Applied Argumentation (Capstone Strand)

This is Unit Nine of Logic: Applied Argumentation. It is the last one. No new machinery arrives here. Every technique in this lecture was built somewhere in the previous eight units, and the work now is arranging them into an order that survives contact with real prose — which is not the order the course taught them in. That order is the unit's first idea and it is stricter than it looks. Reach for a test before you have finished reconstructing, and you will be testing something the author never wrote — getting a clean verdict on an argument nobody made. Then the standards for what you may take for granted, who owes support to whom, and two checklists for the things arguments most often borrow: reported numbers, and cited experts. And a way of diagnosing bias that never requires guessing at anyone's motives. The unit ends by running the whole course backwards — using the same apparatus to build an argument rather than to take one apart. By the end you will be able to do five things. One: run the reconstruction pipeline in order, and say why the order is not optional. Two: tell a rebuttal from an undercutter, and say what each one establishes. Three: ask the four questions of a reported number, and the four of a cited expert. Four: diagnose motivated reasoning from asymmetric standards. Five: build an argument form-first, and pre-check the twin of the form you chose.

The unit opens by putting every technique in the catalogue into a fixed sequence. The steps themselves are familiar. What is new is the insistence on the order, and the reason for it is specific rather than tidy-minded.

The sequence.

Logic 426 The Reconstruction Pipeline

Find the conclusion, strip the filler, lay out the premises, supply what was left unsaid, decide which kind of argument it is, then test. Move the testing earlier and you deliver a verdict on something the author never actually said.

Two entries about the version of a position you choose to engage. Unit One introduced charity as a principle of interpretation; here it becomes a matter of self-interest, and then gets pushed past where politeness would ever have taken it.

Why charity pays.

Logic 427 Charity Under Fire

Beat the strongest version and it stays beaten. Beat the weakest and it is rebuilt by lunchtime. This is efficiency rather than good manners — and an audience always notices which version you chose to fight.

And further still.

Logic 431 Steelmanning

Build the opposing case better than the person making it did, then take that on. The test of having understood a position is stating it so that its holders say yes, exactly — anything short of that is a version of your own.

Two entries on obligations. Validity testing takes the premises as given, and at some point the assuming has to stop — that is the difference Unit One drew between valid and sound, now a working procedure.

Grant, or flag.

Logic 428 Premise Acceptability

Common knowledge and properly cited findings can be granted; contested claims get flagged rather than quietly waved through. Silently granting a disputed premise is how a sound-looking argument gets built on nothing.

Who owes the support.

Logic 429 Burden of Proof

Whoever asserts something owes support for it, and a bolder claim owes more. Demanding to be disproved is Unit Eight's appeal to ignorance in procedural dress — and burdens follow the claim, not whoever spoke most recently.

One distinction that prevents a very common overclaim. There are two quite different things you can do to an argument, and only one of them says anything about whether the conclusion is true.

The two moves.

Logic 430 Rebuttals vs. Undercutters

Saying the witness lied does not say the defendant is innocent — it removes one line of support and leaves the question open. Demolishing an argument for something is not establishing its opposite, and treating it that way is a routine overclaim.

Unit One's diagrams, enlarged to the size of real writing — and the interesting part of a large map turns out to be where nothing is drawn at all.

Scaling the map up.

Logic 432 Mapping at Essay Scale

An essay is a structure with intermediate conclusions as its junctions. Most contain far less argument than they contain prose, and the blank regions are the assertions nobody ever supported.

One paragraph through the pipeline, and there is a distinction inside it worth holding carefully — because two of the steps look identical from the outside and do opposite things to the argument's strength.

Watch it run.

Logic 433 Full Reconstruction Walkthrough

Find the conclusion, cut the filler, lay the premises out, add what was assumed, map it, and classify it. Cutting filler and cutting a premise you dislike are different operations — filler offers no support, and a weak premise offers some.

Two checklists for the two things arguments borrow most often. Both work the same way: four questions, asked before the borrowed item is allowed to do any work in the argument.

Four questions for a number.

Logic 434 Statistical Claims in Media

Where did it come from, who was measured, how common was it to begin with, and against what is it being compared. A risk that doubled from one in a million reverses its own headline once the fourth question is asked.

Four checks on an expert.

Logic 435 Evaluating Expert Testimony

In their own field, where they sit relative to their colleagues, what their record is, and what they stand to gain. A genuine split among specialists is information about the state of the field, not permission to take whichever side suits.

A way of identifying bias that never requires knowing anything about the person. Motives are not observable and arguing about them goes nowhere; the standards someone applies are on the page and can be compared.

What to look for.

Logic 436 Motivated Reasoning, Structurally

Watch how the same study design is treated when it agrees and when it does not — waved through in one case, taken apart in the other. The finding is the asymmetry itself; nothing here is a claim about what anyone was thinking.

Now the course runs backwards. Everything so far has been for taking arguments apart; the same apparatus builds them, and the order inverts — you settle the conclusion first, then choose a form capable of carrying it.

Construction, inverted.

Logic 437 Building Your Own Argument

Fix what you want to establish, pick a valid form, then supply the premises that form demands and check you can actually defend them. Most weak arguments were never designed at all — they accumulated.

Watch one get built.

Logic 438 Constructing a Deductive Argument

Choose the conclusion, choose a form that will carry it, draft the premises that form requires, then stress-test each one separately. A valid form promises that true premises would deliver the conclusion and promises nothing about whether yours are true.

Two capstones rather than one, because the unit taught two skills. The first takes a complete essay apart with whatever machinery each argument in it calls for. The second builds a position from nothing and defends it.

The first. A four-day week.

Logic 439 Capstone: Full Essay Analysis

Map the essay whole, reconstruct each argument inside it, then test each with the apparatus its type requires — categorical, propositional, statistical. Most of the arguments surviving does not make the essay sound; the main conclusion inherits the weakest support it actually rests on.

The second. Extended library hours.

Logic 440 Capstone: Construct and Defend

Stake the claim, construct the main argument form-first, build the strongest objection to it and answer that, then support one statistical premise properly. Answering an objection is not something added at the end — an objection found afterwards usually means rebuilding.

The last entry in the course puts the whole dependency graph on screen, with Unit One at its root and this entry as the final node — and then says the only sensible thing about what happens next.

The final entry.

Logic 441 The Map of Everything

Four foundational moves, nine units, one graph, and you have been through all of it. The catalogue finishes here and the practice does not — every argument you meet from now on is the examination.

Five things you can now do, and the course is complete. You can run the pipeline in order — conclusion, filler, premises, missing pieces, classification, test — and you know why testing early produces a clean verdict on an argument nobody made. You can tell a rebuttal from an undercutter, and you know that destroying the argument for something leaves the something entirely open. You can ask the four questions of a number — source, sample, base rate, compared to what — and the four of an expert: their field, their colleagues, their record, and their interests. You can diagnose motivated reasoning without claiming to know anyone's motives, by comparing how congenial and uncongenial evidence get treated. And you can build rather than only test: claim first, then a form to carry it, then premises you can defend, with the invalid twin checked before you publish.