Informal Logic and Argumentation

What it asks

Formal logic evaluates arguments whose every step is written down. Almost no argument anyone makes is like that.

Real arguments leave out what the audience already accepts, rely on what it would be unreasonable to doubt, and reach conclusions that are supported rather than entailed. Translating them into formal notation either fails or requires supplying so many missing premises that the reconstruction is the arguer's work rather than theirs.

So this branch asks a different question: what standards apply to arguments that are not deductions, and how do you assess one without first rewriting it into something it was not?

The recurring problems

Defeasibility. A good non-deductive argument can be entirely correct and still be overturned by new information. That is not a flaw in it, and no formal notion of validity captures it.

The missing premises. Every real argument has them. Deciding which to supply is interpretation, and a hostile interpreter can make any argument look invalid by supplying the wrong ones.

What a fallacy is. The standard list is ancient and unstable. Appeal to authority is a fallacy and is also how everyone learns almost everything. Slippery slope arguments are sometimes correct. Most named fallacies are patterns that are bad in some contexts and fine in others, and the list conceals that.

Whether standards are context-relative. What counts as sufficient support differs between a criminal trial, a medical decision and a dinner-table dispute. If the standards vary, saying what they answer to is difficult.

How it developed

The tradition is old. Aristotle wrote on rhetoric and on sophistical refutations as well as on the syllogism, and treated persuasion as a subject with its own standards.

Formal logic's success in the twentieth century pushed it aside almost completely. If validity is a formal property, argument that is not formally valid looks like a defective version of something rather than a subject.

Two books in 1958 reopened it. Toulmin's The Uses of Argument proposed a structure for real reasoning - claim, grounds, warrant, backing, qualifier, rebuttal - that fitted law and everyday dispute far better than the syllogism did. Perelman and Olbrechts-Tyteca's The New Rhetoric recovered the classical tradition as a serious discipline.

The 1970s critical thinking movement gave the field an institutional base by putting it in undergraduate curricula, and pragma-dialectics later supplied a rule-governed model of argument as a procedure for resolving disagreement rather than as a structure.

The positions within it

Toulmin's model treats arguments as field-dependent structures. Pragma-dialectics treats them as moves in a regulated discussion. Bayesian argumentation analyses them in terms of how much a consideration should shift confidence. Rhetorical approaches hold that audience and context are constitutive rather than incidental. All are in the Schools section.

Where it stands now

The field has grown because the problem has. Argument mining, automated fact-checking and content moderation all need machine-tractable standards for reasoning that was never formalised, and this is the only branch that has worked on that directly.

The Bayesian approach has been the most fruitful recently, since treating an argument as something that shifts confidence by a degree handles defeasibility naturally, where validity cannot.