The Self-Portrait Made of Others

Why should I publish a reading list?

A public reading list is the most accurate self-portrait most thinkers can produce, because the texture of how a mind is composed is more legible in who it cites than in what it says about itself. On Rodin, citations are public, named, and optionally noted, which turns the reading list from a private record into an intellectual lineage another mind can recognise.

In 1970, W.H. Auden published a peculiar book called A Certain World: A Commonplace Book, the bulk of which was not, in any conventional sense, written by him. It was an arrangement of passages from other writers, ordered alphabetically by topic, from "Acedia" through "Witches" and well past, with brief connective notes in his own voice that were less commentary than the placement of one sentence next to another. In the preface, Auden made the surprising claim that this book was the best autobiography he was likely to produce. Biographies of writers, he wrote, were almost always superfluous and usually in bad taste; what a reader actually wanted to know about a writer was the shape of the world inside the writing, and that shape was more accurately given by the sentences the writer had loved than by any account the writer could give of his own life.

What Auden was claiming, in effect, was that a reading list assembled with sufficient seriousness is not a list of influences appended to a life; it is the life, considered as a particular configuration of attention, made visible in the only register in which the configuration can actually be seen.

The composition of a mind

The intuition behind the commonplace book, which long predates Auden and runs at least back to the Renaissance practice of keeping a *zibaldone*, is that a mind is not principally what it produces but what it has chosen to keep. The Renaissance reader copied passages by hand into a notebook, in effect curating the canon of sentences that had snagged on something already there, and the notebook accumulated, over years, into a portrait of the curator that no first-person account could have given with the same precision. The reason was not that the curator was being modest; it was that the materials he was choosing among were richer than the language of self-description he had available, and the act of selection encoded a sensibility that was almost impossible to falsify.

What might be described as the compositional view of intellectual identity follows from this. A mind is the configuration of the texts, thinkers, and frameworks it has chosen to take seriously, and the configuration is, in the strict sense, more individual than any opinion the mind holds, because opinions tend to be drawn from a small repertoire that the surrounding culture has already provided, while the configuration of citations is a particular shape that nothing in the surrounding culture had ready-made. This is why, in the long literature on intellectual influence, the reading list of a serious thinker is so often the thing that survives them; the books they wrote may date, but the books they read, considered as a configuration, continue to describe the specific kind of attention they brought to the world.

What the LinkedIn profile cannot do

The contemporary professional profile is, by construction, almost the opposite of a commonplace book. It records what one has done in institutional categories already legible to the surrounding economy: the title, the company, the credential, the bullet-pointed achievement. None of these encode the configuration of attention; they encode only the institutional locations in which the attention was exercised. Two people with identical professional profiles can have entirely different reading lists, which is to say entirely different minds, and the profile would not register the difference.

The result is a public surface on which one is legible as a unit of the labour market and almost wholly illegible as a particular configuration. The fact that the labour market is the only register in which this kind of legibility has, until recently, been available is what produces the curious experience of having a long career and feeling, at the end of it, that almost nothing one has actually thought about has been recorded anywhere a stranger could find it. The professional profile is, in effect, a CV; what the commonplace book offered, and what the LinkedIn profile structurally cannot, is a self-portrait composed of the materials a mind has chosen to keep.

Citation as a different kind of edge

The followership graph that the social platforms have spent fifteen years building is not, despite its surface appearance, a record of anything intellectual. To follow someone is to opt into a stream; the act says almost nothing about whether the follower has read the work, carries it forward, or could give an account of why the work matters. The follow is a low-cost signal of ambient interest, and its currency has, over time, accordingly inflated to the point where the count is mostly noise.

A citation is a different kind of edge. It is named, which means the citer is willing to be associated with the act; it is optionally annotated, which means the citer is invited to say, in a sentence or two, what they have taken from the cited work; and it is public, which means the cited person can see who has chosen to declare an inheritance from them. What this produces, on the citer’s side, is exactly the structure of Auden’s commonplace book: a public reading list assembled by selection rather than by algorithm, in which the configuration of cited minds is the self-portrait. What it produces on the cited side is something the commonplace book could not produce, because the cited authors in Auden’s case were almost entirely dead, and is in effect the structural reversal that turns the whole figure.

What the commonplace book could not be

Auden’s reading list was, in the nature of the form, monologic. He could assemble it; the writers in it could not assemble him in return. The commonplace book is, for this reason, an instrument of intellectual lineage in only one direction; the reader inherits from the read, and the read have, by then, mostly stopped being in any position to inherit anything back. What the commonplace book offered as a portrait it could not, by construction, offer as a relationship.

A public reading list among the living is the same operation with one structural difference, and the difference is not minor. The people you cite can read who is reading them. They can, if they choose, cite you in return. The fact that the citation can be returned introduces into the form a possibility that the commonplace book did not have available to it, which is the possibility that two reading lists, assembled independently and without prior coordination, will turn out to contain each other. This is the rarest configuration the system permits, and it is the configuration the system was, in effect, built to make visible.

The mutual edge that cannot be faked

Mutual citation is one of the few signals on the open internet that resists gaming in any straightforward way. To engineer a mutual edge would require the cited person to find the citer’s reading list compelling enough, on its own terms, to declare an inheritance from it; the cited cannot be paid, persuaded, or socially nudged into the gesture without the gesture losing exactly the property that made it valuable. The mutual edge is, in effect, the empirical residue of two minds having recognised something in each other independently, which is what the long history of intellectual friendship has called, when it has used the word seriously, recognition.

The pairs in which this kind of recognition has been historically visible, the ones that get assembled into anthologies and lineage diagrams long after both parties are gone, were almost always preceded by exactly this structure. Each had been reading the other; each had, in some private register, been carrying the other’s work forward into their own; and at some point the carrying became mutual and known, and the work that followed in both careers was, in effect, a continuation of a conversation that the rest of the field could only watch from outside. What the citation graph proposes is that this configuration, which has historically depended on geographical accident or institutional adjacency or a friend of a friend at the right party, can be constructed deliberately, by the same method that constructed it accidentally, which is two people independently declaring an inheritance and, in the declaration, finding each other.

The portrait as it is now possible

What is left, in the end, is not the recommendation that one publish a reading list because a reading list is a useful marketing artefact, which it is not, or because it might attract followers, which is not the relevant currency. What is left is the older claim, which Auden made in the preface to a book composed almost entirely of other people’s sentences, that the truest self-portrait a thinker can give is the configuration of the minds they have chosen to take seriously, and that this portrait becomes possible only when the configuration is made visible in a register where another mind can read it back.

The commonplace book did this for the dead, and was, in effect, complete in itself. The public reading list does it for the living, and is in effect incomplete, since the portrait is one half of an exchange that the other half can, at any moment, return. What that exchange produces, when it is produced, is the rarest object the form permits, and the only one for which a system of this kind has, in the strict sense, been worth building.