Validity Turns a Good Argument Into More Than a Neat Form

Formal logic arrives with a severe little promise. Grant the premises, and the conclusion will behave.

Argumentation arrives afterward, carrying the premises by the collar. Where did these come from? Who said they were granted? Why is this conclusion in the room at all?

The difference is not a quarrel over whether rigor matters. It is a disagreement about where rigor finishes its shift.

Frege’s Begriffsschrift is a slim apparatus with a title long enough to require its own satchel.

It gave modern logic quantifiers over variables, built from negation, the conditional, and universal quantification. Aristotle’s syllogisms had been doing respectable service for two thousand years, and they could not express relations or nested quantifiers.

Frege gave thought a new set of plumbing.

Russell and Whitehead then tried to derive pure mathematics from logical primitives across the three volumes of Principia Mathematica.

The famous proof that one plus one equals two does not complete until the second volume. The result is occasionally useful. The patience required to obtain it is a separate metaphysical question.

This machinery is built for one clean question. Does the conclusion follow?

Set down two premises. All mammals are warm-blooded. Whales are mammals.

Now set down the conclusion: whales are warm-blooded.

The conclusion follows before either premise has been checked against a whale, a mammal, or a thermometer.

Form has preserved the commitment.

That is validity. No case can make the premises true and the conclusion false.

An ordinary sound argument needs more: valid form and true premises. Neither requirement sends a field researcher to find out whether whales are warm-blooded. The symbols have no field boots.

The available tests are wonderfully strict about their jurisdiction.

A truth table enumerates every assignment of truth values, doubling its rows with each new variable. It catches a counterexample with the impersonal satisfaction of a customs officer.

But it cannot display the inner structure of all or some. A three-circle Venn diagram handles categorical syllogisms. Relations, time, and multi-place predicates walk straight through the wall.

Natural deduction is the larger apparatus.

With introduction and elimination rules, assumptions, and their discharge, it constructs a licensed route from premise to conclusion. A proof has the homely decency of showing its work.

So the temptation is to let the derivation settle everything.

Put an argument into symbols, find a route through, and call the whole business sound. The newspaper columnist has become a theorem, and the premises have received their little formal hats.

It does not work.

Formalization decides what follows from a representation. It does not settle whether the representation captured an ordinary sentence, whether an omitted premise deserves admission, or whether the admitted premises have earned belief.

There may be a uniquely correct formalization for every important ordinary-language argument. That is speculation, not a fact any derivation has established.

Real arguments rarely present themselves as numbered premises wearing name tags.

Therefore, thus, so, and it follows that often announce a conclusion. Because, since, and given that often introduce support.

More often there is an enthymeme. Socrates is human, so Socrates is mortal, with “all humans are mortal” left downstairs holding the coats.

Standard form brings the premise upstairs, numbers the claims, and marks the conclusion last.

Then charity becomes necessary, which is awkward for anyone who had hoped logic would eliminate manners.

The principle of charity asks for the strongest defensible reconstruction the words permit, not the easiest object to destroy.

Dennett’s discipline asks an interpreter to state the position clearly, name agreements, say what has been learned, and only then criticize.

Past that point, charity becomes invention. A better argument, perhaps, but not the one anybody made.

The premise check is where the world returns with paperwork.

One standard asks whether premises are acceptable, relevant, and jointly sufficient. A date, a measurement, or a causal claim cannot be made true by a correctly placed arrow.

Logical entailment says what follows if the premises hold. Empirical confirmation asks whether they hold. Causal inference asks still another question: whether the alleged relation is doing the causing, or merely standing nearby looking correlated.

The burden of proof supplies the social geometry.

The burden lies with the one who asserts, not the one who denies.

But a courtroom’s presumption of innocence, a literature’s null hypothesis, and an everyday preference for the mundane are conventions of different rooms. They are not theorems from a common ceiling.

When a dispute becomes only a fight about who must establish what, argumentation theory calls it burden tennis. Nobody has won. The ball has simply stopped meaning anything.

Toulmin’s model makes the missing furniture visible: claim, grounds, warrant, backing, qualifier, rebuttal.

It is made for the arguments that dominate ordinary life, whose warrants are usually unspoken and whose exceptions turn up late, carrying receipts.

Formal logic can assess a stated deductive skeleton exactly. Argumentation asks whether the skeleton belongs to the creature under discussion, and whether it can stand.

The formal story is not therefore a toy.

Gödel proved first-order completeness before he was twenty-five. In a fixed first-order system, semantic entailment and derivability coincide. If a formula is valid, some finite proof exists.

The guarantee is extraordinary. It has nothing to say about whether a journalist’s evidence was responsibly gathered.

It also does not say that logic is a vending machine for answers.

Church and Turing showed independently that no algorithm always decides arbitrary first-order validity.

A proof search can eventually find a proof for a valid formula. For an invalid one it may keep searching into the heat death of the departmental server.

Completeness and decidability are not synonyms, however much their names appear to have conspired over lunch.

No psychological finding is doing the central work here.

There is no experiment whose replication rate decides whether a conclusion is entailed, and no p-value that converts a causal hypothesis into a premise.

The distinction is conceptual. A valid inference can carry a false empirical claim with flawless discipline. An empirical finding can be well supported without becoming deductively entailed by its background.

What the two disciplines share is not that informal argumentation is logic with feelings, or that symbols have finally disciplined conversation.

It is that both insist commitments must travel without changing their passports.

Formal logic asks whether the route preserves them. Argumentation asks whether the traveler was entitled to board.

One guards the shape of the step. The other makes the ground answer for it.