Environmental evidence is produced by arrangements that sense, classify, narrate, maintain and contest environments rather than by observation alone. A sensor, a resident, a landowner, a community organisation and a computational platform occupy different positions within the same evidential ecology. What matters is not simply whether an environment can be measured, but how the capacity to make environmental conditions perceptible is distributed and with what consequences.

Gabrys’s work repeatedly shifts the object of analysis from isolated information to the relations that enable information to acquire force: programmed urban infrastructures make some forms of citizenship operational; participatory sensors create situated data publics; computational sensing redistributes experience beyond the human observer; landscape regeneration depends on institutional forms capable of sustaining collective action. Evidence becomes a capacity to intervene because it is embedded in material, political and organisational relations. The epistemic ground of this argument is relational rather than representational. Environmental measurements do not stand outside the situations they describe. Sensors are selected, calibrated and placed; data are interpreted through particular purposes; landscapes are perceived through histories of ownership and use; communities are constituted through membership, exclusion and authority. ‘Just good enough data’ makes this explicit by refusing a universal hierarchy in which only regulatory equivalence produces valid knowledge. The question shifts from whether data perfectly represent an external world to what relations they make possible, what distinctions they sustain and which publics they allow to act. This does not abolish standards. It locates standards within an ecology of purposes, instruments and institutional claims.

At the urban scale, this produces a distinctive politics of infrastructure. Smart-city systems promise environmental responsiveness by turning streets, buildings, mobility and energy use into streams of computational information. Yet the same architecture determines who or what is allowed to sense, how feedback is organised and which actions become legible. Environmentality names this distribution of governance through space and technology. The political subject is therefore not prior to infrastructure: forms of citizenship emerge through participation in technical arrangements. A resident asked to monitor air pollution, a device transmitting environmental change and a platform aggregating observations become differently authorised components of a programmed environment.

At landscape scale, distributed evidence must also acquire durability. Restoration projects reveal that participation cannot be reduced to consultation or temporary project involvement. Land rights, local institutions, cooperative structures, livelihood systems and long-term maintenance determine whether ecological intervention becomes a sustained territorial capacity. Network building is epistemic as well as organisational: techniques, monitoring practices and ecological knowledge travel between communities, while external scientific knowledge is accepted, translated or contested. Landscape therefore becomes an infrastructure of memory and responsibility in which ecological recovery depends on the persistence of social relations capable of carrying knowledge beyond the lifecycle of a grant. Socioplastics can operationalise this distributed model by treating every evidential object as simultaneously spatial, technical and institutional. A map is not merely an image of territory; it specifies what can be connected. A sensor trace is not simply a value; it has provenance, thresholds and a public of interpretation. A photographic sequence can record duration and change. A community statement can modify the status of supposedly objective spatial evidence. The corpus becomes more rigorous when relations among these objects remain visible rather than being flattened into a common representational surface. Environmental evidence is strongest when its heterogeneous modes can be compared without being forced into equivalence. The resulting field is neither purely computational nor purely participatory. Its operational strength lies in preserving the passage between sensing and action: who sensed, with what apparatus, under which conditions, at what spatial scale, through which institutional relation, and with what capacity for revision. Such a field transforms environmental research from a collection of descriptions into an ecology of consequential relations. For Socioplastics, this adds a model of knowledge in which environmental legibility is inseparable from distributed agency, and that capacity condenses as EvidenceAmbiguity.