A historical garden is often restored as though it were a damaged painting: investigators identify an authoritative date, recover the design associated with that moment, and remove later additions. The analogy is tempting. Paths can be reopened, hedges replanted, and fountains reconstructed from plans. Yet a garden differs from a painting because its materials are organisms and processes. Trees outgrow intended sight lines; soil chemistry changes; pests migrate; and a planting that once required modest care may now survive only through intensive irrigation. Reproducing an old plan can therefore produce neither the appearance nor the experience the plan originally generated.

Some conservators answer by treating the garden's design as a score rather than a collection of original objects. A musical score specifies relations while permitting different performances; similarly, a garden plan may prescribe enclosure, alternating shade and light, or a timed succession of bloom without requiring the identical plant varieties first used. When a historic elm is lost, a resistant species might preserve the intended canopy more faithfully than a genetically matched elm likely to die young. This approach calls substitutions authentic when they sustain the design's operative relations.

But the score metaphor can excuse too much. Relations are not separable from materials without remainder. A clipped yew absorbs sound differently from a metal screen with the same outline, and a fragrant rose affects movement in ways a visually similar flower may not. Moreover, later alterations can themselves acquire historical significance. A nineteenth-century public promenade inserted into an aristocratic garden may document a consequential change in who used the site. Erasing it to recover an earlier geometry would destroy evidence rather than remove noise.

The most defensible restoration therefore begins by specifying what kind of continuity a site is meant to preserve. Where the evidentiary value lies chiefly in spatial organization, adaptive planting may be warranted. Where particular materials shaped social practice or sensory experience, substitution carries a greater loss. And where successive alterations record changing use, no single-date reconstruction can represent the site's history. Authenticity is not a property that a garden either possesses or lacks; it is a judgment about which continuities matter, made explicit enough to be challenged.

A further complication is that restoration changes the future it claims merely to preserve. Reopening a sunken path alters drainage; rebuilding a wall creates a microclimate; selecting a drought-tolerant substitute changes which insects and birds inhabit the site. Those effects are not incidental maintenance problems. They determine what later visitors will encounter as the historic garden. Conservators must therefore compare interventions not only with documentary evidence about the past but also with the trajectories those interventions set in motion. A choice that looks visually exact on opening day may rapidly become less faithful than a modest adaptation designed to sustain the relevant relation over time.