Liberal Arts and Reasoning: Start Here

A map of the wing: why a modern learner might study reasoning, reading, number and making together, the ten-step operating loop, and which page to read first.

A router drops off the network. You restart it, the lights come back, and you call it fixed. Three days later it drops again. The restart showed one thing: the fault involves some state a restart clears. Calling it a cure skipped a step, the one that checks whether the world did what you expected.

That gap between knowing a method and using it is what this wing trains. It treats logic, reading, number and craft as one practice instead of four school subjects. The aim is a mind that is hard to deceive, and hands that can build something and explain how it works.

Most of what follows is a method, not settled doctrine. The one-line summaries of thinkers below are standard readings of their works. Only the Popper line has been checked against the text here. The way this wing arranges them into a programme of study is the author's design. It is offered as a Provisional framework you can test on yourself.

What the wing is for

The old phrase "liberal arts" meant the arts that suit a free person. In that spirit, the wing aims at intellectual freedom. You should be able to check a claim yourself and spot an argument that is a trick. You should be able to decide calmly when others are alarmed. The second aim is practical. A person who can only argue is as incomplete as a person who can only build. The Philosopher-Engineer page develops that pairing.

Knowing methods versus using them

A person can recite the vocabulary of Bayesian updating, falsification and formal proof and still be poor at an unfamiliar problem. Knowing about methods is not the same as using them under pressure. The deeper skill is to meet a confusing situation and work out what kind of problem it is. Then you represent it so its structure shows, propose explanations, and work out what each predicts. Then you test, and revise without defensiveness.

The core cycle underneath all of this is simple. You observe. You form a model. The model creates an expectation. You compare it with what happens. Where they differ you have found an error, and you revise the model. This is a summary of how the thinkers below describe inquiry, not a measured result. That practising it deliberately makes you better at it is Provisional.

One machine for reducing error

Seen this way, the great traditions are parts of one error-reducing machine. This is an analogy, not a historical claim that these thinkers were collaborating.

Socrates cross-examines beliefs until contradictions appear.
Euclid builds chains of necessity from stated premises.
Aristotle asks what a thing is, what causes it and what follows from its nature.
Bacon names the habits of mind that corrupt observation.
Hume shows how much everyday inference exceeds what reason can strictly prove.
Peirce describes inquiry as abduction, deduction and induction working together.
Popper insists that a theory counts as empirical science only if experience could refute it.
Bayes gives belief revision a quantitative form.
Pearl separates seeing an association from intervening on a cause.
Simon shows that real reasoners work under limits of time and information.

The operating loop

Ten verbs make a loop you can run on any problem. Each step also prevents a characteristic error.

Observe. What you do: Record what was seen, without explaining it. Error it prevents: Treating an interpretation as a fact.
Define. What you do: Say exactly what the terms mean. Error it prevents: Arguing past each other over one word.
Model. What you do: Represent only what matters. Error it prevents: Drowning in detail, or hiding structure.
Hypothesise. What you do: Generate several competing explanations. Error it prevents: Falling for the first plausible story.
Predict. What you do: State what each explanation forbids. Error it prevents: Explanations that cannot lose.
Test. What you do: Choose the check that splits the candidates. Error it prevents: Confirming only what you already believe.
Update. What you do: Let the evidence move your confidence. Error it prevents: Defending a view out of pride.
Act. What you do: Act in proportion to knowledge and stakes. Error it prevents: Paralysis, or reckless certainty.
Verify. What you do: Check that the world responded as expected. Error it prevents: Assuming the fix worked.
Learn. What you do: Record the gap between expectation and result. Error it prevents: Repeating the same mistake.

Run it on a silent guitar amp. Observe: write "no sound from the amp", not "the amp is dead". Hypothesise: the cable, the guitar's volume knob, the amp itself. Predict: a spare cable brings the sound back only if the cable was the fault. Test that first, because it is cheap and splits the candidates. Verify: play for ten minutes, because a loose jack can work and then fail again. Learn: note which part failed, so next time you check it first.

The first three steps are developed in Observe, Define, Represent. Network faults put the whole loop under pressure in Evidence-Based Troubleshooting, whose changed case asks what a restart that restores service for a while actually shows.

Neither antiquarianism nor presentism

Read only old books and you risk treating them as ornaments. You also miss tools their authors never had, such as probability theory and computing. If you use only current techniques, you inherit their fashions and cannot judge them. This wing pairs each ancient idea with a modern counterpart so each corrects the other.

Nine powers that correct one another

A well-rounded programme of study develops nine powers, from ends and proof to computation and moral imagination. Each asks its own question and covers a blind spot in the others. The full list, with how each power checks the rest and where the framework could be wrong, is in The Nine Powers.

Three kinds of page, and where to start

Pages here are of three kinds: ideas (what a thinker claimed), methods (what to do), and maps (how topics relate). A sensible order:

The Philosopher-Engineer: Study and Workshop Together (idea): the ideal behind the wing.
The Weekly Loop: Read, Reconstruct, Make, Defend, Log (method): the weekly practice. Begin here if you want to start doing something today.
How to Read a Great Book: Levels, Modes and the Completion Standard (method): how to read difficult texts.
Observe, Define, Represent: The First Three Moves (method): the first three moves.
A Map of Mathematics: Twenty-Five Areas and What Depends on What (map): where number fits, and what depends on what.
Peirce on the Fixation of Belief (idea): a first primary text to practise on.

How to use the wing

Hold one text, one question and one made thing at a time. Finish the made thing before opening another text. A week with something produced beats a week of wide reading.

Connections across the site

The wing links outward. The monochord and the Pythagorean comma show number and music meeting. Euclid's postulates and the elenchus are hands-on entries to proof and questioning. The Canon of sourced quotations gives primary-source wording you can check.

Where this could be wrong

The strongest objection. Real problems do not arrive in ten tidy steps. A good technician hears a symptom and jumps straight to the likely cause. Walking ten steps for a blown fuse wastes time. And ten is arbitrary: the same process could be cut into four steps or fifteen.

The best reply. Read the loop as a checklist of guards, not a march. The expert who jumps ahead still predicts and verifies, only faster. The right-hand column of the table is the useful part: when a fix fails, it tells you which step you skipped.

What would settle it. Nothing here has tested whether running the loop on purpose improves anyone's judgment, so the claim stays Provisional. A small test: for one week, log each problem, the step you skipped, and whether that step bit you. If skipped steps never matter, the loop is overhead for you.

Try this

Take one belief you hold about your work or health. Write the observation behind it in one sentence, free of interpretation.
The next time something at home stops working, such as the Wi-Fi or a printer, write one line per step before you fix it. Circle the step you were tempted to skip.