he does not flatly say that what is empirically unobservable does not exist, but that we are not logically compelled to believe that unobservable entities exist. This is, however, a rather unclear jump from an ontological and epistemological plane to an epistemic one. Indeed there are no logically cogent reasons that compel one to believe that a certain sensory perceived thing really exist. From our side, we have tried to offer, in the foregoing parts of this work, certain logically cogent reasons for admitting the existence of unobservable entities, based on the analysis of the referential nature of truth.
- 56.
See Sect. 2.8 of this work. This thesis is well defended in McMullin (1984, pp. 10–15) where it is significantly said that “imaginability must not be made the test for ontology. The realist claim is that the scientist is discovering the structures of the world; it is not required in addition that these structures be imaginable in the categories of the macroworld” (ibid. p. 14). In keeping with this correct remark, McMullin shows that the anti-intuitive conclusions of quantum physics did not have an anti-realist meaning even for Bohr, since Bohr was simply maintaining that what can be inferred about the world “is entirely at odds with what the classical world view would have led one to expect” (p. 12).
- 57.
More on this point will be said in the discussion regarding ‘phenomenological theories.’
- 58.
A position that denies the possibility of speaking of truth in the case of theories is advocated by Dilworth: “Scientific theories are seen not to be entities of the sort which are either true or false, but to be structures which are more or less applicable depending on the results of certain measurements.” In this case no radical empiricism dictates such a claim, but the general view that theories are not linguistic entities (and for this reason it is correct to say that they are not the kind of entities of which truth or falsity can be predicated). In fact Dilworth defends a view of scientific theories as applied models that are conceived with the purpose of explaining (rather than merely describing) concrete reality, a view that we also share in part and that we call a Gestalt view (in keeping, by the way, with the intuitive basis of Dilworth’s view, namely his Gestalt Model), of which we have spoken several times in this work. The difference, however, is that Dilworth advances his view as an alternative to the statement view (and develops it so as to provide an account of both theory conflict and scientific progress), whereas we defend it as a complement to the statement view, because we maintain that theories entail the linguistic expression of the content of their corresponding Gestalt, which is obtained by means of a certain number of sentences. Therefore, we are entitled to speak of the truth of theories in an ‘analogical’ sense, that is, by relating their global truth to the truth of their explicitly formulated sentences in a suitably specifiable way.
- 59.
Let us note that we do not qualify common sense realism as ‘naive realism’ since we attribute to common sense a much deeper meaning and significance than what is often meant by many philosophers. The relations between common sense and scientific knowledge are by no means trivial, and we refer to Agazzi (2002) for a discussion of this issue.
- 60.
van Fraassen (1980), p. 12. We want to note an interesting consideration made by Dilworth, that points out an implicit extension of van Fraassen adequacy also to unobservable entities: “what a theory referring to unobservables ‘says about observable things and events’ is that they take the form they do as a consequence of the nature and behaviour of the unobservables. So his own manner of expression suggests that a theory’s being ‘empirically adequate’ implies its also being ‘theoretically adequate,’ i.e. that it correctly depict what is unobservable.” (Dilworth 2007, p. 38). In van Fraassen (2008) the notion of “appearance” is introduced in addition to that of “phenomenon,” and appearances are defined as the result of measurements. We do not find this peculiar convention (which we would attribute to the radical empiricism of the author for reasons that we do not explore here), useful, and we have additional reasons for not using it, since our theory of objectivity, as will be more and more clear, intends to analogically concern all the sciences, and not only physics and, in particular, is totally independent of measurement. On this point see Agazzi (1978e).
- 61.
van Fraassen (1985), p. 255.
- 62.
A more detailed reasoning might be appropriate here. Our application of the parsimony principle is indeed correct only in the case that the conclusions are necessarily entailed by the assumptions. In the case of expansive inferences (that are typical for the introduction of theoretical constructs in science) however, this might not be the case. Therefore, considering van Fraassen’s position, we should rather object that one does not see any justification in the fact that the same logico-inferential procedures, whose legitimacy and soundness he admits as far as they produce inferences whose conclusions do not overstep the phenomenal domain, suddenly become unreliable once this domain is overstepped. This kind of criticism can be found in Alai (2010).
- 63.
See, for example, Fine (1984), p. 89.
- 64.
Let us quote a significant statement by McMullin: “Theory explains by suggesting what might bring about the explananda. It postulates entities, processes, relations, themselves unobserved, that are held to be causally responsible for the empirical regularities to be explained” (McMullin 1984b, p. 210). This view is also central to the conception of explanation accurately presented in Dilworth (2007). Similar considerations are expressed by Rescher:
In attempting answers to our questions about how things stand in the world, science offers (or at any rate, both endeavors and purports to offer) information about the world. The theory of sub-atomic matter is unquestionably a ‘mere theory,’ but it could not help us to explain those all too real atomic explosions if it is not a theory about real substances. If I hypothesise a robber to account for the missing jewelry, it is not a hypothetical robber that I envision but a perfectly real one. Similarly, if I theorise an alpha particle to account for that photographic track, it is a perfectly real physical item I hypothesise and not a hypothetical one. Only real objects can produce real effects …. The theoretical entities of science are introduced not for their own interest but for a utilitarian mission, to furnish the materials of causal explanation for the real comportment of real things. (Rescher 1987, p. 38).
- 65.
Arguments in favour of including truth among the requirements of genuine explanatory efficacy are expressed, for example, in Leplin (1984a, p. 212) and McMullin (1984a, p. 29).
- 66.
van Fraassen (1980, p. 69). A fully developed metaphysical alternative to logico-linguistic philosophy of science (partially different from the one that is being developed in the present work) is offered in Dilworth (2007).
- 67.
van Fraassen, op. cit., pp. 202–203.
- 68.
This point is also made in Glymour (1984), pp. 188–189.
- 69.
This issue is presented with much efficacy in Fine (1986), pp. 167–168:
Despite the uniformity of practice, however, constructive empiricism does feel a need to multiply its interpretation of that practice (going here for acceptance, and only there for belief). That need goes against its deflationist promise. It is generated only by the prior commitment to empiricist epistemology…. What positive arguments or reasons connect the two, providing the grounds to multiply interpretations of the inferential practice? The answer is that the constructive empiricist has no argument. It goes its inflationist way in order to prop up empiricist epistemology. There is no other (or better) reason that supports its chosen path (p. 168).
- 70.
Putnam (1975), p. 73.
- 71.
A deductivist attempt to account for scientific progress may be found in Boyd (1984), where, among other things, a defence of the significance of crucial experiments, even in the presence of a kind of theory-dependence, is presented (p. 59), and a defence is provided of the claim that the technological realisations of science cannot be paradigm-dependent (p. 60).
- 72.
This means that one would not be disturbed even if what Fine says were true: “The plausibility of the explanandum (that the conscientious practice of science leads to abundant instrumental success) is an artefact of our historical perspective” (Fine 1986, p. 152).
- 73.
Consider Laudan’s very strong statement:
For every highly successful theory in the past of science which we now believe to be a genuinely referring theory, one could find half a dozen once successful theories which we now regard as substantially non-referring (Laudan 1984, p. 212).
This affirmation should be backed by the historical studies the author has presented elsewhere (e.g., in Laudan 1977), but it is debatable that the results of these studies really justify this claim, as we shall see.
- 74.
Here our example of the map may provide a useful (though only partial) analogy. If we have a good map of a city, printed a few years ago, and which then proved very accurate in describing the city centre, we may still use it even if we know that in the meanwhile new suburbs have developed, which are either not described, or only very imperfectly described. This means that we can still rely on the accuracy of our map as far as the city centre is concerned, and use it for finding streets and squares which we had not visited or seen at the time when we first bought the map, while we are to be very careful about those parts of the city which are new. With respect to them, it is more than probable that our map will not help us, and that we need a new one. That this new map also contains an accurate description of the city centre might be an advantage, but it is not necessary. We could use both maps, according to our concrete needs.
- 75.
In McMullin (1984) there is an interesting discussion of the features which must characterise a theoretical construction describing ‘structural’ properties of a scientific domain in order for it to be a reliable candidate for ontological reference. In particular, ‘fertility’ is a desired virtue (see pp. 30–34).
- 76.
For this reason we must consider the following statement of Laudan as mistaken (a statement which, however, is indicative of a typical misunderstanding in the interpretation of the historical succession of theories):
After all, presumably many theories which we believe to be false (e.g., Newtonian mechanics, thermodynamics, wave optics) were—and still are—highly successful across a broad range of applications (Laudan 1984, p. 228).
In fact, we must say that Newtonian mechanics, thermodynamics, wave optics are still true, not ‘in general,’ but of their objects, that is, of those aspects of reality which may be investigated by resorting to their criteria of referentiality. Of course, it is true that absolute space, for instance, does not exist, but this was not the kind of ‘entity’ which was implied in the predictions of Newtonian mechanics. It was a ‘redundancy’ which Newton had introduced for certain metaphysical reasons. A confirmation of this unsatisfactory approach, which fails to take into account the ‘relativity’ of scientific truth to the ‘objects’ of a theory, is given by Laudan a few pages later:
It is well known that statistical mechanics has yet to capture the irreversibility of macrothermodynamics as a genuine limiting case. Classical continuum mechanics has not yet been reduced to quantum mechanics or relativity. Contemporary field theory has yet to replicate the classical thesis that physical laws are invariant under reflection in space (p. 238).
This is presented as an objection to the (allegedly) realist claim that new theories must include old theories as limiting cases. However, this is not a requirement of realism (despite certain realists’ perhaps having made it one), since in certain cases it may happen that a new theory includes the old as a limiting case (when the objects remain the same, as we have already explained), but such cases are by far the less frequent. In general, we have (as in the examples quoted by Laudan) a change of objects, and therefore we have truly referring theories which do not preserve the old reference, not because they ‘eliminate’ it or show it to be non-existent, but because they investigate a new one.
- 77.
We are not going to deepen the discussion of Putnam’s internal realism here. (See especially Chap. 3: ‘Two Philosophical Perspectives,’ pp. 49–74 of Putnam 1981.) We have provided such a discussion in Agazzi (2001).
- 78.
Laudan (1984), p. 229.
- 79.
Leplin (1984a), p. 202.
- 80.
For a deepening of this discussion we suggest taking account of the treatment of the relative acceptability of scientific theories, as well as of the criticism of Popper’s ‘verisimilitude,’ contained in Dilworth (2008).
- 81.
For a good discussion of this issue see Alai (2012).
- 82.
Boyd (1984), p. 77.
- 83.
See, e.g. Queraltó (1999).
- 84.
Buzzoni advocates an even more radical connection between the scientific and the technological dimensions (see Buzzoni 1995 and 1997).
- 85.
For greater detail, see Chap. 4 of Agazzi (1992), as well as Agazzi (1999).
- 86.
A linguistic precisation may be advisable. The distinction between technique and technology proposed here sounds rather natural in many modern languages where the corresponding words are present, whereas it may sound a little peculiar in English. Indeed, in this language the word “technology” is normally applied throughout the history of humankind, such that e.g. the development of ancient tools constitutes instances of technological development. Therefore it is certainly possible to cover all the cases one wants to cover using the English terms “technique” and “technology” in their usual senses, and qualifying them when necessary. Nevertheless we considered it meaningful to elaborate this a little in our conventional definition of technology, not only for purely analytic purposes, but also because it helps in understanding the spirit of modern civilisation, as will also appear from further considerations we are going to propose later.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Agazzi, E. (2014). Scientific Realism. In: Scientific Objectivity and Its Contexts. Springer, Cham. https://doi.org/10.1007/978-3-319-04660-0_5
Download citation
DOI: https://doi.org/10.1007/978-3-319-04660-0_5
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-04659-4
Online ISBN: 978-3-319-04660-0
eBook Packages: Humanities, Social Sciences and LawPhilosophy and Religion (R0)