Ghost geographies

What you are looking at

Nothing, at present. The layer is registered and its asset is empty.

Fifteen ghost features are recorded in the corpus, in two groups.

Five come from the nineteenth-century surveys: a coast reported in the Weddell Sea in 1823, a range reported south of the Ross Sea coast in 1841, a stretch of continental coast charted in 1840 and later sailed through, and two islands reported in the southern ocean and eventually removed from charts.

Ten come from a single printed page of 1691. On the last leaf of his polar essay Vincenzo Coronelli names the lands he will not carry, and says why: the Sevarambi, the Isle of Pines, Frisquemore, the Terra di Vista, the Terra de’ Papagalli, the Strait of Anian, Frisland, Nova Albion, the Isle of Saint Borondon and Fonseca. His reasons are recorded as his, not adopted as ours.

Not one of their disputed positions has been located from a source. Every record therefore carries no geometry, and the layer that would draw them holds nothing.

How to read this layer

The empty layer is the honest state, not a bug and not a placeholder.

A ghost feature is a claim about a position. Placing one without knowing the position would mean inventing a coordinate for a feature whose whole significance is that it was invented, which repeats the original error in our own voice. The records exist, they can be read, and they will acquire geometry when the disputed positions are found in the charts that carried them.

That rule bites hardest on the Terra di Vista, the one record here that comes with a figure: Coronelli reports a cape at 42 degrees of southern latitude. A latitude without a longitude is not a position, and a cape drawn for a coast its own source rejected would be the clearest possible instance of the error.

Five of the ten Coronelli records are not southern at all. Frisquemore is placed on the confines of Lapland, the Strait of Anian and Nova Albion are north-west American, Saint Borondon is Atlantic and Fonseca Caribbean. They are held here because the record is his act of judgement, made in one breath in a polar essay, and splitting his list by hemisphere would break the thing being recorded. No row of that group may be read as an Antarctic phantom, and each says so.

The contract is published ahead of the data so that a reader looking for the disproved features finds an account of why none can yet be placed, rather than finding no layer at all.

Historical scope

1504 to 1916, once the layer holds anything. The earliest record is a landfall Coronelli reports of Gonneville and does not believe.

Sources and evidence

Each record keeps four things: who introduced the feature, what was actually reported, why that was a reasonable thing to record at the time, and what later evidence said. A record missing any of those cannot be rendered at all.

Two records are filed unresolved rather than disproved, although the literature reports them as removed from charts. No authority for the removal is held, and a disproof that cannot name what disproved it is not a disproof.

The ten Coronelli records use the two status fields to keep his century apart from ours. laterStatus is disproved and names Coronelli as the source of the disproof, because he is the one who rejected them, in 1691. Every one of them holds currentScholarlyStatus at unresolved, because this corpus carries no modern authority on any of them and a seventeenth-century verdict inherited as a present-day one would be the failure the pair of fields exists to prevent.

No record here attributes a cause. Saying a feature was later removed is a record of non-confirmation. Saying it was a mirage, a refracted image or an ice island misread as land is a much stronger claim that needs its own source, and the two are routinely conflated.

Reconstruction and uncertainty

When positions are located, original historical geometry and any modern corrected geometry will be separate records rather than two states of one line.

Editorial decisions

The most important record in this set was claimed by James Clark Ross, the most careful observer in the whole programme, working in polar air where distance and elevation are famously hard to judge. A ghost attributable to a careless witness teaches nothing. One attributable to Ross is the entire point.

Data fields

evidenceClass, geometryProvenance, confidence, reviewState and act. See the Antarctic data fields.

Data and downloads

CC BY 4.0 for the compiled records. In review, and currently empty.