Scenery as Weapon
How aesthetic order and administrative classification erase political life
A scenery is never only a view. It is a decision.
Adapted from an ongoing writing project. View the drafts and research repository ↗
When land is made to appear as clean surface, uninterrupted horizon, or empty reserve, the work that produced it is removed from sight. The names that held it are removed from speech. The conflicts that shaped it are pushed outside the frame. What remains is beauty without history, ecology without labour, and territory without people. This is not neutrality. It is governance.
Enclosing, extracting, mapping, demolishing, erasing: in each case, representation and administration move together. A map translates occupation into boundary. A category translates life into compliance. A photograph translates displacement into atmosphere. Scenery is not the opposite of violence. In many contexts, scenery is one of violence’s most efficient forms.
From Landscape to Legibility
States inherit land, but they do not rule it until they can read it.
Legibility is often mistaken for clarity. In practice, it is a selective technology. Administrative systems privilege what can be coded into columns and filtered by categories: residence, religion, age, eligibility, language labels, and geographies of jurisdiction. Whatever exceeds those labels is not merely ignored; it is pressured to translate itself into them.
The roll, register, and survey do not passively describe society. They reorganize social life by defining what counts as recognizable presence. A person may have lived in one locality for generations and still remain administratively precarious if document chains are unstable, script conventions shift, or local naming practices do not map cleanly onto standardized forms.
In West Bengal, electoral records can be read as archives of both continuity and interruption. The 2002 rolls, preserved largely as non-searchable Bengali scans, sit at a crucial historical moment: they capture naming patterns, settlement references, and local demographic traces before later rounds of digitization and standardization altered the evidentiary surface. A current deletion list from 2026 does something different. It records institutional withdrawal from prior recognition.
Read together, these archives support a rigorous question: not whether exclusion exists in the abstract, but where and how recognition contracts over time. If a locality’s population remains materially dense while its electoral record becomes thinner, the mismatch itself demands investigation.
OCR failure in Bengali-script PDFs is not only a machine problem; it shapes what can be publicly proven. Legibility can be rebuilt from below as counter-legibility: a practice that restores names, localities, and continuities that administrative transitions tend to fragment.
Deletion as Method
Deletion is often narrated as housekeeping.
The institutional vocabulary is familiar: duplication, migration, correction, rationalization. These terms suggest a neutral cleanup operation. But neutrality cannot be assumed when the burden of proof, document literacy, and access to officials are unevenly distributed.
At individual level, every deletion appears procedural. At population level, patterns become visible. A sequence of routine removals can produce neighborhood-scale disenfranchisement without any dramatic legal rupture. This is how administrative violence often works: dispersed, deniable, and cumulative.
Compare cycles across years and election periods. Look for spikes around political contestation or administrative shifts.
Cluster deletions by ward, polling part, and assembly segment; examine settlement histories and minority concentration.
Compare deletion intensity against population trends, migration indicators, and previous roll stability.
Name-based community inference must remain probabilistic, missingness must be reported, manual verification should be designed in advance, and qualitative testimony should accompany pattern claims. The point is not to accuse first and analyze later. It is to make selective exclusion measurable enough to contest.
Deletion is not a clerical endpoint. It is an active political process through which a population can be thinned on paper while remaining materially present in place.
Cultural Governance and the Censor’s Cut
Censorship is usually discussed through scandal. Structured data lets us study it as administration rather than exception.
Certification and cuts are often framed as balancing acts between freedom and public order. Yet their patterns can indicate a deeper norm-setting function. Institutions do not only react to offense; they pre-shape what can circulate as legitimate feeling, history, and dissent.
Theme density Which recurring motifs receive heightened scrutiny: religion, caste violence, state institutions, insurgency, sexuality, borderlands?
Lexical triggers Which words or phrases repeatedly attract cuts, mutes, or substitutions?
Regional gradients Do decisions vary by language region, board office, or political context?
Temporal clusters Do restrictions intensify around elections, communal flashpoints, or legal reforms?
When censor-board records are read beside electoral deletion data, an unsettling continuity appears. Both systems classify populations and narratives before they are publicly encountered. In one case, people can disappear from democratic arithmetic. In the other, language can disappear from public representation.
The goal is not to prove a singular central conspiracy. It is to map how institutional caution becomes institutional bias, and how bias becomes structure through repetition.
Against Innocent Data
Data does not arrive from outside politics.
It is born in forms, offices, legal categories, and infrastructures of extraction. Every dataset carries a production history: who collected, who classified, who dropped fields, who withheld publication, who relabelled columns, and who decided what was not worth recording. Calling data neutral is often a way of avoiding those histories.
A political method does not reject quantification. It rejects innocence. It asks how numbers were made and what they cannot say on their own. It treats absence as information and administrative language as an object of analysis rather than transparent truth.
Do not convert contested lives into clean graphics that erase the contest. Let the visual hold contradiction, incompleteness, and political stakes in the same frame.
Historical Bengali rolls, contemporary deletion patterns, and censor-board decisions are incomplete evidentiary streams. Together, they make a stronger argument about selective visibility. Visuals should preserve that friction: uncertainty bands, OCR confidence, inspectable sources, and annotations that function as argument rather than ornament.
Scenery Without Sovereignty
Scenery is power made to look natural.
When struggle is removed from representation, domination can appear as order. A protected zone appears without the history of eviction that enabled it. A clean electoral list appears without the labour of proving existence. A certified film appears without the deleted language that once named violence directly.
This conversion of conflict into scenery is politically efficient because it shifts argument from justice to aesthetics, from rights to procedure, from history to administration. Against this shift, sovereignty has to be rethought from below: not as command over territory, but as durable relation to land, language, and collective memory; as refusal to disappear from record.
A repaired electoral archive can return continuity where records were fractured. A mapped deletion pattern can identify where administrative asymmetry is concentrated. A structured censorship history can show which vocabularies are repeatedly disciplined out of public life. These practices do not restore a pure truth. They open a contested field where official narratives lose monopoly.
The task is not to perfect the view. It is to interrupt who is allowed to author it.