The Mass and its propers

Most of a Mass never changes: the same Kyrie, the same Creed, the same Canon, every day of every year. That invariant part is the Ordinary. Woven through it are about ten short texts that change with the day — a different opening chant, a different prayer, a different pair of readings — and those are the propers. Knowing what Mass is being said on a given day means knowing which propers are appointed for it.

Two books of propers are in use. The 1962 Roman Missal is the last edition before the Second Vatican Council; the postconciliar Missal (here the editio typica tertia of 2008) replaced it. They share a rite, a language, and a great many actual texts, but they cut the year differently, name the slots differently, and read scripture on entirely different schedules. This repository models both, side by side, and refuses to harmonize them where they genuinely differ.

Every count below is derived from the repository's own files by tools/tpt mass-propers census, and lands in one table under What is actually here. That command is the only thing that writes the table, and the prose around it states no figure the table does not. It had to become one derived table because it was three typed ones: this page, guidance/propers-for-agents.md and the calendars' own README each carried a retyped copy of the same census, and all three disagreed.

Browse the propers — every Mass of both calendars, with each appointed passage rendered in the translation you choose.

Part one — the two books

The concrete case: the First Sunday of Advent

The day opens the liturgical year in both books, and both books open it with the same chant.

In the 1962 Missal it is Dominica Prima Adventus. Its first proper is the Introit, the chant sung as the priest approaches the altar; its incipit — the first words, which is how a chant is identified — is Ad te levavi, built out of Psalm 24, verses 1 to 3, with verse 4 as its psalm verse. In the Douay-Rheims that this repository has indexed, those verses read:

To thee, O Lord, have I lifted up my soul. In thee, O my God, I put my trust; let me not be ashamed. Neither let my enemies laugh at me: for none of them that wait on thee shall be confounded.

The whole 1962 Mass of that day has ten propers:

The postconciliar Mass of the same Sunday also has ten, with the same three chants under the same incipits:

Four things in that pair are each a subject below.

1. Psalm 24 and Psalm 25 are the same psalm. So are 84 and 85. The two books number the psalter differently — see Two numberings.
2. The 1962 Epistle is Romans 13:11-14; so is the postconciliar Second Reading, but only in Year A. The postconciliar readings rotate on a three-year cycle — see One cycle against three.
3. The Collects are different prayers. Chants often survived the reform unchanged; orations were frequently rewritten or replaced.
4. The Douay text above is the psalm, not the antiphon. The Introit's actual sung words are a lightly adapted form of those verses, and this repository does not carry them. It records what a chant is made of, not what it says.

Chants, orations, readings

The Ordinary is out of scope here entirely: the calendar indexes record no part of it, and record no prefaces either. An index of what varies has nothing to say about what does not.

What varies falls into three kinds:

• Chants — Introit, Gradual, Alleluia, Tract, Offertory, Communion; in the postconciliar books, the Entrance and Communion Antiphons and the Responsorial Psalm and Gospel Acclamation. Almost all are scripture, usually a psalm, cut to a few verses.
• Readings — Epistle and Gospel in 1962; First Reading, Second Reading and Gospel in the postconciliar books — scripture outright.
• Orations — Collect, Secret, Postcommunion (1962); Collect, Prayer over the Offerings, Prayer after Communion (postconciliar). These are composed Latin prayers. They quote scripture only incidentally and are not built from it.

Only scripture-bearing propers contribute anything to a passage-keyed index; the orations contribute nothing, because their words are their own. Set the placeholders aside and rather more than half of what remains carries scripture in either book; the census below counts both.

Propers are also not a fixed template. The repository stores each Mass's propers in the order the edition appoints them, so Tracts, Sequences, the palm-rite antiphons, the Improperia, the Exsultet, the Holy Saturday prophecies and the litanies appear only where they are actually appointed, and a ten-row Sunday template is never imposed on a day that does not have ten rows. Each file uses scores of distinct slot names — the census below counts them — and many occur exactly once: First Antiphon at the Imposition of Ashes, Exsultet (Praeconium paschale), Improperia, Washing of Feet, Antiphon 4, Chant at the Place of Reposition.

The temporal cycle and the sanctoral cycle

Every Roman missal is organized around two overlapping cycles. The temporal cycle (the Proprium de Tempore) follows the life of Christ through the seasons: Advent, Christmas, Lent, Easter, and the long stretch after Pentecost. Its days are movable, because they hang off Easter, which moves. The sanctoral cycle (the Proprium Sanctorum) is a fixed-date list of feasts of the Lord, of Our Lady, and of the saints: 25 December, 25 March, 29 June, and so on through the year.

The two run simultaneously, which is why a single calendar date can carry more than one candidate celebration. 26 December is both the second day within the Octave of the Nativity (temporal) and the feast of Saint Stephen (sanctoral). Which wins is a question of precedence, settled by rank and by rubric, and it is a separate question from which formularies exist.

This repository classifies every celebration by kind — seasonal, christological, marian, or sanctoral — and the classification follows the edition's own printed title, not a modern judgement. The 1962 books title 2 February and 25 March as feasts of Our Lady (In Purificatione, Annuntiatio Beatae Mariae Virginis) while the postconciliar books title them as feasts of the Lord, and the repository reproduces that disagreement rather than resolving it.

Where the two books diverge

Septuagesima is the cleanest case of a season existing in one book and not the other. In the 1962 data it is a season in its own right holding three Masses; in the postconciliar data there is no such season and no such key.

The postconciliar Ordinary Time has its own arithmetic quirk: there is no First Sunday in Ordinary Time at all. Week I is the week in which the Baptism of the Lord falls, and the Baptism replaces that Sunday, so the numbered Sundays run from the Second to the Thirty-fourth. The repository's data agrees: its Ordinary Time Masses are keyed ot-2 through ot-33, thirty-two of them, with the Thirty-fourth Sunday carried separately as Christ the King.

One cycle against three

The single largest structural difference is the Lectionary. The 1962 system appoints formulary and readings together: the First Sunday of Advent has one Epistle and one Gospel, the same every year, forever. The postconciliar system separates them — a Sunday has one set of chants and orations but three sets of readings, and which set is used depends on whether the liturgical year is Year A, B or C.

The letter is fixed by the civil year in which the liturgical year ends, not the year it begins. Call that year Y; then Y mod 3 = 1 is Year A, 2 is Year B, 0 is Year C, and the cycle turns at the First Sunday of Advent and nowhere else. Weekdays in Ordinary Time run a separate two-year cycle, Year I in odd years and Year II in even ones, which is independent: with periods 3 and 2 the pairing repeats only every six years, and neither letter can ever be derived from the other.

In the repository's data the propers that carry a cycles mapping are the Gospels, the Second and First Readings, the Responsorial Psalms and the Gospel Acclamations, plus a handful of Holy Week items; the census below counts them. Every one of them is postconciliar. No 1962 proper carries a cycle, and none ever should.

The calendar arithmetic, and its hard limit

guidance/liturgy/calendar-computation.md is the single authoritative statement of the arithmetic. Nothing else in the repository may restate it; everything else references it.

Easter is the Sunday after the first ecclesiastical full moon falling on or after 21 March, by the Gregorian computus — the tabular ecclesiastical rule, not an astronomical full moon. Every other movable day is a fixed offset in days from Easter:

The liturgical year itself turns at the First Sunday of Advent, the fourth Sunday before 25 December, which falls between 27 November and 3 December.

And then the constraint that governs all of it:

Computation is a finding aid, never an authority.

The competent General, territorial, diocesan, religious, parish and church calendar, the approved books, and the competent annual Ordo control what is actually celebrated. The rule here is: compute, then verify against a dated official witness, record both, and fail closed on disagreement — state the divergence and resolve nothing by preference. A computed date is never published as its own occurrence witness.

The sharpest form of this is the territorial transfer. Epiphany, the Ascension, and the Body and Blood of Christ are moved to a Sunday in some places and not in others, and that is a decision of the competent authority, not derivable from any calendar arithmetic whatsoever. The offsets table gives both the Thursday and the transferred Sunday form of Ascension and Corpus Christi precisely because no formula can choose between them.

Two smaller cases are recorded as deliberately unresolved rather than guessed: a 1962 year with only twenty-three Sundays after Pentecost (neither identified normative text provides for the shortfall), and any proper-calendar overlay that has not actually been checked.

Two numberings

The Latin psalter is numbered two ways, and over most of its length the Vulgate number runs one behind the Hebrew one: Vulgate 24 is Hebrew 25, Vulgate 84 is Hebrew 85, Vulgate 118 is Hebrew 119. That is why the 1962 Introit cites Psalm 24 and the postconciliar Entrance Antiphon cites Psalm 25 for the same words. The correspondence is not a formula — six psalms divide between the two systems — so the repository converts through a tracked verse-level concordance built from the Challoner edition, and never from a typed table.

Each calendar file declares its own numbering. roman-1962 declares vulgate and means it throughout. postconciliar declares hebrew — and here the books themselves are not consistent. The file's own recovered citation_convention states the rule:

Lectionary readings and antiphons preserve the numbering in the controlling missal

Which is to say: a postconciliar antiphon reproduces the number printed in the Missal, and a responsorial psalm or acclamation takes the number printed in the Lectionary, and those are not always the same system. The Twenty-third Sunday in Ordinary Time shows it inside a single Mass: its Entrance Antiphon is cited Psalm 118:137, 124 and its Year C Gospel Acclamation Psalm 119:135. Both are the psalm Beati immaculati. One is numbered Vulgate, the other Hebrew.

Part two — what this repository holds

How a calendar file is shaped

One file holds each calendar, under a header that declares the controlling edition, the ordering rule, the psalm numbering, the citation convention, and the verification state. Each Mass carries a stable key, the edition's own catalog name, its identity in the registry that numbers it, and its propers in the order the edition appoints them.

Every proper declares whether its text is scripture, composed, or both, and the rest of its shape follows: a scripture proper carries the passages it is built from and no text, a composed proper carries the Latin text and no passages, and the validator refuses either one carrying the other. A proper whose content varies by Lectionary year carries the A, B and C forms separately. A celebration with several complete formularies — the Nativity's four Masses, the Ember Saturdays the Missal prints in a longer and a shorter form — carries each formulary whole.

References are stored as structured extents rather than as strings to be re-parsed: one or more contiguous ranges under a canonical book name, with the edition's own printed display string kept beside them and never treated as authoritative. The encoding carries what liturgical books actually print — a selection within a chapter, a range crossing a chapter boundary as the Passions do, a chant cited whole by chapter, a printed part-verse letter — and the encoder refuses a citation it cannot encode without guessing rather than writing a wrong range.

What is actually here

The seasonal sections carry real propers. Most of the rest is a name, not a formulary: a mass that holds nothing but a placeholder establishes that the mass exists and where it falls, and records nothing of what is said at it. Coverage is wide and shallow by design, and a reader who sees a Mass listed should not infer that its texts are here.

The table below is derived, not typed. tools/tpt mass-propers census --write rewrites it in place, make check-propers-census fails when it has drifted from the calendars, and nothing else on this page states a figure it gives.

Counted from src/sources/calendars/*/propers.yaml and written here by tools/mass-propers census --write, which is the only thing that writes the block above; make check-propers-census refuses a copy that has drifted. What each row counts, because two honest counts of “propers” differ by hundreds when they key differently: a mass is one entry under sections[*].masses. A proper is one entry in a mass's propers, or in the propers of one of its forms; a proper carrying cycles counts once, not three times. Placeholders are inside the proper and mass totals, and are also given their own rows. A mass holds only placeholders when every proper it holds, those inside forms included, is named Placeholder — keying on the mass's own propers alone undercounts, because it misses the masses whose placeholders sit inside a forms block. Scripture-bearing means a source of scripture or mixed, or a cycles entry that is. Encoded passages and distinct books are tools/citations check's own counts, one passage per encoded citation entry and books counted distinct within a file. Distinct slot names counts distinct proper name values, with Placeholder among them. The two taking rows count the entries that name where their text is printed instead of printing it — a feria taking the preceding Sunday, a saint taking a Mass of the Common. Such an entry holds few propers or none, so it lowers the proper count while raising what the calendar can actually show: every row above counts what a file carries, and these two count what it appoints from elsewhere. Neither is a placeholder.

The rank rows count entries, and a calendar's rank rows sum to its Masses above. (no rank) is a row and not an omission: the temporal cycle prints no rank at all, so unranked entries are a large group in both indexes, and a table that dropped them would invite the ranks it does show to be read as the whole book. Celebrations counts what those entries name. One entry can print more than one celebration — a pair of optional memorials falling on the same day, or the Greater Litanies kept with Saint Mark — and the index records the join only in the entry's name, joined with ;, so that is where it is read from; the plural rank word marks some such entries and not others, and it never says how many. Where the two columns differ the index is keeping more celebrations than it holds entries: reading the entry count as a count of celebrations understates the calendar, and folding the celebrations into the entry count would break the sum against Masses, so both are given and neither replaces the other. A rank is reproduced exactly as the file prints it. Optional memorial and Optional memorials are two rows because they are two different words in the index, and the two Missals' rank vocabularies are not one scale — no row of this table is comparable across calendars, which is why rank is tabulated down the page and not across it. This table answers what an index is scoped to; it does not answer which entries are Sundays, because no entry states that and it is not derived here.

The 1962 temporal expansion, and what it deliberately omits

The 1962 seasonal section began as 59 Masses — the Sunday run plus the Triduum and Palm Sunday. Its largest single expansion added 63 temporal days, and the three Christmas octave placeholders came with it; a few more entries have landed since, and the census above holds the current figure. The 63 cover the days the Missal keeps and the file did not — the Ember Days of all four seasons, Ash Wednesday and the days after it, the ferias of Lent and Passiontide and of Holy Week to Wednesday, the Chrism Mass, the Easter and Whitsun octaves, the Rogation Mass, the Vigil and feast of the Ascension, the Vigil of Pentecost, the Mass of the First Sunday after Pentecost, Corpus Christi, and the Sacred Heart.

Every one of those 63 entries carries only its scripture-bearing propers, and says so in its own notes:

Only the scripture-bearing propers are recorded; the day's orations and other composed chants are appointed but not yet transcribed.

This is a deliberate refusal. The orations of those days exist and are appointed; they were not transcribed because they were not read. The alternative would have been to generate plausible Latin, and a plausible wrong oration is worse than an absent one.

Identity, rank and the printed scripture references were read from an OCR text layer of the CMAA 1962 facsimile, not from the facsimile images. Confidence in each day's identity and rank is high; confidence in each citation is only as high as an unproofed scan allows; nothing has been visually collated.

Eleven antiphons that fell between the numberings

Eleven antiphons carried the Missal's printed Vulgate number inside a file declaring Hebrew numbering. Because Hebrew 118 ends at verse 29, Psalm 118:137 is not a resolvable verse at all under the file's own declaration — and for a long time nothing noticed, because the psalter bounds check held verse ceilings only for the six psalms that split between the systems. The eleven are:

Current state. Each proper may now declare its own numbering rather than being forced into its calendar's, and all eleven carry that declaration; the validation path has not caught up, so all eleven still report as out of bounds and are set aside through a listed exceptions ledger. That ledger is self-cleaning in both directions: an out-of-bounds psalm at an unlisted locus still fails the build, so a new leak cannot hide behind the known ones, and a listed locus that has stopped breaching also fails, so an entry cannot outlive the defect that earned it. Correcting the eleven — deciding, per slot, whether the number moves or the declaration does — is tracked as TASK-32.

Two other references cannot resolve for reasons upstream of any of this: 4 Esdras 2:36-37 is not among the Douay-Rheims' 73 books, and Malachi 3:19-20a is Hebrew numbering where the Vulgate prints Malachi 4:1-6. Only psalms are converted between systems today.

Divergences outside the psalter, recorded by hand

Only the psalter has a concordance; nothing else converts. The postconciliar Lectionary cites the Nova Vulgata, no witness of that versification is tracked here, and every bible in this library follows the Vulgate division. Where the two divide a book differently — Joel, Malachi, and single chapters of Isaiah and Micah — the reference does not fail. It resolves, to different words: Joel 3:1-5 returned the valley of Josaphat instead of the outpoured spirit, in Latin and English alike, and appeared in no error count.

So a calendar records the correspondence by hand, citation by citation, in a citation_divergences list. A correspondence that holds unchanged is written out anyway rather than omitted, on the principle that inside a divergent locus, silence is not evidence that anyone checked. A citation reaching a divergent locus with no correspondence recorded is refused — left out of every index and reported unresolved — because a missing passage is a question and a plausible wrong one is an answer.

Three days that fell between two records

29, 30 and 31 December 1962 — the fifth, sixth and seventh days within the Octave of the Nativity — sat in the date-ordered list of celebrations and in no section of the propers file at all. They are there now, as placeholders, and every celebration the date list carries must now have a Mass, filed under the kind that list assigns it, or the build fails.

Closing the gap meant settling how they classify, and the answer was in the edition's own punctuation. Within an octave:

• a colon introduces the day's own principal celebration — "Second day within the octave: S. Stephani Protomartyris" is Saint Stephen's feast, set within the octave, and classifies sanctoral;
• a semicolon introduces a mere commemoration — "Fifth day within the octave; comm. S. Thomae Episcopi et Martyris" is the octave day itself, and classifies seasonal.

So in the 1962 data, 26–28 December are filed sanctoral and 29–31 December seasonal.

The formulary those three days actually use is still absent. It is one of three the Missal carries that this file does not — the others being D. N. Iesu Christi Regis and Sanctissimi Nominis Iesu.

Everything here is an unverified lead

A calendar index is a planning and cross-reference spine, not a source of record. It carries no artifact hash and proves nothing on its own; a publication still binds the edition and artifact that control each text through its own research/source-bindings.toml.

Both files say so in their verification header. The 1962 file states three tiers of confidence; the postconciliar file states two. Every citation and every text is an unverified lead until collated against the controlling edition, and each file tracks its known problems in open_collation_items rather than silently harmonizing them away. The instruction is explicit: fix a divergence by collation, not by making the file falsely uniform. open_collation_items in both files ends with the line all entries in this file are placeholders pending source-backed completion.

Open questions

Each of these is recorded in the repository as unresolved, not silently decided.

Where to look next

To read a Mass:

tools/tpt mass-propers show --calendar roman-1962 --mass advent-1 --bible douay-rheims
tools/tpt mass-propers list --calendar postconciliar