Aristotle's rhetorical proofs and Toulmin's model of argument

Intuition Beginner

When a mathematician proves a theorem, certainty is the goal: every step must follow by rule from the one before. A persuader faces a different task. The audience will not wait for certainty, and a decision must be reached before all the evidence is in. The study of how reason works under that pressure — convincing without proving — is the art of argument, and Aristotle named its toolkit rhetoric.

Aristotle observed that a persuader has three means of proof to build with. Ethos establishes the speaker's character and credibility. Pathos puts the audience into the frame of mind that will receive the argument. Logos supplies the reasoning itself — the example and the enthymeme, rhetoric's versions of induction and deduction. These three are the artistic proofs: the speaker constructs them. Evidence merely handed over — a contract, a witness — Aristotle called inartistic.

Stephen Toulmin, twenty-three centuries later, asked what an everyday argument is made of once you crack it open. He found six recurring parts. A claim is what you want accepted; data are the facts you cite; a warrant explains why those data support the claim. Behind the warrant stands a backing, a qualifier says how strong the claim is, and a rebuttal names the exception. Mapping an argument onto these six parts shows exactly where reasonable people part company.

Visual Beginner

Toulmin's layout of an argument — six parts, several of them often unstated:

   DATA  ──"so"──>  CLAIM
     |                 ^
     | ("since")       |  QUALIFIER: probably / necessarily / presumably
     v                 |
   WARRANT  <─"on the strength of"─  BACKING        REBUTTAL: "unless..."

Aristotle's three species of rhetoric — match the argument to its occasion:

   DELIBERATIVE   future   urge / warn against     end: the expedient
   FORENSIC       past     accuse / defend          end: the just
   EPIDEICTIC     present  praise / blame           end: the honourable

The three artistic proofs — what the speaker builds (not what is found):
   ETHOS (character)    PATHOS (emotion)    LOGOS (reason)

Worked example Beginner

Here is a one-sentence argument from a town-hall debate:

"The council should approve the solar contract this Tuesday, because the vendor locks in a fixed price of 4 cents per kilowatt-hour for twenty years while our current rate has climbed 18 percent in three years."

This is a deliberative argument: it urges a future action (approve, this Tuesday), so its proper end is what is expedient for the town. We decompose it with the Toulmin model.

• Claim — the council should approve the solar contract Tuesday.
• Data — the bid fixes 4 cents per kilowatt-hour for twenty years; the current rate is up 18 percent in three years.
• Warrant — a long-term fixed price below a rising trend is a financially prudent deal a council ought to accept.
• Backing — councils owe taxpayers stable, predictable energy costs.
• Qualifier — almost certainly.
• Rebuttal — unless the vendor's credit is weak or the contract hides escalator clauses.

The dominant proof is logos: the figures do the work. A skeptic who grants the data can still attack the warrant, arguing that solar prices keep falling, so a twenty-year lock-in is not prudent after all. The model locates that disagreement instead of leaving it as a vague clash of opinions.

Check your understanding Beginner

Formal definition Intermediate+

Rhetoric, in Aristotle's definition, is "the faculty of observing, in any given case, the available means of persuasion" ^{[Aristotle 350 BCE]}. The means divide into two kinds. Inartistic proofs (pisteis atechnoi) are evidence the speaker merely uses — laws, witnesses, contracts, documents; they are found, not made. Artistic proofs (pisteis entechnoi) are constructed by the speaker, and Aristotle assigns them three families.

The three artistic pisteis are ethos, the character the speaker exhibits; pathos, the emotional disposition into which the audience is brought; and logos, the argument proper. Aristotle locates the persuasive power of ethos in the speaker's appearing to possess phronesis (practical wisdom), arete (virtue), and eunoia (goodwill toward the audience); a deficiency in any one undermines the other two. Pathos is not decoration but the alignment of the audience's judgment with the situation, since the judgment of a frightened or angry or ashamed audience differs from that of a calm one.

The three species (eide) of rhetoric are fixed by the audience and the temporal stance of the speech. Deliberative rhetoric (symbouleutikon) addresses an assembly deciding a future action; its end (telos) is the expedient and the harmful. Forensic rhetoric (dikanikon) addresses a court judging a past deed; its end is the just and the unjust. Epideictic rhetoric (epideiktikon) addresses a spectator of a present person or act, assigning praise or blame; its end is the noble and the base. The species are not interchangeable: a forensic proof misapplied in a deliberative setting misses the question the audience is actually deciding.

Each species draws its arguments from a stock of repeatable inference-templates called topics (topoi) or commonplaces. The koinoi topoi, the common topics of the Rhetoric, include the more-and-the-less, the possible-and-impossible, past fact and future fact, degree, and definition; from these a speaker generates lines of argument before the specific case is known. A special topic (idios topos) belongs to a single field — the topics of law, of politics, of ethics. The topics are thus a generative apparatus: standing loci from which warranted transitions can be assembled on demand.

Two vehicles carry the logos. An enthymeme is the rhetorical syllogism: a deduction whose premises the audience supplies or accepts on the strength of probability and signs, so that it is typically compressed, omitting a premise the audience can reconstruct for itself. An argument from example (paradeigma) is the rhetorical induction: a particular case or parallel is offered to ground a general claim. Aristotle calls the enthymeme the body of persuasion (soma tes pisteos) and example its chief support, and both are defeasible where a demonstrative syllogism is not.

Stephen Toulmin's layout of argument ^{[Toulmin 1958]} refactors this vehicle into six components. An argument is a tuple

$$(C,D,W,B,Q,R)$$

in which the data $D$ are put forward in support of a claim $C$; the warrant $W$ is the license authorizing the step from $D$ to $C$; the backing $B$ is the field-specific authority on which $W$ rests; the qualifier $Q$ fixes the modal force — necessarily, probably, presumably — of the supported claim; and the rebuttal $R$ names the conditions under which $W$ fails to license the conclusion. In symbols the warrant licenses the inference $D{\Rightarrow }_{Q}C$ unless $R$ obtains. The warrant is itself a claim open to examination, and different argument fields admit different warrants.

The field-invariance thesis is Toulmin's central philosophical claim: the form of the layout — the six components and their relations — is invariant across argument fields, while the content of the warrants and backings is field-dependent. The warrant that counts in a courtroom (statutory authority) differs from the warrant that counts in a laboratory (experimental reproducibility) or in ethics (consistency with a principle). A modal term, the force of the argument, distinguishes a conclusion reached necessarily from one reached only probably, and Toulmin insists that this force be recorded as a component of the analysis rather than treated as an afterthought.

Perelman and Olbrechts-Tyteca's new rhetoric ^{[Perelman 1958]} adds the audience structure that classical rhetoric had left largely tacit. A particular audience is the concrete group addressed here and now. The universal audience is a constructed ideal — the collectivity of all reasonable people — used as a standard to separate arguments that depend on local commitments from those any competent reasoner should grant. The universal audience is a regulating ideal rather than an empirical population; the strength of an argument varies with the audience to which it is addressed, and an argument that compels only a sect or party is weaker than one that would compel the universal audience.

Rhetorical analysis Intermediate+

Close reading: a deliberative argument

Consider the claim "The council should adopt a four-day school week, because the three districts that piloted it cut transport costs by twenty percent without losing instructional hours." The species is deliberative: it urges a future action, and its end is the expedient. Decomposed by Toulmin, the claim is adoption of the four-day week; the data are the twenty-percent saving and the intact instructional hours observed in three pilots; the unstated warrant is that a policy replicated successfully in comparable districts should be adopted locally; the backing is a fiscal-and-pedagogical standard of evidence the council accepts; the qualifier is probably; and the rebuttal is that local logistics — rural routes, working-parent childcare — may defeat the comparison. The dominant proof is logos, carried by an argument from example that generalizes from three pilots to a rule.

The common topics in action

The same debate can be generated from the common topics rather than from a single example. The more-and-the-less yields: if the county funds the larger library expansion, then a fortiori it should fund the smaller branch repair. The possible-and-impossible yields: if peer districts reversed a comparable decline, then this district can as well. The topic of degree yields: the savings, while real, are modest enough that they must be weighed against offsetting costs. The topics make the invention of arguments systematic: each commonplace is a slot into which the facts of the case can be fitted to produce a warranted transition.

Field-dependence: where the disagreement lives

Return to the solar example. The Toulmin model predicts that data are testable while warrants are ideological: two reasonable readers can accept identical rate figures yet disagree on whether a twenty-year lock-in is prudent. The warrant a financial officer applies (locking in sub-market cost is sound treasury management) differs from the warrant a procurement lawyer applies (a single-vendor, no-exit contract is presumptively suspect). Same layout, same data, different fields, opposite conclusions — and the model converts that clash into a precise, arguable locus rather than a vague impression of disagreement.

Enthymeme versus example

An enthymeme and an argument from example can reach the same claim by different routes, and the route matters for how the argument is attacked. "This solar deal should be approved, because a fixed sub-market price is prudent" is an enthymeme: a suppressed major premise (fixed sub-market prices are prudent) is reconstructed by the audience. "This deal should be approved, because the three comparable towns that took similar deals all saved money" is an argument from example: it generalizes from particulars. The enthymeme is defeated by challenging the warrant; the example is defeated by producing a counter-example. Recognizing which vehicle carries the logos tells the analyst where to press.

Bridge. This six-part anatomy builds toward the modern argumentation-theory claim that reasoning is field-dependent yet structurally uniform, and it appears again in the formal study of inference, where the warrant is recast as a rule of derivation. This is exactly the move that generalises from a single persuasive paragraph to the whole taxonomy of argumentation schemes; putting these together, the central insight is that an argument is not a claim plus noise but a licensed transition whose license can be named, tested, and retracted. The bridge is that every discipline is, at bottom, a community that agrees on which warrants count, and the foundational reason the apparatus travels from law to science to ethics is that the six-part form is field-invariant even when the warrants it licenses are not.

Exercises Intermediate+

Advanced results Master

Classical rhetoric was treated for centuries as a closed inheritance from Aristotle, Cicero, and Quintilian, but the twentieth century rebuilt argument theory three times over — as Perelman and Olbrechts-Tyteca's new rhetoric, as Toulmin's philosophy of argument, and as the contemporary schools of pragma-dialectics, informal logic, and argumentation schemes — and the three reconstructions together define the modern field.

Perelman and Olbrechts-Tyteca's Traité de l'argumentation ^{[Perelman 1958]} restored rhetoric as a theory of reasoning that is neither formal deduction nor raw persuasion. Beyond the universal-versus-particular audience distinction, their apparatus sorts arguments into two families. Liaison arguments link elements the audience already accepts to a conclusion (a quasi-logical argument wraps a substantive claim in mathematical-seeming form, such as an appeal to a rule of justice treating like cases alike; an argument grounded in the structure of reality trades on accepted causal or hierarchical relations; an argument that establishes the structure of reality appeals to a model case). Dissociation arguments split a single concept into a term-I (the appearance) and a term-II (the deeper reality) to resolve a tension the liaison cannot reach — as when "justice" is dissociated into mere legal conformity and real equity. The distinction between liaison and dissociation is the principal technical innovation of the Traité and the resource that lets argument analysis describe conceptual change, not merely the derivation of conclusions from fixed premises.

Toulmin's model, treated above as an analytic tool, is also a philosophical argument against the positivist identification of genuine reasoning with formal proof ^{[Toulmin 1958]}. Its dual thesis — field-invariant form, field-dependent warrants — was directed against the positivist claim that ethics, policy, and historical judgment are non-cognitive because they fail to meet demonstrative standards. Toulmin's later work generalized the point: the modalities (necessary, possible, impossible, probable) are not logical decoration but constitutive of how arguments in different fields carry force, and a theory of argument must record modal force as a first-class component. The practical consequence is that a strong argument in constitutional law and a strong argument in particle physics share a form while differing in every warrant, and both are genuine reasoning.

The pragma-dialectical school of Frans van Eemeren and Rob Grootendorst ^{[van Eemeren 2004]} recasts argument as a regulated critical discussion aimed at the resolution of a difference of opinion. A discussion proceeds through four stages — confrontation (the dispute is identified), opening (the parties and their shared starting points are fixed), argumentation (the defense of a standpoint is mounted and tested), and concluding (the outcome is established) — and is governed by rules distributing the burdens of proof and prohibiting force, ad hominem moves, misrepresentation, and the shifting of the goalposts. On this account a fallacy is not a named rhetorical sin but a violation of a rule that obstructs the resolution of the dispute; the analysis is functional rather than taxonomic, and it unifies the classical catalogue of fallacies under a single normative standard.

Douglas Walton's theory of argumentation schemes ^{[Walton 2008]} supplies the defeasible inference-forms that classical rhetoric called topics and that pragma-dialectics leaves unspecified. Each scheme — argument from expert opinion, from position-to-know, from consequences, from analogy, from commitment, from popular opinion — is paired with a set of critical questions that must be answerable for the scheme to license its conclusion. The scheme for expert opinion, for instance, licenses the claim only if the source is a competent authority on the point, the point falls within that competence, and there is no equally authoritative dissent; an unanswered critical question defeats the inference. The schemes are defeasible: they confer provisional assent, never certainty, and they provide a principled stopping rule for how much testing an argument must survive.

The informal-logic tradition of Ralph Johnson and J. Anthony Blair evaluates arguments by three criteria applied to the premise–conclusion relation. Acceptability asks whether each premise is rationally acceptable to the audience; relevance asks whether the premises bear on the truth of the conclusion; sufficiency asks whether the premises, taken together, supply enough support. A defective argument violates one or more of these, and the fallacies are mapped onto the violations: an ad hominem attacks relevance, a hasty generalization fails sufficiency, a false premise fails acceptability. The three criteria are the field's working replacement for the single formal-logic notion of validity, and they apply directly to the Toulmin layout as a test of whether the data, the warrant, and its backing jointly suffice for the qualified claim.

Synthesis. Putting these together, the long arc of argumentation theory runs from Aristotle's three artistic proofs and the common topics, through Toulmin's field-invariant layout of claim, data, warrant, backing, qualifier, and rebuttal, to Perelman's universal audience, the pragma-dialectic critical discussion, Walton's schemes with their critical questions, and the informal-logic criteria of acceptability, relevance, and sufficiency. This is exactly the central insight that unifies them: argument has a stable form that generalises across two and a half millennia and across every field from law to science to ethics, while the warrants it admits are dual to the community that accepts them. The foundational reason the apparatus still teaches is that it is the bridge between a claim a speaker holds and the audience whose assent is sought, and the bridge is that every genuine argument is a licensed transition that can be named, examined, and retracted when its license fails.

Connections Master

• This unit is the argumentation-theory deepening of the rhetoric-and-composition overview, and it presupposes that overview's account of the rhetorical situation and the five canons as its starting frame 22.05.01 pending. Where the overview maps the field, this unit reconstructs the internal anatomy of a single argument and the modern schools that govern its evaluation.

• The warrant in a Toulmin layout is the informal cousin of a rule of inference, and the questions of when an everyday argument counts as valid, when a premise is acceptable, and when support is sufficient belong equally to argumentation theory and to the philosophy of language and informal logic 20.12.01.

• Rhetorical proofs do their work through language, and the literary chapter's treatment of tone, mood, and figurative language shows how pathos and compressed logos are carried by diction and figure rather than by explicit statement 22.03.07.

• Every argument is finally made of sentences, so the clause-structure analysis of the grammar sequence underwrites the construction of compound enthymemes that a reader can reconstruct without losing the thread 22.01.16.

Historical & philosophical context Master

Western argument theory begins in the Greek law courts and assemblies of the fifth century BCE, where citizens pleaded their own causes and teachers — the Sophists — sold instruction in speaking well. Plato's Gorgias attacks those teachers for making the weaker argument seem the stronger, fixing "sophistry" as a term of abuse; the Phaedrus concedes that rhetoric aimed at truth could be a genuine art. Aristotle's Rhetoric (c. 350 BCE) ^{[Aristotle 350 BCE]} is the constructive answer: a systematic treatise that treats persuasion as a neutral technique, distinguishes artistic from inartistic proofs, names the three pisteis, fixes the three species by audience and time, and catalogs the common topics from which enthymemes are generated.

The Roman orators consolidated the inheritance. Cicero's De Inventione and Topica transmitted the topics and the stasis system; Quintilian's Institutio Oratoria grounded the whole curriculum in the moral character of the speaker. The medieval schools preserved this apparatus as one of the liberal arts, but the early modern period fractured it. Peter Ramus's sixteenth-century reassignment of invention and arrangement to logic left rhetoric with only style and delivery, and the scientific revolution's prestige of demonstration reduced the classical theory of probable reasoning to ornament. By the nineteenth century, the study of substantive argument had retreated to the margins of logic, surviving mainly as composition pedagogy.

The field's modern revival has two independent sources in 1958. Toulmin's The Uses of Argument ^{[Toulmin 1958]} argued, against the positivist tradition, that the reasoning of law, science, and ethics shared an invariant form despite differing warrants, and that formal logic could not exhaust the rational. In the same year Perelman and Olbrechts-Tyteca's Traité de l'argumentation ^{[Perelman 1958]} reconstructed rhetoric as a theory of the values and hierarchies that reasoning under uncertainty actually invokes, anchored in the universal audience. The convergence of two such different projects on the legitimacy of non-demonstrative reasoning reopened argument as a philosophical domain.

The late twentieth century built the schools that now dominate the field. Pragma-dialectics, developed in Amsterdam by van Eemeren and Grootendorst, formalized argument as a critical discussion governed by rules; informal logic, developed in Canada by Johnson and Blair, supplied the acceptability–relevance–sufficiency criteria; Walton's argumentation schemes, developed across four decades, catalogued the defeasible inference-forms with their critical questions and reshaped the analysis of fallacy and of legal and political reasoning. The recurring philosophical question is whether rhetoric is a legitimate art of reasoning or a technique of manipulation aimed at the irrational part of the audience; the Aristotelian answer — that persuasion operates in the unavoidable space where demonstration cannot reach, because decision must precede certainty — remains the working defense, and the modern schools confirm it by showing that the analysis of defeasible argument admits norms every bit as precise as those of formal proof.

Bibliography Master

1. Aristotle. Rhetoric (c. 350 BCE). Trans. W. Rhys Roberts. The artistic versus inartistic proofs; the three species (deliberative, forensic, epideictic); the common topics; the enthymeme and example as vehicles of logos.
2. Toulmin, Stephen E. The Uses of Argument. Cambridge University Press, 1958. The claim–data–warrant–backing–qualifier–rebuttal layout, and the field-invariance and field-dependence theses.
3. Perelman, Chaim, and Lucie Olbrechts-Tyteca. The New Rhetoric: A Treatise on Argumentation. University of Notre Dame Press, 1969 (trans. of the 1958 Traité de l'argumentation). Universal versus particular audience; liaison and dissociation; quasi-logical arguments.
4. van Eemeren, Frans H., and Rob Grootendorst. A Systematic Theory of Argumentation: The Pragma-Dialectical Approach. Cambridge University Press, 2004. The critical-discussion model, its four stages, and the ten rules.
5. Walton, Douglas, Christopher Reed, and Fabrizio Macagno. Argumentation Schemes. Cambridge University Press, 2008. The catalogue of defeasible schemes and their matching critical questions.
6. Johnson, Ralph H., and J. Anthony Blair. Logical Self-Defense. McGraw-Hill Ryerson, 1977 (and later editions). The acceptability–relevance–sufficiency criteria for premise–conclusion support.
7. Corbett, Edward P. J. Classical Rhetoric for the Modern Student. Oxford University Press, 1965. The re-importation of the Aristotelian apparatus, including the topics and Toulmin, into modern composition.
8. Burke, Kenneth. A Rhetoric of Motives. University of California Press, 1969. Identification and consubstantiality as an alternative to the persuasive model.

@book{aristotle_rhetoric,
  author    = {Aristotle},
  title     = {Rhetoric},
  year      = {c. 350 BCE},
  translator= {W. Rhys Roberts},
  note      = {Books I--III; artistic vs inartistic proofs, the three species, the common topics, the enthymeme and example}
}

@book{toulmin1958uses,
  author    = {Toulmin, Stephen E.},
  title     = {The Uses of Argument},
  publisher = {Cambridge University Press},
  year      = {1958},
  note      = {The claim--data--warrant--backing--qualifier--rebuttal layout; field-invariance and field-dependence}
}

@book{perelman1969new,
  author    = {Perelman, Chaim and Olbrechts-Tyteca, Lucie},
  title     = {The New Rhetoric: A Treatise on Argumentation},
  publisher = {University of Notre Dame Press},
  year      = {1969},
  note      = {Translation of the 1958 Trait\'e de l'argumentation; universal vs particular audience, liaison and dissociation}
}

@book{vaneemeren2004systematic,
  author    = {van Eemeren, Frans H. and Grootendorst, Rob},
  title     = {A Systematic Theory of Argumentation: The Pragma-Dialectical Approach},
  publisher = {Cambridge University Press},
  year      = {2004},
  note      = {The critical-discussion model, four stages, and the ten rules}
}

@book{walton2008schemes,
  author    = {Walton, Douglas and Reed, Christopher and Macagno, Fabrizio},
  title     = {Argumentation Schemes},
  publisher = {Cambridge University Press},
  year      = {2008},
  note      = {The catalogue of defeasible schemes with matching critical questions}
}

@book{johnson1977logical,
  author    = {Johnson, Ralph H. and Blair, J. Anthony},
  title     = {Logical Self-Defense},
  publisher = {McGraw-Hill Ryerson},
  year      = {1977},
  note      = {Acceptability, relevance, and sufficiency as criteria of argument adequacy}
}

@book{corbett1965classical,
  author    = {Corbett, Edward P. J.},
  title     = {Classical Rhetoric for the Modern Student},
  publisher = {Oxford University Press},
  year      = {1965}
}

@book{burke1969rhetoric,
  author    = {Burke, Kenneth},
  title     = {A Rhetoric of Motives},
  publisher = {University of California Press},
  year      = {1969},
  note      = {Identification and consubstantiality}
}