ISSA Proceedings 2006 ~ A Discussion Of Habermas’ Reading And Use Of Toulmin’s Model Of Argumentation

Introduction
It will probably be useful to start with some contextualization of the discussion. Apart from a desire to better understand both Habermas and Toulmin, the starting point here is a research and teaching interest in applied ethics. Difficult epistemological problems are raised especially concerning environmental issues. We find ourselves in need of seeing excellent environmental practices adopted if possible by all the major players, whether it be companies or countries, while at the same time the decision makers are affected by risk assessment problems and uncertainties that require modal judgments, using probability or the like. As we will see, those issues are central to the discussion of Habermas’ reading and use of Toulmin.

Habermas contributed in an extraordinary way to normative thinking since the 1960s. He asserted the foundational nature of his theoretical work, especially in the period of the Diskursethik and his Theory of Communicative Action (beginning of the 1980s). Still quite recently, he presents himself as part of what he calls pragmatic Kantianism (Habermas, 2003, 2, p. 16). Throughout his work, he is looking for some rational validation of moral principles, and in a kantian manner he wants to arrive at this justification by means of a universalizing procedure, but according to him, this should proceed in considering first and foremost discourse practices. Any norm implies that some action is required, or forbidden; in considering the consequences, could all persons affected by the norm agree on it? (Habermas, 1983). We must remember also that he adopts a cognitivist perspective on moral issues, which means he does not want to let moral evaluation or prescription rely on emotion or on the will, but wishes to understand normative problems as susceptible of rational solutions. Those elements explain in part why he refers to the so-called Toulmin model, especially between 1972 and 1983: it refers to rationality, to argumentation and seems to permit universalization. In 1972, Habermas’ purpose is not to give a precise presentation. In fact, even in 1981, his use of Toulmin is rather rhetorical and selective. He is mobilizing Toulmin to serve his foundational project, which should not come as a surprise.

On the other side, it is also clear why Toulmin still has an enormous importance today in ethics, in particular (among other things) in decision making situations. Modals and rebuttals are of the foremost importance if we are to make decisions and judgments while taking into account context and possible exceptions. In the problem domain of environmental ethics, where risk issues and decision processes in quite uncertain situations are regularly required, this is even more the case. Those are difficult issues to treat and the available knowledge is far from the level of certainty most people would prefer for making decisions.

Here I will briefly recall some of Jürgen Habermas’ theoretical work, and limit the focus to the use by Habermas of Stephen Toulmin’s model of argumentation, especially in the article Wahrheitstheorien that is not translated in English yet.[i] The article will be looked at in detail.[ii] I will conclude by looking at the 1981 treatment and to what happens with the issue in later work (especially Habermas, 1983 and Habermas, 2003).

What effect has this use of Toulmin on Habermas’ theory, and what is the meaning of his obviously limited take on Toulmin? Toulmin’s model has on one side the effect of supporting Habermas’ rationalist view of argumentation, according to which to argue or plead is essentially to give reasons that justify the speaker to hold a specific claim in a discussion. Habermas might have selected Toulmin’s theory precisely with the purpose of reinforcing his general position on normative theory, according to which we should think the validity claims in the normative sphere (about rightness) in analogy with what happens on the descriptive or connotative sphere (the question of truth). But this limited reading of Toulmin by Habermas has also for consequence to present a simplified and radicalized Toulmin, in a direction that is not coherent with Toulmin’s intentions, as can be seen in his further work, especially the Argumentation handbook written with Rieke and Janik. I argue that Habermas’ access to the 1958 famous book, The uses of argument, is thoroughly directed by his interest for the theory of validity in the normative sphere, even if Habermas does introduce in 1972 some revolutionary notions that we do not find in Toulmin (1958). The question is to know if Habermas fully considered the implications of Toulmin’s theory. It might be that he failed in that direction in reason of his foundational project, i.e. by lack of a sufficient preoccupation for application questions (in which we face nowadays an urgent need to take into account modals and rebuttals).

1. Habermas’ concept of argumentation
The main characteristics of Habermas’ theory of argumentation are constant from Vorstudien und Ergänzungen zur TKH and later work on Truth and Justification or Truth and Ethical theory, even if some development occurs.[iii] We will not present it in detail: the theory of the different validity claims (rightness, truth, sincerity), the presupposition of orientation towards mutual understanding supposedly required in rational argumentation, the idea of the mutual recognition of the validity claims in the Ideal speech situation, the counterfactual and foundational perspective of this Ideal speech situation are notions that are now well known. Truth is defined by justification, a valid claim is a claim that the arguer is ready to back up with arguments, and for him to enter into discussion means to be ready to justify one’s claims. We find this line of thought at each turn, whether it be in the context of MoralBewußtsein, i.e the Diskursethik period, or later in what he calls a theory of discussion (around Between Facts and Norms and The Inclusion of the Other). Habermas is proposing a rational, procedural, universalizing, consensualist theory of truth, especially in a critical discussion with Tarski. But it is in Toulmin that he will turn to give some procedural or specific context to argumentation as such, over and above the reference to Tarski and the T. model, seen as part of a more formal approach to logic. In Truth and Justification he still holds his consensualist theory of the truth. He is in an explicit discussion with R. Rorty. He refutes his contextualism because of the need of keeping a truth theory, in two complementary ways: 1) as a simple sanction of cognitive value or validation for a statement, in the context of a semantic theory of truth inspired by Tarski and 2) as a transcendental criterium allowing for the critique of utterances, for distinguishing knowledge and belief, and as a useful upper limit to knowledge. This last idea was already in Warheitstheorien.

Whenever Habermas refers to or discusses argumentation, it is always to assert that the statements recognized as valid, whether it be the truth or the rightness variety of validity, are justifiable by giving reasons. This element is related to a theory of reason that can not be detached from language, and according to which such a rational argumentation would help the partners of discussion to transcend at least in the discussion their more immediate context of interest. This transcending character of argumentation is asserted as late as 2003 (Habermas, 2003, 2, p. 71-74). Furthermore, we should also stress an important point, the linkage of these elements with the speech act theory, by reference to the works of Austin and Searle. Specifically it is in the assertive speech act that something is produced that can be said to be true or false.

As we might recall, speech acts are seen by Habermas as quasi-transcendental ways to ground discussion processes in the validity claims and the mutual expectations that they presuppose among partners of discussion. Habermas does present argumentation as speech acts, and this was certainly a relatively new perspective at the time. But he does not show how the fact that they are speech acts might have an important effect on the arguments and their reception; on the contrary, he unties the question of truth from the conditions of the act in which the assertion takes place. Instead of looking at how the context of a speech act might have effects on arguments, he insists to consider the arguments in relation to normative elements like truth, rightness, or sincerity. This linkage is clearly present also in the Theory of Communicative Action. The reference to a pragmatic conception of truth and the other validity claims, present both in Wahrheitstheorien and in Truth and Justification, goes hand in hand with the reference to the consensual theory of truth, without looking at the importance of dissent or analysis of the embeddedness of argumentation in speech acts; the connexion between pragmatics and consensus is asserted only at a theoretical level.

The proposition is not susceptible to be true or false, only the act of assertion or statement (Austin): this is a first level of distinction. Habermas then says that the truth category applies to the statement, not to the utterances (in a reading of Strawson). He is distinguishing the different levels of the sentence, statement, proposition, with assertion as the required quality of the statements. Further on, he will say that truth is a propriety that do not pertain to the given information, but to the statement (Aussage), being therefore independent of the context and universalizable. Truth can not be measured by the “probability of prognostics”, but by an unambiguous alternative: “ob der Geltungsansprüchen von Behauptungen diskursiv einlösbar oder nicht einlösbar ist”, “if the validity claim of assertions can or cannot be honoured in discussions” (Habermas, 1984, p. 136). In that manner, it seems that consensus of the discussants is required precisely to overcome the uncertainties of assertions that are merely probable, while the notion of truth is supposed to give us or permit a clear cut situation that helps to make a decision.

2. Toulmin in Wahrheitstheorien
In Wahrheitstheorien, while discussing argumentation, Habermas mentions Chaïm Perelman and Bar Hillel, but he mostly refers to Stephen E. Toulmin. It is in the 4th section of this very important article in the whole of his work, while treating the logic of discourse, that he discusses objections to what he presents as his consensual theory of the truth: he draws explicitly on Ch. S. Peirce on that topic. Let’s recall what the context and the meaning of this is. He asserts that the theory of truth should not fall back on empiricism or transcendentalism per se, as criteria to decide of the validity claims of the statements. The consensus theory helps to put back truth discussions in the scientific community which builds statements and theories in the first place, where justifications will be required to honour validity claims. He offers some arguments to avoid circularity objections (the conditions that permit to judge a consensus cannot be themselves subject to consensus), he holds that such a truth theory serves to explain the binding character of the arguments without coercion by means of the « formal properties of the discussion ». These seem to be the backing reasons for choosing Toulmin instead of Bar Hillel or Perelman, because as he expresses it this author « chooses the adequate level of investigation for a logic of discussion » (Habermas, 1984, p. 161, “Ich werde mich auf St. Toulmins Analyse des Gebrauchs von Argumenten stützen, weil Toulmin die für eine Logik des Diskurses angemessene Untersuchungsebene wählt”.). But here the “formal” leaves some doubt since we will see Habermas exclude from any consideration some important parts of Toulmin’s relatively light formalism, namely the rebuttal (R) and the qualifier (Q).

Thus his reading of Toulmin reinforces some classical way to see argumentation as being more in continuity with logics than with literature. This makes sense in a way, because it is true that Toulmin does not help us with style effects, literary effects, figures of speech or other rhetorical uses like metaphors etc, even if we always can, as analysts, take the liberty of putting these elements anywhere inside the Toulmin schema, under G or W etc. Toulmin’s work was in good part a critique of the classical, formal logic in the syllogistic form. It seems that it is an essential part of his contribution to introduce modalities in reasoning, in a finitist and fallibilist context, conducing to such interesting cases like « A Swede is almost certainly not a catholic » or « So, presumably, Harry is a British subject ». These carefully limited conclusions can be grouped under a theory of the value and validity of verisimilar affirmations. This might not be equivalent to a theory of truth.

Habermas saw in Toulmin a cognitivist and was right about it. For Toulmin a modal statement attributing a strong positive probability to some future event, can not be said not true if the probable event asserted does not happen; for it was true that the statement was correct when it was made, even if the speaker was not certain of the outcome (Toulmin, (2003) [1958], p. p. 59). Toulmin’s attitude on that respect is in contrast with that of Kneale, with which he has a precise discussion: he pleads implicitly for taking the statement asserted as probable as something that can be taken as true. This having been said, Toulmin surpasses this fascination towards truth that we find typically in logics and also in Habermas.

Let us remind that the Toulmin model permits to come to claims presented as conclusions (claims, C), on the basis of some data (ground, G or data, D), by means of certain generalizing statements that are seen as guarantees (warrants, W), themselves being based in some larger semantic context, like the texts of law of this and that country or like a scientific treaty, the backing being always content dependant (backings, B). While helping him to take a distance from formal logic, the introduction of modalities and rebuttals helped to take exceptions into account in a generally valid deduction that still has a lot in common with a syllogism (Q and R). It is the main aspect of Toulmin’s contribution to create an opening that amounts to recognizing the cognitive value of statements that imply no certitude in their material content, while they may have an important cognitive value at the same time. These statements are rationally acceptable, under the condition of an appropriate modalization with the rebuttals explicitly expressed. Habermas will refer often to rational acceptability, warranted assertability, making it an importance piece of his discourse, but he will not refer to rebuttals of modal qualifiers (Habermas, 1984), p. 160. It is hard to think that this was only a poor choice on his part, a mistake or an accident.

 

We arrive at the conclusion C on the basis of D (later called G), with the help of W and B, but in a qualified manner Q which is n relationship to the rebuttals R. These last elements are obviously essential in the toulminian model. The version given by Habermas is the following (Habermas, 1984, p. 163):

3. Commenting on the habermasian treatment of the toulminian schema
Habermas states he is presenting a simplified version (the German word is vereinfachten) of Toulmin. In many cases, a selective reading can be appropriate in philosophy, but here it is more than a synthetic version : it is the amputation of elements that seem both characteristic of Toulmin’s approach and necessary in it. He mentions modalities in general, especially the possible and the necessary, but without looking at Toulmin’s treatment of these. The rebuttals and the modal qualifier, which have disappeared in Habermas, might have had the effect of obliging the consideration of different types of statements, with modal qualifiers like “almost certainly possible”, “more” or “less probable”, elements that are not fitting well inside Habermas’ theory of universalization. Habermas prefers a clear cut situation, for the statements to be decidable. Giving a very different outlook, the rebuttal (R, rebuttal) seems to be opening us to a theory of exceptions that does not seem to be useful for Habermas’ foundational purposes.

Habermas’s does not look at statements like those in Toulmin, of the kind « A Swede is almost certainly not a catholic ». Such a statement is practically acceptable on the basis of the important proportion of Swedes being Protestants (again, according to Toulmin, 1958): in that sense, Habermas would qualify it of being true. But what is the receivable proportion for such an assertion to be admissible? Should the proportion be 98% of the population, or 95%? Obviously, the number has itself a probabilistic quality. But is it decidable, since a statement needs to be decidable if it is to be able to be true or false? There is incertitude, a blurring of the frontier between true or false, a margin of error that might or might not be important. Any sentence which is deemed to be true has the important characteristic that it “ascribes a property to sentences that is preserved through valid inference” (Heath, 2002, p. 287).[iv] Since such inferences seem essential if we are to follow a complex reasoning, we can see the problem it causes if we are dealing with risk issues of a (more or less) probable nature. For instance, the fact that environmentalists will put together, in a long chain of reasons, a series of arguments that all have a certain degree of probability, might not be helpful in convincing some of the receivers of their discourse. Even if a statement can be trivially true while still being probable, it seems doubtful that this quality can be taken together with other similar statements in valid inferences in any conditions.

Also, we should note that Habermas’ choice of ignoring the rebuttal and the modal qualifier is a move that seems to be rhetorical. He does mention probability, but moves beyond it on the basis of the necessity to apply some truth category, and he refers the tough job of discussion and decision to people to which he ascribes to obtain consensus! In effect his evacuation of the modality qualifier and of the rebuttal permits him to put in place his bridging principle, which is used as a means to obtain universally valid statements on moral issues, for any possible partner in the discussion. Whereas for somebody like Toulmin, famous also for his writings with Jonsen on the casuistic tradition (Jonsen & Toulmin, 1988), ethical and moral questions or issues always have to be treated in specific and contextualized settings.

4. A look at details in his treatment
Habermas briefly presents the Toumin model and gives some examples. He treats backing, B, essentially as casuistic evidence, to sustain an hypothesis in the case of an assertion, or to put forward a norm in the case of a prescription (Habermas, 1984, p. 165). There is nothing of the kind in Toulmin. For him the backing B is a factual statement, for instance the law in its content, details on the period of its proclamation (Toulmin, 2003 [1958], p. 131. In other examples, it will be the classification of living beings, the Linné system for plants, other laws formulated by science. The backing is always in dependence with a specific field of argumentation. In Toulmin, casuistic evidence could be admitted only at the level of the data or grounds (D or G). Obviously Habermas fuses together the data with the semantic context of foundation or backing (B), since he explicitly holds that some future consequences or some facts can be used at the level of the backing or B. According to Habermas, the warrant (W) is more than a repetition of the facts, it is a « general moral with practical character, concerning the way we can argument without risking to make a mistake ».[v] Habermas underscores the important relationship between the warrant (W) and the backings (B). But it seems that from inside the toulminian schema’s perspective, what is important instead is the justified passage from D to C by means of W and B, with the reservations R according to possible exceptions, the assertion being done with the proper modal qualifier Q that enables the judgement to be adequate, while at the same time specifying the limits of the discourse. Discussing modalities, Habermas seems to be content with some general reflection on the relationship between parts of the arguments, and with reflection on the possible, the necessary and their negative counterparts, without telling how this could affect his theory.

It is precisely on the basis of this so-called importance of passing from B to W that Habermas will take the toulminian expression of a « bridging principle », an expression he will briefly refer again in Moralbewußtsein und Kommunikativen Handelns in particular, where it plays an important part (Habermas, 1983, p. 67; 73). This principle becomes for him the means of universalization, on the basis of the importance of having arguments valid in any time and independently of the context. We find here again his requirement of some transcendental characteristic of argumentation. This is completely independent of a toulminian perspective.

Let us recall the famous sentences of his 1972 article: « Even if there is no deductive relations between the statements figuring in the Warrant and the Backing, a statement gets its consensual strength from the legitimacy of the passage from B to W.” (Habermas, 1984, p. 166). If we make a generous interpretation of Habermas, we might assert that this is because of the social and intersubjective nature of the backing, since it is always a semantic context, a theoretical construct or a collection of texts. But he himself does not make such an interpretation.

While discussing his moral reading of Toulmin’s model, he asserts that universalization is for normative issues what induction is for empirical questions:« … universalization serves as a bridging principle (Brückenprinzip) for legitimizing the passage from descriptive indications (noticing the direct and secondary consequences of the norm for the satisfaction of universally recognized needs) to the norm » (Habermas, 1984, p. 167). As we might recall, induction gives a way of generalizing, from particular instances to a general or universal law; that is what is requires for practical reason. But it is important to recall that the toulminian discourse, especially the 1958 model, does not address the issue of universalization, even if Toulmin does have a cognitivist position on moral issues. On the contrary, it is the explicit role of the modality and of the rebuttal to help against the spontaneous tendency of classical logical thinking to universalize. As we know, in syllogistic thinking, having clarified that all A’s are B’s, and admitting that C is an A, then it follows in all possible circumstances that C is a B. Toulmin’s essential contribution was to make important modifications to this model, which are completely erased by Habermas.

We should then fairly note that Habermas brings important elements that are not in the 1958 book and that are complementary, especially his reflection on terminology and its importance for the selection of relevant facts (Habermas, 1984, p. 166 f.). These remarks are contributions to a (then) developing theory of framing, before Erving Goffmann’s book on the subject (1974). But we must ask whether these elements should be understood as part of Habermas’ transcendental logic. It is true that he wants to assert a distance both from a logic of statements, i.e. a formal logic and from transcendental logic, in what he calls a pragmatic approach. The transcendental logic would examine « the fundamental concepts (categories) needed for constituting the objects of a possible experience » [relevanten Grundbegriffe (kategorien)] (Habermas, 1984, p. 161-162). Habermas tells us also, that over W, B and the rest, there still is the system of language, of which the validity is decided as a whole, in terms of cohesion of the statements towards one another, and not in singular each time towards specific referents (it is a consensualist holism, in a semantic and pragmatic conception of language). These categories, that are intervening in a sense before the data and the transition laws (like W) permit the selection of elements in a quasi-transcendental manner. It is the fundamental concepts of the language system that make it possible to deduce from D and W while also providing with B a sufficient reason to accept W and therefore the claim C. It is by way of the categories and grundbegriffe that we put together a justification or an explanation to an object domain: we can also say that by choosing a terminological system, we assign a domain of objects to that system (Habermas, 1984, p. 166). Moreover, it is the terminological system that will decide what class of facts are admitted in the argumentation. It is then a transcendental conception of categories and terms that permit what will be called framing especially after Goffman: it is the case that to operate, data, warrant and backing (D, W, B) all need the terms selected inside a language system, that can also (in part) be selected: Toulmin will discuss these elements concerning framing later on, especially in his handbook (Toulmin, Rieke & Janike, 1979). Habermas refers to Cicourel to hold that facts, interpretations of data and needs depend on the « categorical frame proper to the chosen terminological system » (Habermas, 1984, p. 166).

It is while introducing the notion of a cognitive schema, referring to Piaget and to the project of a materialistic epistemology, that Habermas will use the vocabulary of the a priori, typical of the transcendental method. He writes that even if they come from the practical experience, these cognitive schemas “have an a priori value towards the experiences they organize as experiences” (Habermas, 1984, p. 167). As in other papers from the same period, this materialist epistemology takes us back to social work as a synthesis (Habermas, 1984, p. 167, “ … welche die gesellschaftliche Arbeit als Synthesis versteht. ” Even if it comes from experience and from a cultural work of formation, the language of justification precedes experience and work. At the time, it is probably to transcend these obvious historic limitations that universalization is required. It seems that the habermasian statements there are very close to the linguistic (in the sense of sprachlich) self-foundation, as we can see in Knowledge and Human Interests, and in Science and technique as ideology.

5. A discussion of the relevant theoretical issues
It is very important in my view to distinguish between the cognitive value of a statement and its decidability in the sense of the formal logic. In practical life, we have to make decisions in a context of uncertainty, using descriptions of complex states of affairs which are themselves more or less certain, with at best probable consequences. When we consider the domain of future events and asserted levels of their likeliness, difficulties arise by the simple fact that many levels of discussion are fused together and blur the issues at hand.

Consider some basic distinctions that might help clarify the matter discussed. If we assert A= ‘It is likely that event X will occur’, we have to distinguish to levels: one concerns the facts discussed, here event X, the other one proposition A in itself. One the first level, A is telling something about an event, that might occur; on the second level, there is an evaluation of the probability (in a general sense here) of the event X occurring, the second level being included in the first. We can then distinguish between 1a, the eventual fact of event X, and 1b, the truthfulness of A in relationship to its asserted level of likeliness… which is not the same thing as a level of certainty (Sproule, 1980).

Let’s look at a statement probably made by some people in the early summer of 2006, “Germany might win the Football World Cup of 2006 ». It asserted a possibility, a very interesting one at the time for most people in Germany and elsewhere. It is on the basis of what was possible or not in our world that some opinion could be held on the subject. From a non specialist’s point of view, there was no important reason to assert the contrary, i.e. the impossibility of Germany winning that championship. So it seems to be true in an habermasian way: it could be acceptable, it could have been backed by some arguments (even if that team did not statistically surface as substantially better than the other leading teams). To decide if an asserted possibility like the one in this statement was true or false at the time of its production, it was not necessary to know what was due to happen in July 2006. We did not even need to know the actual degree of probability asserted. There is two levels in the discussion: the level of the statement, i.e. the pragmatic act of asserting the positive probability, and the level of the facts or events discussed; the level of the facts was not known before the events took place, but we could already discuss the subject before (and make an informed opinion, or an educated guess, even if it was not impartial!). As we know, statistics and probabilities are very popular these days. Let us suppose that in February 2006 a person said: “We can bet that Germany will win the FIFA competition, because this event has a probability of .89 on a total of 1.00», this could have looked fascinating for some people. Eventual gamblers might have placed their money on the team because of their belief in that statement. But what was interesting for the common gamblers was to know whether or not the team of their bet might win, and not to know if the asserted probability was the right one. Furthermore, the German Mannschaft might have lost or won, we would never have known if that number was correct, even by adding an eventual margin of errors.

Those issues might seem trivial, but what about the following: “If we do not enact some radicalized version of Kyoto (called, let’s say, K3) now and in every country, there is a 95% probability that the GHG will ruin Earth’s atmosphere by 2025”? And what about “Because of wind and other elements, among which extension of the current agricultural practices and competition between markets and local economies, the total surface of the earth covered by GM crops around the world will double in the next 10 years”? These are the kinds of issues for which Habermas might help us to plead for the necessity of debate and deliberation, but it is Toulmin (and the subsequent risk thinkers (Beck, 1986; Leiss, 2001, and already Kahneman, Slovic & Tversky, 1982) that might help us to correctly evaluate such difficult statements that require a lot of specialized, complex and domain-specific knowledge. Without such a knowledge (with all its limits) the habermasian discussion will get nowhere… especially when deciders require of the previsions a level of certainty that seems by definition impossible to obtain, in a context where the knowledge value of an assertion about the strong likeliness of some future events does not have to give certitude about the specifics of the events in question to still be useful and valid.

Conclusion: How Habermas treats Toulmin in later years
If we look at the Theorie des Kommunikativen Handelns (Habermas, 1981), the least that can be said is that Habermas’ concept of argumentation as articulated there is even more strongly normative than elsewhere, presupposing the reflexive participants do thematize their claims to validity. Taking into account more recent work from Toulmin (Toulmin, Rieke & Janik, 1979) he certainly finds in Toulmin a reflexive and non absolutist concept of argumentation, Toulmin is presented as adopting a non deductive stance towards normative issues and rejecting relativism (Habermas, 1981, p. 47). But this reading serves to develop what he calls a logic of argumentation. Argumentation is the pursuit by reflexive means of actions oriented towards mutual understanding (Habermas, 1981, p. 48). He does quote Toulmin mentioning the modal qualifier in the foot notes (quoting, as he usually does, in English), but in his text he will refer to it only once, and as a ‘modifyer’ serving to restrict or modify the validity claims (Habermas, 1981, p. 49). Nothing important seems to come out of that important dimension of Toulmin’s work. And he does not comment on the moral examples Toulmin and others offer (Toulmin, Rieke & Janik, 1979, p. 309 f.). After discussing Klein’s reading of Perelman and Olberechts-Tyteca, he asserts that Toulmin offers a superior theory by differenciating validity claims (Habermas, 1981, p. 56) while admitting the critical and transcending quality of validity. But Toulmin lacks having sufficiently mediated the empirical and logical levels of abstraction…and Habermas criticizes his insistence on field dependency of argumentation, presented as having institutional criteria. On that basis, he develops more fully his own theory of the three general genres of validity claims…that have not much to do with Toulmin.

The references to the Toulmin model that we find in Moralbewußtsein und Kommunikativen Handelns are there only to introduce Habermas’ discourse on the Brückenprinzip, a bridging principle that permits the process of universalization on moral issues, a process that appear then grounded in Toulmin’s informal logic (Habermas, 1983, p. 73). But since we have shown that in Toulmin, W and/or B do not have that role, it follows that the role of the reference to Toulmin here is purely rhetorical. Habermas satisfies himself there in referring to precisions given in Wahrheitstheorien, declaring in the context of the work at hand that an informal logic is required for argumentation theory, and that a moral principle similar to what induction does in empirical science is needed. The discussion continues with moral theoreticians like Kant and Hare. Some misunderstandings are treated, then the principles U and D are introduced, that taken together are the real bridging principle of his Diskursethik (Habermas, 1983, p. 76). As we can understand, the very insistence on universalization goes against taking specifics or exceptions into consideration.

In one of the self-critical stances of which he is capable, he later admits that his older conception of the truth as epistemic and discursive might have been due to an overgeneralization of the special case of the validity of normative judgments (Habermas, 2003, p 8). It is the case that in Truth and Justification, the process of detranscendentalization that is contemporary to the linguistic turn does have to do with uncoupling truth and assertability. And there Habermas does give some manoeuvring space to a modified conception of knowledge in admitting a plurality of ways to correctly consider a certain state of affairs (Habermas, 2003, p. 227-229). It is also true that in that book, he takes more clearly into account issues of risk and difficulties of radical claims to truth by speakers, especially on normative issues (Habermas, 2003, p. 273). But even in that book, he does not come back to Toulmin’s work, in the sense indicated here (or in any other sense).

If Habermas gave a tremendous contribution to normative thinking in the late 20th century by giving to it some rational foundation, it seems that the 21st century will be in need of Toulmin’s thought and nuances to go further in taking into account exceptions and reservations before arriving to judgments that, if not universal, might be susceptible of large approval by numerous people.

NOTES
i. Wahrheitstheorien (=Theories of the truth) was first published in a festschrift for W. Schütz: Fahrenbach (1972). It was then put in Habermas (1984) and later translated in french, see Habermas/Roschlitz (1987), under the title « Théories relatives à la vérité », a translation I previously used. In the absence of an English official version, the translations of sentences given here are mine.
ii. It should be noted that Habermas discusses Toulmin’s early work on ethics, Toulmin (1950), in Habermas (1983), p. 60-61. He looks especially at Toulmin’s idea of comparing the relationship between moral argument to attitudes, with the one between theoretical argument and the flux of perception; Toulmin’s book is seen as a good example of asking the good question without finding the good answers. This book is seen as relatively independent, and as less interesting than the 1958 classic.
iii. Habermas (1999). See also, more recently, a short discussion in french with Alain Renaut and Pascal Engel, in Habermas, (2003). At the time of submitting this paper, that book was not translated into English.
iv. Heath presents “designatedness” as the property common to the different validity claims, truth, rightness and sincerity.
v. In the chapter of the book on probability, Toulmin’s discussion with Kneale and Carnap does inform us of the importance that Toulmin puts on the affirmative nature of a statement made as probable: for him; “probable but not true” is not a tenable position. This might be in part why Habermas feels justified to back his theory with Toulmin.

References
Beck, Ulrich (1986). Risikogesellschaft. Frankfurt am Main: Suhrkamp Verlag.
Fahrenbach, H. (1972). Wirklichkeit und Reflexion. Walter Schulz zum 60. Geburtstag. Pfullingen: Gunther Neske, p. 211-266.
Goffman, E. (1974). Frame analysis: an essay on the organization of experience. Cambridge: Harvard University Press.
Gregersen, T. (2004). Publications by Jürgen Habermas, since 1992. http://www.helsinki.fi/~amkauppi/habbib.htm.
Habermas (2003). Truth and Justification. Cambridge: MIT Press.
Habermas, J. (2003, 2). L’éthique de la discussion et la question de la vérité. Paris: Grasset.
Habermas, J. (1999). Wahrheit und Rechtfertigung. Philosophische Aufsätze. Francfort, Suhrkamp.
Habermas, J. (1998). Between Facts and Norms. Cambridge: MIT Press.
Habermas, J. (2000). The inclusion of the other. Cambridge: MIT Press.
Habermas, J. (1984). Vorstudien und Ergänzungen zur Theorie des Kommunikatives Handelns. Francfurt am Main: Suhrkamp Verlag.
Habermas, J.; Rochlitz, R. (tr.) (1987). Logique des sciences sociales et autres essais. Paris : PUF.
Habermas, J. (1983). Moralbewusstsein und Kommunikativen Handelns. Frankfurt: Suhrkamp Verlag.
Heath, J. (2002). What is a validity claim?, in Rasmussen, D. J. M. & Swindal, J. Jürgen Habermas, v. IV, London – Thousand Oaks: Sage.
Jonsen, A. R & Toulmin, S. E. (1988). The abuse of casuistry. A history of moral reasoning. Berkeley: University of California Press.
Kahneman, D, Slovic, P & Tversky, A. (eds.) (1982). Judgment under uncertainty: heuristic and biases. Cambridge: Cambridge University Press.
Leiss, W. (2001). In the chamber of risks. Understanding risk controversies. Montreal-Kingston: McGill-Queen’s University.
Sproule, M. J. (1980). Argument. Language and its influence. New York: McGraw-Hill.
Toulmin, S. E., Rieke, R. & Janik, A. (1979). An introduction to reasoning. New York: Prentice Hall.
Toulmin, S. E. (2003) [1958]. The Uses of Argument. Cambridge: Cambridge University Press.
Toulmin, S.E. (1950). An examination of the place of reason in ethics. Cambridge: Cambridge University Press.