This is a bold and provocative book. The author seeks to challenge the generally accepted account of the origin of Matthew, and indeed of Mark and Luke as well. Q is dispensed with, as is the hypothesis of Aramaic originals behind Q and M; for good measure, many form-critical assumptions are quietly ignored. The author argues that Matthew is a provincial scribe who has adapted and expanded Mark by means of midrash, a Jewish method of exegesis and interpretation. Matthew depends on oral tradition only to a very limited extent; Mark is his only written source. In short, about half of Matthew’s Gospel has been written by the evangelist himself. Why did Matthew want to expand and expound Mark in this way: ‘The Gospel was written to be read round the Jewish-Christian Year, and was formed by repeated preaching round the Jewish-Christian Year, by Matthew’ (p. 193).

Source, form and redaction criticism have all been under renewed scrutiny recently; a most welcome trend. Hence one turns to a fresh approach to the origins of Matthew with high expectations. This is not however a book which will change the course of current study of the Gospels. It is extremely well-written and its main arguments clearly and vigorously stated. The author’s breadth of learning and his large store of fresh ideas are revealed on nearly every page. But in spite of the book’s length and apparent thoroughness many awkward issues are neatly side-stepped and more seriously, there are several major weaknesses in the author’s position which not only damage his case but ultimately turn out to be fatal.

The opening chapters discuss the distinctive vocabulary, style and method of Matthew. This section of the book includes an interesting examination of the parables in Matthew. Goulder argues that Matthew’s typical method of expounding Mark is by means of parables. Parables in the OT, in each of the three Synoptic Gospels and in the rabbinic traditions, are compared and contrasted. Matthew, the Christian scribe, has used the very methods developed by rabbis to expound sayings and parables of their own teachers. Matthew’s ‘non-Marcan’ parables are so ‘Matthaean’ that it is impossible but that their author is Matthew. Goulder makes a number of fascinating observations and does show how very different the parables in Matthew and the parables in Luke are. But there are too many exceptions to the ‘pattern’ discerned by the author to support his confident wide-ranging conclusions. There are only five OT parables; the rabbinic parables are widely scattered and need to be studied much more rigorously before generalisations can be drawn from them. Whereas a high proportion of the teaching of Jesus is in parables, and in Goulder’s view an even higher proportion of Matthew’s, very rarely do we find more than one or two parables associated with any one rabbi. Some of the rabbinic parables are more ‘Lucan’ than ‘Matthaean’, and if we accept Goulder’s presuppositions, Luke has redacted some of the Matthaean parables in a more ‘rabbinic’ direction!

Questions also need to be asked about the author’s use of the term midrash to describe Matthew’s method of rewriting Mark. If we use the term sufficiently loosely, it becomes possible to argue that almost the whole of the NT is midrash, either on OT or earlier Christian traditions! Goulder wisely distinguishes between different genres of midrash. Matthew’s method is said to be closer to re-writings like Chronicles or Jubilees than paraphrastic expansions like the Targums or the LXX, or juristic commentaries, like the Mekilta and Sifre, or homiletic midrash.

Is re-writing of earlier traditions or works a distinctively Jewish phenomenon? Goulder illustrates a number of midrashic methods used by Matthew: e.g. transcription of material with very little alteration, omission, abbreviation, removal of inconsistencies, explanatory changes etc. But many of these methods can be paralleled in Livy’s use of Polybius, or even in Mozart’s adaptation of Handel’s Messiah! Goulder would reply, quite correctly, that the midrashic activity he has in mind is closely bound up with a firm belief in the inspiration and authority of Scripture. So on Goulder’s view, within a decade Mark has become sufficiently ‘authoritative’ to be re-written with very considerable creativity. But we do not have any other examples of similar midrashic activity within such a short space of lime. Matthew has carefully retained and abbreviated nearly all of Mark, and yet, according to Goulder, he has at the same time expanded creatively some parts of his only source to form the ‘non-Marcan’ traditions in Matthew. As Goulder recognises, the kind of abbreviation of Mark which we find in Matthew is not characteristic of midrash. He is forced to concede that while Mark is authoritative for Matthew, it is not yet scriptural (p. 35). Midrashic re-writing seeks to expound for a new age not merely an authoritative text, but an inspired biblical text. If Mark is not yet ‘Scripture’ for Matthew, how do we account for his creative expansion?

And so we come back to the key question: why did Matthew want to expand Mark with material which stems very largely from his own hand? Goulder’s answer is I think, the lynch-pin of the whole book. Matthew’s Gospel was developed liturgically and was intended to be used liturgically; its order is liturgically significant, for the evangelist has taken the Jewish Festal Year and its pattern of lections as his base. A Gospel is a lectionary book, a series of ‘Gospels’ used in worship week by week (p. 172). Mark is a lectionary book for half a year—after all, Mark only promised to give us ‘the beginning of the Gospel’! (p. 201). Matthew has expanded Mark into a lectionary book for a full year. Not to be outdone, Luke has also written a lectionary for a full year, but has preferred the annual sabbath cycle of lections to the festal cycle followed by Matthew. At last we can understand why Luke has re-written and in particular, re-ordered Matthew. With a wave of the lectionary wand, Q is consigned to oblivion.

Two questions must be pressed vigorously. Were the Scriptures read in first century synagogues in the way Goulder supposes? Does this reconstruction square with such evidence as we have of Christian worship in the first and early second centuries? In neither case will evidence drawn from later centuries do, for we know that there was considerable development in both Jewish and Christian worship.

The relevant rabbinic evidence suggests that in the first century the Pentateuch was read neither over one year nor over three years. Several ‘triennial’ cycles probably date from the end of the second century A.D. Goulder does not claim to be able to trace a triennial cycle behind Matthew, but the annual cycle which he does discern behind Luke’s Gospel and behind the Matthaean infancy narratives is almost certainly even later than the triennial lectionaries. And so we are left with the set of readings for the six feasts and two major fasts, on which Goulder’s case for the origin of Matthew very largely rests. The Matthaean church, the author argues, continued to read the lections for these occasions; in their light the evangelist developed an annual lectionary of Christian readings. Particular Scriptural readings probably were associated with the feasts in the first century A.D., but there is no evidence that there was a set ‘festal lectionary’: indeed, there seems to have been considerable variation in the passages used. Do we know how the feasts and fasts were celebrated in Palestine or Syria prior to the destruction of the temple? Goulder assumes, without argument, that the readings associated with temple worship on these occasions were also used in synagogues. As far as I am aware, there is no evidence that this was so.

Even if we were confident that first century synagogues used a festal lectionary, further awkward questions remain. If Matthew’s church did keep the feasts and fasts, particular traditions about Jesus may well have become associated with these important occasions and understood in the light of traditional OT lections. But why should Matthew have wanted to develop an annual cycle of readings from a festal lectionary? Goulder anticipates the question and argues that the work of the Chronicler provides a precedent as he wrote a liturgical book for use over one year. But here conjecture triumphs over evidence and argument.

Some early Christian communities did continue to follow closely traditional Jewish patterns of worship. And Matthew’s church may have been more ‘conservative’ than other Christian communities. But there is no interest in the Jewish feasts in Matthew and very little evidence that Christian observance of any of the feasts or fasts survived for long. Goulder wants us to believe that Luke developed a Christian lectionary in the light of an annual cycle of OT readings. Justin Martyr’s famous comment on Christian worship surely rules this out. Justin, who almost certainly knew Luke’s Gospel, notes that the OT Scriptures and the ‘memoirs of the apostles’ were read ‘for as long as time allows’ (Apology I, 67). In Justin’s day, at least, there was no prescribed length for the readings and hence no lectionary.

If we refuse to accept Goulder’s lectionary hypotheses Q must be allowed back in court, for his objections to Q lean heavily on his claim that Luke re-wrote Matthew in order to provide a different kind of lectionary. Once we allow Q as Matthew’s second source, whether oral or written, midrashic expansion becomes an inappropriate description of the evangelist’s activity. Goulder’s book will interest and stimulate all serious students of the Gospels, but the old question remains unanswered: why did Matthew write his Gospel?

---