# Children’s liminal horror: a historical and theoretical lineage

This is a bounded research ledger of 41 children’s and family-facing film and television works released between 1903 and 2025. It asks a narrow question: when does a familiar place, route, object, institution, or world remain recognizable while the relation that makes it usable becomes unreliable?

The complete machine-readable ledger is [lineage.json](./lineage.json). It contains the 41 work records, three historical phases, the source register, and the Atlas’s editorial theory metadata.

## The argument

Children’s media makes liminal horror’s threshold contract unusually visible. A door should connect rooms. A train should carry passengers toward a destination. A home should shelter. A village should contain differentiated social life. A school, hotel, or park should have an intelligible schedule and boundary. These are narrative affordances, not claims about a universal child psychology.

The catalogue becomes liminal horror when a work keeps the familiar object or setting in view while damaging the relation it promises. The door opens, but the room does not follow. The train moves, but destination disappears. The home remains recognizable, but shelter becomes capture. The village remains inhabited, but individual identity and ordinary time no longer hold.

This distinction separates a portal from a failed threshold. Portal fantasy ordinarily crosses into a destination whose rules become legible after arrival. Liminal horror keeps the ordinary invitation in view while making its guarantee deceptive, recursive, unavailable, or coercive. The aperture is not enough. The breach decides the classification.

## The three phases

The ledger contains exactly three phases, with 11, 15, and 15 works.

1. **1903–1969: portals, dreams, and first uncanny worlds.** Dream passages, magical objects, wardrobes, windows, trains, and tunnels make crossing visible. The threshold is often a passage into an elsewhere whose rules are still more coherent than frightening.
2. **1970–1999: familiar worlds lose their rules.** Infrastructure, villages, institutions, hotels, houses, schools, amusement spaces, and child-scale domestic settings remain recognizable while their social or spatial guarantees fail.
3. **2000–2025: the space becomes the monster.** Houses, trains, towns, hotels, apartments, parks, and dream environments increasingly perform the hostile operation themselves. Space is no longer merely where the threat occurs. Space routes, processes, captures, duplicates, or refuses the subject.

## The 1975–1977 cluster

*The Changes* (1975), *Sky* (1975), and *Children of the Stones* (1977) form the ledger’s sharpest historical cluster. Each moves beyond the self-contained fantasy portal. The works preserve recognizable British roads, machinery, landscape, villages, and prehistoric sites while withdrawing the rules that make those places socially and temporally readable.

In *The Changes*, machinery remains present after the shared order that made it usable has collapsed. In *Sky*, Stonehenge becomes a temporal border. In *Children of the Stones*, the village’s population, landscape, and time become a circular system that presses against individual identity. The cluster therefore shifts the carrier of horror from “another world” to the compromised world already under the child’s feet.

## Evidence and inference

The supplied historical catalogue establishes the bounded 41-work census. A separate source audit then checked each title against work-specific records from institutions, broadcasters, studios, distributors, archives, classification bodies, or production scholarship. The JSON contains 69 title-specific records and eight theory sources. Every work cites at least one title-specific record in `sourceIds`; records with two distinct evidentiary jobs retain both.

The audit did not force every claim into certainty. It removed unsupported details, narrowed claims to what the retrieved source establishes, and preserved specific limits: the unverified Grand Central mechanism in *Mr. I. Magination*; the difference between production and broadcast dates for *Round the Twist*; the territorial scope of Netflix classifications; the absence of an official rating for *Infinity Train*; and the difference between a 2023 start with two seasons and a closed 2023–2024 run for *Curses!*.

The Atlas adds an explicit editorial layer: `expectedRelation`, `breach`, `liminalOperation`, `relation`, `mechanisms`, `domains`, `featured`, and the 0–3 scare assessment. Each work also has `factualRecord` and `evidenceLimit`. The first field states the bounded work-level record supplied or anchored by the source audit. The second states what the source does not establish and keeps interpretation from masquerading as fact. The JSON keeps those layers separate and states the boundary in `metadata.evidenceNote`.

Official classifications are a separate field, `officialClassification`, and appear in 17 records where the audit found a platform, distributor, studio, or classification record. Each one names its territory or platform and resolves to the cited source. The ledger does not infer a classification from the scare key, a work’s reputation, or a neighboring title.

The source register also contains eight established Atlas theory sources used to define governed passage, domestic space, heterotopia, eerie agency, horror affect, mediation, and infrastructure breakdown. Those sources support the analytic lens. They are not used as substitutes for title-specific evidence.

## Scare key

The 0–3 `scareLevel` is an editorial intensity judgment. It is not an age rating, content rating, or viewing recommendation. The scale runs from essentially non-scary (0), through eerie (1) and deliberately spooky (2), to unusually frightening by children’s or family-media standards (3). Official classifications are recorded only where the supplied research identified one; the scare key never substitutes for a classification authority.

## Use in the Atlas

The ledger is designed to connect the childhood corpus to the Atlas’s existing theory and network. Its five explicit mechanisms—routing, recognition, authority, mediation, and care—are the same mechanisms used elsewhere in the universal theory. A work can therefore be read as children’s media and as a case of failed passage across art, mind, systems, or life without collapsing those registers into one category.

The record is additive. It does not replace the film catalogue, literature catalogue, architecture and technology research, glossary, or existing theory. It supplies a developmental and historical lens through which those sections can be read together.
