The Last Page

You can finish a book about making better decisions without making a single decision. This is a peculiar feature of something you might have bought to change how you think.

Suppose you read a chapter about sunk costs. It explains why money already spent shouldn’t determine whether you keep spending. Someone sits through a terrible film because they’ve paid for the ticket. You see the mistake immediately. By the end of the chapter, the idea seems almost too obvious to need explaining.

Now suppose you’ve spent six months building a product nobody wants. You’re deciding whether to give it another month. This time, the problem arrives as a promising conversation with a potential customer, a feature you haven’t tried, and the unpleasant prospect of telling people you’ve stopped.

Continuing might be the right decision. That’s what makes this difficult. You have to separate evidence that another month would be worthwhile from your desire for the previous six months to have been worthwhile.

The cinema example didn’t ask you to do that. The author had already done it.

This is an awkward possibility: some of the understanding we experience while reading belongs to the author. They’ve chosen the relevant facts, identified the principle, and shown how one follows from the other. We can follow their reasoning. But how much of it could we have produced?

Andy Matuschak raises a related problem in “Why books don’t work.” Books leave much of the work of learning to the reader, including noticing confusion and checking understanding. Reading can proceed quite smoothly without those checks. The absence of difficulty doesn’t tell you whether you’ve understood.

It helps to separate a few things that the word “know” conceals. You might recognize an explanation when you see it. You might be able to explain it yourself. You might also notice, in an unfamiliar situation, that this is the explanation you need. These are different accomplishments.

A chapter heading quietly removes the last difficulty. Inside a chapter about sunk costs, you know what to look for. Outside the book, you have to decide whether the problem concerns sunk costs, insufficient evidence, impatience, or something else entirely. Nobody tells you which chapter you’re in.

This makes books about judgment particularly hard to assess. If you read a book about repairing a bicycle, the bicycle offers a fairly direct opinion of what you’ve learned. Decisions are less obliging. A good decision can turn out badly. A foolish one can succeed. You can even use a newly acquired concept to defend what you wanted to do anyway.

Someone determined to quit can invoke sunk costs. Someone determined to continue can invoke perseverance. A larger vocabulary of principles might improve their reasoning, or merely make their explanations more impressive.

That doesn’t mean the principles are useless. It means being able to explain a decision in their language is weak evidence that they improved it.

The obvious objection is that books really do teach people things. A sentence can resolve a confusion you’ve carried for years. An account of an unfamiliar life can change what you notice about your own. You needn’t recall every argument for a book to have mattered.

And sometimes an explanation is enough. If you already understand a subject and are missing one connection, a paragraph can supply it. It would be strange to insist that nothing was learned because all you did was read.

So “books are bad for learning” is too broad. The more interesting question is why reading can produce such a convincing sense of progress when the result remains uncertain.

Part of the answer may be that reading gives us something definite to finish. Before starting, there are three hundred unread pages. Each evening, there are fewer. Whatever is happening to our understanding, something measurable is happening to the bookmark.

The book also supplies its own endpoint. The author has finished explaining, so we feel we’ve finished learning. But those two events needn’t coincide.

This matters most when the reason for reading is practical. If you read for pleasure or curiosity, reaching the last page may be a perfectly satisfactory conclusion. If you read to become better at making decisions, the book’s ending settles very little.

You may now understand the person who should have left the cinema. Whether you understand the project you keep giving another month is still an open question.