Variable geometry names the spatial consequence of this epistemology. Regions, landscapes, metropolitan margins and heritage territories are meaningful because their boundaries are relational and problem-dependent, not because they possess immutable edges. Gómez Mendoza’s regional geography insists that flexibility and rigour are compatible: the rejection of rigid regional classifications does not imply that every spatial cut is equivalent. A geographical unit becomes persuasive when the scale selected reveals relations that are obscured from closer or more distant positions. This is why the medium scale repeatedly returns as a privileged observatory. It is neither merely local nor abstractly global; it allows settlement, land use, cultural representation, political authority and environmental process to become mutually visible. Changing scale is consequently an epistemic transformation. It changes the object because it changes the relations through which the object is constituted. Movement and narrative are inseparable from this scalar work. ‘Andar y ver’ is not simply an empirical slogan. Walking, direct observation and narrative synthesis provide a way of testing whether cartographic, archival or conceptual classifications correspond to the territory through which they claim to operate. Humboldtian visual thinking becomes a model not because the geographer possesses a sovereign eye, but because the gaze moves, orders and continuously rewrites its relations. Narrative has methodological force precisely where a landscape cannot be reduced to one variable: it can retain sequences, contradictions, temporal layers and interactions without pretending that synthesis is mechanical. The geographical account is thus both observation and construction, provided that the construction remains accountable to the field. Time prevents variable geometry from becoming only a spatial technique. Urban trees, periurban agriculture, forest landscapes and patrimonial territories are intelligible through accumulated transformations. The object observed in the present is a temporary configuration of previous practices, institutions, representations and material interventions. Municipal arboriculture records care and failure across decades; periurban agriculture registers successive pressures of urban growth and changing consumption; forest landscapes contain long histories of exploitation, abandonment, management and new cultural valuation; heritage emerges through changing criteria of what deserves transmission. Geography becomes geohistorical when it treats these temporal sequences as constitutive rather than contextual. The landscape is not simply where history happened. It is one of the material forms through which history remains spatially active. This also gives the geographic gaze a political dimension. Territorial representation contributes to territorial production. Naming a region, establishing a heritage landscape, defining a metropolitan fringe or describing a forest as ‘natural’ changes the field of possible intervention. Gómez Mendoza repeatedly refuses the innocence of such categories: language, planning instruments, cultural representations and disciplinary narratives participate in making the spatial realities they describe. Her later historiographic work extends the same insight to geography itself. A discipline is shaped not only by what it observes but by whom it allows to remain visible in its institutional memory. The recovery of women educators and geographers suppressed by political repression exposes knowledge production as a territorial and archival politics of inclusion, exclusion and transmission. For Socioplastics, this offers an unusually exact model for building a spatial research infrastructure. The corpus need not force every object into one stable scale or medium. It can instead preserve the variable geometry through which a phenomenon becomes legible: body, room, street, neighbourhood, city, region and transnational field may each reveal a different relational structure. Photographs, maps, routes, texts, video, statistics and archival fragments should therefore not be treated as interchangeable evidence but as transformations of the geographic gaze. What matters is the traceability of movement between them: what becomes visible, what disappears, what changes scale, what acquires historical depth, and what becomes institutionally legitimate. The archive then ceases to be a warehouse of observations and becomes a record of epistemic transitions. This gives Socioplastics a geographic principle strong enough to connect fieldwork, longitudinality, multimodal evidence and comparative urban research without flattening their differences: variable geometry makes comparison possible by keeping the transformations of the gaze visible.