Critical thinking (CT) is a reliable method for deciding what to believe and what to do.

A method is a set of steps for getting a desired result.

• A person reads competing claims about the age of the Earth, 6,000 years versus 4.6 billion years, and wants to decide which is more likely correct.
• A juror wants to decide correctly whether the accused is guilty.
• A company wants to decide which marketing plan is best.
• A shopper wants to determine whether organic food is healthier.
• A mechanic wants to figure out what is causing an engine noise.
• An oncologist needs to say which treatment is best for a patient.
• A homeowner wants to decide which house paint is best.
• A family wants to figure out whether to pay off the mortgage or invest the money.
• A group of friends wants to choose the best place for a ski trip.
• An engineering team wants to figure out the best way to solve a technical problem.

In cases where CT is useful, a person or group wants to get a claim right.

1. Explain critical thinking well enough that others can learn to apply it.
2. Make or accept only claims that are likely to be:
   1. True, for claims about how things are
   2. Right, for claims about values or actions

• Right actions
  • You and your groups raise your odds of taking the right action when you make a choice, solve a problem, or plan, such as the right job, treatment, or plan.
• True beliefs
  • You and your groups believe more of what’s true and less of what’s false.
• Not fooled
  • Scams, misinformation, and honest mistakes have less hold on you and your groups.
• Trusted by others
  • Claims from you and your groups hold up, so others listen when you speak.

• Time and effort
  • Checking a claim takes longer than accepting it.
• Discomfort
  • Sometimes you have to accept what you wish weren’t true, take an action you’d rather not take, or say “I don’t know yet.”
• Standing apart
  • Your groups may believe or do things you don’t. Going your own way can make you an outlier and cost you comfort, standing, or friendships.
• Can’t check everything
  • You have to pick which claims are worth checking.

• Every day, many times a day, since you’re always choosing what to believe and do.
• Most when a lot rides on a claim: your health, money, safety, job, or vote.
• When a group decides what to do or believe, such as a family, team, company, or country. More people bear the results, and groupthink makes mistakes easier to miss.
• When someone wants something from you: your money, your trust, or your agreement.

• Everyone, since everyone decides what to believe and what to do.
• Most of all, people whose choices affect others: parents, managers, doctors, jurors, voters, and leaders.
• Groups that want to make good decisions together: families, teams, companies, and governments.

An exam question says, “A fund fell 50% last year and rose 50% this year. What is the net return?”

1. Key Question: Over the two years, what is the fund’s net return, as a percent of its starting value?
2. Standards: The answer will be a T-claim, so use the standard for any T-claim. There is no standard to create for this case.
3. Information: The question gives two facts: the fund fell 50% in year 1 and rose 50% in year 2. You bring one principle: a percent change is a share of the value at the start of that change.
4. Logic and Argument: Start with 100 dollars. A 50% fall leaves 50 dollars, and a 50% rise on 50 dollars gives 75 dollars, which is 25 dollars less than the start. So over the two years, the fund lost 25% of its starting value.
5. Evaluation: Judged by the standard for any T-claim, the conclusion holds. The facts are correct and complete: both percentages are given, and the principle is what a percent change means. The answer follows from them: 0.5 × 1.5 = 0.75, so the result is the same for any starting value. The rival answer, “back to even,” fails because it treats both percentages as shares of the same amount, and they aren’t. The claim is yours and it holds, so state it: the net return is a loss of 25%.

You are on a jury. A man is charged with stealing a laptop from his office.

1. Key Question: Did the accused take the laptop?
2. Standards: The answer will be a T-claim, so use the standard for any T-claim. There is no standard to create for this case, but the court sets how sure you must be: sure beyond a reasonable doubt.
3. Information: His badge opened the office door at 7 p.m. A coworker saw him leave that evening carrying a laptop bag, but she saw a bag, not the laptop, and from across a parking lot. He often works late, and he owns a laptop bag. The door log shows that two other people came in that evening. The laptop was never found.
4. Logic and Argument: The badge record shows that he was there and could have taken the laptop. The door log shows the same for two other people. The bag fits a theft, and it also fits an ordinary evening for someone who carries a laptop bag. Nothing ties the laptop itself to him. So the information doesn’t show, beyond a reasonable doubt, that he took it.
5. Evaluation: Judged by the standard for any T-claim, the conclusion holds. The facts are reliable, and they include the ones that count against a theft. The conclusion claims no more than the facts show. The rival answer, “he did it,” fails because it takes for granted that nobody else could have, and two other people could. The claim is yours as a juror, so act on it: vote not guilty. If the laptop turned up in his car, you would think again.

You are saving for retirement and have to choose where to put the money. A coworker says Bitcoin beats an index fund.

1. Key Question: For my retirement savings, which is the better investment, Bitcoin or an index fund?
2. Standards: The answer will be an S-claim, so use the standard for any S-claim, and find a standard for this case: a good retirement investment grows over decades, with little risk that the money isn’t there when you need it.
3. Information: Over the past ten years, Bitcoin’s price has grown far more than the stock market. It has also fallen by more than half several times. An index fund spreads your money across hundreds of companies, and Bitcoin is one asset. Nobody knows what either will do over the next thirty years.
4. Logic and Argument: Compare each choice to the standard. On growth, Bitcoin has done better so far, but past growth doesn’t promise future growth. On risk, the index fund is far safer, because its swings are smaller and it doesn’t depend on one asset. A retirement investment has to do well on both, so the index fund is the better choice.
5. Evaluation: Judged by the standard for any S-claim, the conclusion holds. The standard for this case fits the purpose, the facts are reliable, and the conclusion comes from comparing both choices to that standard. The rival answer, “Bitcoin is better,” fails because it treats “has grown the most” as the whole standard. The claim is yours, so act on it: put your retirement savings in the index fund. Someone with a different purpose, such as investing money they can afford to lose, would need a different standard and could fairly decide otherwise.

Your electric bill keeps rising, and an installer has given you a quote for solar panels.

1. Key Question: Should we install solar panels on our house?
2. Standards: The answer will be an S-claim, so use the standard for any S-claim, and create a standard for this case: the panels should pay for themselves while we own the house.
3. Information: The installer quotes 18,000 dollars and says the panels will save 1,800 dollars a year. The installer profits from the sale, so you check that figure. Your past bills and a neighbor’s panels point to savings of about 1,500 dollars a year. You expect to stay in the house about 8 years.
4. Logic and Argument: At 1,500 dollars a year, the panels take 12 years to pay back 18,000 dollars. You expect to own the house for 8 years, and 12 is more than 8. So the panels wouldn’t pay for themselves, and you shouldn’t install them now.
5. Evaluation: Judged by the standard for any S-claim, the conclusion holds. The standard for this case fits what you care about, the savings figure was checked against sources with nothing to sell, and the arithmetic follows. The rival answer, “install them,” fails because it takes the installer’s figure and a long stay for granted. The claim is yours, so state it no more strongly than its grounds support: “We shouldn’t install solar panels unless we plan to stay at least 12 years.” Then bring it to your family.

The same recipe gives results like these.

• A supplement ad
  • “This pill improves memory” turns out to rest on one small study paid for by the maker, so you withhold judgment and keep your money.
• Two job offers
  • Once you say what matters most to you, such as pay, growth, or time at home, one offer comes out clearly ahead.
• A news story
  • “Crime doubles” turns out to mean a rise from 3 cases to 6 in one neighborhood, so you accept the count and reject the alarm.

• Each starts with a question, not a claim.
• Each follows the same five steps.
• Each names the type of claim before any information is gathered, so you know what the answer will be judged by. An S-claim also needs a standard for the case.
• In each, the answer is the conclusion of an argument built from the information.
• In each, the rival answer failed because of something it took for granted.
• Each ends with a claim you can act on, held as firmly as the grounds allow.

Nobody gets good at CT by reading about it. You get good by running the recipe many times and improving each time.

1. Try: Run the recipe on a real claim.
2. Look back: See how your judgment held up, and find the step where things went wrong.
3. Improve: Fix that one thing, then run the recipe again on a new claim.

Repeat the loop until you reliably meet the standard.

A TwF (task with feedback) is a task you attempt, paired with feedback that compares your attempt with a standard.

• TwFs are how you run the loop: you work through a claim, find your gaps, close them, and repeat until you reliably meet the standard.
• They show whether you’ve learned: you meet the standard when you can pass the Mastery TwFs on your own, months after you stop practicing.
• They keep you from fooling yourself. Reading about CT makes it feel easy; a TwF shows whether you can do it.
• They save time. You work on your gaps, not on what you already know.

TwFs come in four levels, from a first try to the full standard: Basics, Apply, Range, and Mastery.

• Claims are how people say what is true, what is better or best, and what should be done and by whom.
• Your beliefs and choices rest on claims, and claims can be false, so spotting them shows you what to check.

1. Explain what a claim is, and how it differs from a question, a command, or an exclamation, well enough that others can learn to apply it.
2. Spot the claims in what you read and hear, such as an ad, a news story, or a conversation, and split each compound claim into its atomic claims.

The key question is the main question at issue, which a claim or argument sets out to answer.

A claim can only be judged against the question it answers, so naming that question keeps you from judging it on the wrong issue.

1. Explain the key question, and how it differs from the topic, well enough that others can learn to apply it.
2. State the question that a claim answers, such as the question behind an ad or a disagreement with a friend, so you judge the claim against the right question.

An argument is a set of claims in which some (the reasons, also called premises) are offered as support for another (the conclusion).

Most claims come with reasons, and seeing the argument shows you what the claim rests on, so you can check that and not just the claim.

1. Explain what an argument is, and how it differs from a quarrel, well enough that others can learn to apply it.
2. Find the argument behind a claim and pick out its reasons and conclusion, even when the conclusion comes first or isn’t marked.

A reason is a claim offered as an answer to the question, “Why believe the conclusion?”

Only reasons support a conclusion, so separating them from the rest keeps you from being swayed by background, repetition, or a good story.

1. Explain what a reason is, and how it differs from background, a restatement of the conclusion, and the story of how someone came to believe it, well enough that others can learn to apply it.
2. Pick out the reasons in an argument, and leave out everything that isn’t one.

Standard form is a way of writing an argument that states the key question, the conclusion, and each reason on a line of its own.

In standard form, an argument shows at a glance what is claimed and why, so gaps and weak reasons have nowhere to hide.

1. Explain standard form well enough that others can learn to apply it.
2. Write an argument in standard form, so that a reader sees at a glance what is claimed and why.

Stating someone else’s argument is putting it into standard form so that it says what the arguer meant.

When you state an argument as the arguer meant it, you judge what they really said, and they’ll listen when you answer it.

1. Explain stating someone else’s argument, and how it differs from restating it in a weaker form, well enough that others can learn to apply it.
2. State the argument in an ad, an article, or a conversation so that the arguer would agree it is theirs, recording only what they said.

• You know how to settle it: with logic or evidence, not with anyone’s taste.
• A disagreement has an answer: if two people disagree about a T-claim, at least one of them is wrong, so it’s worth finding out who.
  • “Agree to disagree” and “that’s just your opinion” don’t fit a T-claim.
• You can often look it up: many T-claims have been checked by people who study them, such as doctors, engineers, and historians, so you can check what they found instead of arguing.

1. Explain T-claims well enough that others can learn to apply them.
2. Say whether a claim you read or hear, such as a news headline or a friend’s remark, is a T-claim, in clear cases.

• You know to find the standard first: an S-claim can’t be checked until you know which standard it’s judged by, so you ask, “By what standard?”
• You stop arguing about the wrong thing: many disagreements are about standards, not facts. Two people can agree on every fact and still disagree.
• You can disagree fairly: someone who reaches a different S-claim may have a different purpose, not worse facts.
  • Example: someone investing money they can afford to lose can fairly pick a different investment than someone saving for retirement.
• But not anything goes: once the standard is named, the claim can be checked, and the standard itself can be judged by whether it fits the purpose.

1. Explain S-claims well enough that others can learn to apply them.
2. Say whether a claim is an S-claim, in clear cases, and name the kind of standard it rests on, even when the speaker leaves it unstated.

Most claims are easy to sort. This lesson is about the ones that aren’t.

• The type picks the method: sort a claim wrong and you judge it by the wrong standard. You look for evidence where a standard is needed, or treat a fact as a matter of taste.
• You see what a disagreement is about: facts, standards, or both.
• You aren’t fooled by wording: ads and arguments dress S-claims as facts (“experts agree it’s the best”) and T-claims as opinions (“that’s just your view of the data”).

1. Explain how to tell T-claims from S-claims well enough that others can learn to apply it.
2. Classify any claim as a T-claim or an S-claim, including look-alikes such as reports of values, claims that name their standard, and compound claims that mix the two.

• It tells you the method: check the logic of a definition-type claim; weigh the evidence for an evidence-type claim.
• It tells you how sure you can be: a definition-type claim whose logic checks out is certain. An evidence-type claim can only be likely, so you hold it as firmly as its evidence supports.
• You don’t demand proof where only evidence will do, and you don’t settle for evidence where proof is possible.

1. Explain definition-type and evidence-type T-claims well enough that others can learn to apply them.
2. Name the sub-type of any T-claim, and say whether it can be certain or only likely.

• Each needs different checks: a value claim needs a standard for the case; an action claim also needs the real options and the facts about each.
• You find the claim behind the claim: most action claims rest on a value claim, and checking it shows where you agree and where you don’t.
• You catch a push to act: ads and advice often state an action claim and skip the comparison it rests on.

1. Explain value-type and action-type S-claims well enough that others can learn to apply them.
2. Name the sub-type of any S-claim, and for an action claim, state the value claim it rests on.

• You can use what others write: books, articles, and other CT courses use these terms instead of T-claim and S-claim.
• You catch loose matches: some of these terms don’t mean quite what a T-claim or S-claim does, and mixing them up leads to wrong sorting.

1. Explain normative and non-normative claims well enough that others can learn to apply them.
2. When a source uses one of these terms, say which type of claim it means, and notice when the match is loose.

Logic is the study of the principles that determine whether conclusions follow from premises.

Knowing what logic can and can’t settle keeps you from looking for proof where only evidence or a standard can decide, and from missing it where logic can.

1. Explain logic well enough that others can learn to apply it.
2. Tell when logic alone settles a claim (definition-type T-claims) and when the claim goes to evidence or a standard for its type.

Modus Ponens is the rule of inference that from “if A, then B” and “A,” you may conclude “B.”

Modus Ponens is the rule behind most conclusions you draw, including the last step of every evaluation, so using it well keeps your conclusions sound.

1. Explain Modus Ponens, and how it differs from affirming the consequent (“if A, then B; B; so A”), well enough that others can learn to apply it.
2. Use Modus Ponens to show that a conclusion follows from its premises, and spot arguments that only look like it.

An argument is valid when its conclusion must be true if its premises are true.

A valid argument carries the truth of its premises to its conclusion, so checking validity tells you whether checking the premises is enough.

1. Explain validity, and how it differs from soundness (valid with true premises), well enough that others can learn to apply it.
2. Judge whether an argument is valid.

An assumption is a premise that an argument relies on but doesn’t state.

Arguments often rest on premises nobody states, and the unstated one is often the weakest, so finding it shows you what really needs checking.

1. Explain what an assumption is well enough that others can learn to apply it.
2. Spot the assumptions an argument depends on, in your own arguments and in others’.

A definition is a statement of what a term means.

A claim’s key terms decide what it says, so defining them makes the claim clear and keeps people from arguing past each other.

1. Explain what a definition is well enough that others can learn to apply it.
2. Make a claim clear by defining its key terms, such as what “net return” means in a claim about a fund.

Deriving a claim from definitions is showing, step by step, that it follows from definitions and given facts by logic.

When definitions and given facts settle a claim, deriving it gives you a certain answer instead of a guess.

1. Explain deriving a claim from definitions well enough that others can learn to apply it.
2. Derive a claim, or show that it can’t be derived, such as the net return of a fund that falls 50% and then rises 50%.

A counterexample is a single case that shows a general claim is false.

One counterexample is enough to show that a general claim is false, which makes it the quickest way to test one.

1. Explain what a counterexample is well enough that others can learn to apply it.
2. Test a definition-type claim by looking for a counterexample, such as starting with 100 dollars to test “a 50% fall and a 50% rise bring you back to even.”

The pattern of an evaluation is the form that every evaluation shares: a deductive argument in which one reason must be checked against definitions, evidence, or a standard, depending on the type of claim.

Every evaluation has the same shape, so once you know the pattern, you know what to check for any claim.

1. Explain the pattern of an evaluation well enough that others can learn to apply it.
2. Lay out the evaluation of any claim in this pattern, and say which reason needs checking and what to check it against.

Probability is how likely a claim is to be true, from 0 (certainly false) to 1 (certainly true).

Few claims are certainly true or false, so saying how likely a claim is lets you match your belief to its grounds instead of forcing a yes or no.

1. Explain probability well enough that others can learn to apply it.
2. Estimate how likely a claim is to be true and state it as a probability.

The required probability is the bar you set for how likely a claim must be before you accept it.

Setting the bar before you look keeps you from accepting a risky claim on thin grounds, and from demanding certainty when little is at stake.

1. Explain the required probability, and how it differs from the probability of the claim, well enough that others can learn to apply it.
2. Set the required probability for a case, higher when more rides on being wrong, such as “beyond a reasonable doubt” in a trial.

Withholding judgment is declining to accept or reject a claim until there are better grounds.

Withholding judgment when the grounds are weak keeps you from believing too soon and from rejecting too soon.

1. Explain withholding judgment, and how it differs from rejecting a claim, well enough that others can learn to apply it.
2. Withhold judgment when the grounds for a claim are too weak to accept or reject it.

Deciding on a claim is choosing, after evaluating it, to accept it, reject it, or withhold judgment.

An evaluation pays off only when it ends in a decision you can act on, and saying what would change it keeps you open to new information.

1. Explain deciding on a claim well enough that others can learn to apply it.
2. Decide on a claim once you have evaluated it, and say what new information would change your decision.

Information is the facts, definitions, and principles that bear on the answer to a key question.

An answer can be no better than the information behind it, so getting good information is where a sound claim starts.

1. Explain information well enough that others can learn to apply it.
2. Get the information a key question needs, and judge whether it is high-quality.

A standard is a description of what is most desirable in something.

Every value or action claim is judged against a standard, so stating the standard shows what the claim rests on and where people disagree.

1. Explain what a standard is well enough that others can learn to apply it.
2. State the standard behind a value or action claim, such as what makes one job offer better than another.

The standard for any T-claim is a description of what is most desirable in a T-claim, whatever it is about.

One standard applies to every T-claim, whatever it is about, so learning it once gives you a check for any claim about how things are.

1. Explain the standard for any T-claim well enough that others can learn to apply it.
2. Judge a T-claim against the standard for any T-claim, such as the answer to an exam question or the charge in a jury trial.

Framing your key question is turning what you want to know into one clear question that a claim can answer.

Starting with a question instead of an answer keeps you from settling on a conclusion before you’ve looked.

1. Explain framing your key question well enough that others can learn to apply it.
2. Frame a high-quality key question for something you need to decide or find out, before you settle on an answer.

Getting information is finding the facts, definitions, and principles that your key question needs.

Your claim can be no better than the information behind it, and seeking information against the answer you expect keeps you honest.

1. Explain getting information well enough that others can learn to apply it.
2. Get high-quality information for your key question, including information that counts against the answer you expect.

Building your argument is organizing your information into reasons that lead to an answer to your key question.

Building your argument shows whether your answer really follows from what you know, before anyone else has to point it out.

1. Explain building your argument well enough that others can learn to apply it.
2. Build an argument in standard form from your information, so that the answer follows from the reasons.

Checking your own claim is evaluating it by the same standard you would use for a claim from someone else.

Your own claims are the easiest to get wrong because you want them to be true, so checking them as strictly as anyone else’s keeps you from fooling yourself.

1. Explain checking your own claim, and how it differs from looking for support for what you already believe, well enough that others can learn to apply it.
2. Check your own claim before you hold or state it, looking hardest where you most want it to be true.

Wording a claim to fit its grounds is stating it so that it says no more than its reasons support.

A claim that says more than its grounds support is likely false even when its core is right, and people stop trusting you when your claims overreach.

1. Explain wording a claim to fit its grounds well enough that others can learn to apply it.
2. Reword a claim that says too much, such as changing “we shouldn’t install solar panels” to “we shouldn’t install them unless we plan to stay at least 12 years.”

Choosing which claims to check is deciding which of the claims you meet deserve close checking.

You can’t check every claim you meet, so choosing well puts your effort where being wrong would cost the most.

1. Explain choosing which claims to check well enough that others can learn to apply it.
2. Choose which of the claims you meet to check closely, and let the rest go without accepting or rejecting them.

The interest of a source is what the source stands to gain or lose if you accept its claim.

A source that gains when you accept its claim may be biased or may mislead you, so knowing its interest tells you how closely to check.

1. Explain the interest of a source, and how it differs from the source being wrong, well enough that others can learn to apply it.
2. Name what a source stands to gain, such as an installer who profits from a sale, and check its claim more closely when the interest is large.

Relying on an expert is accepting a claim you can’t check yourself because a qualified source makes it.

You can’t check most claims yourself, so relying on the right experts lets you believe more truths than you could check alone.

1. Explain relying on experts well enough that others can learn to apply it.
2. Decide when to rely on an expert, such as a doctor or a scientist quoted in the news, and when not to.

A fallacy is a common pattern of reasoning that seems persuasive but doesn’t support its conclusion.

Fallacies persuade without supporting their conclusions, so spotting them keeps you from being moved by bad arguments, including your own.

1. Explain what a fallacy is, and how it differs from an argument with a false premise, well enough that others can learn to apply it.
2. Spot common fallacies in the arguments you meet, your own included, and say why they fail.

A threat to sound judgment is anything that pushes people to accept or reject a claim on grounds other than its merits.

Biases, misinformation, and pressure tactics push you to accept or reject claims on grounds other than their merits, often without your noticing.

1. Explain threats to sound judgment, and how they differ from good reasons you dislike, well enough that others can learn to apply them.
2. Spot threats to sound judgment, such as biases, misinformation, and pressure tactics, in your own thinking and in claims from others, so they don’t sway what you accept.