Download a PDF version of this review.
The Cartesian syndrome
How is it that the traditional image of science has been completely overturn? Only half a century ago science was seen as the paragon of rationality, the most palpable form of rational knowledge (with logic and mathematics being the impalpable ones). Marcello Pera begins his book by explaining this transition. A “philosophical tragedy” in three acts captures the main events that led to the collapse of an ideology, that of scientism.
Act 1. In this first state, science is certain, infallible, universal and objective. Due to these characteristics, Pera concludes this incipient act might be called science as demonstration. There are two components making up the thick, reliable texture of scientific knowledge: (a) the epistemic component, (b) the methodological component. The epistemic component tells us that the scientist’s data are acquired directly and objectively from reality. These data can be experimental (like Galileo’s “sensory experiences”) or intellectual (like Descartes’ “clear and distinct ideas”). Whatever they are, they are a reflection of reality – and this itself is a truth that needs to be taken for granted. We can thus speak of two dogmas: the empiricist dogma, that of immaculate perception, and the rationalist dogma, that of immaculate conception.
The methodological component states that science provides (new) knowledge by making use of a method which guarantees that if the information is correct – which it is, due to the previous component – then so will the conclusion. This method has been portrayed differently by Bacon (organon), Descartes (regulae ad directionem ingenii), Newton (regulae philosophandi), Leibniz (a libra that weighs probabilities) and many other scholars following the same path. Nothing represents philosophizing around this component better than Leibniz’ slightly enthusiastic prophecy of a scientific debate: “when controversies arise, there is no more need for discussion between two philosophers than there is between two calculators. All the two need to do is to sit down at a table, pen in hand (having called a friend if they wish) and mutually declare: let us calculate”.
Act 2. The epistemic component crumbles. After a series of impressive successes, the epistemic pillar of science-as-demonstration receives the repeated blows from: non-Euclidian geometry, crisis of the foundation of mathematics, rejection of associationist psychology and eventually quantum physics. As the epistemic component is falling, the methodological one tends to be reinforced. Certainty is replaced by truth, truth by probability, probability by verisimilitude, but all along this path philosophers stick to the central ideas of the methodological component. “From this point of view”, Pera writes, “there is little substantive difference between Bacon, Descartes, Leibniz, Newton, Whewell, Mill etc., and between Popper, Lakatos, Laudan etc.” (p. 4) What brings them together is a common effort to carry out the Cartesian project, which Pera summarizes as follows:
First thesis: There is a universal and precise method that demarcates science from any other intellectual disciplineFrom Kant to Lakatos, many philosophers have “suckled the Cartesian project with the milk of their philosophical training” (p. 4). The most interesting thesis in this tragedy is the third one. In reality, it gives rise to a dilemma: an endeavour is either dominated by rules or dominated by irrationality. There is no middle ground.
Second thesis: The rigorous application of this method guarantees the achievement of the aim of science
Third thesis: If science possessed no method, it would not be a cognitive and rational endeavour.
Act 3. After a period of relative calm, with the epistemic component’s final collapse, the “new philosophy of science” started attacking the methodological component too. As a result, the first two theses came to be rejected. Despite what Cartesians told them, the new philosophers averred that there is (a) no universal method and, even if there were such a method, (b) it wouldn’t guarantee the achievement of any kind of scientific aim. Science, they will say, just doesn’t work like that. In this context, the third thesis received special attention. More precisely, the dilemma it gives rise to has been fully embraced by philosophers of science (and scientists) for whom science is just another form of culture like basketball and parliamentary debates. Along these lines, Pera notes that even those who opposed the Cartesian propensity for laying out methodology (Feyerabend, 1975) were still working within the Cartesian project; their conclusions regarding the rationality of science (actually, the lack thereof) were based on the acceptance of the same dilemma.
In this picture, Kuhn was an unusual apparition. Studying the kind of Gestalt switch going on within scientific revolutions (Kuhn, 1996), Kuhn tried to “see both the duck and the rabbit” of scientific activity. Pera uses this analogy and separates Kuhn’s duck (the view that scientific progress occurs outside the reach of pre-defined rules) and Kuhn’s rabbit (the view that within the process of choosing between alternative paradigms, the scientific community maintains its rationality).
From method to rhetoric
If one tries to search for a scientific method, one runs into various sorts of troubles. Before one begins, a distinction needs to be made between procedure (an ordered series of moves governed by a global strategy oriented toward a goal), rules (prescriptions and norms governing the steps in the procedure), and techniques (the actual moves required by the procedure). So, when we ask ourselves what the method of science is, we ask (or we should ask) what the procedure, the rules and the techniques of science are? These are the things that are in need of explanation – explicandum – and the business of the philosopher of science is to pair them with adequate and precise explicata.
As he advances, Pera runs into various “paradoxes”. Maybe the word “paradox” is not the best choice here, but in any case the way they are described makes them insurmountable. Each element of “method” seems to be subject to these difficulties, so there is a paradox of scientific procedure, a paradox of scientific rules and a paradox of scientific techniques and they run as follows:
Paradox of scientific procedure: given an adequate scientific procedure, it is possible to find inquiries considered pseudo-scientific which will satisfy the procedure.To sum up, wherever we look for, the two criteria of adequacy and precision work against one another: a very precise method (read: procedure, technique or rule) will not be adequate and, vice versa, a very adequate method will not be very precise. Pera concludes: “This paradox expresses an intrinsic limitation of every scientific code. It is something like a principle of methodological indeterminacy: adequacy and precision are two properties of scientific method whose product cannot go beyond a certain limit” (p. 28). This result suggests that the first thesis of the Cartesian project is untenable.
Paradox of scientific techniques: a scientific discipline can legitimately adopt the same techniques used by pseudo-scientific disciplines.
Paradox of scientific rules: given any methodological rule, there are always scientific inquiries in which it is violated.
However, before “raising the flag and admitting that the idea of method is doomed to failure”, maybe there is another line of defence. Maybe one should not look for general rules but for local ones.
Where does one look for these local explicata? History of science. But surely, history of science does not hand in ready-made methodologies, so the question of choosing between different ones will once again arise. As Pera rightly observes, “this raises the suspicion that historical meta-methodology is as circular as Cartesian methodology: the former finds in the history of science the very method it favours, the latter finds in the mind (or in actual practice) those rules it considers most desirable” (p. 33). This suspicion aside, however, there are other problems encountered by such “historical methodologists”. The problem is analogous to that of a judge in a court of law. The rules (like the law-codes) will contain lacunae and since these lacunae can be filled only with case-by-case decisions, there will eventually be no methodology – not even local ones – but a host of them. In fact, Pera concludes, “the absence of a single overarching methodology for science only complicates the task of the methodologist” (p. 39)
Just because the attempt to reduce the variety of possible methods fails, however, does not mean that there are no constraints in science – as is the view of (Feyerabend, 1975). Pera’s claim is that these constraints will be noticeable and become practical once one “moves” science from the domain of demonstration to that of argumentation. He urges a return to Aristotle for which dialectic was regulating rhetorical argumentation.
This return is necessary because methodological rules have an “open texture” and it is neither induction nor deduction but rhetorical argumentation that is being employed in cases where disagreement arises.
In other words, the problem so far was that there were only two entities on the scientific stage: facts and theories. The question thus inevitably became: how do facts support theories? But as soon as one sees that facts do not speak for themselves, that one’s activity cannot be mechanically assessed by reference to this or that methodology, the speaker – or interpreter – appears on the stage. The scientific code, just as the legal one, needs to be interpreted by the proponent in front of the judge. The whole setting resembles that of a law court and the dialectics governing it will be applied accordingly. Pera’s philosophical research program seems clear: “From now on, I shall reserve the term scientific rhetoric for those persuasive forms of reasoning or argumentation that aim at changing the belief system of an audience in scientific debates and the term scientific dialectics for the logic or canon of validation of those forms. In the case of law, a juridical logic already exists that studies specifically the validation of juridical arguments. In the case of science, a deductive logic and, although more precariously, fragments of inductive logic, exist, but we still know very little about scientific dialectics” (p. 58).
The aim of the next chapter is twofold: (1) to document the fact that scientists do in fact use rhetorical argumentation in situations where neither empirical data nor methodological principles suffice and (2) provide an understanding of the function of academic debate works as such. The first aim is easily achieved with a few examples from Galileo’s Dialogue Concerning the Two Chief World Systems, Darwin’s Origin of species and a multi-party debate in contemporary cosmology. Pera shows that scholars often explicitly restrict themselves to, say, only using “direct observations” and “mathematical demonstrations”, as Galileo did, but in reality use all kinds of rhetorical strategies to promote their theories. Moreover, a scientist condemning rhetoric and oratory will most probably make use of it a couple of pages later – as Darwin did in numerous instances.
This comes as no surprise, but it should be noted that Pera’s ease in finding examples of rhetorical argumentation is also a consequence of his rather loose definition of the concept. His examples are chosen more or less at random: “argument by retort”, “argument ad hominem”, “argument by counter-example”, “argument of parts and the whole”, “argument from a model”, “absurdity and ridicule” (pp. 62-69). Pera concludes: “Not unlike Galileo, who proceeded not only on the basis of ‘sensory experiences and necessary demonstration, Darwin did not rely uniquely on ‘true Baconian principles’ or on the principles of the hypothetico-deductive method. Darwin too preached one thing and practiced another. If not, his Origins … would have fallen prey to the first objection raised” (p. 88). In other words, it is common practice of scientific discourse to stretch the meaning of “method” to one’s own convenience.
The next question concerns the role(s) which academic rhetoric plays in scientific discourse. According to Pera, the main functions of these rhetorical arguments in scientific debate are:
(1) Choosing a suitable methodological procedure.Each of these functions is explained and illustrated. For some of them, Pera explains the reason why neither deduction nor induction would be a proper tool to resolve the appearance of doubt around these points. These functions are not meant to “exhaust all the possible roles rhetoric plays in science”, but according to Pera they are “sufficient to conclude that this role is not merely ornamental” (p. 102).
(2) Interpreting a methodological rule (establishing the exact prescriptive content of a rule)
(3) Deciding whether to apply a rule to a concrete case.
(4) Justifying a starting point
(5) Attributing to a hypothesis a positive degree of plausibility
(6) Criticizing/discrediting rival hypotheses
(7) Rejecting objections against a hypothesis
All rhetorical arguments aim at “convincing an audience”, that is, at obtaining consensus for a certain claim – be it the plausibility of a hypothesis, the intellectual and pragmatic advantages of a research program, the explanatory merits of a theory or something else. But what kind of logic governs these arguments? First, Pera dismisses logicism from a classical informal-logic-style vantage point. A few examples are offered where deductive and inductive logic – as systems designed to define the rationality of derivation and generalizations, respectively – prove inadequate for the assessment of scientific arguments (or arguments, generally, for that matter). The next step goes as follows:
With dialectics the situation is different. Since it aims at establishing whether arguments are good or bad in specific situations for specific audiences, it must deal not with arguments in themselves but with arguments in a debate. An argument may be valid or correct when taken out of context but bad when considered in a debate; conversely, it may be invalid and incorrect when taken out of context but good when considered in a debate. The fact is that, as part of a debate, an argument is submitted to certain constraints or rules governing the debate and establishing which moves are prohibited or permitted. Dialectics fix such rules. (p. 108)
A simple example is a deductively valid derivation like “p or q, not p, therefore q” in a debate where an alternative r has been put forward for consideration. In that case, the proponent would be denying a starting point – or, in some sense, refusing a reaspnable collaboration.
Since Pera is committed a contextualist view of argument assessment (“only the context can provide the necessary information”, p. 109), he draws the conclusion that it is not only their special function that connects rhetorical argumentation with dialectics (instead of, say, formal logic) but it is also their field. This Toulmin-like model implies that the concrete, real-life situation in which an argument in put forward is consequential for the argument’s appraisal. To simplify, we would say that an argument put forward by Scientist A in a discussion with Scientist B will be assessed based on the (1) minor premise which is Scientist B concedes to Scientist A, and (2) major premise which is provided by the context. Pera’s attention focuses on (2); in fact, on certain categories of bridging premises – resembling Aristotle’s loci – which can be abstracted away from their varying content. The question is, then, whether there is a useful typology of bridging premises as they are used in scientific debates.
Here, Pera makes use of a distinction introduced by Perelman – and nowadays very much in use in pragma-dialectics, (Van Eemeren & Grootendorst, 2004) – between substantive factors and procedural factors. The former are the bridging premises we are looking for, the latter are “the rules that govern debates”. In scientific debates, the most common substantive factors can be grouped into these categories: (1) facts, (2) theories, (3) assumptions, (4) values, (5) commonplaces of preference, (6) presumptions. These categories should be understood as blank-fillers for “appeal t0 …” rhetorical argumentation. When attempting to convince, the scientist will appeal to (1)-(6) – of course, to the extent that he orients his argumentation to the context at hand, i.e. the scientific context. A scientist’s non-scientific argumentation will be outside the boundaries of (1)-(6), and can still be reasonable, but the analyst would observe that it is not attuned to the context in which it occurs.
On the basis of these notions, the main normative framework is introduced. A scientific argument is evaluated according to the following parameters (pp. 118-121):
Pertinence: “A scientific argument is pertinent if the reasons supporting its conclusion belong to the substantive factors of scientific dialectics admitted in that field and for that function”Pera gives examples of procedural rules that enter within the same framework, rules such as “The debate is adjudicated in A’s favour if B does not offer reasons in support of his thesis belonging to the admitted substantive basis” (p. 124). The examples, however, are admittedly vague and unpolished. In any case, within this framework, Pera commits himself to a powerful analytical idea, namely, that every instance of scientific argumentation is following the strategy of confutation: “finding one or more concessions made by the interlocutor which, united with a shared substantive factor that acts as a bridge-premise, leads to the negation of that thesis” (p. 123). Interesting enough, this means that every scientific debate comprises mutual attempts at using the right modus tollens.
Validity: “A scientific argument is valid if in favour of its conclusion a winning dialectic strategy exists on the basis of the substantive factors of scientific dialectics”
Strength: “A scientific argument is strong if in favour of its conclusion a winning dialectical strategy exists on the basis of both the premises conceded in the dialectical situation and the configuration of the substantive factors of scientific dialectics in force in that situation”
Efficiency: “A scientific argument is efficient for an interlocutor if the reasons adduced in support of its conclusion belong to the configuration of substantive factors of scientific dialectics that the interlocutor considers optimal in that situation”
Conclusion: Science within a dialectical model
The shift of perspective, from the traditional methodological picture, is total and unreserved: “Bringing dialectics into science is not just a matter of making small adjustments here and there; the founding fathers’ very image of science is irrevocably altered” (p. 131). What the Fathers had in mind can be represented as a game with two players: the Inquiring mind (I) and Nature (N). (I) asks (N) questions and (N) answers, forced to reveal its secrets. Method (M) is not a player, but rather the “big brother” of the whole enterprise. Its set of rules watch over this conversation between (I) and (N).
In the new perspective, science becomes a game with three players: an individual or group from the scientific community (C1), nature (N) and another group from the scientific community (C2). “In this dialectical model, nature reacts and scientist agree upon its correct answer throughout a debate based on the factors of scientific dialectics.
As a philosopher, Pera is steadily paving the work for further dialectical interest in scientific (or, generally, academic) communication. Nevertheless, as a rhetorician (or linguist) his analyses of instances of scientific argumentation, as well as the normative suggestions he offers, are rather rudimentary. He shows that the dialectical perspective “provides us with an image of scientific practice that is perhaps less sever than that of the methodological, and less elastic than that of the counter-methodological model, but more realistic than both” (p. 136).
Feyerabend, P. (1975). Against method. London: New Left Books. Kuhn, T. S. (1996). The structure of scientific revolutions. Chicago: University of Chicago Press. Van Eemeren, F. H., & Grootendorst, R. (2004). A systematic theory of argumentation: The pragma-dialectical approach. Cambridge: Cambridge University Press.
 Note the empirical dimension! Pera writes: “With regard to these factors, the attitude I adopt is nonprescriptive [his emphasis]. As much as his personal proposals may be appreciated, the philosopher of science is not free to construct systems or models artificially, because he is constrained by the history and practice of science” (p. 112)