Class B — the spacing record against our unit boundaries
Class B is "Intervalla sive spatia, quæ mentem auctoris sæpius indicant" — the intervals and spaces, which often show the author's mind. It is the manuscript's own record of where Andrewes left air on the page, and it is the highest-value class in the apparatus for the edition's structure, as CONVENTIONS §13 anticipated.
Scope: printed pp. 2–250 only — the whole extent of the apograph.
⚠ First, the negative result the brief asked for
The brief asked that class B be weighed against the boundaries we chose ourselves — naming §28's and Part III's §2/§3 cut.
Class B cannot speak to either. §28 is at printed 361–367 and the Part III §2/§3 cut is at printed 416, both far outside the apograph's usque ad pag. 250. Those two boundaries remain exactly as well or badly supported as they were, and this pass adds nothing to them. They stand on our own reasoning — which, in §3's case, was independently confirmed from the plate (the half-blank page at 424) and does not need the manuscript.
Everything below therefore concerns Part I only.
1. Where the manuscript's silences fall, against our files
Our Part I files divide: front 1–29 · day1 30–69 · day2 70–91 · day3 92–105 · day4 106–129 · day5 130–147 · day6 148–167 · day7 168–195 · deprecation 196–219 · evening 220–251 · meditations 252–263.
The manuscript's blank pages and blank leaves — its major articulations, as distinct from ordinary line-spaces:
| printed | the MS does | our boundary | verdict |
|---|---|---|---|
| after 184 | ⚠ tria sunt folia non scripta — three blank leaves, six blank pages, the largest silence in the manuscript | mid-day7 | ⚠⚠ MISMATCH — see §2 |
| 190, after line 4 | duæ paginæ non scriptæ | mid-day7 | ⚠ MISMATCH — see §2 |
| 196 | pagina non scripta sequitur | deprecation opens at 196 | near-match, offset by a page |
| 207 | pagina non scripta sequitur | mid-deprecation | ⚠ and see §3 — this is the only odd page in the whole class |
| 212 | pagina non scripta sequitur | mid-deprecation | unmarked by us |
| after 218 | duæ paginæ non scriptæ | deprecation ends 219, evening opens 220 | ✅ MATCH |
| after 236 | duæ paginæ non scriptæ | mid-evening | unmarked by us |
| after 248 | pagina non scripta | mid-evening | unmarked by us |
| 250, after line 16 | nova pagina incipit cum verbis Γνῶσιν καὶ πίστιν | mid-evening | the apograph ends here |
One clean match, two clear mismatches, five divisions we do not mark at all.
2. ⚠⚠ The three blank leaves after printed 184
This is the finding of the class-B pass.
What our transcript has at the foot of printed 184:
καὶ ὁ θάνατος οὐκ ἔσται ἔτι, / οὔτε κραυγὴ, / οὔτε πόνος οὐκ ἔσται ἔτι, / ὅτι τὰ πρῶτα ἀπῆλθον.
Revelation xxi. 4 — and there shall be no more death, neither sorrow, nor crying, neither shall there be any more pain: for the former things are passed away. A terminal text if the book has one. And the manuscript answers it with six blank pages, more air than it gives anything else.
What follows, at printed 186: ΑΓΑΘΥΝΟΝ, ΚΥΡΙΕ — the great personal intercession, Barking and Pembroke Hall and Southwell and Westminster, Andrewes' own churches under Greek names.
⚠ Our own transcript already noticed that this is a different kind of thing. The note at the head of printed 186 in day7-transcript.md reads "the great personal intercession (cf. Day 3 pp. 98–101)" — it cross-references it out of Day 7 to a different day, because it does not belong where it sits. The manuscript agrees, and says so with three blank leaves; our file runs straight through it.
Likewise the two blank pages after 190. 4, which fall immediately after the close of a list (προαιρέσεις, / ἐπιχειρήσεις, / ἔξοδον, καὶ εἴσοδον, / κάθεδραν, καὶ ἔγερσιν.) — a clean end, and the MS marks it.
What this does NOT mean. It does not mean day7-transcript.md is wrong. The file records the 1853 plate faithfully and the plate does run continuously. The manuscript and the printed book disagree about where Day 7 ends, and that disagreement is a fact about the edition, not an error in our files. Whether to re-cut the file, or to record the division in the apparatus and leave the file alone, is Wilson's call — and it is the same question the Hebrew raises in CLASS-A-ledger.md §2: is this edition a faithful reprint of the 1853, or a critical text of Andrewes?
A pragmatic note for whoever decides: re-cutting a transcript re-cuts its English, and Part I's English is line-keyed page by page. The cost is real. The cheaper and probably better move is to record the MS's divisions in the apparatus and in the print design — Prototype B has spreads to spend, and six blank manuscript pages are exactly the sort of thing a mirror edition can honour typographically without touching a transcript.
3. ⚠ p. 207 — the only odd page in class B
Every other citation in classes A and B is an even printed page, because Part I sets Greek on the verso (even) and Latin on the recto (odd), and the apograph is Greek only. p. 207. pagina non scripta sequitur is the sole exception in the entire apparatus.
Three possibilities, and this file does not choose between them:
- a misprint for 206 or 208 — the likeliest, and the volume has proved itself capable of exactly this class of slip repeatedly;
- the editor briefly citing by his own Latin page because that is where the leaf-break falls;
- something about the manuscript's own foliation at that point.
Left as a crux. It costs nothing and resolves nothing to guess.
4. What class B corroborates rather than challenges
The seven days. Class B does not mark the day-divisions with spacing — and it should not, because the preface's principle iv says Dies sua quisque nota astronomica distinguitur: each day is distinguished by its own astronomical sign, a different device entirely. Our day boundaries at printed 30 · 70 · 92 · 106 · 130 · 148 · 168 are corroborated by principle iv and are untouched by class B's silence about them. This is worth stating plainly, because a naive reading of class B would take the absence of spacing at those pages as evidence against our cuts. It is not.
The deprecation/evening cut at 218–220 is confirmed outright by two blank pages.
5. The ordinary spacing record — for the typesetter, not the editor
The bulk of class B (apparatus vii–ix) is a line-by-line register of ordinary spaces, larger spaces, and lines run together without space: e.g. p. 124. lin. 3, 4, 6, 7, 11, 12, 17, 18, 24, 25, sine intervallo; p. 172. post lin. 4, 5, 8, 9, 10, 11, 12, et post 14 et 25 majus; p. 36. post lin. 1, 5 majus, 18, 22, 25, 29. It distinguishes at least three weights — spatium · majus spatium / majus intervallum · breve spatium — and records where lines the plate separates are joined (colligantur, 192. 11–13 and 14–16) or where lines the plate joins are separated.
⚠ This is the single most useful document in the apparatus for Prototype B, and it is entirely unexploited. The mirror's whole difficulty is page-fit: every spread must fit or the verso/recto mirror silently flips. Class B is a page-by-page account, from the author's own book, of where the vertical air is meant to go and how much of it — which is precisely the variable a typesetter has to guess at otherwise.
Recommendation, for a later print session and not for now: when transcript2tex.py is finally generalised, feed class B in as a spacing table keyed to printed page and line. It will not solve page-fit, but it converts an arbitrary choice into a documented one, and where a spread will not fit, it says which spaces Andrewes cared about and which he did not.
What is owed
- Nothing has been changed. No transcript re-cut, no boundary moved, no English touched.
- Wilson's calls, both of them the same question in two coats: whether the manuscript's divisions after printed 184 and 190 should re-cut
day7, and whether the Hebrew ofCLASS-A-ledger.md§2 should be restored. Both ask: faithful reprint of the 1853, or critical text of Andrewes? The project has never had to answer that, because until this pass the two coincided. p. 207goes on the pre-print list as a crux.