The day has spoken, and it pointed midweek. But one objection lingers, and it is not the least: the church has held to the Friday from her earliest days, and the year she names is not AD 31. A reading with so many centuries against it bears the burden of proof, and that burden is heavy. So before the year comes up, this question must be faced, and faced as honestly as the subject deserves. It falls into two parts, for the tradition hands down two things: a year, and a day. And of both the question is: is this a memory, or a computation? Did anyone carry this knowledge independently, or was it inherited and passed on?
I state the outcome in advance, for it is the heart of this chapter. The tradition does not remember the year. It calculates it. And it calculates, down to the last witness, with the same two ingredients: the fifteenth year of the emperor Tiberius from Luke 3:1, and a guessed length of the ministry. That is no malice and no conspiracy; it is simply how a church that had preserved no year-count of the crucifixion reconstructed one anyway. But it means that the sentence "but the tradition says AD 33" — or AD 29, or any year whatever — carries no weight that could overrule the astronomy or the text. The tradition has measured nothing here to overrule them with.
See how clean the pattern is. Walk through the ancient writers who name a crucifixion year, set beside each the length of ministry he assumes, and the whole fan of dates falls into place. Whoever assumes a ministry of zero years — Jesus baptised and dying shortly after — arrives at AD 29 (Tertullian, Lactantius, Augustine). Whoever assumes one year, at AD 30 or 31 (Clement, Africanus). Whoever assumes three years, at AD 32 (Hippolytus, Epiphanius). Whoever assumes three and a half to four years, at AD 33 (Eusebius). There is no residual scatter, not a single source that gives a year that does not follow neatly from this sum. And Clement of Alexandria lets you watch while he makes it: "fifteen years of Tiberius and fifteen years of Augustus, so the thirty years were completed until He suffered" (Clement, Stromata I.21). That is not testimony. That is subtraction and addition, out loud.
Dossierthe two family trees
The family tree of the year. The spread of the patristic year-dates correlates completely with the assumed length of ministry, and with nothing else:
assumed ministry
outcome
witnesses
~0 years
AD 29
Tertullian, Lactantius, Augustine
1 year
AD 30/31
Clement, Africanus
3 years
AD 32
Hippolytus, Epiphanius, Jerome
3½–4 years
AD 33
Eusebius
The year is thus everywhere a derivative of Luke 3:1 plus a guess — never a preserved date. What looks archival is not: "the two Gemini" (AD 29) is merely Tiberius's fifteenth year in consular form, and the Acta Pilati confess the confusion in their own text.
The family tree of the day. Two lines run through the centuries. The Friday line is old and broad: Justin (155), Hippolytus (±230), Victorinus (±300), the Didascalia (3rd century), Hilarianus (397) — unanimously paraskeuê. The midweek line is narrower but tough: the Wednesday fast — alongside the Friday fast — which the Didache already prescribes around the year 100 (Didache 8:1), a century before the Didascalia, though it does not yet name the reason; the Didascalia's Tuesday-evening supper and Wednesday arrest (±225); the "many" who according to Gregory of Tours (±590) held a Saturday resurrection; and in the modern age Chwolson (1908) and Jaubert (1957). Both lines exist; the question is which carries a memory and which a chain of reasoning. [DOCUMENTED]
That is the first half, and it exonerates no one and burdens no one — except whoever thought the tradition settled the matter. For see what happens to the three candidate years. AD 33 is the choice of Eusebius, but his family tree is one man deep and hangs on the longest ministry. AD 31 surfaces in Africanus, but his year is internally contradictory — it follows not from his own Tiberius year but from an Olympiad that clashes with it — and it rests on a one-year ministry that Irenaeus, older and closer to the source, expressly rejects. A second patristic line does run to 31 — Orosius (around 417) and Gregory of Tours (around 590) date the passion to the seventeenth year of Tiberius, which falls in the spring of 31 — but it too is a derivation from Luke 3:1 with an assumed length, not an independent memory. None of the years, then, has an independent ancient witness. The patristic field is empty — for everyone. And an empty field does not refute the case; it gives the case back to the place where it belongs — the astronomy and the text.
Then the second half, the day, and here I must be honest in a way that seems to help the other side. Every ancient writer who names a weekday names the Friday, and early. The oldest weekday witness is Justin Martyr, around the year 155: "on the day before that of Kronos they crucified Him" — the day before Saturday, hence Friday (Justin, 1 Apology 67). Half a century before Tertullian, and crystal clear. Not one surviving source proclaims, in its own voice, a Wednesday crucifixion; in the ancient church the Wednesday is the day of the betrayal, not of the cross, and that is why one fasted on Wednesday. That must be said, by me, before a critic says it: on the weekday, the early church, so far as her texts survive, does not stand on this book's side.
And yet "unanimous" is the wrong word, and it is precisely the sources that show why. For the other reading was fought — and no one fights what no one holds. Gregory of Tours must still, in the sixth century, refute the Saturday resurrection that "many" held; the Apostolic Constitutions cut the Tuesday evening out of the Didascalia; Epiphanius ranks among his heretics the "Alogi," who read the Johannine Passover chronology differently. That is not silence — that is defence, and defence is the trace of a living opponent. Here applies the rule that underlies this whole chapter: a suppressed reading leaves no texts of its own; it leaves behind the refutations of its opponents. What has come down to us is therefore not "the unanimous church" but the voice of the party that won — and the refutations by which it won are at the same time the evidence that there was something to refute. The Friday has been handed down; the debate lies preserved beneath it. We know of that debate only because the Friday had to defend itself against it — and one does not defend oneself against a phantom.
But it is the same diagnosis as with the year. They remember nothing; they read — and they read through a language that had meanwhile shifted. In the first century the Greek paraskeuê meant "preparation" — the day before a day of rest, whether that was the weekly sabbath or an annual festival sabbath. But in Christian Greek the word became the fixed synonym for Friday; to this day paraskevi is the ordinary Greek word for that day. The fathers thus read back into the gospels what their own language had meanwhile made of the word. And the opening they passed over stands plainly in the text: John writes that the day of that sabbath "was a great day" (John 19:31) — the festival sabbath, not necessarily the weekly one. The fathers are numerous, but they are not independent witnesses; they are later readers of a single shifted word.
Experthow paraskeuê became "Friday"
In the New Testament paraskeuê (παρασκευή) is still simply "preparation, making ready": the day on which one prepares for the day of rest that follows (Mark 15:42, where it is glossed as prosabbaton, "the day before the sabbath"). So too — relatively, not as a name — the oldest non-Christian witness uses the word: the Augustan decree exempting the Jews from legal proceedings "on the sabbath, or on the preparation day before it, from the ninth hour" (Josephus, Ant. 16.163) — older than the gospels, with a time of day, and yet no weekday name: the "before it" needs the day of rest beside it in order to point to anything (the full weighing of this strongest counter-witness stands in ch. 13). In first-century Judaism that day of rest could be the weekly sabbath or an annual festival sabbath. In later Christian Greek the word narrowed into the fixed name of the sixth weekday — a semantic shift that runs on into modern Greek (paraskevi = Friday). Whoever reads the gospels with the later meaning hears "Friday" where the text says "preparation day"; and a preparation day also preceded the great festival sabbath of 15 Nisan (ch. 4). The fathers read the later language back into the oldest text — no bad faith, but no independent memory either.
And yet the midweek line cannot simply be waved away, for it has left a trace too concrete to ignore: the Didascalia. This is no heretical writing but a catholic, liturgical church order from the third century, and it tells the passion week longer than the ordinary scheme allows. "When we had eaten the Passover on the Tuesday evening, we went out to the Mount of Olives, and in the night they seized our Lord; and the next day, which was Wednesday, He remained in custody" (Didascalia XXI; Connolly, 1929). A Tuesday-evening supper, an arrest in the night from Tuesday to Wednesday — a liturgical memory of a week that stretched over more than the eighteen hours of the Friday scheme. And it preserves a detail no third-century redactor invents: that Judas received his money on the tenth of the month, the Monday — the tenth of Nisan, the day on which the lamb is set apart (Exodus 12:3). "How ancient must that be," wrote Merx beside that line, and he was right: it is a liturgical fossil.
Excursusthe seam that betrayed the redaction
The finest trace that something was smoothed away here is not the Didascalia itself but what a later hand did with it. The Apostolic Constitutions (±380) are a reworking of the Didascalia, and they have cut the Tuesday evening out: the Passover meal now stands on Thursday. But they left the scaffolding standing. In their text the sanhedrin still meets on Monday, Tuesday and Wednesday — three council days that are utterly pointless in a Friday chronology in which nothing happens until Thursday evening. Their Passover fast still begins on Monday — six days, though the suffering would begin only on Thursday evening. And most revealing of all: immediately after their Thursday–Friday chronology they cite Matthew 12:40, "three days and three nights in the heart of the earth," as fulfilled — though Friday evening to Sunday morning counts two nights, not three (Constitutiones Apostolorum V.14–18). They kept the text and broke the sum. The beam has been taken out; the scaffolding still stands. And Epiphanius says in so many words that one fasts on Wednesday "because the Lord was arrested on Wednesday morning" — the arrest, not the decision (Epiphanius, De Fide 22.7; GCS 37, Holl). That is not the proof of a Wednesday crucifixion; it is the trace of a longer week that was later cut short.
But — and this is where this chapter draws its own limit, out loud, so that no one else has to — the Didascalia delivers no Wednesday crucifixion. Three sentences after the Tuesday meal it says that they crucified Him "on that same Friday." It lengthens the custody, not the time in the tomb. Whoever deploys it as a witness for a Wednesday crucifixion abuses the source — and this book does not, even though it would be convenient. The Didascalia is a breach, not proof: it shows that the early church itself wrestled with the length of the week and with the three nights, but it does not share this book's conclusion.
The same honest reservation holds for the two witnesses who come closest to this book's reading. Gregory of Tours reports, around 590, that "many" believed the resurrection took place on the seventh day, the Saturday — a view that presupposes a midweek crucifixion, and that was evidently tough enough to force a bishop to refute it (Gregory, Historiae I.23). But he names no group, gives no reason, writes five hundred years after the fact, and there is an innocent explanation that cannot be excluded. And the modern forerunners each go only part of the way: Daniel Chwolson argued a death on the fourteenth of Nisan (ch. 5) — but kept the Friday; Annie Jaubert reconstructed from the solar calendar of Qumran a Tuesday-evening supper that reconciles the synoptic and Johannine chronologies (Jaubert, 1957) — but she too kept the Friday. It is honest to say so: the midweek line has a real family tree, down into the best modern scholarship, but even its champions stopped at the Friday. This book goes one step beyond them, to the Wednesday, and it carries that step itself — it does not lean on them as though they had already taken it.
Countercasethe tradition at its strongest
Let the objection keep its full weight. The Friday is no mere late habit: it is attested from Justin in 155, within a century and a half of the event; it stands in the creeds; it carries the liturgy of the whole church, east and west, to this day. Not a single ancient source teaches a Wednesday crucifixion. A reading with such credentials against it, it will be said, cannot be right; the burden of proof is crushing. This is the strongest form of the argument, and I do not soften it.
The answer is not that the fathers lie, but that they are not what they are taken to be. On the year they demonstrably calculate — the spread betrays the sum. On the day they read through a word whose meaning had meanwhile shifted. They are numerous, but they are not independent: it is one stream that branched, not a multitude of witnesses confirming one another. And precisely beneath their apparent unanimity lies — in the Didascalia, in the seam of the Constitutions, and in the polemic the Friday had to wage against the midweek reading — the preserved trace of a longer week; for a reading one defends oneself against was a reading that lived. The tradition is not a wall that closes the case; it is a palimpsest, and beneath the top layer an older one shows through.
Where does the investigation stand, then, at the end of the day-axis? The earliest voices do not decide the matter — in either direction. They do not remember the year but calculate it, and they read the day through a shifted language; and where they seem most independent, in the liturgical week of the Didascalia, they preserve precisely the trace of a longer week that they later cut short. That is the honest balance: on the text, the day points midweek; the tradition sets no independent weight against it; and the midweek reading is old enough to be no modern notion. But a day is not yet a date. Wednesday — which Wednesday, in which year? For that the investigation must leave the text and go up into the sky. To the moon over Jerusalem, and to the years it allows.
HeartHEART
There is something moving about a fossil in an order of prayer — a Monday left standing because the lamb is set apart on the tenth, a fast that begins too early because it is older than the redating it was made to carry. The church has, without knowing it, preserved in her own liturgy the memory her scholars later smoothed away. That is no indictment of the fathers; it is a sign of how faithful habit is, even where the explanation shifts. The days are counted here to find the truth. But beneath the counting lies a prayer older than any reconstruction, and it fasts anew each week, on the day of the betrayal — as though the body of the church remembers something her head has forgotten.
Note on the sources
The patristic witnesses have been consulted in the source itself or through a named scholarly translation, with location; Justin (1 Apol. 67), the Didascalia (Connolly, 1929, alongside the Lagarde text and the searchable translation of Gibson, 1903, p. 94, which gives the same Tuesday evening and the same Wednesday in custody: "when we had eaten the Passover on Tuesday in the evening, we went out to the Mount of Olives, and in the night they took our Lord Jesus; and on the next day, which was Wednesday, He remained in prison"), the Constitutions (ANF) and Gregory (Historiae I.23, Latin) have been checked in the primary text. Epiphanius' Wednesday fast is located in De Fide 22.7 (GCS 37, Holl); the Greek text itself has not been checked, and it carries here as corroboration, not as primary carrier. The correlation between the patristic year-dates and the assumed length of ministry can be recomputed from the witness table. This chapter is a corroborating ring: it shows that the tradition does not measure the year but calculates it, and that the midweek reading has an old family tree — but it delivers no Wednesday witness, and it is not advanced as one. The Didascalia is a breach, not proof; Chwolson and Jaubert themselves kept the Friday. This chapter claims a proof for neither side.
Literature (APA)
Chwolson, D. A. (1908). Das letzte Passamahl Christi und der Tag seines Todes (2nd ed.). Leipzig.
Connolly, R. H. (Trans.). (1929). Didascalia Apostolorum: The Syriac version translated and accompanied by the Verona Latin fragments (ch. 21). Clarendon Press.
Donaldson, J., & Roberts, A. (Eds.). (1886). The Ante-Nicene Fathers (Vol. VII: Constitutiones Apostolorum, V.14–18; VII.23). Christian Literature Company.
Gibson, M. D. (Trans.). (1903). The Didascalia Apostolorum in English (Horae Semiticae II, p. 94). Cambridge University Press.
Gregory of Tours. (1951). Libri historiarum X (B. Krusch & W. Levison, Eds.; Book I, ch. 23) (MGH SS rer. Merov. I.1). Hahn.
Holl, K. (Ed.). (1933). Epiphanius, Band III: Panarion haer. 65–80; De fide (Die griechischen christlichen Schriftsteller 37; De Fide 22.7). Hinrichs. [Location; Greek text not checked.]
Jaubert, A. (1957). La date de la Cène: Calendrier biblique et liturgie chrétienne. J. Gabalda. [English ed.: The date of the Last Supper, Alba House, 1965.]
Justin Martyr. (1885). The first apology (ch. 67; Ante-Nicene Fathers, Vol. I). Christian Literature Company.
Merx, A. (1902). Die vier kanonischen Evangelien nach ihrem ältesten bekannten Texte (Vol. II, Part 1: Erläuterung — Matthäus; the Didascalia/calendar excursus pp. 372–381, the cited footnote at p. 377). Georg Reimer.
Primary texts: Exodus 12:3 · Mark 15:42 · John 19:31 · Josephus, Antiquitates Judaicae 16.163 (read in translation; see ch. 13) · Clement of Alexandria, Stromata I.21 · Irenaeus, Adversus Haereses II.22.3, 22.6 · Tertullian, Adversus Iudaeos 8 · Julius Africanus (via Syncellus) · Eusebius, Chronicon · Epiphanius, De Fide 22.7 (GCS 37, Holl); Panarion 51 · Didache 8:1 · Orosius, Historiae adversus paganos VII.4. Greek/Latin/Syriac via the editions named; author's own working translation.
The day has spoken, and it pointed midweek. But one objection lingers, and it is not the least: the church has held to the Friday from her earliest days, and the year she names is not AD 31. A reading with so many centuries against it bears the burden of proof, and that burden is heavy. So before the year comes up, this question must be faced, and faced as honestly as the subject deserves. It falls into two parts, for the tradition hands down two things: a year, and a day. And of both the question is: is this a memory, or a computation? Did anyone carry this knowledge independently, or was it inherited and passed on?
I state the outcome in advance, for it is the heart of this chapter. The tradition does not remember the year. It calculates it. And it calculates, down to the last witness, with the same two ingredients: the fifteenth year of the emperor Tiberius from Luke 3:1, and a guessed length of the ministry. That is no malice and no conspiracy; it is simply how a church that had preserved no year-count of the crucifixion reconstructed one anyway. But it means that the sentence "but the tradition says AD 33" — or AD 29, or any year whatever — carries no weight that could overrule the astronomy or the text. The tradition has measured nothing here to overrule them with.
See how clean the pattern is. Walk through the ancient writers who name a crucifixion year, set beside each the length of ministry he assumes, and the whole fan of dates falls into place. Whoever assumes a ministry of zero years — Jesus baptised and dying shortly after — arrives at AD 29 (Tertullian, Lactantius, Augustine). Whoever assumes one year, at AD 30 or 31 (Clement, Africanus). Whoever assumes three years, at AD 32 (Hippolytus, Epiphanius). Whoever assumes three and a half to four years, at AD 33 (Eusebius). There is no residual scatter, not a single source that gives a year that does not follow neatly from this sum. And Clement of Alexandria lets you watch while he makes it: "fifteen years of Tiberius and fifteen years of Augustus, so the thirty years were completed until He suffered" (Clement, Stromata I.21). That is not testimony. That is subtraction and addition, out loud.
Dossierthe two family trees
The family tree of the year. The spread of the patristic year-dates correlates completely with the assumed length of ministry, and with nothing else:
The year is thus everywhere a derivative of Luke 3:1 plus a guess — never a preserved date. What looks archival is not: "the two Gemini" (AD 29) is merely Tiberius's fifteenth year in consular form, and the Acta Pilati confess the confusion in their own text.
The family tree of the day. Two lines run through the centuries. The Friday line is old and broad: Justin (155), Hippolytus (±230), Victorinus (±300), the Didascalia (3rd century), Hilarianus (397) — unanimously paraskeuê. The midweek line is narrower but tough: the Wednesday fast — alongside the Friday fast — which the Didache already prescribes around the year 100 (Didache 8:1), a century before the Didascalia, though it does not yet name the reason; the Didascalia's Tuesday-evening supper and Wednesday arrest (±225); the "many" who according to Gregory of Tours (±590) held a Saturday resurrection; and in the modern age Chwolson (1908) and Jaubert (1957). Both lines exist; the question is which carries a memory and which a chain of reasoning. [DOCUMENTED]
That is the first half, and it exonerates no one and burdens no one — except whoever thought the tradition settled the matter. For see what happens to the three candidate years. AD 33 is the choice of Eusebius, but his family tree is one man deep and hangs on the longest ministry. AD 31 surfaces in Africanus, but his year is internally contradictory — it follows not from his own Tiberius year but from an Olympiad that clashes with it — and it rests on a one-year ministry that Irenaeus, older and closer to the source, expressly rejects. A second patristic line does run to 31 — Orosius (around 417) and Gregory of Tours (around 590) date the passion to the seventeenth year of Tiberius, which falls in the spring of 31 — but it too is a derivation from Luke 3:1 with an assumed length, not an independent memory. None of the years, then, has an independent ancient witness. The patristic field is empty — for everyone. And an empty field does not refute the case; it gives the case back to the place where it belongs — the astronomy and the text.
Then the second half, the day, and here I must be honest in a way that seems to help the other side. Every ancient writer who names a weekday names the Friday, and early. The oldest weekday witness is Justin Martyr, around the year 155: "on the day before that of Kronos they crucified Him" — the day before Saturday, hence Friday (Justin, 1 Apology 67). Half a century before Tertullian, and crystal clear. Not one surviving source proclaims, in its own voice, a Wednesday crucifixion; in the ancient church the Wednesday is the day of the betrayal, not of the cross, and that is why one fasted on Wednesday. That must be said, by me, before a critic says it: on the weekday, the early church, so far as her texts survive, does not stand on this book's side.
And yet "unanimous" is the wrong word, and it is precisely the sources that show why. For the other reading was fought — and no one fights what no one holds. Gregory of Tours must still, in the sixth century, refute the Saturday resurrection that "many" held; the Apostolic Constitutions cut the Tuesday evening out of the Didascalia; Epiphanius ranks among his heretics the "Alogi," who read the Johannine Passover chronology differently. That is not silence — that is defence, and defence is the trace of a living opponent. Here applies the rule that underlies this whole chapter: a suppressed reading leaves no texts of its own; it leaves behind the refutations of its opponents. What has come down to us is therefore not "the unanimous church" but the voice of the party that won — and the refutations by which it won are at the same time the evidence that there was something to refute. The Friday has been handed down; the debate lies preserved beneath it. We know of that debate only because the Friday had to defend itself against it — and one does not defend oneself against a phantom.
But it is the same diagnosis as with the year. They remember nothing; they read — and they read through a language that had meanwhile shifted. In the first century the Greek paraskeuê meant "preparation" — the day before a day of rest, whether that was the weekly sabbath or an annual festival sabbath. But in Christian Greek the word became the fixed synonym for Friday; to this day paraskevi is the ordinary Greek word for that day. The fathers thus read back into the gospels what their own language had meanwhile made of the word. And the opening they passed over stands plainly in the text: John writes that the day of that sabbath "was a great day" (John 19:31) — the festival sabbath, not necessarily the weekly one. The fathers are numerous, but they are not independent witnesses; they are later readers of a single shifted word.
Experthow paraskeuê became "Friday"
In the New Testament paraskeuê (παρασκευή) is still simply "preparation, making ready": the day on which one prepares for the day of rest that follows (Mark 15:42, where it is glossed as prosabbaton, "the day before the sabbath"). So too — relatively, not as a name — the oldest non-Christian witness uses the word: the Augustan decree exempting the Jews from legal proceedings "on the sabbath, or on the preparation day before it, from the ninth hour" (Josephus, Ant. 16.163) — older than the gospels, with a time of day, and yet no weekday name: the "before it" needs the day of rest beside it in order to point to anything (the full weighing of this strongest counter-witness stands in ch. 13). In first-century Judaism that day of rest could be the weekly sabbath or an annual festival sabbath. In later Christian Greek the word narrowed into the fixed name of the sixth weekday — a semantic shift that runs on into modern Greek (paraskevi = Friday). Whoever reads the gospels with the later meaning hears "Friday" where the text says "preparation day"; and a preparation day also preceded the great festival sabbath of 15 Nisan (ch. 4). The fathers read the later language back into the oldest text — no bad faith, but no independent memory either.
And yet the midweek line cannot simply be waved away, for it has left a trace too concrete to ignore: the Didascalia. This is no heretical writing but a catholic, liturgical church order from the third century, and it tells the passion week longer than the ordinary scheme allows. "When we had eaten the Passover on the Tuesday evening, we went out to the Mount of Olives, and in the night they seized our Lord; and the next day, which was Wednesday, He remained in custody" (Didascalia XXI; Connolly, 1929). A Tuesday-evening supper, an arrest in the night from Tuesday to Wednesday — a liturgical memory of a week that stretched over more than the eighteen hours of the Friday scheme. And it preserves a detail no third-century redactor invents: that Judas received his money on the tenth of the month, the Monday — the tenth of Nisan, the day on which the lamb is set apart (Exodus 12:3). "How ancient must that be," wrote Merx beside that line, and he was right: it is a liturgical fossil.
Excursusthe seam that betrayed the redaction
The finest trace that something was smoothed away here is not the Didascalia itself but what a later hand did with it. The Apostolic Constitutions (±380) are a reworking of the Didascalia, and they have cut the Tuesday evening out: the Passover meal now stands on Thursday. But they left the scaffolding standing. In their text the sanhedrin still meets on Monday, Tuesday and Wednesday — three council days that are utterly pointless in a Friday chronology in which nothing happens until Thursday evening. Their Passover fast still begins on Monday — six days, though the suffering would begin only on Thursday evening. And most revealing of all: immediately after their Thursday–Friday chronology they cite Matthew 12:40, "three days and three nights in the heart of the earth," as fulfilled — though Friday evening to Sunday morning counts two nights, not three (Constitutiones Apostolorum V.14–18). They kept the text and broke the sum. The beam has been taken out; the scaffolding still stands. And Epiphanius says in so many words that one fasts on Wednesday "because the Lord was arrested on Wednesday morning" — the arrest, not the decision (Epiphanius, De Fide 22.7; GCS 37, Holl). That is not the proof of a Wednesday crucifixion; it is the trace of a longer week that was later cut short.
But — and this is where this chapter draws its own limit, out loud, so that no one else has to — the Didascalia delivers no Wednesday crucifixion. Three sentences after the Tuesday meal it says that they crucified Him "on that same Friday." It lengthens the custody, not the time in the tomb. Whoever deploys it as a witness for a Wednesday crucifixion abuses the source — and this book does not, even though it would be convenient. The Didascalia is a breach, not proof: it shows that the early church itself wrestled with the length of the week and with the three nights, but it does not share this book's conclusion.
The same honest reservation holds for the two witnesses who come closest to this book's reading. Gregory of Tours reports, around 590, that "many" believed the resurrection took place on the seventh day, the Saturday — a view that presupposes a midweek crucifixion, and that was evidently tough enough to force a bishop to refute it (Gregory, Historiae I.23). But he names no group, gives no reason, writes five hundred years after the fact, and there is an innocent explanation that cannot be excluded. And the modern forerunners each go only part of the way: Daniel Chwolson argued a death on the fourteenth of Nisan (ch. 5) — but kept the Friday; Annie Jaubert reconstructed from the solar calendar of Qumran a Tuesday-evening supper that reconciles the synoptic and Johannine chronologies (Jaubert, 1957) — but she too kept the Friday. It is honest to say so: the midweek line has a real family tree, down into the best modern scholarship, but even its champions stopped at the Friday. This book goes one step beyond them, to the Wednesday, and it carries that step itself — it does not lean on them as though they had already taken it.
Countercasethe tradition at its strongest
Let the objection keep its full weight. The Friday is no mere late habit: it is attested from Justin in 155, within a century and a half of the event; it stands in the creeds; it carries the liturgy of the whole church, east and west, to this day. Not a single ancient source teaches a Wednesday crucifixion. A reading with such credentials against it, it will be said, cannot be right; the burden of proof is crushing. This is the strongest form of the argument, and I do not soften it.
The answer is not that the fathers lie, but that they are not what they are taken to be. On the year they demonstrably calculate — the spread betrays the sum. On the day they read through a word whose meaning had meanwhile shifted. They are numerous, but they are not independent: it is one stream that branched, not a multitude of witnesses confirming one another. And precisely beneath their apparent unanimity lies — in the Didascalia, in the seam of the Constitutions, and in the polemic the Friday had to wage against the midweek reading — the preserved trace of a longer week; for a reading one defends oneself against was a reading that lived. The tradition is not a wall that closes the case; it is a palimpsest, and beneath the top layer an older one shows through.
Where does the investigation stand, then, at the end of the day-axis? The earliest voices do not decide the matter — in either direction. They do not remember the year but calculate it, and they read the day through a shifted language; and where they seem most independent, in the liturgical week of the Didascalia, they preserve precisely the trace of a longer week that they later cut short. That is the honest balance: on the text, the day points midweek; the tradition sets no independent weight against it; and the midweek reading is old enough to be no modern notion. But a day is not yet a date. Wednesday — which Wednesday, in which year? For that the investigation must leave the text and go up into the sky. To the moon over Jerusalem, and to the years it allows.
HeartHEART
There is something moving about a fossil in an order of prayer — a Monday left standing because the lamb is set apart on the tenth, a fast that begins too early because it is older than the redating it was made to carry. The church has, without knowing it, preserved in her own liturgy the memory her scholars later smoothed away. That is no indictment of the fathers; it is a sign of how faithful habit is, even where the explanation shifts. The days are counted here to find the truth. But beneath the counting lies a prayer older than any reconstruction, and it fasts anew each week, on the day of the betrayal — as though the body of the church remembers something her head has forgotten.
Note on the sources
The patristic witnesses have been consulted in the source itself or through a named scholarly translation, with location; Justin (1 Apol. 67), the Didascalia (Connolly, 1929, alongside the Lagarde text and the searchable translation of Gibson, 1903, p. 94, which gives the same Tuesday evening and the same Wednesday in custody: "when we had eaten the Passover on Tuesday in the evening, we went out to the Mount of Olives, and in the night they took our Lord Jesus; and on the next day, which was Wednesday, He remained in prison"), the Constitutions (ANF) and Gregory (Historiae I.23, Latin) have been checked in the primary text. Epiphanius' Wednesday fast is located in De Fide 22.7 (GCS 37, Holl); the Greek text itself has not been checked, and it carries here as corroboration, not as primary carrier. The correlation between the patristic year-dates and the assumed length of ministry can be recomputed from the witness table. This chapter is a corroborating ring: it shows that the tradition does not measure the year but calculates it, and that the midweek reading has an old family tree — but it delivers no Wednesday witness, and it is not advanced as one. The Didascalia is a breach, not proof; Chwolson and Jaubert themselves kept the Friday. This chapter claims a proof for neither side.
Literature (APA)
Chwolson, D. A. (1908). Das letzte Passamahl Christi und der Tag seines Todes (2nd ed.). Leipzig.
Connolly, R. H. (Trans.). (1929). Didascalia Apostolorum: The Syriac version translated and accompanied by the Verona Latin fragments (ch. 21). Clarendon Press.
Donaldson, J., & Roberts, A. (Eds.). (1886). The Ante-Nicene Fathers (Vol. VII: Constitutiones Apostolorum, V.14–18; VII.23). Christian Literature Company.
Gibson, M. D. (Trans.). (1903). The Didascalia Apostolorum in English (Horae Semiticae II, p. 94). Cambridge University Press.
Gregory of Tours. (1951). Libri historiarum X (B. Krusch & W. Levison, Eds.; Book I, ch. 23) (MGH SS rer. Merov. I.1). Hahn.
Holl, K. (Ed.). (1933). Epiphanius, Band III: Panarion haer. 65–80; De fide (Die griechischen christlichen Schriftsteller 37; De Fide 22.7). Hinrichs. [Location; Greek text not checked.]
Jaubert, A. (1957). La date de la Cène: Calendrier biblique et liturgie chrétienne. J. Gabalda. [English ed.: The date of the Last Supper, Alba House, 1965.]
Justin Martyr. (1885). The first apology (ch. 67; Ante-Nicene Fathers, Vol. I). Christian Literature Company.
Merx, A. (1902). Die vier kanonischen Evangelien nach ihrem ältesten bekannten Texte (Vol. II, Part 1: Erläuterung — Matthäus; the Didascalia/calendar excursus pp. 372–381, the cited footnote at p. 377). Georg Reimer.
Primary texts: Exodus 12:3 · Mark 15:42 · John 19:31 · Josephus, Antiquitates Judaicae 16.163 (read in translation; see ch. 13) · Clement of Alexandria, Stromata I.21 · Irenaeus, Adversus Haereses II.22.3, 22.6 · Tertullian, Adversus Iudaeos 8 · Julius Africanus (via Syncellus) · Eusebius, Chronicon · Epiphanius, De Fide 22.7 (GCS 37, Holl); Panarion 51 · Didache 8:1 · Orosius, Historiae adversus paganos VII.4. Greek/Latin/Syriac via the editions named; author's own working translation.