Baker Academic

Friday, January 8, 2016

Was the Last Supper on Wednesday? (A Review of Colin Humphreys' The Mystery of the Last Supper)

Since the publication of Jesus and the Last Supper, I have gotten more than one request from readers for my thoughts on Colin J. Humphreys' hypothesis that Jesus celebrated the Last Supper on Wednesday night according to a special pre-exilic lunar calendar. Although my chapter on the date of the Last Supper is 120 pages, I still didn't have space for an extended treatment of his theory. Below is a brief review of Humphreys' theory. I thought it might be interesting to readers of the Jesus blog intrigued by the debate over the chronology of Jesus' last days in Jerusalem. Enjoy.

Colin J. Humphreys. The Mystery of the Last Supper: Reconstructing the Final Days of Jesus. (Cambridge: Cambridge University Press, 2011). A Review by Brant Pitre.

This contribution to the age-old debate over the date of Jesus’ Last Supper is truly unique. For one thing, Humphreys is not a biblical scholar but a professor of Materials Science and Metallurgy at the University of Cambridge, who has published previously in calendrical and biblical studies. Moreover, unlike many other treatments of the date of the Last Supper, the book draws heavily on up-to-date astronomical calculations of Graeme Waddington, an Oxford astrophysicist, with whom Humphreys has co-authored several articles in refereed journals. Finally, Humphreys proposes a truly novel solution to the apparent chronological contradiction between the Synoptic and Johannine chronologies of the death of Jesus. He argues that Jesus neither celebrated the Last Supper on Thursday evening (the traditional date) nor on Tuesday evening (as proposed by Annie Jaubert and others). Instead, Humphreys contends that Jesus celebrated the Last Supper on Wednesday evening of his last week. According to Humphreys, Jesus was following a special pre-exilic lunar calendar inherited by the Israelites from ancient Egypt and later used by Samaritans (and possibly Galileans) in the Second Temple period.

Summary of Contents
The book can be roughly divided into four parts: First, Humphreys begins with an overview of the date of the Last Supper debate and highlights the weaknesses of previous solutions (chs. 1-3). Second, he then draws on the astronomical calculations of Waddington in order to propose an exact calendar date of the crucifixion (chs. 4-5). After a painstaking analysis of multiple options, Humphreys concludes that Friday, April 3, AD 33 is “the only possible date” (p. 72). In support of this conclusion, he contends that that Peter’s speech at Pentecost about “wonders in the heaven” and “the moon” being “turned to blood” (Acts 2:19-21) is a historical reference to an actual lunar eclipse that was visible in Jerusalem on the same day as the crucifixion (ch. 6). Third, Humphreys deals with the date of the Last Supper proper by devoting an entire chapter to outlining and rejecting the theory that Jesus followed the Qumran solar calendar (ch. 7). He then goes on to suggest that while the official Jewish calendar at the time of Jesus—a sunset-to-sunset lunar calendar—originated in the Babylonian exile, there existed a pre-exilic sunrise-to-sunrise lunar calendar, akin to the religious calendar of ancient Egypt and the lunar calendar used by the Samaritans (ch. 9-10). The existence of these two calendars side-by-side led to confusion within the Hebrew Bible itself over whether, for example, the feast of Unleavened Bread started on 14 Nisan (Exod 12:17-19; Ezek 45:21) or 15 Nisan (Lev 23:6; Num 28:17). Humphreys suggests that not only the Samaritans, but perhaps also Galilean Jews, may have continued to use this pre-exilic lunar calendar, leading to at least a one-day divergence in the dating of the Passover (pp. 147-48). He then concludes by arguing that (in part) because Jesus saw himself as a new Moses, he deliberately followed the pre-exilic calendar. Moreover, the Gospels, when read carefully, suggest that more than one day transpired between the Last Supper (Wednesday night) and the crucifixion (Friday morning) (chs. 11-13).

Strengths: Clarity, Depth, Critique of Essene Hypothesis
The strengths of Humphreys’ study are several. The book is very clearly written, and makes the complex subject of ancient calendars and the date of the Last Supper remarkably accessible. Moreover, the author demonstrates a firm grasp of the secondary literature in English on the date of the Last Supper. Finally, Humphreys offers an intriguing critique of Jaubert’s theory that Jesus followed the Qumran solar calendar. He points out that in the book of Enoch (which reflects the solar calendar), the first day of the first month (1 Nisan) always falls after the spring equinox, with the significant result that the solar Passover would consistently take place after the lunar Passover (see 1 Enoch 72). Hence, according to Humphreys: “Jesus could not have used the Qumran calendar to celebrate his last supper as a Passover meal since, whether it was intercalated or not, Passover in the Qumran calendar did not fall in the same week as Passover in the official Jewish calendar” (p. 109). If correct, this is a significant point, since even recent defenses of Jaubert continue to assume without demonstration that the solar and lunar Passovers took place during the same week. See, e.g., Stéphane Saulnier, Calendrical Variations in Second Temple Judaism: New Perspectives on the 'Date of the Last Supper' Debate (JSJSup 159; Leiden: Brill, 2012). In other words, there is no way for Jesus to have celebrated the solar Passover of the Essene if the Essene Passover did not take place the same week that he was crucified.

Weaknesses: Observation, Intercalation, Agricultural Conditions, and Gospel Exegesis
On the other hand, there are several serious weaknesses with Humphreys’ overall hypothesis.
First and foremost, Humphreys’ study operates on the dubious assumption that we can use contemporary astronomical calculation to determine the calendar dates of ancient Jewish liturgical feasts which were based on human observation of the new moon. Cf. R. T. Beckwith, Calendar and Chronology (Leiden: Brill, 2001), 276-96. For example, according to the Mishnah, “while the Temple still stood” the determination of the exact date of the Passover feast was completely based on whether or not the new moon is “manifestly visible or not” (Mishnah, Rosh Hashanah 1:1-9). As John Meir pointed out some time ago, this means that the declaration of the new moon (and hence the date of Passover) “depended not on whether the new light actually existed, but on whether human beings had seen it” (Meier, A Marginal Jew, 1.401-402). As Humphreys himself admits, when it comes to an observation-based lunar calendar, there is always at least a one-day margin of error involved in the ancient method of detecting a new moon, since “poor atmospheric transparency” can never be ruled out (p. 50). It should go without saying that using calculations with a 24-hour margin of error cannot be conclusive for solving the problem of an apparent 24-hour chronological discrepancy. 
     Second, Humphreys study also largely ignores the enormous problem of ancient Jewish intercalation (e.g., leap years and leap months). Humphreys admits again that we have no clear idea how first-century Jews, whether in the solar or lunar calendars, dealt with intercalation and leap years (p. 56).  To be sure, the the Tosefta records a lengthy rabbinic debate about  a number of competing and contradictory ideas about when or when not to intercalate (see Tosefta, Sanhedrin 2:9-12). But even if we assume that the multiplicity of views reflected in the Tosefta were in play in the first-century A.D. (a questionable assumption), we still have no idea which method of intercalation was being employed by the first century Temple authorities, to say nothing of whether and how it affected the liturgical calendar in the particular year that Jesus died. Apart from knowledge of what method of intercalation was being used by Temple authorities in the exact year Jesus died, the correspondence between contemporary calculations and ancient liturgical practice must always remain speculation and thus cannot be used to solve the question of the date of the Last Supper. This is particularly important for Humphreys’ novel suggestion of a Wednesday Last Supper, since it is based almost entirely on an argument from astronomical calculations that the “only possible date” 14 Nisan in the pre-exilic sunrise-to-sunrise lunar calendar could have occurred was Wednesday, April 1, AD 33 (p. 163-64).
    Third, Humphreys’ hypothesis fails to reckon with the fact that, according to most ancient evidence we possess, the Jewish system of intercalation for the Passover feast was also contingent on weather and agricultural conditions that were variable each year. For example, the Tosefta attributes a saying to Rabbi Gamaliel to the effect that if that year was a long winter and the “lambs were too thin” and “the first-ripe grain has not yet appeared,” then “thirty days” were to be “add[ed]” to the year in the Spring, to give the lambs and the grain time to grow (Tosefta, Sanhedrin 2:5-6). This means that if there was a long winter the year Jesus died, then Jerusalem authorities could have added an entire month to the calendar in order to give the Passover yearling lambs and the first shoots of barley time to ripen. This would obviously cause any modern-day calculations based solely on astronomy to fall to pieces, since we have no way of knowing what the weather conditions were like in the year Jesus crucified (especially when we remain unsure about exactly what year that was). 
     Fourth, and most important of all, although the astronomical sections in the book are impressive and meticulous, Humphreys’ discussion of the actual exegesis of the Gospel accounts of the Last Supper and passion of Jesus have serious weaknesses. It goes without saying that the Gospel accounts themselves say nothing about Jesus following a special pre-exilic lunar calendar, despite Humphreys' valiant attempt to transform Mark's description of "the first day of Unleavened Bread (azymōn)" as the day when "the Passover lamb (pascha) was sacrificed" (Mark 14:12) into evidence that Jesus and his disciples were following a different calendar from the Jewish Temple. However, this supposed "anomaly" in Mark 14:12 is easily explained from by the linguistic fact that, by the first-century A.D., the distinction between "Passover" (pascha) (14 Nisan) and "Unleavened Bread" (azymōn) (15-21 Nisan) often disappeared in common parlance. Indeed, both Luke and Josephus make clear that the two terms could be used interchangeably: 

"The feast of Unleavened Bread, which is called Passover" (Luke 22:1)

"The feast of Unleavened Bread, which we call Passover" (Josephus, Antiquities 14:21; 18:29)

 "We keep for eight days a feast called the feast of Unleavened Bread" (Josephus, Antiquities 2:317)

In other words, either term could be used to refer to the entire Passover octave (See Pitre, Jesus and the Last Supper, 338-39). Hence, contrary to what some scholars contend, it is not Mark here who is ignorant of things Jewish; it is we who are ignorant of and sloppy in our assertions about first-century Jewish Passover terminology. For what it's worth, a similar fluidity exists today in Catholic circles with the word "Easter," which can refer to (1) the Easter vigil on Saturday evening, (2) Easter Sunday, (3) the Easter Octave, or even (4) the seven week Easter season. It is no coincidence that the Latin word for "Easter" is pascha!
      Moreover, although a Wednesday Last Supper would indeed provide more time for the number of events recounted in the Gospels, it highly questionable that a multi-day chronology of Jesus’ passion is “the natural reading of all four Gospels” (p. 185). The overwhelming majority of Gospel interpreters over the centuries, from the earliest patristic writers to today, have discovered only one night transpiring between Jesus’ Last Supper on Thursday night and his crucifixion on Friday morning (see Justin Martyr, Dialogue with Trypho, 111.3; Irenaeus, Against Heresies 2.22.3). Indeed, the accounts of Peter’s denials and the morning “cock-crow” in the Synoptics and John clearly suggest the passing of only one night before Jesus execution (see esp. Mark 14:30, 66, 72; 15:1; John 13:38; 18:26-28). This is perhaps why Humphreys rather awkwardly and unconvincingly tries to explain these Gospel passages away (pp. 178-80). Also unconvincing is Humphreys' idiosyncratic suggestion that the Gospels narrate two "successive" daybreak "trials"of Jesus--a "first" trial on Thursday morning (supposedly recorded in Luke 22:66-71) and a "second" on Friday morning (cf. Mark 15:1; Matt 27:1)(see pp. 225-36 n. 28). Although correlating the trial narratives has well-known difficulties, the much more natural reading is that Luke is expanding Mark's brief reference to a single daybreak gathering of the Sanhedrin on Friday morning in which Jesus is bound and led away (Mark 15:1). 

In short, despite a truly impressive knowledge of astronomy and a truly unique contribution to the debate, Humphreys’ attempt offer a solution to the date of the Last Supper fails to reckon with the fact that ancient Jewish liturgical calendar appears to have been contingent upon observation, seasonal shifts, and an unknown method of intercalation. Most important of all, the overarching hypothesis fails at the level that any viable theory must succeed: the actual exegesis of the Gospel texts.  Not least because a Wednesday Last Supper cannot be found in or reconciled with the Gospel accounts themselves, Humphrey’s astronomical solution, however ingenious, is ultimately unconvincing. The point bears repeating: "For better or for worse, the question of when the Last Supper took place and whether or not it coincided with the Jewish Passover meal must be settled on the basis of the textual evidence in the Gospels" (Pitre, Jesus and the Last Supper, 313). 


  1. Excellent treatment, Brant. Clear and very helpful.

    I posted some of my own thoughts a couple of years ago on my blog here, and there is some similarity, especially on the question of the last supper / cock crow / this night material: .

  2. Thank you Mark! I appreciate the kind words and the link. I missed that post, and look forward to reading it.

  3. A larger methodological point, relevant to memory theory and historical Jesus research, might be made here. If finally in the Bible, the facts of astronomy are at times modified by human frailties, and the subjectivity of human perception.

    No matter when the moon actually was first there, a major element determining what was finally done, was the moment when fallible human beings finally perceived it.

    So a human subjectivity is evident.

  4. Thank you for your review of The Mystery of the Last Supper by Colin Humphreys.

    After reading your review, I started studying the writings of Josephus more carefully and discovered that Josephus seems to flatly contradict himself on the subject of the length of the Feast of Unleavened Bread and apparently also on the date of the First Day of the Feast of Unleavened Bread.

    In Antiquities 2.317, Josephus writes:
    Thus it is that, in memory of the want we were then in, we keep a feast for eight days, which is called the Feast of Unleavened Bread.

    However, in Antiquities 3.249, he writes:
    The Feast of Unleavened Bread succeeds that of the Passover, and falls on the fifteenth day of the month, and continues seven days, wherein they feed on unleavened bread; … But on the second day of unleavened bread, which is the sixteenth day of the month, they first eat of the fruits of the earth …

    The first passage can be used to support your viewpoint and the second one seems to lend support to Dr. Humphreys’ viewpoint.

    I have not found a place where Josephus gives an explanation for this contradiction. As a result, it seems to me that the writings of Josephus too ambiguous and indefinite to settle the questions that need to be answered. It looks to me like they are another dead end.

    If this is correct, then we cannot use Josephus’s writings to determine whether the wording in Mark 14:12 is “anomalous” or not. We will have to look for other clues in other places if we want to find a way to answer the questions that we have about the Last Supper.

    Luke 22:7 is not decisive either, because it is fully consistent with your theory and Dr. Humphreys’ theory too.

  5. Regardless of the differences in possible calculations, the fact is that the OT points to Nisan (Babylonian name) 14, which makes sense according to the cycles of 7 and apparent Sabbath identity of God. Christ likewise, most certainly celebrated Passover when it was correct to practice it, according to the time it was given, and not by a contrived amendment or evolutionary change brought about by Jewish historical changes in the day. Nisan 14 at the time God gave it, was still Nisan 14 at the time of Christ. This is the critical point in understanding the behavior of Christ at that time. Likewise, it would seem to me, that the Jewish changes to Passover, etc. must be considered in light of the Jewish calendar, and not the Roman calendar. Adding Roman calendar references to a sacred event would seem to be misleading in terms of the way the sacred record actually views the events in terms of time. Is that not so?

  6. For those who are wondering whether or not to read Humphreys’ book, I would personally recommend it. The issues mentioned in the review above do not, in my view, represent genuine weaknesses in Humphreys’ work. Humphreys repeatedly stresses the vagaries of our knowledge of the Jewish calendar. Hence, he continually works with ranges of dates rather than absolute dates; that is to say, he quotes, for instance, a date when a new moon could have been cited, and a second later date when it might have been cited given poor weather conditions, and so on. While, therefore, claims like, “It should go without saying that using calculations with a 24-hour margin of error cannot be conclusive for solving the problem of an apparent 24-hour chronological discrepancy” have a certain rhetorical flourish, they are not, in my view, substantive. If one can establish the range of possibilities for when a month might have begun, then one can rule out certain options, and hence decide between two options separated by only 24 hours. Suppose, by way of analogy, we have two different estimates for today’s temperature (23 and 24 degrees), and suppose it’s actually 25 degrees, and suppose we have a thermometer which measures the temperature with a 1 degree margin of error. We can clearly then decide between our estimates (23 and 24), since we can establish ranges of possibility. And that is the procedure that Humphreys employs here, to good effect. (Note: I think Humphreys may have overlooked one important point, namely that intercalated months are, acc. to Beckwith, of necessity 30 days in length, which would require Humphreys to increase his ranges of possibility and to give a 31 AD crucifixion further consideration.)

    As for Humphreys’ exegesis of the Gospels’ accounts of Jesus’ trial, Humphreys gives a detailed argument for his position, which strikes me as quite persuasive. The situation as far as the arrest and trial narratives are concerned is not straightforward, as is implied by the reviewer’s statement, “The Synoptics and John clearly suggest the passing of only one night before Jesus execution”. (Note here the notion of a ‘clear suggestion’.) The “much more natural reading” of the Gospels proposed by the reviewer is quite possible, but then so is Humphreys’. For my part, I would prefer simply to say that both readings are possible, rather than strongly come down on one side or the other.

    Ultimately, for what it’s worth, I believe Humphreys’ view to be a very plausible--persuasive even--and undoubtedly fascinating and thought-provoking one. I also believe the reviewer’s view to be plausible. (I have read his book on the matter, which I would also recommend. It contains a wealth of information, all of which is clearly documented and discussed, and goes beyond a mere rehash of old arguments due to its thorough treatment of Second Temple literature, although, in my view, it overlooks the odd reference here and there, such as a Mishnaic reference that implies a time for the Passover sacrifice prior to 3 p.m.)

    James Bejon.