The open verdicts, the loose threads and the debt this book still has outstanding.
Status: 12 August 2026.
THE RULE
This book makes one promise it must keep everywhere: what stands here does not stand in the main text. A question that stands open in this register may only be used as an argument after it has been discharged here — with a date and with a reason.
That is no gesture of modesty. It is the only way to prevent a book from, over time, mistaking its own conjectures for findings. A loose thread you write down, you cannot unnoticeably weave back in.
There are three sorts of entries in this register, and they each require something different.
- Open verdicts are questions to which the text gives no unambiguous answer. They remain open, and the book shows them as open. Research does not resolve them. - Debts are matters that can be decided, but for which the work has not yet been done — an edition that has not been compared, a book that has not been consulted, a plate that has not been examined. They are dischargeable and ought to be discharged. - Instrument defects are faults in the measuring tool itself, measured and noted, not repaired within the current round.
I. THE OPEN VERDICTS
Questions to which this book gives no answer, and why it does not.
D1 · Why does בַּר stand in Psalm 2:12 and not בֵּן?
The ordinary Hebrew word for "son" is בֵּן, and it stands five verses earlier in the same psalm. In verse 12 stands בַּר. The old translators visibly diverge here, and four emendations have been proposed.
Remains open. The book gives the reading it follows, with the others beside it, and does not decide. See Appendix B, B4.
D2 · The subject of Psalm 110:7
Who drinks from the brook along the way, and who lifts up the head? The grammar allows both answers, and the choice changes the close of the psalm drastically.
Open verdict, and that is an authorial decision that has been made: the verdict remains open. So it stands too in chapter 8.
D3 · The center of Psalm 45 hangs on the colometry
The fine pass finds nested frames around מֵחֲבֵרֶיךָ — the word that Hebrews 1:9 cites. But the center cannot be established without first making a colometric choice, and that choice is precisely what is under discussion.
Not resolvable without a choice beforehand. Presented as such.
D4 · The five archaic suffixes of Psalm 2
The psalm bears five -mô-forms. That is striking, but the book draws no conclusion from it — not before the literature on archaisms in psalm poetry has been seen. That is a debt; see §IV.
D5 · The companions of Hebrews 1:9 and 3:14
The word μέτοχος occurs five times in Hebrews, once from the quotation of Psalm 45 and four times thereafter. Whether the letter deliberately lets those later places reflect back on the quotation is attractive to think and not to be substantiated.
Do not claim.See Appendix C, §I.3.
D6 · The עֲבֹת of Psalm 2:3 and the priestly garment
Two psalms that each share one word with the high-priestly garment is too little; and with עֲבֹת the ordinary meaning is more probable.
Do not claim — but noted, because with a third find it could suddenly become something after all.
D7 · Psalm 103:4 — how far does it carry?
Open. Usable as an Expert box at the kinsman-redeemer line, not as proof.
D8 · What has happened with the Hebrew of Deuteronomy 32:43?
The Greek text is longer than the Masoretic, and the scroll from cave 4 stands on the side of the Greek. Whether something has dropped out or been added, and why, is not decided — and the two explanations do not exclude each other.
Open verdict; taken up thus in chapter 2, mark weighed.
D9 · What the form-profile cannot distinguish
The calibrated form-profile test places Psalm 110 and Psalm 8 robustly on either side, but on the middle section it is blind. A text in which the instrument detects nothing lands not on zero but on a fixed line higher — and Psalm 45 lies thirteen thousandths above that line. Its measured position is therefore an artifact and no finding. And Psalm 2 and Psalm 45 cannot be distinguished from each other at all by this instrument.
Open — and the book does not use the middle group as argument. About Psalm 45 it says nothing on form-critical grounds; about Psalm 2 only the philology of verse 7. See Appendix A §V-bis.
II. THE TEXT-CRITICAL DEBT
What Appendix B reports as "not yet checked". Each of these entries is dischargeable.
#
what stands open
where it comes from
D10
A complete inventory of the proposed emendations at Psalm 2:12, with their literature. The four most cited are there; the rest not.
B4
D11
The edition comparison at Hebrews 2:8. For 1:13 it has been done, for 2:8 only partially. On that hangs moreover the question whether the line "and you have set him over the works of your hands" in Hebrews 2:7 is original — the critical text omits it, the majority text has it.
B6 · Appendix C §VI
D12
What the old translations and the revisers do with עֲנָוָה in Psalm 45:4. Not checked.
B9
D13
How old the distinction לַאדֹנִי / אֲדֹנָי is, and what the other old translations do with it. The distinction in the Masoretic tradition is established; its age is not. This is the heaviest of the seven, for on that distinction rests the dispute of Matthew 22.
B11
D14
The plate of 4QDeut^q itself. The reading rests on the published edition as the literature renders it; the edition itself has not been consulted.
B13
D15
The motive for the broken-off clause of 2 Samuel 7:14.That Hebrews omits the discipline-clause is established; why is weighed and no more.
B14 · image bank no. 18
D16
The commentary debt at 2 Corinthians 3:7, 13 — the fading radiance.
B20
D17
~~The shorter reading of Hebrews 7:21.~~ DISCHARGED 13 August 2026 — see below.
Appendix C §I.4
D18
~~The Western reading of Luke 3:22.~~ DISCHARGED 13 August 2026 — see below.
Appendix C §VI
III. THE RECEPTION DEBT
What Appendix C has not been able to establish.
D19 · The indexes themselves have not been consulted
Neither the list loci citati vel allegati of Nestle-Aland, nor the "Index of Quotations" of the UBS edition. Everything the book says about how those lists count rests on secondary description. That is the biggest gap of Appendix C, and it touches directly the assertion about Psalm 110:1 as most-cited verse.
D20 · Two negative assertions have been made plausible, not established
The absence of Psalm 8 and Psalm 45 in Revelation rests on targeted checking of the chapters that come into consideration — not on a word-for-word search of all twenty-two. A negative statement is stronger than that research.
Likewise Psalm 45 in the rest of the New Testament has not been searched exhaustively: Matthew 22:1-14, 2 Corinthians 11:2 and Ephesians 5:25-27 have not been checked.
D21 · Psalm 110:2 and 110:5-7
Nothing has been established, and nothing has been targeted in the search either. For an appendix that claims that two of the seven verses carry the whole New Testament reception, this must be nailed down.
D22 · The counter-candidate for the superlative
Leviticus 19:18 is named as a verse that, on a pure quotation-count, possibly exceeds the five of Psalm 110:1. That count has not been verified, and the assertion therefore stands as disputed.
D23 · The image-convergence at Revelation 19:16
The name on the thigh — Psalm 45:3 and Revelation 19:16 share ἐπὶ τὸν μηρόν, and μηρός is in the New Testament a hapax. The Greek texts are verified; the attribution has not been tested against the Revelation commentaries (Fekkes, Beale, Aune). As long as that has not been done, the entry remains an own judgment. See Appendix C §II.4.
D24 · Two loose attributions
Whether Revelation 1:5 goes back to Psalm 2:2 or to Psalm 89:27 — the Greek Psalm 88 has not been retrieved. And whether anyone in the literature connects John 19:39 with Psalm 45:8; the book rejects the connection, but has not investigated how it has been defended.
IV. THE LITERATURE DEBT
This is the debt of step 9, and it is the largest entry of this register. On these points the book cites at second hand.
Not consulted, and it matters:
work
what it is needed for
Meier, two articles in Biblica 66 (1985)
his own division of Hebrews 1:5-14, and his position on the inauguration model
Attridge, Hebrews (Hermeneia)
idem — wholly unknown whether he accepts or rejects the model
Grässer, An die Hebräer I
here lies the key note where the inauguration model would come from. As long as that note has not been seen, the origin of the model is unknown
Ellingworth, NIGTC
only via a verbatim quotation in a third party; the pages not seen firsthand
Bateman 1997 (the book)
only the journal article of 1995 has been seen
Beale (NIGTC), Moyise, Fekkes, Aune
the Revelation reception, and in particular the absence of Melchizedek and the thigh of 19:16
Hay, Glory at the Right Hand; Hengel, Studies in Early Christology
the numbers 33 and 21. That the own count of this book comes out at 21 and coincides with Hengel may be coincidence — it is not known whether it concerns the same places
Koester, Lane, Johnson, Cockerill, deSilva
their structural division of Hebrews 1 has not been established
the literature on archaisms in psalm poetry
for D4
DJD for 4QDeut^q
for D14
The rule that goes with this: where the book names a name in an Expert box, it must be clear from the formulation whether the source itself has been seen. Where that is not so, it is noted.
V. THE INSTRUMENT DEBT
Faults in the measuring tool, measured and not repaired — in accordance with the rule that a prescribed run runs once.
D25 · A registered marker did not fire. The announcement-הִנְנִי bears in the coding used Strong 2005 and not 2009; the routine searched on the wrong lemma. Not repaired, but reported. The fault works conservatively: it shortens the axis and thus works against the outcome.
D26 · The direction-determination has a measured error fraction of 20%, with a confidence interval of 10.5 to 34.8 percent, established on a sample of forty laid by hand against the text. The fault sits almost entirely in one marker and likewise works conservatively.
D27 · The emptiness-line is not processed into the measure. Twenty-four units bear no hallmark whatsoever and nonetheless receive a position. As long as the measure gives an empty text a number, it lies politely. The next round must report such units as unclassifiable instead of positioning them.
All three to be preregistered separately before running again.
VI. DISCHARGED
What has been dealt with, with the date and the reason. This paragraph is there to show that the register moves.
The formula layer had been assessed on the wrong corpus — discharged, 12 August 2026
The criticism was that the test had been run on corpora that were demarcated by place in the canon and not by genre. That criticism was correct. The revision was carried out in three rounds: a blind re-test on genre, an inventory of the full prophetic form-repertoire, and a calibrated form-profile that met its calibration threshold.
What the book may now indeed say as a result stands in Appendix A §V-bis — and what it may not stands above as D9.
The word "dropped" and the bare p-values — processed, and twice
An outcome that does not reach the threshold set in advance is thereby not refuted; a p-value says nothing about the size of an effect and proves no hypothesis. The formulations have been replaced.
And here sits a lesson worth mentioning. This entry had already been signed off once — but only in the document where the correction was conceived. At the review of August the "governing" structure dossier turned out still to carry the abolished formulations in three places. A signed-off entry must be carried through in every governing document, and that must be checked and not assumed.
Why is Melchizedek missing in Revelation? — answered, verification open
The question is now answered, and the answer is sharper than expected: Revelation draws its priesthood from Exodus 19:6 and not from Psalm 110:4, it gives it to the congregation instead of to Christ, and it develops no priestly christology at all — lexically neither, for its Lamb bears no sacrificial word but a diminutive.
Moreover Revelation turns out not to cite Exodus 19:6 but to reformulate it; the verbatim quotation stands in 1 Peter 2:9. See Appendix C §IV.
What remains open is the verification against the commentaries — see §IV.
The two structural models of Hebrews 1 — decided: do not choose, and the claim to consensus lapses
The danger was that the book would choose the model that suits it best. That danger is now smaller, for there is less to choose than it seemed: the tripartition is broadly held, but on grammatical grounds — and the inauguration ritual as explanation is a minority hypothesis of which even proponents write that the indications are meager. The three principal renderings moreover disagree about where the first phase ends.
The formulation in chapter 5 has been adjusted accordingly.See Appendix C §III.
Four of the nine text-critical entries — discharged, 13 August 2026
In full in `DATA — tekstkritische schuld (afwerking D10-D18).md`. They were the four heaviest.
D18 — the Western reading of Luke 3:22. The entry does not shift. Codex Bezae reads at the baptism the full Psalm 2:7, and that reading is carried by Latin manuscripts and by patristic witnesses down into the sixth century — Justin cites it twice and adds that it stands "in the memoirs of the apostles", and Augustine already knew that the older Greek manuscripts did not have it. But none of the nine printed editions takes it into the text, and externally it stands very weak: one Greek manuscript.
Psalm 2:7 thus remains at Luke 3:22 an echo and not a quotation — with a note that honestly reports the Western reading. That the textual criticism holds it for assimilation and Ehrman holds it precisely for the older reading are two camps; the book does not choose. Acts 13:33 remains the place where Luke does undisputedly cite the psalm.
D17 — Hebrews 7:21. The shorter reading is established and has been verified firsthand in Codex Sinaiticus. Melchizedek remains eight times in the New Testament on the critical text, nine times on that of the Textus Receptus.
D11 — the missing line in Hebrews 2:7 is a later supplement from the Greek psalm; this too has been checked in Sinaiticus itself. And it yielded a second find: Sinaiticus reads in 2:8 ὑποκάτω τῶν ποδῶν αὐτοῦ — the preposition of Psalm 8, precisely as Appendix B states. Only the eightfold edition comparison remains as a residual entry.
D13 — the age of the distinction לַאדֹנִי / אֲדֹנָי. This was the heaviest entry of the register, and it is largely closed: the Masoretic tradition marks Psalm 110:5 expressly as a divine name and Psalm 110:1 not. Verse 5 stands on the list of places where the written אדני is the holy Name; verse 1 does not stand on it. That that list stands beside the points-system and not as a note in the margin indicates that the distinction is older than the points.
And one warning at the use, which must work through into the book. Part of the sources that adduce this datum do so with an interest: they use it to argue that the second "lord" of verse 1 is no divine figure. The observation and the inference are two different things. The observation stands and has been independently checked in four old translations. The inference is precisely what makes Matthew 22 a dispute: that David calls him "my lord" is the riddle, not the solution. The book adopts the observation and not the inference.
The Qumran question is answered, and the answer is a silence. This was the entry of which it stood here that it would deliver "the decisive piece of evidence". Gentry gives in his exhaustive witness table for Psalm 110:3 one line that closes the matter: "Dead Sea Scrolls — Not Extant for Psalm 110:3". There is thus no scroll that can weigh the Greek reading "I have begotten you" against the Masoretic "your youth". For the crux of this book that is definitive — and it is a negative finding that the book may simply report.
Open remains within D13: the margin of the Leningrad codex, Ginsburg's own volumes, and the question whether any scroll carries Psalm 110:1 or 110:5 — for that Flint's list is needed of which psalms have been transmitted at Qumran.
And the source that answered this resolved a second entry along the way. Gentry's appendix gives the complete witness list for Psalm 110:3 — Septuagint, Secunda, Aquila, Symmachus, Theodotion, Quinta, Sexta, both Vulgate psalters, Peshitta and Targum, each with text and translation. The revisers, who stood in this register as "not established", stand there all five written out. His conclusion moreover supports the choice of this book: the vocalization behind the Septuagint rests presumably on a defective reading tradition at a difficult verse, whereas all other early witnesses support the Masoretic text.
And one observation that the book can use. No New Testament writer cites Psalm 110:3 — and in Hebrews 1:5 and 5:5 the writer sets the "today I have begotten you" of Psalm 2:7 beside Psalm 110:1 and 110:4, whereas the Greek Psalm 110:3 offered those same words too. He could have taken it from this psalm and did not. That is a choice, and it confirms along an independent route what Appendix C established mechanically: the reception of Psalm 110 rests on two verses, and verse 3 is not one of them.
Four counts that were too strong — corrected
The finger of God stood at four places and the word at thirty-one; רָדָה was called rare and stands twenty-seven times; the companions stood at fifteen and are twelve; the kinsman-redeemer stood at eighty-four and is one hundred three. All four recounted and reduced to what they carry. See Appendix A §IV.
And one lesson from August 2026, the same as at R16 above. The kinsman-redeemer correction had been carried through in the manuscript but not in the `DOSSIER SYNTHESE`, where the old number still stood. A corrected count must be gone over in every document — the manuscripts, the dossiers and the appendices — and that must be done mechanically, not from memory.
VII. WHAT THIS REGISTER IS NOT
It is no list of weak spots for which the book apologizes. Every entry here is a place where something is to be gained — and three of them have in the past weeks been gained, two because the criticism turned out to be correct and one because the answer turned out sharper than the question.
What the register does do is guard the boundary. A book that is about revelation may not permit itself to mistake its own conjectures for findings. The four marks of this book — well documented, weighed, disputed, speculative — are worth nothing if there is no place where the fourth is honestly written down.
This is that place.
Appendix D, 12 August 2026. The working register with the full reasonings per entry stands in `RAFEL-REGISTER — losse draden en onopgeloste vragen.md` and travels along as a living document.
The open verdicts, the loose threads and the debt this book still has outstanding.
Status: 12 August 2026.
THE RULE
This book makes one promise it must keep everywhere: what stands here does not stand in the main text. A question that stands open in this register may only be used as an argument after it has been discharged here — with a date and with a reason.
That is no gesture of modesty. It is the only way to prevent a book from, over time, mistaking its own conjectures for findings. A loose thread you write down, you cannot unnoticeably weave back in.
There are three sorts of entries in this register, and they each require something different.
- Open verdicts are questions to which the text gives no unambiguous answer. They remain open, and the book shows them as open. Research does not resolve them. - Debts are matters that can be decided, but for which the work has not yet been done — an edition that has not been compared, a book that has not been consulted, a plate that has not been examined. They are dischargeable and ought to be discharged. - Instrument defects are faults in the measuring tool itself, measured and noted, not repaired within the current round.
I. THE OPEN VERDICTS
Questions to which this book gives no answer, and why it does not.
D1 · Why does בַּר stand in Psalm 2:12 and not בֵּן?
The ordinary Hebrew word for "son" is בֵּן, and it stands five verses earlier in the same psalm. In verse 12 stands בַּר. The old translators visibly diverge here, and four emendations have been proposed.
Remains open. The book gives the reading it follows, with the others beside it, and does not decide. See Appendix B, B4.
D2 · The subject of Psalm 110:7
Who drinks from the brook along the way, and who lifts up the head? The grammar allows both answers, and the choice changes the close of the psalm drastically.
Open verdict, and that is an authorial decision that has been made: the verdict remains open. So it stands too in chapter 8.
D3 · The center of Psalm 45 hangs on the colometry
The fine pass finds nested frames around מֵחֲבֵרֶיךָ — the word that Hebrews 1:9 cites. But the center cannot be established without first making a colometric choice, and that choice is precisely what is under discussion.
Not resolvable without a choice beforehand. Presented as such.
D4 · The five archaic suffixes of Psalm 2
The psalm bears five -mô-forms. That is striking, but the book draws no conclusion from it — not before the literature on archaisms in psalm poetry has been seen. That is a debt; see §IV.
D5 · The companions of Hebrews 1:9 and 3:14
The word μέτοχος occurs five times in Hebrews, once from the quotation of Psalm 45 and four times thereafter. Whether the letter deliberately lets those later places reflect back on the quotation is attractive to think and not to be substantiated.
Do not claim. See Appendix C, §I.3.
D6 · The עֲבֹת of Psalm 2:3 and the priestly garment
Two psalms that each share one word with the high-priestly garment is too little; and with עֲבֹת the ordinary meaning is more probable.
Do not claim — but noted, because with a third find it could suddenly become something after all.
D7 · Psalm 103:4 — how far does it carry?
Open. Usable as an Expert box at the kinsman-redeemer line, not as proof.
D8 · What has happened with the Hebrew of Deuteronomy 32:43?
The Greek text is longer than the Masoretic, and the scroll from cave 4 stands on the side of the Greek. Whether something has dropped out or been added, and why, is not decided — and the two explanations do not exclude each other.
Open verdict; taken up thus in chapter 2, mark weighed.
D9 · What the form-profile cannot distinguish
The calibrated form-profile test places Psalm 110 and Psalm 8 robustly on either side, but on the middle section it is blind. A text in which the instrument detects nothing lands not on zero but on a fixed line higher — and Psalm 45 lies thirteen thousandths above that line. Its measured position is therefore an artifact and no finding. And Psalm 2 and Psalm 45 cannot be distinguished from each other at all by this instrument.
Open — and the book does not use the middle group as argument. About Psalm 45 it says nothing on form-critical grounds; about Psalm 2 only the philology of verse 7. See Appendix A §V-bis.
II. THE TEXT-CRITICAL DEBT
What Appendix B reports as "not yet checked". Each of these entries is dischargeable.
III. THE RECEPTION DEBT
What Appendix C has not been able to establish.
D19 · The indexes themselves have not been consulted
Neither the list loci citati vel allegati of Nestle-Aland, nor the "Index of Quotations" of the UBS edition. Everything the book says about how those lists count rests on secondary description. That is the biggest gap of Appendix C, and it touches directly the assertion about Psalm 110:1 as most-cited verse.
D20 · Two negative assertions have been made plausible, not established
The absence of Psalm 8 and Psalm 45 in Revelation rests on targeted checking of the chapters that come into consideration — not on a word-for-word search of all twenty-two. A negative statement is stronger than that research.
Likewise Psalm 45 in the rest of the New Testament has not been searched exhaustively: Matthew 22:1-14, 2 Corinthians 11:2 and Ephesians 5:25-27 have not been checked.
D21 · Psalm 110:2 and 110:5-7
Nothing has been established, and nothing has been targeted in the search either. For an appendix that claims that two of the seven verses carry the whole New Testament reception, this must be nailed down.
D22 · The counter-candidate for the superlative
Leviticus 19:18 is named as a verse that, on a pure quotation-count, possibly exceeds the five of Psalm 110:1. That count has not been verified, and the assertion therefore stands as disputed.
D23 · The image-convergence at Revelation 19:16
The name on the thigh — Psalm 45:3 and Revelation 19:16 share ἐπὶ τὸν μηρόν, and μηρός is in the New Testament a hapax. The Greek texts are verified; the attribution has not been tested against the Revelation commentaries (Fekkes, Beale, Aune). As long as that has not been done, the entry remains an own judgment. See Appendix C §II.4.
D24 · Two loose attributions
Whether Revelation 1:5 goes back to Psalm 2:2 or to Psalm 89:27 — the Greek Psalm 88 has not been retrieved. And whether anyone in the literature connects John 19:39 with Psalm 45:8; the book rejects the connection, but has not investigated how it has been defended.
IV. THE LITERATURE DEBT
This is the debt of step 9, and it is the largest entry of this register. On these points the book cites at second hand.
Not consulted, and it matters:
The rule that goes with this: where the book names a name in an Expert box, it must be clear from the formulation whether the source itself has been seen. Where that is not so, it is noted.
V. THE INSTRUMENT DEBT
Faults in the measuring tool, measured and not repaired — in accordance with the rule that a prescribed run runs once.
D25 · A registered marker did not fire. The announcement-הִנְנִי bears in the coding used Strong 2005 and not 2009; the routine searched on the wrong lemma. Not repaired, but reported. The fault works conservatively: it shortens the axis and thus works against the outcome.
D26 · The direction-determination has a measured error fraction of 20%, with a confidence interval of 10.5 to 34.8 percent, established on a sample of forty laid by hand against the text. The fault sits almost entirely in one marker and likewise works conservatively.
D27 · The emptiness-line is not processed into the measure. Twenty-four units bear no hallmark whatsoever and nonetheless receive a position. As long as the measure gives an empty text a number, it lies politely. The next round must report such units as unclassifiable instead of positioning them.
All three to be preregistered separately before running again.
VI. DISCHARGED
What has been dealt with, with the date and the reason. This paragraph is there to show that the register moves.
The formula layer had been assessed on the wrong corpus — discharged, 12 August 2026
The criticism was that the test had been run on corpora that were demarcated by place in the canon and not by genre. That criticism was correct. The revision was carried out in three rounds: a blind re-test on genre, an inventory of the full prophetic form-repertoire, and a calibrated form-profile that met its calibration threshold.
What the book may now indeed say as a result stands in Appendix A §V-bis — and what it may not stands above as D9.
The word "dropped" and the bare p-values — processed, and twice
An outcome that does not reach the threshold set in advance is thereby not refuted; a p-value says nothing about the size of an effect and proves no hypothesis. The formulations have been replaced.
Why is Melchizedek missing in Revelation? — answered, verification open
The question is now answered, and the answer is sharper than expected: Revelation draws its priesthood from Exodus 19:6 and not from Psalm 110:4, it gives it to the congregation instead of to Christ, and it develops no priestly christology at all — lexically neither, for its Lamb bears no sacrificial word but a diminutive.
Moreover Revelation turns out not to cite Exodus 19:6 but to reformulate it; the verbatim quotation stands in 1 Peter 2:9. See Appendix C §IV.
What remains open is the verification against the commentaries — see §IV.
The two structural models of Hebrews 1 — decided: do not choose, and the claim to consensus lapses
The danger was that the book would choose the model that suits it best. That danger is now smaller, for there is less to choose than it seemed: the tripartition is broadly held, but on grammatical grounds — and the inauguration ritual as explanation is a minority hypothesis of which even proponents write that the indications are meager. The three principal renderings moreover disagree about where the first phase ends.
The formulation in chapter 5 has been adjusted accordingly. See Appendix C §III.
Four of the nine text-critical entries — discharged, 13 August 2026
In full in `DATA — tekstkritische schuld (afwerking D10-D18).md`. They were the four heaviest.
D18 — the Western reading of Luke 3:22. The entry does not shift. Codex Bezae reads at the baptism the full Psalm 2:7, and that reading is carried by Latin manuscripts and by patristic witnesses down into the sixth century — Justin cites it twice and adds that it stands "in the memoirs of the apostles", and Augustine already knew that the older Greek manuscripts did not have it. But none of the nine printed editions takes it into the text, and externally it stands very weak: one Greek manuscript.
Psalm 2:7 thus remains at Luke 3:22 an echo and not a quotation — with a note that honestly reports the Western reading. That the textual criticism holds it for assimilation and Ehrman holds it precisely for the older reading are two camps; the book does not choose. Acts 13:33 remains the place where Luke does undisputedly cite the psalm.
D17 — Hebrews 7:21. The shorter reading is established and has been verified firsthand in Codex Sinaiticus. Melchizedek remains eight times in the New Testament on the critical text, nine times on that of the Textus Receptus.
D11 — the missing line in Hebrews 2:7 is a later supplement from the Greek psalm; this too has been checked in Sinaiticus itself. And it yielded a second find: Sinaiticus reads in 2:8 ὑποκάτω τῶν ποδῶν αὐτοῦ — the preposition of Psalm 8, precisely as Appendix B states. Only the eightfold edition comparison remains as a residual entry.
D13 — the age of the distinction לַאדֹנִי / אֲדֹנָי. This was the heaviest entry of the register, and it is largely closed: the Masoretic tradition marks Psalm 110:5 expressly as a divine name and Psalm 110:1 not. Verse 5 stands on the list of places where the written אדני is the holy Name; verse 1 does not stand on it. That that list stands beside the points-system and not as a note in the margin indicates that the distinction is older than the points.
The Qumran question is answered, and the answer is a silence. This was the entry of which it stood here that it would deliver "the decisive piece of evidence". Gentry gives in his exhaustive witness table for Psalm 110:3 one line that closes the matter: "Dead Sea Scrolls — Not Extant for Psalm 110:3". There is thus no scroll that can weigh the Greek reading "I have begotten you" against the Masoretic "your youth". For the crux of this book that is definitive — and it is a negative finding that the book may simply report.
Open remains within D13: the margin of the Leningrad codex, Ginsburg's own volumes, and the question whether any scroll carries Psalm 110:1 or 110:5 — for that Flint's list is needed of which psalms have been transmitted at Qumran.
Four counts that were too strong — corrected
The finger of God stood at four places and the word at thirty-one; רָדָה was called rare and stands twenty-seven times; the companions stood at fifteen and are twelve; the kinsman-redeemer stood at eighty-four and is one hundred three. All four recounted and reduced to what they carry. See Appendix A §IV.
VII. WHAT THIS REGISTER IS NOT
It is no list of weak spots for which the book apologizes. Every entry here is a place where something is to be gained — and three of them have in the past weeks been gained, two because the criticism turned out to be correct and one because the answer turned out sharper than the question.
What the register does do is guard the boundary. A book that is about revelation may not permit itself to mistake its own conjectures for findings. The four marks of this book — well documented, weighed, disputed, speculative — are worth nothing if there is no place where the fourth is honestly written down.
This is that place.
Appendix D, 12 August 2026. The working register with the full reasonings per entry stands in `RAFEL-REGISTER — losse draden en onopgeloste vragen.md` and travels along as a living document.