The Antilibrary: Why You Should Buy Books You Will Never Read

Books. What I read, what I buy, and what I never get to.

I am at the till with four books and I know I will read one of them. Maybe one and a half. This is not a lapse of control, not guilt, and not a collector's tic. It is the purchase I regret least often.

At home they stand there. Spine by spine, read next to unread, unmarked, unsorted by progress. Sooner or later somebody comes over, looks up the wall, and asks the question everybody asks.

Have you read all of these?

No. And the question is the wrong one.

A shelf is not a certificate. It is a tool. A certificate shows what you can do. A tool shows what is still possible.

The question that gives the game away

Umberto Eco owned around 30,000 books. Nassim Taleb turned that into a term in The Black Swan: the antilibrary. It means the unread part, and the claim is that the unread part is where the value sits.

I do not want to retell that, I want to use it. Because the interesting bit of Eco's story is not the number. It is that he sorted his visitors by their reaction: some are impressed by the quantity, the others get that a private library is not an ego-boosting appendage but a research tool.

That sorting still works, and it works on yourself. If you look at your shelf and think "I got through all of that", you own a trophy cabinet. If you think "there is plenty in there I cannot reach yet", you own a workshop.

The Japanese have a word for the practice: 積ん読, tsundoku. Buying books and letting them pile up. It shows up in a satirical text as early as 1879, and it is still meant without malice. We are the ones who treat the pile like an unpaid invoice.

A book is an option, not a task

Here is the framing that settled the whole thing for me.

A book at 25 euros is not a purchase of reading time. It is the purchase of an option: the right, on some unknown day in the future, to reach for something immediately instead of searching, ordering and waiting for it then.

Options usually expire worthless. That is not a defect in the system, that is how the instrument works. Nobody calls an insurance policy a failure because the house did not burn down. With books we invert that logic anyway and call every unread copy a small defeat.

Do the arithmetic. Forty books a year is about 1,000 euros. You read twelve. Read as "I wasted 28 books", you burned 700 euros. Read as "I bought forty options and exercised twelve", you have a 30 percent hit rate on positions that can never cost more than 25 euros each.

And now the side nobody writes down: what does the book you did not buy cost you?

Not 25 euros. It costs you the question you never asked. That cost appears nowhere. You do not notice it, you cannot put a number on it, so you quietly set it to zero. Which is exactly why people buy too few books, not too many.

One side of the ledger is visible and small: 25 euros. The other is invisible and unknown in size. Optimise on the visible side alone and you optimise yourself into a corner.

The shelf · scroll

On the shelf12

Read12

Reached into the unread0

A shelf holding only what you have read is finished. It cannot tell you anything you do not already know. The twelve hatched spines stay twelve, however large the case gets. What grows is the part that can still answer back. The whole stock pays for itself on the day you reach into the row you never read.

"Then buy it when you need it"

That is the best objection to everything above, and it holds right up to the question of how you are supposed to know that you need it.

Amazon will deliver any book by tomorrow. What Amazon cannot deliver is the occasion. You can only order what you can name, and you can only name what you have already brushed against somewhere. A search is an index onto your existing vocabulary. It does not extend it.

A shelf works the other way round. It sits in your field of view without being asked. You walk past it twenty times a day, your eye catches titles you bought three years ago and have not touched since, and eventually one of those titles collides with a problem you happen to have. That is not luck. That is one hit per thousand glances, at a thousand glances a month.

A folder holding 400 EPUBs produces zero such encounters. Not few. Zero. It has no periphery. It shows you a list when you open it, and nothing at all otherwise.

The periphery · scroll

Because you asked1

In view1

Without you asking0

Left, the search: one question, one hit, blank paper around it. An index returns only what you were able to name. Right, the same person in front of the same quantity of books, but visible. Twelve of them reach them unasked. That is the only difference between a shelf and a folder, and it is the difference that matters.

That is the entire difference, and it is a functional one, not a romantic one. This is not about the smell of paper. It is about a book you see without looking being able to offer you a question you did not have.

What the e-book does well, and what it cannot do

I use e-books. Travelling, for 900-page bricks, for anything I want to search rather than read, for reference texts I have to quote. They are cheaper, instant, weightless, and the notes export. On those points the e-book wins outright.

It loses on two, and both are structural.

First, the periphery. See above. An e-book exists only while it is open. The rest of the time it is a filename nobody reads.

Second, ownership. A purchased e-book is not yours the way a book is yours. In 2009 Amazon remotely deleted copies of Orwell's 1984 from its customers' Kindles because the seller did not hold the rights. Money back, book gone, along with the annotations of a high-school student who sued over it. The irony was so complete that Bezos called the move "stupid" himself. In February 2025 Amazon then switched off "Download & Transfer via USB", the feature that let you keep your purchased titles as a file.

None of this is a conspiracy. It is simply what follows from having bought a licence rather than an object. Nobody can delete an object out of your flat.

And one more thing that cuts against the usual story: the e-book did not win. In 2025 its share of the German consumer book market was 6.3 percent. After a decade and a half. That is not a gap still closing, that is an equilibrium. Two formats, two jobs.

The summary was never the value

Now the part that touches my day job.

I build systems that summarise text. They work well enough by now that the summary as a product is dead. Any book will give you its core theses, its structure, its three best quotes and a critique in eight seconds. For free.

So when somebody argues that a shelf is redundant because the knowledge is cheaper and faster elsewhere, they are right about the knowledge. The knowledge was never the scarce part.

A language model answers the question you ask. It is brilliant at that and it gets more brilliant every year. But it sits behind an input field, and that field is empty until you type into it. It cannot offer. It cannot nudge. It does not know that on a Tuesday in October you will walk past a spine that names your current problem in a word you have never used.

The answer got cheap. The question did not. A shelf is a stock of questions you set out for yourself years ago.

Which is why buying a book you will not read is not a nostalgic act right now. It is the rational one. The moment every answer is available on demand, the ability to arrive at the right question becomes the actual bottleneck. And you do not train that at an input field. You train it on material you did not order.

How to tell this is not superstition

I like arguments that survive contact with numbers, so here is the most relevant study.

Sikora, Evans and Kelley analysed PIAAC data across 31 societies in Social Science Research in 2019. The finding: growing up with books in the home predicts measurably better adult literacy, numeracy and technological problem solving. And it does so on top of what parental education and your own qualifications already explain. The effect is loglinear, with the biggest jump in the first few metres of shelf.

Read honestly, that is a correlation, not causation. Households with books differ in many ways. But the study controls for the obvious suspects and the effect survives. That is good enough for a decision that costs 25 euros.

Related, since format is the subject: the meta-analysis by Delgado and colleagues (2018, 54 studies, more than 170,000 participants) finds a small but stable advantage for paper over screen in comprehension of expository texts. Not for narrative texts. And the advantage has not shrunk over the years, which kills the "digital natives will grow out of it" line.

Five rules I buy by

One: buy against yourself, not for yourself. A shelf full of books that confirm your view is expensive wallpaper. The only reason to own a book whose thesis you reject is the best reason there is.

Two: buy from a position of ignorance. The right buying signal is not "I want to read that", it is "I understand none of this". The first produces a reading list. The second produces a library.

Three: never sort by read and unread. Two separate piles turn the unread one into a reminder notice. Shelved together it is an offer. Same wall, different tone of voice.

Four: put it where you can see it, not where it looks good. A shelf sorted by colour has stopped being a tool. The spines have to be legible, at eye level, in the room where you work. Periphery only works where you actually walk past.

Five: stop counting the pile. The pile is not a debt. If it shrinks because you are disciplined about working through it, you have stopped buying, and that is the real problem.

In closing

The honest answer to the question at the wall is not "no, maybe a quarter of them".

It is: that is not the point. What stands there is not a record of achievement. It is a list of open questions I set out for myself, in a form I cannot dismiss and do not have to search for. Most of them I will never answer. A few will save me on a day I do not know about yet.

For that, 25 euros is nothing.

Do not fill your shelf with what you know. You already know it.

Sources

Antilibrary and tsundoku

Ownership versus licence

Book market

Studies

Books. What I read, what I buy, and what I never get to.

At home they stand there. Spine by spine, read next to unread, unmarked, unsorted by progress. Sooner or later somebody comes over, looks up the wall, and asks the question everybody asks.

Have you read all of these?

No. And the question is the wrong one.

A shelf is not a certificate. It is a tool. A certificate shows what you can do. A tool shows what is still possible.

The question that gives the game away

A book is an option, not a task

Here is the framing that settled the whole thing for me.

And now the side nobody writes down: what does the book you did not buy cost you?

One side of the ledger is visible and small: 25 euros. The other is invisible and unknown in size. Optimise on the visible side alone and you optimise yourself into a corner.

The shelf · scroll

On the shelf12

Read12

Reached into the unread0

"Then buy it when you need it"

That is the best objection to everything above, and it holds right up to the question of how you are supposed to know that you need it.

A folder holding 400 EPUBs produces zero such encounters. Not few. Zero. It has no periphery. It shows you a list when you open it, and nothing at all otherwise.

The periphery · scroll

Because you asked1

In view1

Without you asking0

What the e-book does well, and what it cannot do

It loses on two, and both are structural.

First, the periphery. See above. An e-book exists only while it is open. The rest of the time it is a filename nobody reads.

None of this is a conspiracy. It is simply what follows from having bought a licence rather than an object. Nobody can delete an object out of your flat.

The summary was never the value

Now the part that touches my day job.

So when somebody argues that a shelf is redundant because the knowledge is cheaper and faster elsewhere, they are right about the knowledge. The knowledge was never the scarce part.

The answer got cheap. The question did not. A shelf is a stock of questions you set out for yourself years ago.

How to tell this is not superstition

I like arguments that survive contact with numbers, so here is the most relevant study.

Five rules I buy by

Two: buy from a position of ignorance. The right buying signal is not "I want to read that", it is "I understand none of this". The first produces a reading list. The second produces a library.

Three: never sort by read and unread. Two separate piles turn the unread one into a reminder notice. Shelved together it is an offer. Same wall, different tone of voice.

Five: stop counting the pile. The pile is not a debt. If it shrinks because you are disciplined about working through it, you have stopped buying, and that is the real problem.

In closing

The honest answer to the question at the wall is not "no, maybe a quarter of them".

For that, 25 euros is nothing.

Do not fill your shelf with what you know. You already know it.

Sources

Antilibrary and tsundoku

Ownership versus licence

Book market

Studies

Buy the Books You Will Never Read

The question that gives the game away

A book is an option, not a task

"Then buy it when you need it"

What the e-book does well, and what it cannot do

The summary was never the value

How to tell this is not superstition

Five rules I buy by

In closing

Sources

Related Articles

The Sense of Beauty — Beauty Is Pleasure Objectified (and Your Feed Knows It)

The Smartest Guys in the Room: why the smartest guys are always the last to notice

The Picture of Dorian Gray — and Whether a Book Can Be Immoral

Let's talk.

Buy the Books You Will Never Read

The question that gives the game away

A book is an option, not a task

"Then buy it when you need it"

What the e-book does well, and what it cannot do

The summary was never the value

How to tell this is not superstition

Five rules I buy by

In closing

Sources

Related Articles

The Sense of Beauty — Beauty Is Pleasure Objectified (and Your Feed Knows It)

The Smartest Guys in the Room: why the smartest guys are always the last to notice

The Picture of Dorian Gray — and Whether a Book Can Be Immoral

Let's talk.