The vocabulary of the haunted
3,516 dossier names, broken into10,521 words — 3,890of them unique. Read together they say less about ghosts than about what America builds: houses, graves, hotels and forts.
The commonest words
- 01house323 names
- 02site250 names
- 03cemetery176 names
- 04hotel137 names
- 05fort136 names
- 06state110 names
- 07park110 names
- 08historic99 names
- 09museum98 names
- 10old92 names
- 11church67 names
- 12hospital60 names
- 13island57 names
- 14theatre56 names
- 15hall55 names
- 16national54 names
- 17inn53 names
- 18mansion52 names
- 19city48 names
- 20hill47 names
- 21bridge46 names
- 22burial46 names
- 23plantation44 names
- 24lighthouse44 names
Only grammatical glue is stripped (the, of, at, and …). "Old" — 92 dossiers, rank 10 — stays in on purpose: it is part of the vocabulary, not noise.
Name families · the last word
- … house255
- … site206
- … cemetery159
- … hotel104
- … park88
- … museum84
- … inn52
- … theatre47
- … hospital46
- … mansion45
- … battlefield42
- … lighthouse41
- … hall38
- … church38
- … plantation36
- … bridge34
What the name settles on. A third of the archive ends in a building, a burying ground or a room you can rent.
The same word, unrelated places
- Devil14 files · 13 states
- Massacre14 files · 10 states
- Witch9 files · 8 states
- Evergreen8 files · 7 states
- Golden7 files · 7 states
- Liberty7 files · 7 states
- Lost7 files · 6 states
- Monster7 files · 5 states
- Orpheum6 files · 4 states
- Waverly2 files · 2 states
Words that recur across dossiers with no connection beyond the name.
The same name, filed twice
- Mount Hope Cemetery×3
- Orpheum Theatre×3
- Belle Isle×2
- Bethel African Methodist Episcopal Church×2
- Chloride×2
- Colfax×2
- Elmwood Cemetery×2
- Evergreen×2
- Evergreen Cemetery×2
- Florence×2
- Fort Union×2
- Franklin Hotel×2
Identical names in different states — different places, different files.
The archive names the building, not the ghost
The top of the vocabulary is real estate. House (323 dossiers) outranks every other word; site (250), cemetery(176), hotel (137) and fort(136) fill out the first rank — a census of structures, not spirits. The word haunted appears in 3 names and ghost in12; the archive almost never puts the claim in the title.
Some frequencies are the collection's own history leaking through. Escapeappears in 27 names — Underground Railroad flight sites, filed as dark history rather than haunting — and burial site runs close behind.Old (92 dossiers) nearly outranks the itself; in this archive age is a filing term.
The collisions are the stranger finding. Waverly belongs to a Kentucky tuberculosis sanatorium and a Mississippi mansion that share nothing else.Mount Hope Cemetery is filed three times — Maine, New York, Georgia — and Orpheum Theatre three more. The mining West reused its own names: Galena, Golden, Chloride, Sumpter and Virginia City each sit on the map twice. Naming a town after the last one emptied was apparently common practice.
Read the frequency table as a sketch of what got documented. Houses, cemeteries and hotels are the places Americans built, abandoned and — eventually — told stories about. The vocabulary of the haunted is the vocabulary of what was already there.