In this paper, I wish to present and defend the thesis that the impasse at which the philosophy and history of science find themselves in the last couple of decades is due, to a large extent, either to the complete neglect or to a misguided treatment of t
he role of scientific controversies in the evolution of science.
In order to do so, I will first provide a preliminary clarification of the impasse to which I refer (section 1). Then, I will explain why I see the study of controversies as a fundamental step in solving it (section 2), I will locate controversies within the set of empirical phenomena of the class of `polemical discourses' (section 3), and I will single out the properties of controversies which explain their potential role for solving the impasse (section 4). I will then show how the extant epistemologic al options are unable to handle controversies in a satisfactory form, which explains their inability to solve the impasse (section 5). I will then formulate an essential desideratum for the solution of the impasse (section 6), and will suggest how it can be satisfied through the use of pragmatics both as an instrument and as a model (section 7).
What seemed to me, when I began writing, a relatively short and schematic article, has grown larger and larger, despite my efforts to be concise. In spite of its size, there are some affirmations here which may seem dogmatic because they are insufficient ly argued, as well as some suggestions that seem merely programmatic. I take upon myself the responsibility for providing better justifications, as well as for developing the program here sketched in a larger essay in preparation.
I take as a starting point the more or less universally admitted failure of logical positivism qua philosophy of science. According to its critics, this philosophy has failed both at the normative and at the descriptive levels.
On the normative level, because its main notion, that of `justification' of a theory, which is essentially inductivist and based on the notion of `confirmation', is untenable. According to Popper, the main critic of the normative component of the positiv ist model, it is not possible to build an inductive logic which is capable of satisfactorily formalizing the idea that it is possible to attribute to each theory a `degree of confirmation', defined in terms of its conditional probability relative to the a vailable evidence. If one insists on using such a notion, the result is that all theories will have, at best, the same degree of confirmation, namely zero. Hence, this notion does not permit us to formulate a criterion for the acceptance or abandonment of a theory. The alternative normative criterion proposed by Popper is based -- as is well-known -- on the refutability of theories, rather than on their confirmability. Every scientific theory should be considered as a refutable conjecture. Those theories that have survived `serious' or `severe' efforts of refutation are `accepted'. Popper's criterion contains, therefore not only a principle for evaluating theories, but also a normative recommendation for the scientists to act upon: they are advised, on th e one hand, to produce refutable conjectures and, on the other, to try to refute conjectures (theirs or others'); furthermore, they should be ready to abandon a theory, however cherished, as soon as it is refuted.
On the descriptive level, the critique of positivism has consisted mainly in indicating that the `growth of (scientific) knowledge' does not occur through progressive incrementation of the degree of confirmation of the proposed theories, as the positivis t model suggests. Critics like Kuhn, from the vantage point of the history of science, argue that, at best, such a conception of scientific progress applies to certain relatively homogeneous periods of `normal science'. But such periods are abruptly inter rupted by `crises' which engender radical ruptures in the evolution of a science, which in turn do not permit even the measurement of the progress made from one period of normal science to another. The growth of knowledge, if sense can be made of this phr ase, occurs not linearly or incrementally, but rather in zig-zag. Furthermore, it is characterized by an `incommensurability' between its different phases, which affects not only the theories proposed in each phase, but also the methodological norms preva iling in each of them.
More radical critics, such as Feyerabend, argue that, even in `normal' periods, scientists are in fact `methodological opportunists', who make use of all available methods in order to develop and promote their theories, without swearing faithfulness to a ny one of them. In this way, not only is the positivistic norm called into question, but also any normative methodology. The outcome of this type of critique is the denial of the possibility of universal criteria of legitimization of scientific knowledge, i.e., of criteria based exclusively on principles of rationality. Ultimately, what is at stake is the very project of formulating an `epistemology' in the sense of a theory capable of justifying the valid modes of knowledge, rather than merely describing the modes considered valid by this or that community, culture, or historical period.
The two faces of the onslaught on positivism -- the normative and the descriptive -- are intimately related. The arguments of the descriptivist and normativist critics make this clear: the former point out the uselessness of norms nobody follows as well as the historical and social variability of these norms; the latter, the necessity of adopting norms that bear some descriptive adequacy, without falling into the relativism of pure descriptivism, which would make it impossible to explain the rationality of scientific progress. Consequently, in post-positivist philosophy of science, where "the descriptive and the normative are inextricably mixed" (Kuhn 1970: 233), the positions no longer range themselves along the categories normative vs. descriptive, but rather according to the way in which these two aspects are combined. When Lakatos (1970: 177) describes his differences with Kuhn saying that "Kuhn's conceptual framework for dealing with continuity in science is sociopsychological: mine is normative" (t hereby provoking vehement protests from Kuhn), no doubt he exaggerates the polarization of the two positions for rhetorical purposes. After all, he himself stresses, in the next sentence, the word `also': "Where Kuhn sees `paradigms', I also see rational `research programmes'" (ibid.). He also stresses the need to account for the interaction between the three Popperian `worlds' (Lakatos 1970: 180), and in another text he emphasizes the need of explaining "how the historiography of science should learn fro m the philosophy of science and vice-versa" (Lakatos 1981: 107).
The problem then, is not to choose between normativism and descriptivism, but to find an adequate way of combining them, in the light of the criticism they address to each other. The impasse in the philosophy and history of science results from the fact that the proposed combinations have so far tended to subordinate one of these components to the other, without really attempting to harmonize them.
The way proposed by Lakatos in his "methodology of scientific research programmes" tries to take into account the diachronic dimension, especially the fact that in most cases theories withstand the `instant death' their empirical refutation ought to brin g about (according to the position Lakatos dubs `naive falsificationism'). His main concern is to justify rationally this fact. That is to say, to show how the frequent violation of what seems to be a fundamental norm of scientific activity can be underst ood in terms of a broader and deeper notion of rationality. His solution consists in taking as units of rational assessment not isolated theories but series of theories -- `research programmes' -- sharing a "core" of guiding principles and theses. Histori city thus becomes essential for the application of the norm of rationality. Nevertheless, in order to fulfill this function, it must be -- according to Lakatos -- an `internal' historicity, grounded in the `rational reconstruction' of the series of theori es, whence `external', allegedly irrelevant factors are excluded a priori. The normative eye of `philosophy' is responsible for determining the selection of what is relevant, while the descriptive eye of history (or of sociology, psychology, etc.) has at best an auxiliary function. No wonder, therefore, that for Lakatos the drama of science's evolution occurs not in real history, but rather in World Three, where normativity reigns, and whose characters are not the knowing subjects but `articulate knowledg e', independent of them.  In my opinion, Lakatos never satisfies his own demand for elaborating a theory of the growth of knowledge that is really able to integrate the three worlds.
Nor does descriptivism, in its known versions. The ambivalence between prescription and description detected by Feyerabend (1970: 198) in Kuhn's work simply confirms that here too there is a concern for combining the two components. And Kuhn (1970: 237) i s certainly right when he replies that his work tends to be read both ways simultaneously. But his argument to this effect -- which he considers simple and unobjectionable -- clearly reveals the priority he assigns to the descriptive and the consequent `w eak' normativity he allows for:
[a] Scientists act in certain ways.
[b] These ways of acting fulfill certain essential functions.
[c] There is no other way of acting that fulfills the same functions.
[d] Scientists wish to improve scientific knowledge.
[e] Scientists must act as they do.
No doubt there is in this argument a deontic premise [d] which formally permits, without committing the so-called `naturalist fallacy', the drawing of the normative conclusion [e]. However, it is a `weak' normativity, applicable only under condition [c]. But this condition is satisfied only within a `paradigm': scientists trained within it are simply unable to discern other `scientific' ways of acting. But, according to Kuhn himself, a shift of paradigm involves, among other things, a change in the ways o f acting that scientists deem conducive to the "flourishing of their enterprise". That is to say, with the change in paradigm other ways of acting capable of fulfilling the essential `functions' of science necessarily arise. In other w ords, the phrase `must act as they do' acquires different contents with each shift of paradigm, without there being any possibility of justifying or criticizing this norm other than `internally'. The normative conclusion [e] is weak because, since conditi on [c] is either trivially true (within a paradigm) or else false (transparadigmatically), we don't have any criterion that really permits us to compare different ways of acting; thus, there is no way to attribute to them a stronger normativity, that is, a normativity grounded in criteria which apply, although not universally, at least beyond the limits of each period of normal science.
The strict separation Kuhn makes between periods of normal science where a rigorous normativity reigns and of extraordinary science where normativity is practically suspended and "anything goes" is symptomatic of the difficulty this project faces in harm onizing the two components. More radical versions of descriptivism -- be they historiographic, sociological, or psychological -- tend in turn to renounce completely the normative component, resting content with offering descriptions or, at most, causal an alyses. In this way, they are led -- paradoxically -- to a sort of positivism (in the sense this word is used in the philosophy of law): the only criterion of validity of norms is their existence as such, and their acceptance and evolution can only be exp lained in terms of external processes or forces that act as causes but not as reasons. 
If extant theories have been unable to harmonize a normative rationality with the description of the facts of scientific praxis, it is because they have not been able to bring them closer to each other. Both descriptivism and normativism have accepted wi thout question the opposition and consequent separation between a regulative or formal rationality and a descriptivity which, although often declaring itself reconstructive, aims at capturing the facts so to speak in their raw state, i.e., prior to any no rmativity. This mutual alienation is what prevents the finding of a point of contact between them. 
What is needed in order to overcome this impasse is to go beyond the opposition on which it depends. On the one hand, one must bring rationality down from its regulative heights, making it constitutive of content, sensitive to the context of its applicat ion, even if this is accompanied by the danger of its `contamination' by contingency and the consequent abandonment of its universalist pretensions. On the other hand, it must be acknowledged that scientific praxis, qua human activity guided by reasons, c annot be described in terms entirely alien to the constitutive role of these reasons. In my opinion, what has to be demonstrated is the possibility of this rapprochement between norm and description, between rationality and `reality'. I believe that the p oint of convergence par excellence of this encounter is scientific controversy. Therefore, what has to be done is to try to understand why scientific controversy is impossible according to the extant epistemological theories, and to try to develop an argu ment demonstrating its possibility. Before doing so, I must explain why the role controversies play is so crucial in my view and what my conception of a controversy is.
THESIS A: Controversies are indispensable for the formation, evolution and evaluation of (scientific) theories, because it is through them that `serious' criticism -- i.e., the kind that allows for engendering, improving and controlling both th
e `well-formedness' and the `empirical content' of scientific theories -- is performed.
One can immediately note in this thesis what it shares with Popperian normativism, namely the importance I assign to criticism. I think this starting point is reasonable and I won't try to defend it here. Starting here, what is essential is to try to unve il the conditions of possibility of criticism. As opposed to Popper, who considers these conditions as belonging exclusively to the logical plane, for me it is indispensable to take into account also those conditions that permit critical activity, exercis ed mainly in controversies. In the latter it is not observations vs. conjectures, nor theories vs. theories, nor arguments vs. arguments, nor even research programmes vs. research programmes that are in opposition, but rather scientists who perform experi ments and observations, formulate theories and arguments, and persevere in their research programmes or abandon them -- all of them vs. scientists who oppose them by performing the same kinds of activity. Concepts such as `seriousness' (Popper) or `tenaci ty' (Lakatos), which are essential for criticism to fulfill its function according to the normativists, apply first and foremost to the activities of scientists and only derivatively to their results. These human activities take place in World Two, rather than in World Three. It is only by virtue of an unjustified assumption, which creates an insurmountable gap between these two worlds, that one is led to believe that all that which in critical activity does not lend itself to a direct `translation' into components of World Three is (a) irrelevant for explaining the scientificity of science, (b) corresponds to `irrational' or `nonepistemic' factors, and therefore (c) has nothing to do with the explanation of the growth of scientific knowledge. But none of these can be a priori affirmed. If criticism is so essential -- and I agree with Popper that it is -- it is necessary to study the set of its conditions of possibility as they manifest themselves in its `natural place', i.e., scientific controversies, i n order to be able to decide which of these conditions are relevant for explaining the `rationality' of science.
Thesis B: The rigorous study of controversies is an indispensable means for providing an adequate description of the history and praxis of science. For controversies are the natural `dialogical context' where theories are elaborated and where thei r meaning progressively crystallizes. Furthermore, the study of controversies will permit us to determine empirically, on the one hand the precise nature of those `crises' and `ruptures' that allegedly introduce an element of irrationality in the evolutio n of science, and on the other, wherein lies the `continuity' which allegedly is required as the background for conceptual `change' and `innovation'.
In this thesis one can note what is shared with descriptivism, namely the importance I attribute to the careful description of the historical process of scientific evolution, where controversies play a decisive role. As opposed to Kuhn and other descripti vists, however, I stress the fact that not all `context' has the same weight. The dialogical context of controversy is the `direct' context of constitution of the meaning of a theory and it is relative to it that other aspects of context acquire their rel evance. Furthermore, I believe that the touchstone for understanding the alleged differences between normal and extraordinary science consists precisely in finding out whether there are fundamental differences between the controversies that occur in each of them. This would enable us to say something significant about the existence and nature of phenomenona such as incommensurability, radical changes in methodological norms, etc. 
In short: science manifests itself in its history as a sequence of controversies; these are, therefore, not anomalies but the `natural state' of science; controversies are the locus where critical activity is exercised, where the meaning of theories is d ialogically shaped, where changes and innovations arise, and where the rationality or irrationality of the scientific enterprise manifests itself; for all these reasons, to ignore them in the philosophy and history of science is a capital mistake which mu st be corrected.
But what are -- the impatient reader will justly ask -- scientific controversies? In the present section I will try to give an answer, locating controversy properly so called, and in particular scientific controversy, within the family of phenomena to
which it belongs.
This family is, evidently, that of discursive dialogical polemical phenomenona. There is no controversy, strictly speaking, without there being at least two persons who employ language to address each other, in a confrontation of opinions, arguments, the ories, and so forth. In this respect, the definition of controversy provided by Leibniz (VOR, p. 18) in one of his fragments on the subject -- "controversia est quaestio circa quam opiniones contrariae habentur" [a controversy is a question over which con trary opinions are held] -- is clearly insufficient (except perhaps for Popper), since it reduces controversy to its logico-semantical aspect. I also propose to exclude from the extension of the term `controversy' (although not necessarily from the field of polemical discourse) what is called by the German tradition `critical reception' -- for instance, Popper's diatribes against Marx, Hegel, and Plato, in his The Open Society and its Enemies. Similarly, I would exclude prefabricated `dialogues', such as the one between Leibniz and Locke in the New Essays.
The reason for excluding these cases has to do with my insistence on viewing controversy as an activity, and -- more importantly -- an activity that always contains an element of imprevisibility. In controversy it is essential to preserve the possible an d the actual use of the right to contest the opponent's views granted to each of the disputants. For a live, real, and active (i.e., neither dead, nor imaginary, nor silent) opponent is unpredictable in his/her reactions. Although we may anticipate to som e extent our opponent's reactions, and even undertake to manoeuver her to react in a certain way, the controversy game is essentially a `strategical game', that is a game where our capacity for predicting the adversary's move is limited. This fact is responsible to a great extent for the controversy's capacity to call attention to confusions and produce clarifications, to force conceptual, methodological or theoretical changes, and thereby to lead to innovations.
Another reason for not including critical reception, prefabricated dialogue, and the like among controversies is the fact that they are not really dialogical. To be sure, every text is, at least implicitly, addressed to some audience and designed accord ingly (cf. Clark 1992: xviii). But in dialogue the audience is very explicitly defined, the roles of speaker and hearer alternate with high frequency, and the `conversational demands' (cf. Dascal 1977) shift progressively with each intervention by each pa rticipant. These properties of dialogue in general, which do not apply only to oral debates but also to written controversy (cf. Dascal 1989a), permit us to understand some of its aspects: the variable or ad hoc nature of the arguments used in it (dependi ng on the opponent, the stage of the controversy, etc.), the dynamic rather than static character of the definition of the `problem situation' in the course of the controversy, the thematic shifts it undergoes, etc.
Conceived of as a dialogical phenomenon, controversy thus primarily consists in those texts or utterances directly addressed by each disputant to the other (or others), privately or publicly. In addition to this `primary text', there is in general a vast `secondary text' which, at least partially, belongs to the controversy. It includes, for instance, works by the disputants where the controversy is reflected directly or indirectly, as well as letters to third parties where allusion is made to the contro versy. A broader circle of texts that are pertinent for the controversy form its `co-text' which includes, for example, works by prior or contemporary authors quoted and relied upon by both disputants. Finally, every dialogue unfolds within a nondiscursiv e `context', whose various aspects and levels have always a more or less important role in the content and development of controversies.
All that was just said applies to the whole family of polemical dialogues. We must now try to specify controversy's position within this microfamily which includes, among its many members, verbal quarrels between couples, political debates, round tables in scientific congresses, critical reviews of books and replies to them, medieval disputationes, etc. Several classifications, more or less ad hoc, have been proposed, which I will not be able to discuss here. I will offer a taxonomy which has proved useful in the course of my work on controversies.
Let us reserve the term `polemics' for denoting the set of discursive dialogical polemical phenomena, whose general characteristics have been described above. Within this set, I propose to distinguish between three `ideal types', members of the subfamily to which controversies belong. I will call them `discussion', `dispute', and `controversy'.
A discussion is a polemics whose object is a well-circumscribed topic or problem. As the discussion develops, the contenders tend to acknowledge that the root of the problem is a mistake relating to some important concept or procedure within a well-defin ed field (even though they disagree regarding the nature of the mistake in question and about who commits it). Discussions allow for solutions, which consist in correcting the mistake thanks to the application of procedures accepted in the field (e.g., pr oof, computation, repetition of experiments, etc.).
A dispute is a polemics which also seems to have as its object a well-defined divergence. But at no point do the contenders accept its definition as grounded in some mistake. Rather, it is rooted in differences of attitude, feelings, or preferences. Ther e are no mutually accepted procedures for deciding the dispute, that is, a dispute has no single `solution'. At most they can be dissolved. But in general the underlying divergences tend to recur in disputes over other topics. Some c ontenders see in the position of their opponents symptoms of an illness against which the only reasonable action to take is therapeutic.
A controversy is a polemics that occupies an intermediate position between discussion and dispute. It can begin with a specific problem, but it spreads quickly to other problems and reveals profound divergences. These involve both opposed attitudes and p references and disagreements about the extant methods for problem solving. For this reason, the opposition is not perceived simply as a matter of mistakes, nor are there accepted procedures for deciding them -- which causes the continuation of controversi es and sometimes their recurrence. However, they do not reduce to mere unsolvable conflicts of preferences. The contenders pile up arguments they believe increase the weight of their positions vis a vis the adversaries' objections, thereby leading, if not to deciding the matter in question, at least to tilting the `balance of reason' in their favor. Controversies are neither `solved' nor `dissolved'; they are resolved. Their resolution may consist in the acknowledgment (by the conten ders or by their community of reference) that enough weight has been accumulated in favor of one of their positions, or in the emergence (thanks to the controversy) of modified positions acceptable to the contenders, or simply in the mutual clarification of the nature of the differences at stake.
In real polemics, elements of the three ideal types are usually mixed up. But in general, under analysis, it is possible to identify the dominant type. Also in polemics between scientists, the three types are mixed and it is not easy to separate them. A simplistic solution would be to employ the Popperian technique of demarcation. Just as he distinguishes (within science) between `ideological revolutions' and properly `scientific revolutions' (Popper 1981), demarcation would consist of assimilating scien tific polemics with discussions, while considering disputes as being merely and clearly `ideological' (and therefore without interest for science). At the opposite pole, a Kuhnian would say that discussions are typically intra-paradi gmatic and thus characteristic of periods of normal science, whereas in the periods of extraordinary science the polemics are essentially disputes. Neither demarcation leaves room for controversies (which they tend to assimilate either with discussions or with disputes), and therefore they are unable to recognize their essential characteristics and fundamental role. This difficulty is not merely circumstantial but results from certain basic assumptions, which will be discussed in section 5. Before that, I will go into some more detail in describing some of the basic characteristics of controversies.
I. The first important property of controversies is that they do not remain confined to the initial questions that spark them. They tend to spread rapidly, both in extension and in depth.
In one of the cycles of the Malthus vs. Ricardo controversy, for example, the initial problem is that of `corn laws', which restrict the import of corn to England. However, this problem quickly recedes into the background, and the correspondence between the two economists focuses on other topics, such as the relations between profits and interest rates, between salaries and the true price of labor, between effective demand and the consumption of luxuries, as well as on methodological issues, such as the distinction between permanent laws and temporary tendencies, the nature of scientific language, the question of whether economic theory should be monocausal or multicausal, etc.  It is important to note that all these new topics are introduced in the debate without explicit digression markers (cf. Dascal and Katriel 1979) -- which shows that the problem shifts in question are perceived by the participants as directly relevant to the initial topic, now "broadly conceived". That is to say, the thematic evolution of the controversy is ruled, at least partially, by intrinsic conditions of semantic and pragmatic relevance (cf. Dascal 1977).
In the course of a controversy, the initial problem is often completely set aside -- so much so that the controversy may come to an end with the adoption of one of the positions in conflict, even if it is unable to solve the initial problem. In the contr oversy about continental drift, one of the initial problems that Wegener's theory (about the early union of the continents and their subsequent separation) purported to solve was to account for the observed similarities between life forms existing on cont inents that are presently separated. The alternative theories, which deny continental drift, explain this phenomenon by means of ad hoc hypotheses (e.g., the existence of early `intercontinental bridges', whose later disappearance they are unable to expla in). Sixty years after Wegener's theory was first proposed, the controversy ends with the almost universal adoption of the drift theory (albeit in a substantially modified form). But the decisive evidence that supports it comes from another discipline (pa leomagnetism), and the initial paleobiogeographical problem, to which Wegener had attributed decisive importance, remains unsolved.
II. The expansion of the problematics includes the questioning by the contenders of their adversaries' basic factual, methodological, and conceptual presuppositions. Thus, in his effort to undermine Malthus's theory of value, Ricardo simply denies a fact affirmed by Malthus: that "repeated experience" demonstrates that "the money price of labour never falls till many workmen have been for some time out of work" (Ricardo to Malthus, 21 July 1821; WC, IX: 25). The rejection of this alleged fact (rec all that at the time there was no reliable statistical information), in its turn, exemplifies Ricardo's general methodological attitude, according to which it is more important that a hypothesis be supported by "general reasoning" than that it conform to the facts (ibid.). Analogously, in the above mentioned controversy about continental drift, Simpson attacks his opponents for using incorrect data based on the mistaken identification of fossils, for underestimating the capacity of dispersion of several o rganisms, for incorrectly summing up data from other research sources, and for not considering the possibility of parallel or convergent evolution -- i.e., he questions factual, methodological, and conceptual assumptions, respectively. Another example is the expansion and deepening of the controversy between Geoffroy de Saint-Hilaire and Cuvier about the existence of a uniform `plan of organization' underlying the anatomy of all animal species. As pointed out by Gil (1985), this apparently well-defined an d `technical' problem leads to a brief but intense (and even dramatic) controversy, where "a whole series of fundamental epistemological decisions was at stake" (Gil 1985: 358). Whereas Cuvier favors multiplicity, diversity, and discontinuity, demands str ict fidelity to data, requires the use of a precise language, and imposes severe constraints on the use of analogy, Saint-Hilaire adopts the opposite choices: he seeks a unity underlying multiplicity, criticizes empiricism, does not worry about rigorous d efinitions, and employs analogy rather freely.
It is worth noticing the remarkable similarity between the oppositions Cuvier vs. Saint-Hillaire and Malthus vs. Ricardo (for the latter, see Cremaschi and Dascal 1996). While Malthus favored an anti-deductivist and anti-apriorist methodology and sought an `exact' adherence to reality, Ricardo had no qualms in sacrificing such an adherence in favor of the simplicity of universal and `rigorous' laws. In Granger's (1959: 103) terms, one valued above all the spirit of exactness; the other, the spirit of rig or. Both, however, professed fidelity to the `same' scientific paradigm, inspired by Newton -- which shows that even within a paradigm there is room for radically different interpretations (see below, paragraph III). The remarkable coincidence between the methodological oppositions underlying these two controversies suggests that, behind their conscious options, scientists oppose each other also by virtue of different `casts of mind', i.e. different ways of conceiving the very aim of scientific activity, and that such differences have a lasting impact on their actual exercise of such an activity.
III. A third important characteristic of controversies is their hermeneutical aspect. The question of the correct interpretation of data, of language, of theories, of methods, and of the status quaestionis, arises again and again throughout a cont roversy. At each stage the contenders accuse each other of misrepresenting their theses, of using ambiguous language, of not responding to the objections and -- for that reason -- of not addressing themselves to the `true problem' at stake (either at a gi ven stage or in the controversy as a whole). There are endless examples of these kinds of claims in all types of controversy, and I will not illustrate them here (for some examples, see Dascal 1990a).
An interesting question is the reason or function these (alleged or real) `misunderstandings' may have in the controversy. I have ventured the hypothesis (cf. Dascal 1990a) that they are related to the fact that, when they engage in a controversy, the co ntenders take upon themselves a double commitment: to defend their own theories and to criticize those of the adversary. Now, a relatively easy (and current) way of fulfilling these commitments is to invoke (qua defendant) and to use (qua opponent) `disto rtion': I evade my opponent's criticism alleging that it does not really apply to my theory, but to the incorrect version s/he gave of it; I attack more efficiently a `reconstruction' (usually simplistic and sometimes caricatural) I myself propose of my o pponent's theory, rather than the theory `itself', in all its complexity. Notice that this hypothesis is not psychological nor `political', since I am not accusing the contenders of ill faith. I am suggesting that they employ the `strategy of misunderstan ding' by force of the objective obligations they undertake once they engage in a controversy.
IV. The dynamic character of the problematic, the ongoing questioning of the presuppositions, and the `hermeneutic freedom' the contenders grant themselves leads to the fourth -- perhaps the most important -- characteristic of scientific controver sies: their `openness'. What I want to express by this term is this: (a) when we begin a controversy, we do not know where its inherent dynamics will lead us; (b) controversies are rarely confined to a single discipline and, within a discipline, to a well -defined topic; (c) they reveal the existence of deep differences regarding the meaning of concepts, methods, and facts so far accepted without dispute; (d) it is not possible to anticipate all the objections of the opponent; (e) they clear the way for th e emergence of radical innovation -- one might even say that they invite the appearance of `non conventional' ideas, methods, techniques, and interpretations.
V. A special aspect of the openness of controversies, which it is convenient to single out as their fifth characteristic, has to do with their closure -- a topic that has been the focus of the volume compiled by Engelhardt Jr. and Caplan (1987). I n this volume, Beauchamp (1987: 28-35) and McMullin (1987: 77-82) propose two similar typologies of `closures'. Leaving aside `natural death' (Beauchamp) and `abandonment' (McMullin), one can clearly recognize in the two typologies t he types of closure I have indicated as typical of discussions and disputes. `Resolution' (McMullin) and `sound argument closure' (Beauchamp) are instances of what I have called the `solution' (of a discussion), whereas `closure' (McMullin) and `procedura l closure' (Beauchamp) are typical of disputes, which cannot be terminated without the intervention of some umpire endowed with the required authority, even if this authority at best `dissolves' the opposition for practical purposes, without in fact resol ving the conflict of attitudes.
Curiously, what these two typologies lack are the intermediate types of termination, more open than the algorithms that solve discussions, but not arbitrary as the means for dissolving disputes are. Beauchamp's `negotiation' and `consensus' might exempli fy these intermediate means, in so far as their decision power is contingent (hence provisional), depends upon the contenders themselves, and is acknowledged as such by them. No doubt these forms of `closing without closure' correspond to the typical open ness of controversies. Nevertheless, they stress excessively the element of `agreement'. Controversies in fact can terminate without agreement, without thereby becoming disputes. What is interesting is that even in such cases they are perceived by the par ticipants and analysts alike as `productive'. That is, although they do not permit even a tilting of the balance in favor of one of the positions in conflict, they make a cognitive or epistemic contribution: they clarify the problem, they permit one to id entify conceptual or methodological divergencies or difficulties, they reorient the research effort, or they simply produce `understanding'. As a whole these contributions are decisive for clearing the ground for radical innovation -- often achieved in te rms of an alternative quite different from the positions held by the adversaries in the controversy itself.
VI. The openness of controversies does not mean they are anarchical. Although they are not governed by implicitly or explicitly codified rules, as discussions are, they do not fall in the `anything goes' extreme. Although they are not decided by a `tribunal of Reason' (Kant) nor by an impartial `judge of controversies' (as the one often dreamt of by Leibniz), they are not terminated by invectives, censorship, or fist fights. Although their thematic evolution and problem shift s may be influenced by external events (discoveries in other disciplines, new technologies, pressing practical needs, etc.), they also obey relations of relevance inherent to the contents of the theories in conflict. In short, scientific controversies man ifest some sort of order or systematicity, which is sufficiently weak so as not to deprive them of their essential openness, and yet sufficient for ensuring that their development is not entirely arbitrary. In this kind of order, I believe, a special form of rationality is operative -- a rationality which it is up to the theory of controversies to elucidate.
I will now connect the impasse described in the first section with what has been said about controversies so far. My hypothesis is that the impasse results from the fact that the main extant positions in the philosophy and history of science are unable
to acknowledge and explain the role of controversies in the formation, evolution, and evaluation of scientific theories. That is, I will defend the following thesis:
THESIS C: Neither the known forms of normativism, nor the known forms of descriptivism permit the existence of scientific controversies displaying their role and characteristics as described in sections 2-4.
In order to complete my argument, it would be necessary to defend also the following thesis:
THESIS C': So far, no alternatives to normativism and descriptivism, capable of accounting for the role and characteristics of controversies, have been proposed.
Nevertheless, I will not defend C' here, except by way of some marginal comments on some of these `alternatives', because (a) I do not pretend to be knowledgeable about the vast recent literature on the subject and (b) it doesn't seem to me -- by the exam ples I know of -- that the authors who propose such alternatives have recognized the basic difficulty I identify as the deep cause of the impasse (namely, the fact that the importance of controversies has been overlooked), so that they have not even attem pted to overcome it.
My argumentation in support of C will consist in showing how the more or less `classical' descriptivist and normativist positions exclude the possibility of the existence of controversies (in my sense of the term) because they do not permit the re alization of one or more of their essential properties. This exclusion leads to a polarization that reduces the field of scientific polemics to only two types -- discussions and disputes -- with their respective modes of treatment. I will also argue that one can find in Kant an anticipation of this situation. In the next section I will argue that the solution of the impasse requires the demonstration of the possibility of transcending such a polarization.
I. The predominantly descriptivist positions, in their many variants, stress the variety and variability of factors that intervene in scientific polemics. For them, the distinction between `epistemic' and `non-epistemic' factors is contingent: dif ferent disciplines or `communities of experts', at different historical moments, draw such a distinction in different ways. Hence, in order to have epistemic value, critical activity must conform to the way in which a certain community establishes this va lue. This means that, necessarily, only controversies that are `internal' to the community can be said to have epistemic value. That is to say, the space of, say, `legitimate' controversy is pre-determined by a set of contingently valid norms given for ea ch `debate community'. To attempt to engage in a trans-communitary debate is like attempting to play two different `language games' applying the same rules. The rules defining each `epistemic game' cannot explain such an endeavor; in order to account for it -- when it occurs -- it is necessary to appeal to non epistemic factors. In other words, trans-communitary polemics is either absurd and produces only equivocation, or else -- if it has any value -- such a value can only be non-epistemic, unless, contr ary to descriptivism's fundamental presupposition, one admits the existence of supra-communitary epistemic norms.
Therefore, the internal logic of descriptivism, forces us to conceive of controversies either as `well-behaved' debates conforming to the intra-communitary norms that ensure their epistemic value, or else as conflicts governed only by non-epistemic facto rs. In the former case, they are reduced to what I have called discussions; in the latter, to disputes. No room is left for controversies properly speaking. For no room is left for radical innovation -- the kind of innovation that, precisely because it is grounded on deep criticism which includes the contestation and revision of the contingently given demarcations between `epistemic' and `non-epistemic', cannot be included in either category.
II. Another descriptivist tendency, which prevents it from being able to account for controversies, is its `contextualism'. In so far as it tends to explain rather than just describe the norms in force within a scientific community, descriptivism does so in terms of external causes (social, political, historical, psychological, etc.). This means that, ultimately, the true explanation of each of the aspects of scientific activity must be given in terms of one or another type of preferred cause. In this way, any `autonomy' -- even relative -- is denied to such phenomena as the thematic evolution of a controversy. These phenomena must result, ultimately, from the influence of external events. The assumption of an internal logic, of a semantic or prag matic relevance, is viewed as entirely spurious -- a mere illusion.
In its extreme versions, this tendency leads to the thesis that every `rational reconstruction' (which not only Lakatos, but also Kuhn, takes to be necessary in the history of science) of such processes in terms of their alleged internal constraints belo ngs to the order of `rationalization' (Freud), `ideology' (Marx), or disguise of what really happens in all discursive processes, namely a fight for power (Foucault). Accordingly, even well-behaved polemics are nothing but moves in a dispute, for their tr ue purpose is to defend a given `truth regime' against its rivals.
The main problem of such a contextualism lies in the difficulty to establish how contents and relations between contents are determined by their alleged causes. One can of course intimate that Marx's hemorrhoids were responsible for his theory of alienat ion (cf. Bartley III 1987: 430-431), but it is difficult to demonstrate it, except by a vague analogy between the two kinds of phenomena. Likewise, generic Machiavellian arguments invoking the cui bono principle only acquire explanatory power if the conne ction between the alleged interests of an agent and the event that favors them is independently established: the fact that a defendant has a motive for the crime is nothing but a piece of accessory evidence for condemning him/her. Nevertheless, contextual ism generally satisfies itself with these kinds of argument, and hardly manages to establish concrete causal relations of the sort required to ground its claims. In my view, the relation between context and content is not unidirectional, but rather bidire ctional. The context no doubt contributes to the determination of content, but the latter also contributes to the determination of the relevant context, i.e. it allows us to face with success the problem of selecting from the infinity of potentially perti nent contextual factors. Such a bi-directional interaction between context and content is essentially open (rather than pre-determined), and constitutes one of the essential conditions for understanding the recurrence of interpretative or hermeneutic issu es in controversies.
III. If descriptivism is wrong in suppressing or minimizing the autonomy of the internal processes responsible for the evolution of science, normativism is wrong in exaggerating it. Nevertheless, the result is -- paradoxically -- the same: a sort of pre-determination or closure of the space of legitimate criticism, which transforms into arbitrary, cognitively irrelevant, or `ideological' any criticism that does not conform to the limitations of such a space.
As I have already pointed out, in his effort to demonstrate the rationality of scientific revolutions, Popper (1981) introduces a "sharp distinction" (p. 99) between "scientific" and "ideological" revolutions. The former obey a logical criterion of progr ess: the new theory must be able to account for all the facts accounted for by the theories it purports to replace. For this reason, such revolutions are `rational'. In the latter, although they are often associated with the former, there is no rational c riterion of progress; they are simply "intellectual fashions" resulting from processes of social acceptance, which are independent of the rational merits of the theories in conflict (p. 100). According to Popper, although both are often enmeshed, one can and must distinguish clearly between them: the Copernican revolution, for instance, was a scientific revolution because it replaced earlier theories satisfying the logical criterion of progress; but it also had an ideological impact due to the fact that i t questioned theological dogmas and radically changed the position of man in the universe.
The Copernican example might suggest that, for Popper, the `ideological' comprises only ethico-political and `metaphysical' aspects, usually accepted as `non-epistemic'. But this is not the case. When he speaks about the Einsteinian revolution, for examp le, he includes in its ideological aspects methodological attitudes ("the myth that Einstein had reached his result by an essential use of epistemological and especially operationalist methods" - p. 105), interpretations of the new theory ("the dominance of a subjectivist interpretation of quantum mechanics" - p. 106) and interpretations of the extension of the conceptual changes it has introduced (e.g., the radical modification of the concepts of space and time). But not only is it not evident that questions such as these are `non-epistemic' or `ideological'; rather, they are precisely the kinds of questions that occur all the time throughout controversies -- where methodologies, interpretations, and opposed meanings are in constant confrontation. Postulating that these are merely `ideological' questions, devoid of cognitive or rational value because they do not obey a pre-determined and strict norm of rationality is simply to limit a priori and arbitrarily their import. Furthermore , it is to assert -- like the descriptivists -- that there are only two possible types of polemics: discussions or disputes.
Popper himself explicitly draws this consequence when he contrasts the "conservative" character of scientific revolutions with the "open" character of ideological revolutions. The former are necessarily conservative, for him, because they must preserve the success of the preceding theories. For this reason, "a scientific revolution, however radical, cannot really break with tradition" (p. 106). The latter, on the contrary, are those that can effectively break with tradition, because they are not constra ined by this kind of requirement. That is to say, in sharp contrast with his well-known `rhetoric of openness', which hails openness to criticism as the hallmark of scientific thought, Popper in fact considerably restricts the scope of such an openness by describing as non-scientific those forms of openness that do not obey his own criterion of scientificity.
IV. In order to do justice to Popper, one should also analyze the `evolutionary' version of his epistemology, developed by himself (1981, 1987) and his collaborators (cf. Radnitzky and Bartley III (eds.), Bartley III 1987, Campbell 1987). For thi s version contains mechanisms that might permit one to account for some of the properties of controversies. For example, problem shifts could be explained as follows: the adoption of a new theory resolves some problems but creates others, since, as the ap pearance of a new biological species, it changes the environment, thus altering the `selective pressures' that impose adaptive needs on the remaining organisms (cf. Popper 1981: 83 and passim). Likewise, one could attribute the phenomenon of `clearing the ground' for the emergence and eventual flourishing of innovations to the fact that the pressure exerted by criticism, qua mechanism of natural selection, eliminates theories that have been refuted, thus permitting the appearance of new ones.
However, evolutionary epistemology, like the other versions of Popperian epistemology, is above all concerned with fighting against anti-realist conceptions. Its main objective is to ensure the objectivity of those theories that survive critical pressure , i.e. their veracity (or at least verisimilitude) vis-a-vis the real world (World One). Evolutionary epistemology believes itself able to achieve this only by establishing a relation of `direct' representativity between World Three and World One, where t he intervention of World Two factors is minimized or altogether eliminated. This is Popper's well-known notion of "knowledge without knowing subjects". According to this view, any intervention of World Two in the exercise of criticism endangers the objec tivity of criticism and must therefore be considered as `not scientific'. Criticism is thus purged from all its human activity aspects and confined to the logico-semantic relations of World Three. Thereby, it acts as a purely passive selective filter, and loses entirely its dialogical character. In short, it has nothing to do with its real manifestation in controversies.
Finally, one must raise in connection with evolutionary epistemology the inevitable tu quoque. Whereas in the biological theory of evolution the natural selection mechanism has the status of an empirical conjecture, being therefore refutable, it seems th at in Popperian epistemology (including its evolutionary version) there is no possibility of refuting the `natural selection mechanism' of theories it posits. But if this is the case, don't we face here a use of the Darwinian revolution that is as `ideolo gical' as those Popper himself denounces when speaking of the Einsteinian revolution?
V. Popperian `demarcationism' belongs to the family of strategies that attempt to defend rationality by insulating it or by purging it from `extraneous' factors (cf. Dascal 1990b). The result of such a strategy is to ensure rationality's autonomy and supposed normative value. But the price paid by sharply separating the pure from the impure is the creation of a chasm between them, which it is impossible to bridge. The two domains thus created are viewed as belonging to different, unharmonizable `w orlds'. This Popperian strategy is reinforced by what may be called his `naive semantics' (possibly deliberately naive), which is the opposite of the descriptivists' contextualism. In this view, contents are conceived as entirely autonomous and independen t of their textual embodiment, along with its co-textual and contextual environment. Texts are supposed to be transparent and capable of revealing -- as the Scriptures according to Luther -- their meaning without the need of an interpretation process that would take into account co-text and context. There are no problems of `translation' between different conceptual schemes, and the question of their eventual incommensurability doesn't even arise. The philosophy of science not only doe s not need a sophisticated philosophy of language, but it must distance itself from such a thought as from the devil, because the philosophy of language can lead it to relativistic heresies. This stance excludes, of course, any cogni tively significant role for real controversy, where the dialogical context is essential for the determination of the content of theories no less than their `autonomous' evolution.
VI. Lakatos's `sophisticated falsificationism', although it aspires to be `dialectical', ultimately is also victim of an essential `monologism', perhaps resulting from the fact that he shares with Popper both the belief in the priority of normativ ism and in naive semantics. It is worthwhile to pinpoint this often unnoticed monologism of Lakatos's conception, because it implies the denial of the relevance of real dialogue -- hence, of controversy -- to his methodology of scientific research program mes.
According to Lakatos (1970), a research programme (R.P.) consists in a set of `methodological rules' of two types: the `positive' and the `negative' heuristics (p. 132). The latter's job is to protect the `hard core' of the programme, dodging the objecti ons towards auxiliary hypotheses (some of them especially designed for that purpose) or to other marginal components of the R.P.(p. 133). The former's concern is to develop the programme according to its pre-established `research policy', through the succ essive articulation of the various models or theories that form the series the R.P. consists in. This double heuristics is Lakatos's device for accounting for the fact that, contrary to what is implied by Popper's theory, there is no `instant death' of a programme (not even of a single theory) through its `refutation'. On the contrary, R.P.'s in their initial stages should be protected from `destructive' criticism, so as to allow them to flourish (p. 179). The double heuristics thus acts as an insulating mechanism, which creates around the R.P. a sterilized environment, free from the toxic agents (e.g., experimental data contrary to the programme, antagonistic programmes, etc.) which would kill it before it could mature.
The policy of favoring the proliferation of R.P.'s is, of course, a tolerant and `friendly' one, and I have no objection to it. What seems to me objectionable, however, is the `preformism' inherent to this conception, and expressed in the organic and eco logical metaphors it employs. Such a preformism seems to me to be what deprives Lakatos's theory of the possibility of accounting for the role of controversies in the evolution of science. It manifests itself in several ways and at s everal levels:
(a) The historical-dynamic element of a R.P. materializes in the succession of theories of models that realize it. But the `identity' of the R.P. -- which consists in its double heuristics -- is essentially fixed, behind the `surface' changes in i ts `embodiments'. Like the death and replacement of the cells of an organism, which are controlled by its genetic code, so too each successive model within a R.P. "is bound to be replaced during the further development of the programme" (p. 136). For this reason, the refutation of any of these models, just as the death of any particular cell, is irrelevant. The positive heuristics not only predicts (many of) these `refutations' in advance, but provides a strategy for "digesting" them (another organic meta phor) without damage to the organism.
(b) Lakatos attributes to the creators of new R.P.'s planning and predictive capacities that are really extraordinary. Not only do they anticipate refutations; they also anticipate the stages of development of the successive models.[ 31]
(c) Experimental `refutations' -- whether anticipated or not -- are neither decisive nor essential for the progress of the programme (i.e., for the development of new models in the series).
(d) As a result, a theoretician can quietly develop his or her R.P. alone. A "solitary" Bohr can work in his office, "far ahead of the experimenter: we have a period of relative autonomy of theoretical progress" (p. 152).
As a whole, the result of this preformism and insulation is monologism, which prevents the acknowledgment of the real and effective role of controversy, especially in its essentially dialogical capacity of deepening and expanding criticism in non-anticip ated ways. Although Lakatos mentions the importance of "rival programmes", his account simply lets them "emerge" (p. 154) without explaining how. In this, he does not differ from Kuhn. He also leaves unexplained the rise and development of controversies b etween rival programmes. In fact, for him criticism does not necessarily come from a rival programme, since it may be entirely internal -- even `anticipated'. If we recall, finally, Lakatos's unequivocal assignment of the process of the growth of scientif ic knowledge to the autonomous domain of World Three, then we are led to agree with those who have already noticed troubling similarities between Lakatos's and Hegel's `dialectics'.
VII. As indicated explicitly by the prefaces to the first and second editions of the Critique of Pure Reason, Kant considers scandalous the fact that metaphysics, as opposed to physics and mathematics, has been throughout its history a battlefield of endless disputes. The main objective of the Critique is to put an end to such scandals, thus transforming metaphysics into a respectable science. As far as possible, metaphysics must imitate physics and mathematics, which have been able to eradicate c ontroversies "by a single and sudden revolution" (B xvi). For this purpose, it is necessary to understand in depth the source, the nature, and the function of metaphysical controversies -- a task Kant seeks to perform in the section called "The antithetic of pure reason". Even though his analysis is not intended to apply to scientific controversies (he refers to the latter only indirectly), it deserves to be recalled here because it contributes -- in my opinion -- to clarifying the `transcendental' struct ure of the impasse I have been trying to expound.
As always, Kant's technique consists in looking for the presuppositions or `conditions of possibility' of the phenomenon under examination -- in the present case, metaphysical controversies. As a result, the "Antithetic" begins with a distinction between `thetic' and `antithetic':
If thetic be the name for any body of dogmatic doctrines, antithetic may be taken as meaning, not dogmatic assertions of the opposite, but the conflict of the doctrines of seemingly dogmatic knowledge (thesi cum antithesi) in which no one assertion can e stablish superiority over another. The antithetic does not, therefore, deal with one-sided assertions. It treats only of the conflict of the doctrines of reason with one another and the causes of this conflict (A 420-421).
That is to say, the antithetic lifts itself to a level above that of the controversies themselves. Its primary aim is not to lead to a decision in favor of one of the `dogmatic' theses vis-a-vis the other, but rather to understand the structure of these c ontroversies and their deep causes: "The transcendental antithetic is an enquiry into the antinomy of pure reason, its causes and outcome" (A 421). The definition of controversy implicit in this shift to an upper level is completely independent of the his torical occurrence of discursive episodes where the controversy becomes manifest. The conflicts discussed by the antithetic are quite different from those conflicts Kant calls `sophistical'. For, they (a) do not refer to `arbitrary' issues, but rather to questions "which human reason must necessarily encounter in its progress" (A 422), and (b) they do not involve a "mere artificial illusion such as at once vanishes upon detection, but a natural and unavoidable illusion, which even after it has ceased to b eguile still continues to delude though not to deceive us" (A 422, B 450).
As opposed to other conflicts which can be resolved quite easily, the antithetic deals with those pseudo-rational doctrines which can neither hope for confirmation in experience nor fear refutation by it. Each of them is not only in itself free from co ntradiction, but finds conditions of its necessity in the very nature of reason -- only that, unfortunately, the assertion of the opposite has, on its side, grounds that are just as valid and necessary (B 449).
The shift to a meta-level and the inevitability of the conflict are what makes Kant suppose that there is a limited set of possible (metaphysical) controversies, that their causes are essentially the same, and that, once they are understood, reason can i f not avoid them, at least understand its own limits. This assumption lets Kant define the task of the dialectic of pure reason in terms of providing answers to a small number of questions:
(1) In what propositions is pure reason unavoidable subject to an antinomy?
(2) On what causes does this antinomy depend?
(3) Whether and in what way, despite this contradiction, does a path to certainty still remain open?
In order to perform this task, Kant dismisses from the analysis of controversies all their circumstantial aspects (how they arise, how they are terminated, what strategies the contenders employ, which positions they defend, etc.), as well as any interest in "deciding in favour of one or the other side" (B 451). The attitude to be adopted is that of an "impartial umpire" (B 451) -- in fact, of a `meta-umpire' -- whose interest is not to decide the dispute, but to investigate "whether the object of controv ersy is not perhaps a deceptive appearance which each vainly strives to grasp" (B 451). From this overseeing point of view, what is of interest is not the decision of the dispute, but its `deep structure'. Kant calls this attitude the sceptical method, to be sharply distinguished from scepticism. Whereas the latter seeks to undermine "the foundations of all knowledge" (A 424), the former -- like Cartesian doubt -- "aims at certainty" (B 452). Regarding controversies, this methods acts as a prudent legisla tor vis-a-vis the perplexities of judges who undertake to apply the law. Such a legislator sees in these perplexities symptoms of difficulties inherent to the laws -- which leads him to correct them in order to avoid further perplexities. In the same way, the antinomic conflicts of reason act as symptoms that call attention to the need for determining more cautiously the principles and limits of pure reason.
This symptomatic character of controversies is both inevitable and essential for the Kantian enterprise. It is inevitable because it derives from the existence of two types of unity or synthesis, both legitimate (within their scopes), both expressing our aspiration for synthesis: the unity of the understanding and of reason. It is essential because it is the only touchstone (Probstein) that permits us to discover the architecture of our knowledge capacity and to determine what belongs de jure to each of its components -- which is the only way of, if not avoiding these controversies, at least understanding their true nature.
This is why the theory of pure reason needs the `sceptical method' -- the only means to understand and thereby overcome controversy. The other sciences do not need it: "In mathematics its employment would, indeed, be absurd" (B 452), since this science h as methods of proof able to solve any dispute; similarly, "experimental philosophy" also has a "final means of deciding [any] dispute", namely experience (A 425). It is because they can avail themselves of means for resolving controversies that these scie nces have made continuous progress, whereas metaphysics was stuck in a perpetual debate. A form of transforming it in a science and ensuring its progress would be to endow it, as in mathematics and experimental sciences, with a means of solving all its de bates. One might say that this was the aim of the dogmatic philosophers who have looked for a method. But this approach presupposes that it is possible to solve metaphysical debates. Kant, as we have seen, does not believe in such a possibility. The Kanti an way of freeing metaphysics from its scandalous state of endless controversy is, rather, to acknowledge the insolubility of such controversies, and to use this acknowledgment in order to draw secure conclusions about the nature of metaphysical knowledge and its limits. Controversy, from this point of view, has a propedeutic value for metaphysics. On the other hand, the `solution' of controversy -- as that of any other symptom -- requires a therapeutic intervention. The required therapy is essentially pr eventive: the elaboration of a "discipline" for the polemical use of pure reason, that permits the avoidance of those transgressions we inevitably tend to indulge in -- namely, to employ constitutively (i.e., so as to determine content) a faculty which ha s only a regulative function.
Kant, therefore, offers us a `transcendental' typology of controversies which, like those typologies implicit in contemporary epistemology, admits only two possibilities: either they are conflicts that can be solved or not. The former are those for which there are rigorous and accepted methods of decision, which terminate disputes and permit progress. The latter, although inevitable, are nothing but symptoms of transgressions of certain norms of rationality. To attempt to resolve them would be to fall in the trap. What has to be done in these cases is simply to avoid the transgressions that bring about the controversies.
In both cases, controversy, for Kant, neither has nor can have a constitutive role in the progress of knowledge. Like the `selection pressure' resulting from Popperian critique, controversy does not create content; at most it allows for a diagnosis and consequent elimination of `errors'. For Kant, this essential limitation of its role is indispensable for avoiding the return to the state of barbarism where conflicts reflect either power struggles between "despotic empires" or "compl ete anarchy". However, the price to be paid for `civilizing controversy' is to castrate it in its innovative capacity, by submitting it to a supreme and incontestable law.
If the prevailing theories are not only unable to account for scientific controversies in the sense I have outlined, but also in fact purport to demonstrate that they cannot exist, then it is not sufficient to point out their exi
stence in order to demonstrate their possibility. For the prevailing theories interpret what I would point to as instantiating controversies as something different from controversies. For this reason it is necessary, in addition to the criticism undertake
n in the preceding section, to demonstrate positively the possibility of controversy.
If the fundamental reason for the inability of current epistemological theories to account for controversies is the reductive dichotomy in terms of which they conceptualize all polemics -- those which are rational because they can be solved through more or less strict rules vs. those which are essentially aporetical, anarchic, and outside the domain of reason -- what has to be demonstrated is the possibility of a tertium, i.e., of an alternative that shows that between `hard' rationality (purely `calcula tive' or `logical') and arbitrariness there is a possibility of a `soft' rationality, and that controversies belong to this sphere. Furthermore, it is necessary to develop a theory of this kind of rationality, a task which has been n eglected in the past. I consider the elaboration of a `theory of controversies' an essential step in this direction.
If the reductive polarization of the set of polemics derives from a radical opposition between what is `closed' (ruled by rigorous norms) and what is `open' (where all norms are suspended), what has to be demonstrated is the possibility of a dynamical an d dialectical interaction between openness and closure, which connects (rather than disconnects) the components of relative stability (`closure') and instability (`openness') in the evolution of science.
In short, what has to be done is nothing less and nothing more than to resist the dilemma that depicts the admission of the contingency of norms as the first step on a slippery slope leading inevitably to irrationality, which leaves as the only possibili ty of salvation the tenacious defence of their necessary character. The possibility of the existence of scientific controversies with the characteristics here described depends upon the demonstration that this is a false dilemma.
To satisfy this desideratum is, of course, an enormous task. In the last section of this paper, I will limit myself to indicating that there is at least a model and a tool -- directly related to controversies -- which can contribute to the elaboration of the epistemological conception capable of satisfying this desideratum.
The model and instrument I propose is pragmatics -- the theory of language uses.
I. Its ability to serve as an appropriate instrument for the analysis of controversies is evident insofar as controversies are, above all, particular types of language use. However, it is surprising to observe that rarely have scientific controver sies been studied from a pragmatic point of view. I believe this is due to the extraordinary reluctance -- past and present -- to admit the fundamental role played by language in science and in thought, as well as the essentially pra gmatic nature of this role. This reluctance was already apparent in the refusal -- by Descartes, Bacon, and others -- to attribute to language more than a purely external role in thought and in the constitution of scientific knowledg e, a role of mere transmission or at most of organization of a thought formulated in a way entirely independent of language. The same reluctance manifests itself in Kant's deliberate silence about language in his main writings (cf. Dascal and Senderowicz 1992). I think the reason for this reluctance has been mainly the fear that, by admitting a more fundamental role for language in thought and science, one would ipso facto contaminate them with the component of arbitrariness or contingency typical of huma n languages. Such a fear persists in our century, when even those who acknowledge the importance of language attempt to replace natural languages by formalized symbolic systems, the only ones considered appropriate for the ends of science, or else conceiv e of natural languages as if they were formal systems whose semantics is `transparent'.
There has been recently a significant change in this attitude, thanks to the growing emphasis on the importance of the pragmatic aspects of language for science. Perhaps the beginning of this tendency can be traced to Mary Hesse's book (1966) which ackno wledged the cognitive value, the irreducibility and the indispensability of the use of metaphor in science, and was followed by many other researchers.
Granger is another philosopher who has recognized the importance of the analysis of the nonsemantical elements of language in science, since his Essai d'une philosophie du style. For him, "the textual appearances of science are no mere epiphenomena of sc ientific knowledge; they pertain to its very substance" (Granger 1985: 349), whence he concludes that the philosopher of science must be interested in "registering and interpreting actual [linguistic] usage from the combined standpoints of syntax, semanti cs, and pragmatics", and not merely in "eliciting the grammatical rules of a perfect -- or almost perfect, ideal Begriffschrift" (p. 351). Granger distinguishes four types of scientific discourse to which he associates different degrees of `pragmaticity'. The discourses he calls `polemical' are, according to him, those that have the largest range of pragmatic elements. He illustrates this with the example of the well-known letter of Pascal to Father Noel (29 October, 1647) about the problem of the existen ce of a vacuum. Pragmatic analysis allows Granger to identify the underlying intentions in Pascal's letter, which are revealed by his use of irony, by the pessimism he expresses about the value of controversies, and by other indications. According to Gran ger, Pascal's basic intention is not to convince Noel about the mistakes in his arguments about the vacuum, for he doesn't consider Noel intellectually worthy; what matters for him is to establish -- for an audience larger than Noel alone -- a distinction between a methodology or argumentative practice he considers inadequate for science (that of Noel) and the requisites of a methodologically valid argumentation (that of Pascal).
Both the identification of root metaphors that structure a text and of the basic intentions that orient it, are essential in order to understand its `argumentative direction', its conceptual presuppositions, and its essential premises -- without which it is impossible to evaluate its communicative efficacy and the validity of its argument. In addition to this `macro' level, pragmatic analysis is also an essential tool for handling the `micro' level, namely the understanding and assessment of each passage in the text, especially the texts that comprise a controversy. No speech act, in a polemical context can be interpreted only on the basis of its semantics and syntax. For example, in light of the conversational demand established by the preceding utteran ce of the opponent and with the help of other pragmatical indicators it is possible to know whether a response is to be read as an acceptance, a refusal, or a hesitation, since a `no' can always mean in context a `perhaps' or even a `yes'. The `logic' of controversies -- insofar as it exists -- can only be applied through a continuous process of pragmatic interpretation.
In order to fulfill its instrumental function in the analysis of controversies, pragmatics must refine its theoretical apparatus. First of all, it must account for the peculiar fact that polemical discourse is at once cooperative and competitive. This ma y lead to the recognition of special forms of respecting and violating the communicative norms of cooperativity and intelligibility. Furthermore, it must give us the means to distinguish empirically between the different forms of polemical discourse which were distinguished above in theoretical terms. I don't think it will be particularly difficult for pragmatic research to fulfill these and other requisites.
II. I will now describe those aspects of pragmatics that, in my view, permit one to see in it model of a theory capable of satisfying the desideratum formulated in the preceding section.
(a) The relationship semantics-pragmatics
Pragmatics does not replace semantics but rather cohabits with it. The pragmatic interpretation of, say, a metaphorical expression, depends upon the existence and recognition of its literal meaning (see Dascal 1986). Semantics accounts for the codificati ons of meanings in a system of rules. Pragmatics permits one to use the system in a flexible way, occasionally violating the rules, without thereby sacrificing intelligibility. This symbiotic relationship between pragmatics and semantics seems to me to be an excellent model for systems (in the present case, language) where closeness and openness are combined without damage to the efficacy of the system. In particular, it permits one to explain: (i) the possibility of conceptual change or evolution -- even a radical one -- without `incommensurability'; (ii) the difficulties of translating (underestimated by Popper and stressed by Kuhn), as well as the possibility of achieving reasonable translations; (iii) the fact that there can exist semantically closed 'conceptual schemes' without it preventing communication across such schemes, thanks to pragmatic openness.
(b) The nature of pragmatic rules
Pragmatics is ruled by norms which ensure the intelligibility of communicative acts, but such norms, in contrast to semantic rules, are not algorithmic in nature but rather heuristic. They have the character of presumptions, that is, their conclusions ar e accepted on condition that there be no stronger reasons for giving them up (see Dascal 1983). They can be formalized by means of nonmonotonic or gradualist logics (see Pena 1991). This characteristic of its norms makes pragmatics a good model (and perha ps also a means of explanation for nonrigid normativity -- that is, for `soft' reason -- which, in controversies, permits one to justify not abandoning one's position vis-a-vis its apparent refutation by the opponent: what is important is not the isolated logical value of the alleged refutation, but rather its relative weight vis-a-vis the presumption that favors the position allegedly refuted. This, in turn, explains the often ferocious debate between contenders in a controversy aimed at securing favorab le presumptions (which, in turn determine the onus probandi).
c) Reasons as causes
From an ontological point of view, pragmatic interpretation occupies a well-defined niche. It deals with a form of causal explanation, for it consists in attributing to a speaker, on the basis of his linguistic behavior and other co ntextual information, a communicative intention that serves as a cause for that behavior. However, both regarding the way in which a mental state can act as a cause and regarding the inference that allows for such an attribution to an agent, pragmatic int erpretation differs from usual physical causality, because it depends essentially on the fact that the speaker's communicative intention and his corresponding linguistic behavior have a `content', whose recognition by the hearer is indispensable for the s uccess of communication. Since it rests on this special type of causality, which operates at the level of conscious mental states, communicative behavior cannot be reduced to explanation in terms of `deep' causes, upon which the conscious subject has no c ontrol (see Dascal 1992c). Pragmatic interpretation, thereby, escapes the kind of causal determinism underlying `contextualist' descriptivism, while at the same time it suggests a model for harmonizing in a non reductionist way between World Two and World s One and Three.
(d) Non-arbitrary contingency
Pragmatic norms possess instrumental rationality since they can be seen as corollaries of the view of communication as a cooperative process (Grice). We speak coherently and clearly, we say relevant things, and so on, because we wish to be understood (an d thereby, eventually to have the others cooperate with our aims). In this sense, pragmatic norms are not arbitrary. But they are contingent, because the very idea of cooperation can be interpreted (and in fact is) in very different ways in different cult ures and historical periods. They are also contingent because the very idea of instrumentality is interpretable in different ways in different contexts. That is to say, we have here an example of the possibility of an example of a non-arbitrary, contingen t and variable normative system, whose non-arbitrariness, however, need not be guaranteed by a Universal Reason forever immune to change and criticism.
1.My presentation here of the situation of philosophy of science is evidently schematic. I know that there have been several efforts to escape the impasse here described, some of which are very valuable. don't mention them here in order t
o simplify the exposition of what I take to be the basic coordinates of the `problem situation'. Besides its contribution to clarifying the problem, this procedure will allow me to show the specificity of my proposal for its solution.
2.Lakatos begins this text with the following paraphrase of a famous Kantian dictum: "Philosophy of science without history of science is empty; history of science without philosophy of science is blind". As pointed out by Hacking (1981: 1 38n), Kant's dictum does not speak about blindness bur rather about `cyclopic' vision: "Mere polyhistory is a cyclopean erudition that lacks one eye, the eye of philosophy". The distortion of this quotation by Lakatos is quite significant, since, as I wil l show below, one major problem in his own way of combining descriptivism with normativism consists in its `cylopean' or `monological' character, which results -- in my opinion -- of the excessive subordination of the former to the latter.
3. "...the -- rationally reconstructed -- growth of science takes place essentially in the world of ideas, in Plato's and Popper's `third world', in the world of articulated knowledge which is independent of knowing subjects" (Lakatos 197 0: 179-180).
4."[S]cientists behave in the following ways; those modes of behavior have ... the following essential functions; in the absence of an alternate mode that would serve similar functions, scientists should behave essentially as they do if th eir concern is to improve scientific knowledge" (Kuhn 1970: 237). 5.This is another form in which, significantly, Kuhn expresses premise [d]. The weak normativity of Stich's (1994) proposal for the implementation of the program Quine dubs `the naturalization of epistemology' suffers from difficulties similar to those po inted out in the text as well as from others discussed in Dascal (1995b).
5.This is another form in which, significantly, Kuhn expresses premise [d]. The weak normativity of Stich's (1994) proposal for the implementation of the program Quine dubs `the naturalization of epistemology' suffers from difficulties sim ilar to those pointed out in the text as well as from others discussed in Dascal (1995b).
6.All depends, in this argument, on the level of generalization of the descriptions `ways of acting' and `essential functions'. If they are sufficiently general (but not vacuous) so as to apply to different paradigms, they engender a stron ger normativity. But in this case they endanger the thesis of the incommensurability of paradigms.
7.And this `positivism` is obviously easy prey to a tu quoque argument, which would point out that its descriptions and causal analyses, being applied to different paradigms, communities, and historical periods, must obey veracity norms wh ich are trans-paradigmatic.
8.This paragraph sounds to me excessively metaphysical and abstract. However, the allusion to Kantian terminology will be vindicated below (see section 5, paragraph VII).
9.For other differences between my position and Popper's and his disciples see below, section 5 paragraphs III-VI.
10.For other differences with descriptivism, see below, section 5, paragraphs I-II.
11.There was a real dialogue between Leibniz and Locke under the form of a correspondence mediated by third parties such as Lady Masham.
12.For this reason one cannot in general predict all the possible objections to a thesis one defends (unless it is a tautology), which blocks the possibility of using a shortcut which would consist in demonstrating the thesis, thus immuni zing it against any future objection. Leibniz applied this kind of consideration against Bayle in their debate about the mysteries of faith (see Dascal 1975).
13.On the relationship between text, co-text, and context, see Dascal and Weizman (1987). On the notions of primary and secondary text as well as co-text, see Dascal (1990a).
14.See for instance Granger (1985), Gil (1985), and especially Bohler and Katsakoulis (1994).
15.I proposed this taxonomy for the first time in the framework of the research group "Leibniz the Polemicist" at the Institute for Advanced Studies (Jerusalem), in December 1994. It should be considered as a sort of collective product of the group's work. I take this opportunity to thank the group's members for the fruitful discussions and the Institute and its staff for the magnificent support.
16.Except, of course, the recourse to some instituted authority, such as a tribunal. But in these cases, the conflict of opinions or attitudes is not decided but merely repressed, since, as pointed out by Leibniz, no one has the power to force the other to forget or to attend -- both indispensable conditions for having the power to make the other change her opinion (VOR, p. 19).
17.On the use of this fundamental metaphor, see Dascal (in press a).
18.In contrast to Frankel (1987) who distinguishes between ` scientific controversies which are `pure' and those which have political, ethical, and practical implications, Popper considers `ideological' also those `scientific revolutions' where methodological or philosophical considerations predominate or even just intervene. On this topic, see below, section 5, paragraph III.
19.As I have done above, McMullin (1987:88) too claims that a scientific controversy must be looked at as an activity, "a complex human action in which social, political, and psychological factors play an important part". But in contrast to what I propose, he omits the fact that it is primarily a dialogical activity. Furthermore, he hastens to add in the same sentence a significant `but': "but in which epistemic factors are most likely to be determinative". With this he remains essentiall y faithful to the normativist tendency, imposing a priori a category distinction (which he deems obvious) between `epistemic' and `nonepistemic' factors. Mentioning the Kuhn of The Essential Tension for whom `mature' sciences are ruled by `epistemic value s` such as predictive capacity, coherence, and fecundity, McMullin (p. 83) insists that these norms, although they vary from scientist to scientist and from community to community "are not just contingent social conventions", because "they have been found to serve the basic end of science, which is problem solving". Evidently, this (normative) definition of science simply consists in opting for normal science as the true expression of scientific activity. With this option, the possibility of accounting -- in respectable epistemic terms -- for controversies in periods of extraordinary science, as they are described by the Kuhn of The Structure of Scientific Revolutions, is excluded. No wonder that in the typology of forms of closure of controversies propos ed by McMullin (which I discuss in note 22) we find only the dichotomy discussion vs. dispute, with no room left for what I call controversy.
20.For a detailed study of this controversy, see Cremaschi and Dascal (1996 and forthcoming) and Dascal and Cremaschi (Forthcoming).
21.At least according to the paleobiologist George Gaylord Simpson, who throughout the controversy ranked with the opponents of the drift theory. In 1971, Simpson writes: "I now believe that continental drift did occur ... but direct foss il evidence is still curiously scanty or equivocal. The Cenozoic evidence indicates that drift had little or no zoogeographic effect in that era as regards most of the continents, including al for which the evidence is considerable or reasonably adequate ... Now everything has to relate to plate tectonics ... but it is rather amusing that currently there are four completely different plate tectonic `explanations' of the early distribution of marsupials, none of them based on a reasonable balance of eviden ce" (quoted in Frankel 1987: 231-232).
22.For easy reference, I reproduce here the summaries of these typologies, as given by the editors (Engelhardt Jr. y Caplan 1987: 5-6). McMullin: "1. Resolution. A controversy is resolved when an agreement is reached on the merits of the case in terms of what the participants take to be standard epistemic factors. 2. Closure. A controversy reaches closure when it is terminated on the basis of nonepistemic factors, such as the authority of the state, the pride, ambition, or laziness of a c ontroversialist; or the withdrawal of publication facilities.3. Abandonment. Controversies may terminate through participants losing interest". Beauchamp: "1. Sound argument closure: occurs if, and only if, a correct position has been reached ... thereby rendering opposition views incorrect. 2. Consensus closure: occurs if, and only if, a consensus has been reached ... through a means other than sound argument closure or some form of procedure or negotiation, that some position is best and that opposition views are incorrect. 3. Procedural closure: occurs if, and only if, a controversy is terminated by formal, procedurally governed efforts to end the sustained discussion ... 4. Natural death closure: occurs if, and only if, a controversy has come to an en d through a gradual natural death, as by fading away because of waning interest. 5. Negotiation closure: occurs if, and only if, a controversy is settled through an intentionally arranged and morally unobjectionable resolution acceptable to the principals in the controversy".
23.In the case of the controversy about continental drift, Hess's theory of tectonic plates, which was finally accepted by all the contenders, differs substantively from the original drift theory, even though it shares with it the idea th at continents actually drift. The controversy prepared the ground for the tectonic plate theory not only because it called attention to the possibility of continental drift, but also to the absence of any reasonable hypothesis (including in the original d rift theory) about the mechanism responsible for such a drift. As pointed out by Thagard (1992: 161), the possibility of conceiving of such a mechanism required a substantive `conceptual change': the abandonment of the assumption (shared by all earlier th eories) that the earth's crust is uniform, which lead to the admission that there is a basic difference between the material the continents are made of and the material the ocean floor is made of.
24.In the rare occasions when `scientific tribunals' have been constituted in order to resolve scientific controversies, there was no guarantee that the tribunal's decision would terminate the controversy. This is clearly shown by the con tinuation of the controversy over spontaneous generation many years after the decision of the French Academy (1862), which was favorable to Pasteur in his confrontation with Pouchet (cf. McMullin 187: 89; Mendelsohn 1987: 105-113).
25.In Minkovski's words: "The views of space and time I wish to lay before you ... are radical. Henceforth space by itself, and time by itself, are doomed to fade away into mere shadows, and only a kind of union of the two will preserve a n independent reality". After quoting Minkovski, Popper adds: "This is an intellectually thrilling statement. But it is clearly not science: it is ideology. It became part of the ideology of the Einsteinian revolution" (Popper 1981: 104).
26.A similar restriction to openness, in the political sphere, can be found in his `principle of limited tolerance', which I discuss in Dascal (1979, 1989b).
27.Evolutionary epistemology subtly moves from a literal to a metaphorical use of the notion of biological evolution, a move that may be responsible for this `ideologization'. Throughout history, the biological notion of evolution (concei ved in different ways by different theories) has been an important formative metaphor in philosophy. It deserves a study similar to those I have devoted to other root metaphors in philosophy (cf. Dascal in press a and b).
28.Against what he calls "the myth of the framework", which he attributes to Kuhn, Popper (1970: 56) says: "I do admit that at any moment we are prisoners caught in the framework of our theories; our expectations; our past experiences; ou r language. But we are prisoners in a Pickwickian sense: if we try, we can break out of our framework at any time. ... critical discussion and a comparison of the various frameworks is always possible. It is just a ... dangerous dogma that the different f rameworks are like mutually untranslatable languages .... even totally different languages are not untranslatable..."
29.Kuhn (1970: 268), on the contrary, stresses the difficulty (not the impossibility) of translation ("Why is translation, whether between theories or languages, so difficult?") and points out the importance of the philosophy of language for the philosophy of science (p. 235): "... philosophers of science will need to follow other contemporary philosophers in examining, to a previously unprecedented depth, the manner in which language fits the world, asking how terms attach to nature, how those attachments are learned, and how they are transmitted from one generation to another by the members of a language community".
30.For instance: " ... a research programme can challenge a considerable bulk of accepted scientific knowledge: it is planted, as it were, in an inimical environment [subrayado por mi] which, step by step, it can override and transform" ( Lakatos 1970: 140). Evidently, the condition for the programme to be able to overcome this hostile environment is that it be granted a favorable environment, at least at its beginnings.
31."All this was planned right at the start" (p. 146); although "[not] all developments in [Bohr's] programme were foreseen and planned when the positive heuristics was first sketched" (p. 153).
32."In most cases we need no refutations to tell us that the theory is in urgent need of replacement: the positive heuristic of the programme drives us forward anyway" (p. 151).
33."... the -- rationally reconstructed -- growth of science takes place essentially in the world of ideas, in Plato's and Popper's `third world', in the world of articulated knowledge which is independent of knowing subjects" (pp. 179-18 0).
34.The juridical comparison reappears later on in the "The Discipline of Pure Reason in respect of its Polemical Employment". For proposals to adopt in epistemology a `juridical model', see Romano (1989) and Dascal (1990c).
35.Fernando Gil (1980), in a brilliant chapter on "Kant and controversy", attempts to interpret the Kantian texts I have referred to here in terms of the "less Kantian" hypothesis that "controversies are constitutive and not a mere prepar ation for the exercise of thought" (p. 154). Though I fully agree with this hypothesis -- which is precisely the one I defend in the present paper -- and also with its `Leibnizian' formulation ("Controversy is an aspect of an intrinsically complex rationa lity, neither finite nor infinite, bur rather `confused', more leibniziano" -- p. 164), I confess that I do not understand the tour de force through which Gil believes to be able to reconcile it with Kant's text.
36.In the preface to the first edition, Kant describes this scandal by means of an extraordinarily powerful metaphor. Metaphysics, he tells us, has oscillated between two poles: the "despotic empire" of the dogmatists and the "complete an archy" of the sceptics. The former led to the latter because its "legislation still bore traces of the ancient barbarism" (A ix), whereas the latter led to the return of the former because of the primitive methods of "demolition" of the dogmatic fortresse s they employed. In order to overcome this scandalous oscillation, Kant will, on the one hand, improve the `method of demolition' and, on the other, build fortresses grounded on `civilized legislation' (i.e., `regulative principles'), absolutely universa l and therefore withstanding any attempt of demolition.
37.I would include Davidson's (1974) argument against the existence of different "conceptual schemes" -- and in fact against the very notion of "conceptual scheme" -- in the list of arguments which allegedly demonstrate the impossibility of controversies where radically different positions face each other. Unfortunately, I don't have here the necessary space for developing these interpretations of Davidson's paper.
38.The distinction here proposed is similar, though not identical, to that proposed by Granger (1985: 350) between `logic' and `reason': "... the discourse of scientists allows one to perceive and explicitate two avenues to justification and proof. Roughly speaking, one might call the one `logic' and the other `reason'. The first consists in calculating, insofar as abstract concepts and rules of deduction are strictly defined; the second consists in strategic choices, processes of orienta tion, and evaluation. ... If we are to track the non-logical, though rational ways the discoverer has followed, then we must closely study his language".
39.One should not confound a pragmatic with a rhetorical treatment [of controversies]. The latter has become popular in many studies about the `rhetoric of science' (cf. Gross 1990, 1993; Pera 1991) and aroused intense polemic, e.g., in e conomy (cf. McCloskey 1994a, 1994b; Mayer 1994; Maloney 1994). The essential difference between pragmatics and rhetoric, in my opinion, lies in the fact that the latter is basically interested in producing a certain effect in an audience -- the adoption o f a certain opinion. The means to achieve this effect -- linguistic or otherwise -- are chosen exclusively as a function of their causal efficacy. From this point of view, subliminal propaganda, hypnosis, and the use of language according to norms of vera city, informativity, relevance and clarity (i.e., according to Grice's `maxims') are equally legitimate alternatives. Pragmatics, on the other hand, is interested in the properly communicative processes, i.e., those in which the capacity of speakers and hearers to transmit and recognize communicative intentions, are essentially involved. In so far as these processes are of an essentially inferential nature, they involve a very specific kind of causality -- the kind of causality that leads one to accept a n opinion or to adopt an interpretation of an utterance by virtue of the `reasons' we have for that. Pragmatics, in my view, is the study of this very specific type of causality or rationality. See section 7(c) below.
40.For a historical account of the problematic of the relationships between language, science, and thought, see Dascal (1994, 1995). For the importance of the pragmatic perspective for the understanding of these relationships, see Dasca l (1992).
41.The pragmatic analysis in terms of intentions proposed here by Granger seems to me to be in conflict with his own notion of pragmatics, which -- like the one of Apel and Habermas -- tends towards transcendentalization. Granger (1979, 1 985) makes use of the Kantian distinction between "constitutive" and "regulative" principles. In addition to this, he distinguishes between "empirical" and "transcendental" conditions. For him, the empirical constitutive conditions of the use of language are the subject matter of psychology, sociology, and so on, and are of no interest to him. The transcendental constitutive conditions, on the other hand, are "algebraic", i.e. purely syntactical and semantical. What is left for pragmatics, therefore, are those conditions which are not constitutive but regulative and not empirical but transcendental. Granger singles out two of them -- the "conditions de visée" (corresponding to illocutionary forces) and "conditions d'anchrage", (responsible for the inserti on of personal experience in the use of language). But, if the psychological aspects have been discarded from pragmatics, what is `subject' or `individual' whose experiences serve to anchor the utterance?
42.Davidson does not admit talk of pragmatic norms. For him, what characterizes pragmatic interpretation is its absolutely open, nonconventional, character (See Davidson 1978, 1985). Searle, on the other hand, takes as a model for pragmat ic norms the strict constitutive rules he employs in his speech-act theory. For this reason, he is led to the conclusion that there is no pragmatic structure in conversation -- a conclusion I criticize in Dascal (1992b).
43.On this see above, note 39.