Stephen Toulmin, The Uses of Argument

Toulmin's book appeared in the same year (1958) as Perelman's. Like Perelman, he takes issue with the then-dominant model of good argument as demonstration and of logic as a formal science. He proposes to replace what he calls the "geometrical model" of logic with a "jurisprudential model": "Logic (we may say) is generalized jurisprudence." (7) "...we shall be interested in justificatory arguments brought forward in support of assertions, in the structures they may be expected to have, the merits they can claim and the ways in which we can set about grading, assessing and criticising them." (12)

The heart of Toulmin's book is his third chapter, "The Layout of Arguments", and especially the first 14 pages of this chapter, in which he sets out on the basis of his jurisprudential model an analysis of the "macrostructure" of arguments which is an alternative to the traditional analysis into premisses and conclusion which has come down to us from Aristotle. (For a more recent discussion, see James Freeman's Dialectics and the Macrostructure of Arguments (1991).) Toulmin's analysis in terms of claim, data (later grounds), warrant, modal qualifier, rebuttal and backing has become standard in the study of argumentation in many American speech communications departments. It has also found its way into some of the informal logic textbook literature, including a textbook co-authored by Toulmin himself. There are also objections to Toulmin's model, e.g. by Ralph Johnson, by Freeman in the above-mentioned book, and by Frans van Eemeren, Rob Grootendorst and Francisca Snoek Henkemans in Fundamentals of Argumentation Theory.

The latter portions of chapter 3, in which Toulmin develops a rather opaque distinction between analytic and substantial arguments, have not been influential, and should be ignored.

A key thesis of Toulmin's book is that our standards for appraising arguments are field-dependent. In the first chapter he defines the notion of a field in terms of an undefined notion of a logical type to which a statement belongs: "Two arguments will said to belong to the same field when the data and conclusions in each of the two arguments are, respectively, of the same logical type: they will be said to come from different fields when the backing or the conclusions in each of the two arguments are not of the same logical type." (14) As examples of logical types, Toulmin gives: reports of past and present events, predictions about the future, verdicts of criminal guilt, aesthetic commendations, geometrical axioms. (13) Toulmin's field-dependency thesis, which informs his textbook, has been a subject of controversy, especially in discussions of the educational ideal of critical thinking among philosophers of education.

The first chapter also develops the parallel between procedures of rational assessment and legal procedures.

Chapter 2, "Probability", is an analysis of modal terms by the methods of ordinary-language philosophy fashionable in the 1950s, one which provides a background for the discussion of modal qualifiers in chapter 3. Toulmin's speech act interpretation of the word 'probably' in chapter 2 as expressing a guarded commitment to what it qualifies has been vigorously and ably defended by Robert Ennis in a paper published in 2006: 'Probably', in David Hitchcock and Bart Verheij (eds.), Arguing on the Toulmin Model: New Essays in Argument Analysis and Evaluation (Dordrecht: Springer, 2006), Ch. 10, 145-164.

Chapter 4, "Working Logic and Idealised Logic", traces the present divergence between logical practice and logical theory to Aristotle's ideal of logic as a formal science comparable to geometry. Thus for Toulmin Aristotle is a villain, not a saviour.

Chapter 5, "The Origins of Epistemological Theory", argues that the pride of place given in logic to arguments backed by entailments (i.e. deductively valid arguments?) has created spurious problems in epistemology, such as the problem of the justification of induction. "The proper task of epistemology would be not to overcome these imagined deficiencies [of arguments which are not deductively valid-DH], but to discover what actual merits the arguments of scientists, moralists, art critics or theologians can realistically hope to achieve." (10)

Reception: Toulmin's book was almost universally panned by the philosophers who initially reviewed it, who attacked its ideas on probability, defended formal logic and ignored his model. Its ideas were however enthusiastically received in disciplines concerned with teaching argumentation for practical purposes, e.g. jurisprudence and speech communication. They were discussed sympathetically and extended in the scholarly literature in those disciplines, and entered into many elementary textbooks, including one co-authored by Toulmin himself. Toulmin's book has remained popular; a paperback edition came out in 1964, and has been reprinted many times since; an "updated edition" with new pagination, identical in content except for a new preface by Toulmin, was issued in 2003). Despite the criticisms, Toulmin has not altered a word in the book; in the preface to the paperback edition, he acknowledged the criticisms only to the extent of saying that they had "served only to sharpen for me the point of my central thesis-namely, the contrast between the standards and values of practical reasoning ... and the abstract and formal criteria relied on in mathematical logic." (Toulmin 1964: viii) In a keynote address to the Second International Conference on Argumentation in Amsterdam in 1990, he said: "If I were rewriting the book [The Uses of Argument] today, I would broaden the context, and show that it is not just the 'warrants' and 'backing' that vary from field to field: even more, it is the forums of argumentation, the stakes, and the contextual details of 'arguing' as an activity." (Toulmin 1992: 9; italics in original) He maintained his adherence to his controversial field-dependency thesis in his keynote address to a conference in 2005 on the uses of argument. (1) Toulmin's thesis of the field-dependency of standards of argument evaluation has found a receptive ear among some philosophers of education who write about the educational ideal of critical thinking, and who identify it with the epistemology of the disciplines, identified with Toulmin's fields.

Criticisms by van Eemeren et al.: In their Fundamentals of Argumentation Theory: A Hadbook of Historical Backgrounds and Contemporary Developments, van Eemeren et al. raise the following objections to Toulmin's model:

1) He uses the term field ambiguously to denote sometimes the logical type of the statements composing an argument (in which case all forecasts based on data about present conditions belong to a single field) and sometimes the discipline to which the warrants licensing the inference from data to claim belong (in which case economic forecasts belong to a different field than meteorological forecasts). Comment: This objection can be handled by clarification; one suspects that Toulmin would privilege the classification by discipline of the warrant, to judge by the way he distinguishes types of arguments in his examples and in his textbook.

2) He uses the term valid ambiguously, sometimes for formal validity, sometimes for the soundness or cogency or acceptability of an argument. Comment: This objection is simply unfair. Toulmin never uses the word validity on its own to mean formal validity; he always uses it to mean soundness or cogency (with respect to the inferential move from data to claim). If he means formal validity, he uses the expression formal validity. His thesis is that in most cases the validity of an argument (i.e. ultimately, the derivation of the conclusion/claim from the data and the backing) is not a matter of formal validity. Though he has an erroneous definition of formal validity, his point remains correct if one uses a correct definition.

Additions/corrections to the summaries of chapter 3:

Clarification (Khameiel, Nancy): The issue in Chapter 3 is not what structure arguments should have, but what layout of their components is most logically candid, in the sense of revealing the sources of their validity (95). Note that he is not rejecting the notion of logical form, but proposing a new notion that is more perspicuous, one based on a jurisprudential rather than a geometrical model.

What is the point of Toulmin's novel proposal for the layout of arguments? To justify his claim, developed through on the model of dialectical jurisprudence, that the standards for the appraisal of arguments are field-dependent, not a single field-invariant standard of formal validity.

In contrast to Perelman, whose approach is rhetorical (taking as the goal of argumentation to intensify or secure adherence of the audience) and empirical (basing a taxonomy on the types of premisses and argumentation schemes used in philosophy and literature and/or included in the taxonomies of the rhetorical tradition), Toulmin's approach is dialectical (to justify a claim to a single interlocutor who challenges it) and quasi-normative (cf. formal validity of "data; warrant; so conclusion" and justification of warrant by backing and of data by preliminary argument).

Define: claim, data, warrant, qualifier, rebuttal, backing. What is the function of each? (To what question from a challenger is it a response?) Give a new example of each.

C: Claim = what is involved in an assertion, what a speaker is challenged to establish when a challenger asks, "What do you have to go on?" (97, 98)

Clarification: not representative of the merits the arguer seeks to establish (Leslie)

D: Data (later grounds) = the facts adduced as support for the claim (97), the answer to the question "What do you have to go on?" (98)

W: Warrant = the inference-license that entitles us to infer C from D, most candidly expressed in the form, "Data such as D entitle one to make claims such as C"; the answer to the question "How do you get there?" (98)

Clarification: not support for the data (Leslie, Nancy), not data but inference-licenses (Nancy), not a missing premiss (Nancy), more candidly expressed as a general rule that from data such as D you may infer claims such as C rather than as a license to infer C from D (Patrick), Toulmin's point on page 100 is not that rational assessment of arguments in a field requires warrants but that it requires being prepared to work with warrants of some kind (Nancy), backing not warrant is field-variant (Leslie), grounds for acceptability of warrant rather than its acceptability is field-dependent (Nancy)

Q: Qualifier = an indicator of the strength conferred by the warrant on the step from data to claim (101) Examples: presumably, certainly, almost certainly

Clarification (Nancy): qualifier does not establish the relevance of the warrant and is primarily applied to the conclusion not the warrant,

R: Conditions of rebuttal = conditions "in which the general authority of the warrant would have to be set aside" (101)-distinguish overriding defeaters (Pinto) from undermining (Pinto) or undercutting (Pinto) defeaters

Clarification: Conditions of rebuttal don't rebut the warrant but rather override or undermine it in the particular case under discussion (Nancy), they don't necessarily override the claim but might undermine the inference (Patrick)

B: Backing = assurances without which the warrant would lack authority (103), varying in type from one field to another: e.g. system of taxonomical classification, statutes, statistics (104)

Note that warrants are not always appealed to explicitly; they are not the data on which we base our conclusion, but general principles which explain (and justify) our getting from the data to the conclusion.

Clarification: not necessarily specific facts, since it could be a system of taxonomic classification (Leslie, Patrick); backing and warrant not interdependent but warrant dependent on backing (Leslie), Toulmin's point on page 116 that logicians have been wedded to a certain form has to do with the form of statement 'All A's are B's rather than with the traditional premiss-conclusion analysis (Kelly)

Reflection (Leslie, Nancy): How can one distinguish backing from warrants?

Clarification/reflection (Kelly): What is Toulmin's point about the difference between force and backing (113)? How if at all is his point on page 14 about the difference between a singular premiss that expresses information and a universal premiss that expresses a guarantee related to the difference between force and backing?

Reflection (Leslie): Should warrant and backing be placed side by side to show their interdependence?

Clarification (Khameiel): Toulmin's point on page 112 about statements of the form 'All A's are B's" is that the form does not make clear whether the statement is functioning as a warrant or as backing. The habit that he decries is that of making a single statement of this form do double duty, which he complains is not logically candid.

Note too that Toulmin's point (p. 119) that any argument expressed in the form "Data; warrant; so conclusion" is formally valid gives us an important clue as to how we can identify the implicit warrant given an argument in the form "Data; so conclusion".

Points for discussion:

Reflection (Khameiel): The validity of a form of argument depends on context. The fact that a valid form works in one context does not show that it works in another.

Clarification/reflection (Kelly): Why does Toulmin say (p. 108) that "we are interested primarily in arguments by which particular propositions are applied to justify conclusions about individuals"?

Reflection (Nancy): "Overall, the syllogism and the warrant model seem to perform similarly." Is Toulmin then making a lot of fuss about nothing?

Reflection (Nancy): Toulmin's field-dependency thesis means that every argument occurs in a context (architectural, construction, inhabitant), which will influence the lines of qualification and rebuttal.

Reflection (Nancy): Toulmin is correct to point to arguments that are substantial and not analytic.

Wittgensteinian background to Toulmin's approach to the appraisal of arguments.

Reflection (Patrick): An advantage of Toulmin's model is that it avoids the regress argument of Lewis Carroll (in "What Achilles Said to the Tortoise") that not all inference-warrants can be explicit.

Reflection (Patrick): Toulmin's model complicates the analysis of arguments unnecessarily. If someone argues syllogistically, it is more straightforward to take the major premiss as a premiss than as a rule of inference.

Is Toulmin's model too burdensome to apply in practice? What is gained by this finer structure if the validity of an argument becomes indeterminate because of field-variability?

How do we determine whether a warrant is acceptable?

Is what counts as validity simply field-dependent?

If Toulmin's approach is basically rhetorical, judging success of an argument by its acceptance by the audience, how can he discuss the validity of arguments? Since he does not mean by validity formal validity, what does he mean by validity?

When is it necessary to make explicit the warrant or the backing?


van Eemeren, Frans, Rob Grootendorst and Francisca Snoek Henkemans. 1996. Fundamentals of Argumentation Theory: A Handbook of Historical Backgrounds and Contemporary Developments. Mahwah, NJ: Lawrence Erlbaum Associates.

van Eemeren, Frans, Rob Grootendorst, J. Anthony Blair and Charles A. Willard (eds.). 1992. Argumentation Illuminated. Amsterdam: SICSAT.

1. Stephen Toulmin, "Reasoning in theory and practice", in David Hitchcock and Bart Verheij (eds.), Arguing on the Toulmin Model: New Essays in Argument Analysis and Evaluation (Dordrecht, Springer, 2006), Ch. 2, 25-29.