Verify the secondhand attributions that Born of a Virgin's ratings rest on

[deanberris]

Follow-up from #72 and #74.

Several confidence ratings in Testimony/Arguments/BornOfAVirgin/ rest on what one author reports another as saying. The reporting author has been read. The reported author has not, and is not in the bibliography. Each report is attributed honestly in a comment, but a rating that turns on a secondhand report is only as good as the report.

Reported author	Reported by	What rests on it
J. D. W. Watts, Isaiah 1–33	Compton p. 5 n. 13	berryLine's step rated disputed: Watts grants that the chronology is uncertain and still keeps the near-term sign (Lines.lean)
R. E. Clements, Isaiah 1–39	Compton p. 5 n. 12	motyerLine's step rated disputed: identifies the child with Maher-shalal-hash-baz and keeps the near-term sign (Lines.lean, Sources/NearTerm.lean)
W. D. Davies and D. C. Allison, Matthew 1.212	Rhodea p. 71 n. 66	lukeAllusionEdge rated a disputed .allusion of Isaiah 7:14 (Atoms.lean)
J. A. Fitzmyer, Luke I–IX 336	Rhodea p. 71 n. 66	the other side of the same dispute: no allusion
R. E. Brown, The Birth of the Messiah 155–59	Johnson p. 270 n. 6	annunciationFormEdge: the annunciation type-scene (Atoms.lean)
M. Rydelnik	Postell p. 474	ordinarySignLine's step rated disputed: Rydelnik grants both grounds and still reads the pregnancy as a sign as deep as Sheol (Lines.lean, Sources/Wegner.lean)

To do, for each row

1. Find the reported work and verify its edition and pages per .claude/skills/adding-a-citation. Never guess an ISBN.
2. Read the passage. Does it say what the reporting author says it does?
3. If it does, add the work to Bib/Works.lean and cite it directly, keeping the reporting author as supporting.
4. If it does not, fix the rating and follow wherever it leads. The Watts, Clements and Rydelnik rows set disputed ratings on inference steps, and a step that is no longer disputed can change a dispute's extensions. Re-run the dispute proofs and update the docstrings and docs in the same commit.
5. If the work cannot be obtained, leave the secondhand attribution and say so in the comment.