Kurt Aland, Barbara Aland, and Klaus Wachtel, eds., in collaboration with Klaus Witte. Text und Textwert der Griechischen Handschriften des Neuen Testaments. Vol. 4: Die Synoptischen Evangelien. No. 2: Das Matthäusevangelium. 2.1: Handschriftenliste und Vergleichende Beschreibung. 2.2: Resultate der Kollation und Hauptliste sowie Ergänzungen. Arbeiten zur Neutestamentlichen Textforschung, nos. 28-29. Berlin: de Gruyter, 1999. Pp. 24* + 517 (2.1); 327 + 61* (2.2). DM 398; US $249.00. ISBN 3-11-016418-3.

No. 3: Das Lukasevangelium. 3.1: Handschriftenliste und Vergleichende Beschreibung. 3.2: Resultate der Kollation und Hauptliste sowie Ergänzungen. Arbeiten zur Neutestamentlichen Textforschung nos. 30-31. Berlin: de Gruyter, 1999. Pp. 545 (3.1); 353 + 62* (3.2). DM 418; US $261.00. ISBN 3-11-016420-5.

1. The series Text und Textwert now moves into the Synoptic Gospels, and the Arbeiten zur Neutestamentlichen Textforschung pass the thirty mark. Both achievements bear witness not only to the present activity and high standards of the Institut für Neutestamentliche Textforschung, Münster, but also to the solid foundations laid in the early years of the Institut's life, while materials were being acquired and working practices and editorial principles laid down. The forethought and wisdom shown then are amply demonstrated in the remarkable sequence of publications of recent years.

2. Three other publications will be necessary for fully interpreting the material in these volumes. The first is the volumes relating to the Gospel of Mark.1 It contains material introductory to the three Gospels at the beginning, explaining the volumes, including innovations from earlier instalments. The second is the contribution by Professor Aland and Dr Wachtel to the second Metzger Festschrift.2 The third is yet to appear. It will consist of further study of synoptic parallels, 'with the manuscripts which differ most from the majority text being collated thoroughly, not in test passages, but in longer sections in parallel pericopes shared by two or three of the Synoptic Gospels.'3 It should be borne in mind, therefore, that the volumes under review will be complemented by material yet to appear.

3. To remain sensible in length, the review must not attempt to offer a discussion of the overall method of analysing manuscripts used by Münster. Instead, it will describe what is in these particular volumes and offer a few comments on what has caught the reviewer's eye.

4. It needs pointing out, if one is to understand what is being done in Text und Textwert, that the aim is above all to find a way of distinguishing those MSS which are Byzantine from those which contain another form of text. That is, given the vast number of witnesses to be examined, many of which are unknown quantities, a way must be found of identifying as accurately and as economically as possible those witnesses which can be subsumed in an apparatus criticus into the siglum BYZ, and those which need to be analysed further in the process of reconstructing the history of the text and editing the book in question. Various solutions have been attempted to this problem. The INTF method is to select test readings throughout a book, to collect the reading of every witness in those readings, and then to analyse the results, first separating out the Majority witnesses and then analysing the remainder in terms of the number of agreements between witnesses and in relation to the Majority reading.

5. The pattern of the two volumes is the same, except that the Matthew volume contains a list of all Synoptic MSS which differ more than 10% from the Majority Text, as well as a few addenda and corrigenda to Mark (pp. 18*f). Each Gospel consists of the following material in both German and English (the translations here, as in other Münster volumes, were made by Erroll F. Rhodes): a list of symbols and abbreviations; a list of all MSS analysed in that Gospel, with information on those lacunose in all Teststellen and the number differing more than 10% from the Majority Text; analyses of the MSS according to their proportion of (1) agreements with the Majority Text, (2) agreements with the 'old Text',4 (3) singular readings, followed by lists of the BYZ and non-BYZ MSS, and lists of groupings of MSS. The next main section, comprising the remainder and indeed the vast bulk of each first volume, provides the information on the number of readings in the various categories found in each MS. The second volume consists first of the collations of the non-Byzantine MSS for each Teststelle. Then comes the Hauptliste, a profiling MS by MS in which are listed the 33 MSS closest to each, in those readings where that MS has neither a majority text nor a singular reading. The Hauptliste is essential as the link between the statistics and the Teststellen. Finally, in a separate paper cover and held in a folder at the back is the Ergänzungsliste, the Supplementary List. For each MS, all the 155 MSS closest to it are listed, on the basis of all Teststellen, including the readings of the majority text.

6. The task of analysing the Gospels is more considerable than that which has been undertaken in earlier volumes of Text und Textwert, since there are more manuscripts to be sampled. The official numbers presented in these volumes are as follows. The total number of MSS containing some or all of one or more of the Synoptic Gospels is 2,212 (4.2.1, p.18*). The total number available to the editors was 1,997. The number available for the Gospels which are the subject of these volumes is
Of these, some MSS are so fragmentary as to be lacunose at every Teststelle:
Thus, the following numbers of MSS are examined in the studies:
All MSS90+% ByzantineLess than 10% Byzantine
We thus have a statement, backed up by the information, of the number of Byzantine witnesses in each Gospel, and of the remainder.

7. To be successful in its aims and to meet the inevitable restraints of resources, the process of identification requires both the selection of satisfactory test passages and economy in the number chosen. As to the latter need, there are 64 Teststellen for Matthew and 54 for Luke. This works out as just over 2 per chapter, though they are not spread as evenly as that. In Luke, the first comes at 2.14; the distribution by chapter is:
ChapterNo. of Teststellen

8. The kind of variant that is required is really one where the Byzantine text departs from the superior reading, usually found in one of 01 () or 03 (B) (and 75 where it is extant in Luke). This means that only two variants are necessary in a passage. In Luke, this is the situation in eight places, which divide into two groups. In the first, the choice is between a 1: (BYZ) and a 2: ('old text') reading:

TS 14 8.52 <grc>ouk</grc> BYZ | <grc>ou gar</grc> 01 03 etc

TS 17 9.54 <grc>autous ws kai Hlias epoihsen</grc> BYZ | <grc>autous</grc> 01 03 etc

TS 21 10.22 <grc>kai strafeis ... panta</grc> BYZ | <grc>panta</grc> 01 03 etc

TS 49 23.34 read the verse BYZ | omit the verse 01 03 etc

TS 52 24.42 <grc>meros kai apo melissiou khriou</grc> BYZ | <grc>meros</grc> 01 03 etc

9. In all these places except the first, there are in fact sub-readings in the BYZ reading, and these are listed for the MSS concerned in the Verzeichnende Beschreibung. But the important fact is that the Byzantine reading can be detected against the better reading. In the second group, the choice is between a 1/2 reading and a third:

TS 16 9.3 <grc>ana duo</grc> BYZ 02 al | <grc>duo</grc> 01 03 al

TS 32 15.21 SINE ADD. BYZ 75 02 al | <grc>poihson ... sou</grc> 01 03

TS 40 19.25 <grc>kai eipan ... mnas</grc> | BYZ 01 al omit 05 al

10. These are in fact a sub-set of test passages where the first reading is 1/2. There are eight of them in all in Luke. They are TS 16, 23, 32, 37, 38, 40, 42, 43. These readings are there usually because they are places where 01 and 03 are considered by the researchers to be in error; that is, they have reading 3. The exception is at TS 40, where BYZ happens to follow the 01 03 text. If the aim is to distinguish BYZ manuscripts, then it is arguable that this last passage is not much help. But this is to be wise after the event. It is highly probable that a few out of 54 promising passages will turn out less useful than one would have expected.7

11. The most variants in a single Teststelle is 12, and this number is found twice (TS 5 and 36). The greatest number of sub-variants to a BYZ reading is 36, at TS 18. But at TS 47, where there are 7 readings, there are so many sub-readings (including 34 BYZ) that the full total of variants listed is 60.

12. Turning to Matthew, we find that the 64 Teststellen are even more unevenly divided. There are 46 in the first fourteen chapters and only 18 in the second fourteen:
ChapterNo. of Teststellen

13. There are also far more 1/2 readings, 17 in all, of which five are readings in which 03 (and usually 01 as well) agree with BYZ (TS 1, 2, 10, 22, 36). As it happens, all but one of the 17 fall within the first fourteen chapters. Again in contrast to Luke, there are 12 readings where there are only two variants (7 of them also 1/2 readings). Nor do any Teststellen have as many variant readings as the most complex in Luke. The greatest number is 38 in TS 15, followed by 37 at TS 48 and 32 at TS 19. These figures, I assume, reflect the fact that the textual tradition of Matthew is less complex than that of Luke.

14. There emerges very obviously a typical profile for a Byzantine MS. In Luke it is 2: 2; 1/2: 8; 1: 44, and MSS with this shape or one very close to it (a few errors can change the figures slightly) can be seen on every page of the Verzeichnende Beschreibung. In Matthew the numbers are 2:1; 1/2: 18; 1: 42, or 1/2: 19; 1: 45.8

15. It has often been stated that one of the demands on the selection of test passages is that they should be sufficient to detect block mixture, or even what Colwell used to call box car mixture, in a witness. That is, it should pick up witnesses which are Byzantine in one part of a Gospel and not in the rest, or even witnesses which switch from one to the other more than once. It may be that the rather thin spread in the second half of Matthew could make it less likely for a text change to show up. Of course, the percentages are of no help in finding block mixture; one has to go through the Verzeichnende Beschreibung and the Hauptliste.

16. I turn now to one or two random comments which I hope will serve to give something of the flavour of the volumes. There is something of a surprise in looking at the list of manuscripts with 'old Text' readings in Matthew. At the top are two rather fragmentary manuscripts, 0281 and 035. Then come 03 with 47 out of 64 type 1/2: and 2: readings and 01 with 45 out of 64. Next is 892 with 43 out of 64. Even if one excepts the 1/2 readings (where there could be Byzantine influence), the pure 2: readings amount to 25 out of 45. This is ahead of 1582 (22/42 2: readings) and 1 (23/45 2: readings). Forget the Queen of the Cursives (33), with a mere 17/45. Rendel Harris was justified in his description of this MS.9 But in Luke, the quality falls right off, to a mere 11 2: readings.

17. Another interesting way of studying the Lukan data is by comparing it with the results of the Claremont Profile Method (CPM).10 One can do this both to test the results against each other and to find how to use Text und Textwert, by comparison with something different in approach. Three chapters were analysed for profiling by CPM: 1, 10, 20. There are in fact only four variants in common between the 196 of CPM and those in Text und Textwert (ch. 10, CPM Readings 36, 37, 59 = TS 20, 21, 22; ch. 20, CPM Reading 48 = TS 42). Turning again to 892, we find that it is defined as a member of Group B (called after Codex Vaticanus). Searching for an explanation for this marked divergence from the Münster findings, I turned again to the data in Text und Textwert. Looking at the list, one notes that there are quite a few distinctive readings, 11 of them. Going through the Collation Results, one finds in TS 9, 23, 37, 42, 44, and possibly 43 (where it has a sub-variant of the 01* 03 reading) that these are variants which the CPM would count as Group B readings, since it does not use the 'old Text category'. This reduces the difference significantly, but maybe not enough. How big a margin of error is allowable between the two methods?

18. Staying with 892 and looking at the remaining Sonderlesarten, one can then discover how many potentially fruitful lines there are to follow up, for we find that TS 20 is particularly interesting: 892 is in agreement with 0115 2309 2542 only. Of these, only 2542 features in Wisse's study. This MS looks interesting from the data in Text und Textwert. According to CPM, it is mixed in Luke 1, and a weak member of Family 1 in the other two chapters. At TS 28, it shares its reading with only one other MS, 2437. Making a detour to explore this piece of information, we find that 2437 receives the designation BYZ in the Verzeichnende Beschreibung, so it does not feature in the Hauptliste or Ergänzungsliste. From Wisse, we learn that it is a member of Group 1519. Checking that back with the Hauptliste, we find that 2437 does not make it into the list as one of the 33 MSS closest to 1519. It does not even make it into the Ergänzungsliste. It is a pretty sobering fact that of the 19 MSS identified by CPM as belonging to Group 1519, only one (1309) makes it into the Hauptliste, and only four (1211 1321 1481 1566) into the Ergänzungsliste. 1309, 1211 and 1481 are only Group 1519 in Luke 20 (according to CPM), where it will be recalled there is only one TS. Putting this the other way around, of the 155 MSS listed as closest to 1519 in the Ergänzungsliste, only four were identified as associated with 1519 by CPM. It is a relief to begin to find some affinity between the two methods when we look at the MSS listed in the Ergänzungsliste as closest to 1519. The two highest which are extensive are 111 and 281, members of Cluster 281, a cluster within Kx.11 One realises at this point that whilst the presence of the Hauptliste allows one to examine for oneself the evidence on which the conclusions from the Teststellen are based, this is not possible with Wisse's volume, since it gives the profile for groups but not for the individual witnesses. The consequence is that we just cannot tell how strong a member of a group any given witness may be. Comparison between Text und Textwert and the CPM can be only partial, particularly since CPM looks for groupings within BYZ, while Text und Textwert is filtering out BYZ as a whole. These areas have been debated frequently enough in the past, and there is still room for explanation and analysis.

19. The thread of such exploration is practically endless, and can lead one into MS relationships totally unconnected to the ones with which one had begun. The scarcely connected ramblings of the previous few paragraphs are intended to illustrate the thousands which are possible. It is the number of different ways of looking at the data, and the fact that one can pick particular routes through the data, looking at variants, looking at MSS, looking at relationships, that make these volumes so fascinating. Readers of Wisden Cricketers' Almanack will know the pleasure of browsing through the section called 'Cricket Records' and finding out who scored most runs for England against Australia, or the bowlers who have taken ten wickets in an innings. These volumes give similar endless pleasure (I can think of no higher compliment). To turn the pages of Wisden is also to recapture something of long fled summer days. To study Text und Textwert is to appreciate that the scholarship of our own day preserves the traditions of distant generations.

© TC: A Journal of Biblical Textual Criticism, 2000.


1Kurt Aland, Barbara Aland, Klaus Wachtel, eds., in collaboration with Klaus Witte, Text und Textwert der Griechischen Handschriften des Neuen Testaments, vol. 4: Die Synoptischen Evangelien, no. 1: Das Markusevangelium, 1.1: Handschriftenliste und Vergleichende Beschreibung, 1.2: Resultate der Kollation und Hauptliste sowie Ergänzungen, Arbeiten zur Neutestamentlichen Textforschung, nos. 26-27 (Berlin: de Gruyter, 1998).

2Barbara Aland and Klaus Wachtel, "The Greek Minuscules of the New Testament," in The Text of the New Testament in Contemporary Research: Essays on the Status Quaestionis, ed. Bart D. Ehrman and Michael W. Holmes, SD 46 (Grand Rapids: Eerdmans, 1995), 43-60.

3Introduction to the Matthew volume, p. 3*.

4The 'old text' reading is that of Nestle-Aland27.

5The number given on Matthew, p. 4 (bis) is 54, but the number of MSS listed and the arithmetic both come to 55.

6The number given on Luke, p. 3 and p. 4 is 30, but the number of MSS listed and the arithmetic both come to 31.

7This is acknowledged by the authors in their abandonment of the 'Bezeugte Variante' section of the Hauptliste. See the introduction to Mark, p. 23*.

8The single reading in most Byzantine witnesses is test passage 32, at Mt 11.8, and the variation in question is between <grc>basileiwn</grc> and <grc>basilewn</grc>. The former reading is considered to be BYZ, while the latter is read by 01 03 04.

9J. R. Harris, "An Important Manuscript of the New Testament," JBL 9 (1890), 31-59.

10F. Wisse, The Profile Method for Classifying and Evaluating Manuscript Evidence, SD 44 (Grand Rapids: Eerdmans, 1982).

11It is described by Wisse as 'close to Kx in Luke 1 and 10' (p. 109).

D. C. Parker
Reader in New Testament Textual Criticism and Palaeography
Department of Theology
University of Birmingham