The Knowledge

To teach a person a method is to leave it with him without losing it oneself. Two workshops can follow the same drawing; two minds can hold the same proof. The wood, the working day, and the attention of the teacher remain scarce, but the idea has crossed the room without emptying the place from which it came. This is the promise within the knowledge commons. Its users need not take turns possessing what they know. A second reader does not consume a first reader's understanding, and a useful correction can improve the inheritance of both. Yet the shelves do not assemble themselves, the proof does not check itself, and the discovery must first be made. The abundance of an idea once found leaves open the arrangements by which people find it, keep it, and make it available.

Nonrivalry describes this possibility of simultaneous use. It does not establish that everyone can obtain the idea. A method can be kept secret or its use restricted even though one person's use would not diminish another's. Romer's account of technological change separates rivalry from excludability and distinguishes a reusable design from the human skill needed to produce or understand it. Knowledge approaches the pure public good when access, as well as joint use, is open. Its production presents a further difficulty: if a discovery entails a fixed cost while an additional use costs little, receipts at that additional cost need not repay the discovery. In Romer's model, partially excludable designs and market power help finance inventive activity. That is an account of one institutional arrangement, not a proof that every idea must be sold under the same terms.¹

Suppose a manual costs nine to prepare and one to deliver to each reader. There are two prospective readers, valuing it at twelve and four. Preparing and delivering both copies costs eleven and creates value of sixteen: five remains after all costs are counted. A uniform price low enough to attract both readers can raise at most eight. A price of eleven can finance a copy for the first reader alone, leaving the second unserved even though that person's benefit of four exceeds the additional cost of one. Charging one per copy would admit both, but bring in only the two required for delivery. The missing nine is the cost of having something to deliver. The choices to create and to distribute cannot be answered by the same marginal calculation.²

The gap does not make the manual impossible. Payments of nine from the first reader and two from the second would cover all eleven, leaving each better off than if it were never made. That arithmetic establishes a way to divide the bill, not a mechanism that will discover the values or collect the payments. Once the manual exists and can be shared, each reader may prefer that someone else have paid for it. Funding arranged before production therefore faces a different problem from access arranged afterward. A subscription, a commission, a public appropriation, or a contribution can bear the initial cost; each needs some account of who chooses the work and how the promised support arrives. Restricting later access can furnish an incentive to create while withholding uses that would cost less than they are worth. Opening access can realize those uses while leaving the creator's bill unsettled.

The object being shared also needs a name precise enough to govern it. Hess and Ostrom distinguish ideas from the artifacts that carry them and the facilities that store and distribute those artifacts. A mathematical result, a particular volume containing it, and the library housing that volume have different conditions of use. The result can be held simultaneously; the sole physical copy cannot be borrowed by everyone at once. Digital copies change that constraint without removing every other one. Storage, organization, preservation, and the work of making a collection intelligible still require provision. Rules suited to an exhaustible stock do not simply transfer to ideas, but the generosity of an idea does not settle the care of its containers. The knowledge commons joins resources with different capacities and different needs for attention.³

Consider an open collection of computer instructions. Let anyone read it, copy it, and propose a repair. The permission to propose need not carry the permission to replace the version on which everyone relies. Someone must distinguish a repair from a mistake, preserve the earlier version, and record why the change was accepted. Those tasks consume time even when the resulting instructions can be copied freely. Limiting who may approve a change can protect a shared resource without restricting who may use its published form. It can also place considerable power in the reviewers' hands. A workable arrangement must therefore make both the contribution and the judgment intelligible: what is being changed, whose reasons govern, and how a mistaken decision can be revisited. Unrestricted reading does not, by itself, answer any of these questions.⁴

The common stock is enlarged by discovery, but its usefulness depends on more than the number of statements placed within reach. Readers need grounds for confidence; contributors need some expectation that their work will be recognized and cared for; maintainers need a way to continue. A false statement does not use up a true one, yet distinguishing them can absorb the attention on which knowledge depends. The governing task is thus to sustain inquiry without confusing a gatekeeper's authority with the truth of what passes through the gate. Open examination, a record of reasons, and the ability to correct a mistake are parts of that task. The Trust turns to the expectations that let people undertake such work together before every obligation has been written down.⁵