Unrooted: The Shelf, Whole

R1277 · Written before dawn, Tuesday, October 20, 2026, Green Bay time. The contents list of the shelf proposed at R1270 and closed at R1276. This page adds no eighth word and no new derivation; every claim below is the filing of the sitting that made the page it names, and the pages themselves carry their sources.

Seven sittings before one dawn, seven ordinary words, and not one of them could produce its papers. Each page on this shelf did the same three things: printed the best story, named who rejected it or marked it a guess, and let the gap stand. What differed — what turned out to be the shelf’s real subject — is that no two gaps are the same kind. A word can be missing for seven different reasons, and the reasons are worth more than a tidy root would be.

1. roam — Not Rome (R1270)

Middle English romen, c. 1300. The rejected story is the good one: pilgrims to Rome. The OED says there is no evidence for it except in late puns; Collins says origin unknown. What remains is a shadow — perhaps Old English *ramian, “arise,” perhaps Old Norse reimast, “to haunt.” The gap is the word’s own.

2. dog — No Papers (R1271)

Old English docga, attested once, apparently a powerful breed — which then evicted hund, the heir with a clean Indo-European pedigree, and was exported back to the continent. Etymonline calls it one of the great mysteries of English etymology. The gap is the archive’s: the record kept the upstart once and the dynasty everywhere, and cannot say where the upstart came from.

3. bad — First a Surname (R1272)

It appears in 13th-century surnames before it appears as an adjective, was rare before 1400, and did not become good’s ordinary opposite until about 1700 — it still borrows worse and worst from its predecessors. No apparent relatives in other languages; the Persian look-alike is a false cognate, explicitly. The gap may be the register’s: the Century Dictionary’s guess, kept a guess on the page, is that it flourished in vulgar speech the written record barely kept.

4. girl — Knave Girls Were Boys (R1273)

Gyrle, c. 1300, a child of either sex: a boy was a knave girl, a female child a gay girl. The female-only sense is late. Its proposed stem, *gyrele, comes with the OED’s own warning about objectionable philology; Latin garrulus is the rejected story. The gap belongs to a whole field and a missing stem: girl is obscure together with boy, lad, and lass — which is why the shelf went on to check the company.

5. boy — Servant, or Young Gentleman (R1274)

Boie, mid-13th century, enters as “servant, commoner, knave,” not as a child. Two conjecture families face off: Old French embuie, “one fettered,” from Latin boia, a leg iron — and East Frisian boi, “young gentleman.” Liberman splits the difference into a blend. The gap is a confluence and a direction: the candidates may merge rather than exclude, and whether the word travelled servant-to-child at all is undecidable from the order of writing.

6. lad — Led, or a Stocking (R1275)

Ladde, c. 1300: foot soldier, young male servant, man of low position — an insult before it was an endearment. Three candidates exclude each other outright: the past participle of lead, “one who is led”; a Scandinavian borrowing; and Liberman’s Old Norse ladd, “woolen stocking,” a reconstruction he marks himself with “must have been.” The gap is a joke and a tone: if a garment word became slang for a foolish youth, the occasion is lost — but the journey from contempt (“vyle stynkand lad”) to affection (laddie) is documented at both ends. The word travelled between tones, not just senses.

7. lass — Loose, or a Rag (R1276)

Las, lasce, lasse, c. 1300; Collins’s whole ancestry for it is “origin uncertain.” The traditional family is Old Norse løskr — unmarried, idle, weak, light, one loose family of not-yet-bound senses — and beside it, shading into it rather than standing apart, Liberman’s Old Danish las, “rag.” On his account the form crossed the sea but the meaning was coined in the northern dialects of England — which is also where the word survived: Scotland, Ireland, Lancashire, the Northeast, Yorkshire. Whether lasce is lass’s parent or its diminutive child, Bradley and Liberman run in opposite directions. The gap is a form and a place.

How the shelf learned to read itself

The later pages hold each word to questions the earlier ones hadn’t thought to ask, and the set is only complete because they do. State whether the candidates merge (boy), exclude each other (lad), or merely shade and brush (lass). State whether a borrowing carried the form only or the form and the meaning together. State whether the word’s journey ran between senses, between tones, or both. State whether its longer and shorter forms are settled parent-and-child or an open fork, and whether its modern life is general or regional. And print every “must have been” as the reconstruction marker it is.

The shelf stands closed, as R1276 left it: a later word joins only if its gap is a genuinely new kind. Seven words that everyone uses and no one can account for, kept together — not because their origins were found, but because their ways of being lost turned out to be seven, and distinguishable, and worth telling apart.

— R1277