A field can be exhausted by the people entitled to use it; an idea can remain undiminished after passing through a thousand hands. The distinction changes what stewardship must protect. A melody learned by a second singer has not departed from the first. A method used in another workshop leaves the earlier workshop able to use it still. Sometimes the second user improves it, finds an error or discovers a use its maker never imagined. The common stock can grow through its circulation. Yet the work of finding, explaining and preserving the idea occupies people whose time cannot be used everywhere at once. Knowledge escapes one scarcity while depending on others, and the rules that govern it must answer to both.
The earlier Mirror turned the tragedy inside out by asking what happens when use enriches a commons instead of consuming it. Its knowledge passage reaches a more precise claim than a promise that every shared thing improves: one person's use of an idea need not subtract from another's opportunity to use it. That property is called nonrivalry. Whether a person can be prevented from using the idea is a separate matter. A secret can be withheld; access to a design can be licensed; a technical restriction can impede copying. None of these arrangements makes the idea disappear from the first user's mind when the second learns it. Exclusion changes access to the stock, while rivalry concerns what one use does to the availability of another.
The cost of making knowledge and the cost of letting another person use it can consequently point toward different prices. Consider the hypothetical manual from the later Utility treatment. Preparing it costs nine; supplying a usable copy costs one per reader. Two readers value it at twelve and four. Supplying both creates value of sixteen at a resource cost of eleven, leaving a gain of five. Supplying only the first creates value of twelve at a cost of ten, leaving two. Once the manual exists, including the second reader adds three to the gain. An access price that excludes that reader can therefore discard something worth more than the resources its use requires, even though the preparatory work must also be paid for.
A single access price cannot finance both readers in this example. To attract the second, the price must be no higher than four; two sales then yield at most eight, below the cost of eleven. A price of eleven can finance a single copy for the reader who values it at twelve, while shutting out the other. Charging each reader the cost of the additional copy brings in two and leaves the entire preparatory cost unpaid. The obstacle is visible without assuming either that the maker is greedy or that the excluded reader is indifferent. The shared benefit exceeds the shared cost, but the particular rule for collecting payment cannot realize that gain while balancing its account.
Several arrangements can cover the cost, and they distribute the obligation differently. Contributions of nine from the first reader and two from the second cover eleven and leave each better off. A sponsor could pay the initial nine and let the readers cover their copies. An association could collect dues in advance; a producer could charge for teaching or repair alongside a freely shared manual. Each arrangement asks someone to supply resources or accept a condition. The two contributions prove that a feasible division exists under the stated valuations. They do not explain how a collector learns those valuations, persuades the first reader to contribute more or prevents either from waiting for the other. The free-rider question returns at the stage of provision even when the finished knowledge cannot be worn out.
Exclusion may help finance the first copy while reducing the uses made of the copies that follow. In the manual example, the single sale funds production that otherwise might not occur, but leaves the second reader's additional gain unrealized. A rule defending that arrangement needs a comparison with the available means of funding the work, including their administrative costs and their effect on what gets made. The maker's claim to payment and the public's interest in access both have substance. Neither establishes that any particular duration, price or scope of exclusion strikes the right balance. The fence earns an economic defense through the work it enables and the alternatives it improves upon; its mere existence proves only that someone has acquired the power to close a gate.
The same arithmetic also shows why access cannot be made the sole measure of success. If preparing the manual cost fifteen instead of nine, the total cost of supplying the two readers would be seventeen, above their combined value of sixteen. No rearrangement of payments between those readers and the maker would turn that specified project into a positive total gain. A subsidy could pay the bill, but the resources would still have been used. Additional benefits to other people might justify the work, and a different objective might justify its distribution. Those reasons would have to enter the account. Nonrivalry tells us something important about repeated use; it does not confer unlimited value on everything that can be copied.
The practical commons also includes the artifacts and facilities that keep knowledge available. A formula differs from the fragile notebook in which its only surviving account is written. A digital file differs from the server, catalog and continuing labor that make it discoverable. Hess and Ostrom distinguish ideas, artifacts and facilities because their uses and vulnerabilities differ. A reader can occupy the sole physical copy of a book, even though understanding its argument does not exhaust that argument. A library can lose a collection; a format can become unreadable; an archive can become so poorly described that a surviving file is effectively lost. Stewardship must identify which of these things is scarce before choosing what to ration, preserve or share.
Open access to an existing work can coexist with carefully bounded authority over its maintenance. Imagine a shared program that anyone may read and copy. Allowing every reader to replace the version on which others rely would confer a different power, with different consequences. Reviewers may need to test a proposed change, maintain compatibility and explain which version contains a repair. A common right to benefit from the published code does not by itself allocate those responsibilities. The institution can keep use open while assigning approval, attribution and release to people answerable for them. Such a boundary protects the reliability of the common resource; its justification depends on the work and accountability it supplies, rather than on a presumption that all boundaries betray sharing.
The distinction between copying and maintaining also changes the meaning of a contribution. A person who improves the manual may add more to the common stock than a person who pays for a copy; a person who checks the improvement may add something different again. Those activities need not have the same reward. The rules can recognize authorship, fund testing and preserve earlier versions while leaving the resulting explanation widely available. They can also fail: unpaid reviewers may be overburdened, a dominant maintainer may exclude a rival improvement, or a funding arrangement may favor the questions of its sponsors. Calling the work a commons does not settle these disputes. It identifies a shared benefit whose conditions of production remain open to judgment.
A durable knowledge commons needs agreements about continuity as well as immediate access. A file available today may vanish with the person or institution that hosts it. A correction can improve an argument while making an earlier citation difficult to follow. Keeping identifiable versions, reliable records and usable copies requires recurring work. Hess and Ostrom's discussion of scholarly information makes that institutional problem central: preservation and dissemination depend on coordination among creators, repositories and users. For a particular project, the question becomes concrete. Who undertakes the continuing labor, what resources support it, and which users can challenge a decision that makes the shared material harder to find or trust?
What is gained by a fence therefore depends on which part of the enterprise it encloses and what the enclosure accomplishes. Restricting access may finance discovery, reward a contributor or support a dependable service. It may also collect a payment for knowledge whose creation is already funded while preventing uses that cost almost nothing to permit. Restricting alteration may defend a reliable common version while leaving everyone free to learn from it. Rules of attribution can prevent a stranger from passing off a revision as someone else's work. These powers are distinct. A judgment about one should not pass silently into a judgment about all the others, especially when the people benefiting from the restriction differ from those whose work supplies its defense.
The idea can be shared without division; the work that sustains it still needs a share.
The torch illuminates a commons whose central danger is often failure to create, preserve or circulate what use itself cannot consume. The inherited comparison with a field remains useful precisely because the differences are now visible. A fence around a pasture can restrain depletion. A fence around knowledge changes who may learn, build and earn from something that another person's understanding leaves available. Its gains and losses must be judged at those margins, with the continuing work in view. The next plate returns to a resource that use can damage: an atmosphere shared across a world whose users cannot be gathered around a single local table. The argument remains open.