85 Tools of Philosophical Argument

What puts a method in this section

A method is here because it can be shown working. The spine of the section is the worked argument: every entry takes its method and runs it on a real question — a proposition, the steps in order, and where it lands. Modus ponens is not defined and left there; it is put to work and made to produce a conclusion you can agree with or reject.

That test was chosen over importance because it is checkable. A claim that a method matters cannot be verified or refuted; a claim that a method takes these steps and yields this conclusion can be, and the argument either runs or it breaks in front of you.

The three shapes

Every entry carries one of three shapes, and the numbers are what they are rather than what would look tidiest. Forty-three are sequences — procedures with steps, performed in order. Thirty-one are inferences, where the move is a form and the form can be written out in numbered premises. Eleven are devices, and they are the odd ones: Ockham's razor, the burden of proof, the veil of ignorance. A device is not a procedure at all. It is a rule about what an argument is allowed to do, and it is filed as such rather than dressed up as a method to keep the columns even.

The six bands

Forms of inference · Classical procedures · Analytic tools · Continental procedures · Non-Western procedures · Devices and constraints.

The Non-Western band holds fifteen entries and is not an appendix to the others. Anumāna, the Indian inference, runs to five steps where Aristotle's syllogism runs to three, and it was set out around 200 BCE, a century and a half after the Prior Analytics, with no evidence that either side knew of the other. The catuṣkoṭi allows four positions where the logic of the first band allows two. These are not variants of the Greek material. They are separate solutions to the same problem, and placing them in a band of their own is the only arrangement that does not quietly make one of them the standard.

The oldest and the newest

Neti neti — the discipline of saying what a thing is not, because saying what it is will always say too much — is in the Upanishads by roughly 700 BCE. Conceptual engineering, which stops asking what a concept means and asks what it is for, arrives in 2000. Twenty-seven centuries separate them, and almost everything between is still in use: only two entries in the section are marked historical rather than live.

Where it fails

Every entry carries a block under that heading, and it is the block to read first if you are deciding whether to trust a method. It names the objection, names who raised it, and gives the case that raised it — Hume against induction, Duhem and Quine against the clean refutation, the false dilemma against the dilemma. A method you cannot break is a method you do not understand, and an atlas that presented eighty-five techniques as eighty-five reliable instruments would be lying about all of them.

What the section does not claim

It is a working set, not an inventory. Philosophy has no agreed list of its own methods, and any eighty-five are a selection with a hand behind it.

The line between a method and a doctrine is genuinely unclear in places, and where an entry sits close to it that is said in the entry rather than hidden. Dialectical materialism is a method of reading history and also a claim about history. The veil of ignorance is a device for testing rules and also an argument for particular ones.

Some words do double duty and the section does not tidy that away. Epoché names two different moves here — the sceptic's suspension of judgement and Husserl's bracketing of the world — and each entry says plainly which one it is and that the other exists.

Dates, and why some of them are hedged

Many of these dates mark when a method was first described, not when it was first used. Aristotle did not invent deduction in 350 BCE; he was the first to write down what people had been doing for centuries, and the entry says so. Where a date is a scholarly reconstruction rather than a record it is given as approximate, and it never carries an arithmetic that it cannot bear.

The arguments, and where they come from

The worked arguments are written for this Atlas. They are illustrations of the move rather than the historical example, except where an entry says otherwise — and where the historical case is the point, it is named and dated.

Entries are written against the standard reference literature, the Stanford Encyclopedia of Philosophy, the Internet Encyclopedia of Philosophy and Britannica, and checked against the primary texts wherever a method has a locus classicus. Corrections are welcome through the contact page.