ECM Commentary Introduction[1]


Preliminary Note

This part of the NTVMR forum is dedicated to text-critical comments on variants listed in the apparatus of ECM Acts. As a start, we reproduced the commentary as printed in the Studies volume of the edition (ECM III.3, p. 1-38).

All readers are invited to discuss existing comments using the “Reply” function. Any registered expert user of the NTVMR may additionally register as a commentator and thereby obtain the right to open a new commentary thread on any passage of the online apparatus for which there is no comment so far. To register as a commentator, send an email to <>.

We presume that contributors are willing to subscribe to a scholarly methodology. This prerequisite boils down to the understanding that the initial text from which the manuscript tradition started is not simply given to us in one preserved form, but has to be reconstructed from the documents that came down to us.


A presentation of basic terms and procedures on which comments of ECM editors are based was published in two contributions to TC 2015:

G. Gäbel, A. Hüffmeier, G. Mink, H. Strutwolf, K. Wachtel: “The CBGM Applied to Variants from Acts: Methodological Background” (cited as “CBGM Background”)

K. Wachtel: “Constructing Local Stemmata for the ECM of Acts: Examples” (cited as “Examples”).

The TC articles refer to “Genealogical Queries Acts (Phase 2)”, a suite of tools available online at <>. In phase 3 of our work on the reconstruction of the initial text, these tools were updated to <>. The textual commentary as printed in ECM III.3 refers to phase 3. The results achieved in phase 3 open up phase 4 and are available with a new interface at <>. The online commentary and links to examples in the present introduction refer to the new interface which is based on phase 4 data.


1. The initial selection of variant passages to be commented on

A textual commentary is given for each passage where the editors of ECM Acts reconstructed the initial text different than NA28 and UBS5. A commentary is also included for passages with a split guiding line. For other passages a commentary is given if the editors’ assessment needs additional explanation to supplement the guidelines below.

In addition, the online commentary contains discussions of examples referred to in the guidelines and rules below. The commentary text for these passages is based on “Examples.

2. Guidelines and rules for assessing variants and their Greek manuscript attestations[2]

    Each example given below is linked to a comment. From there follow the link to the respective online ECM entry. Click on the arrow icon beside the ECM entry to access the relevant CBGM diagrams.

1. Singular readings and unique readings of small groups which differ from the mainstream of transmission are secondary. Exceptions to this rule require strong support from internal criteria. However, as was the case with the Catholic Letters, such variants are systematically subjected to text critical analysis if they are supported by witnesses closely related to A.

    Examples: 10:30/56, 10:43/22, 11:26/46-52.

2. An attestation lacking coherence is a sign of multiple emergence i.e. posteriority of a variant. Multiple emergence weakens the force of internal criteria which might be used to account for the priority of the variant.

    Examples: 10:30/56, 10:31/16-18, 10:31/24,11:13/35.

3. Good coherence of an attestation is primarily a sign of unfractured transmission. Good coherence is a valid argument for the priority of a variant only if supported by internal criteria. In reverse, disturbed coherence is not an irrefutable argument against the priority of a variant. TP can be decisive in both cases.

    Examples: 10:37/20, 11:8/12-18.

4. A strong argument for assessing a variant as initial text is provided by an attestation which combines coherence and a broad range of diverse witnesses closely related to A. In such cases strong coherence often only materializes if A is part of the attestation.

    Examples: 9:3/20-22, 9:15/22-24, 10:36/6, 12:25/10-14.

5. The priority of a majority reading is indicated if it is linguistically more difficult or contextually less suitable and thus atypical of the majority text. This may be valid even if the competing variant has a broad range of witnesses closely related to A.

    Example: 10:36/6

6. The source of a variant is likely to be a similar variant. If the attestation of a variant indicates that two or more other variants need to be considered as possible sources then Transcriptional Probability (TP) suggests that the one which requires the least change to be transformed into the variant in question is preferable.

7. The source of a variant is questionable:

    a) if Genealogical Coherence (GC) and TP point to different potential source variants or cannot be aligned with each other for other reasons;

    b) if we cannot decide which of two or more variants is the prior one because neither GC nor TP provides a convincing argument.

    Examples: 15:7/6 (for a), 17:3/34-40 (for b)

8. Consciously introduced editorial variants are exceptional. If possible, variation should be explained with reference to the process of copying itself and to known causes of error.

    Example: 4:27/6-26.

9. If the witnesses of a variant have to be assigned to different source variants, then the attestation should split accordingly.

    Example: 1:22/34-38, 10:1/7

R(ule) 1 and R(ule) 2 leading to a split guiding line

Commentaries for passages with a split guiding line may consist only of a reference to one of the following scenarios:

R1: TP provides no good argument that applies to the variants in question. At the same time, one of them is supported by a coherent majority of witnesses, while the other’s attestation has a wider range of early A-related witnesses and may be lacking overall coherence.[3]

     Examples: 1:25/18, 3:7/24-28

R2: TP provides no good argument that applies to the variants in question, while there are coherent cores of A-related witnesses on either side.

     Examples: 13:46/36, 13:49/14

In R1, a wider range of early A-related witnesses outweighs a coherent majority attestation that has a smaller range of A-related witnesses at its top. In R2, the coherent cores of A-related witnesses are given approximately equal weight, regardless of the rest of the attestations.

Usually, simply stating “R1” or “R2” is sufficient to explain the situation, but sometimes an explanation why TP does not present a sufficient argument is also provided.


3. “Genealogical Queries – Acts” <>

To document our work on the local stemmata and the text of ECM Acts, we provide access to the data we used in phase 3 of our work (the third iteration of the material). The database behind phase 3 reflects the state of work that resulted from phase 2. The printed text, including the split guiding lines, reflect the result of phase 3. This allows the user to see the basis for our work on the text and the local stemmata in phase 3 representing the state from which the printed edition was produced. The state of work reached at the end of phase 3 is available as “Genealogical Queries – Acts (Phase 4)” at <>.


There are Guides available on the websites for phase 3 and phase 4 providing more information about how “Genealogical Queries – Acts” functions. For this introduction, a few basic recommendations about efficient use of the “Coherence in Attestations” module (Phase 3 and 4) may suffice.

For Acts the default value of connectivity is “Lower (1-5 ancestors)” (phase 3) or 5 on the “Conn” slider (phase 4). This means that we acted on the assumption that for this writing a more cautious approach is appropriate. The connectivity of variants in Acts appears to be generally lower than in the Catholic Letters. In most cases, the user wants to assess the coherence of one attestation in comparison with another one. For this purpose it is crucial that the same parameters are used for both. For example, if one of the variants in question is a, and a is regarded as initial text, then the user should set Initial Reading (IR) to a (or A=a in phase 4), in order to see how the witnesses are related to A. If the coherence of the c attestation, for instance, is to be compared, then IR should be set to c accordingly. If the user wants to factor A out of the equation, IR should be set to ? for the respective variants. (The latter is default in phase 4, click on A in the navigation bar to include A in a diagram or a list of relatives.)

If the coherence appears weak with lower connectivity, it is often interesting to ascertain at which rank a strand or a single witness is integrated into one textual flow with the other witnesses of a variant. For this purpose the user may tick “Connectivity: Absolute” (or click on the extreme right of the “Conn” slider in phase 4), which will connect all witnesses to one consistent graph. Typically, a higher ranking number points to a weaker connection between the witnesses in question. Clicking the “Show Table” feature above the graph will open the table of ranking numbers and percentages behind the graph. Data can be further differentiated by the lists of potential ancestors and descendants. (In phase 4, “Show Table” has become obsolete, because a click on any witness number will bring up its list of relatives.)

For example, if “Connectivity: Absolute” is set for 16:12/12-16a, 218 is connected to 1509 as a seventh potential ancestor. The “Show Table” feature tells us that the respective percentage of agreement is 90.794. Is this relatively high or relatively low? The list of potential ancestors of 218 (with “Exclude Fragments and Undirected Relationships”) displays that there are only three potential ancestors that 218 agrees with more than with the majority text: 18, 468, and 35. Between the third and the fourth potential ancestor, 1735, there is a gap of about 1.7 percentage points, and it takes three more steps to reach ranking number 7. An agreement rate of 90.794 is still higher than the average (AA) and the median agreement (MA), but the closely related witnesses are obviously found in the first three positions. Hence the binding force attributed to the variant had to be fairly high to accept the relationship of 1509 and 218 as coherent. In this case, a TP of the independent emergence of the variant is enforced by the fact that the first six potential ancestors of 218 support variant b.

For 1842, whose seventh potential ancestor is 218, this conclusion cannot be drawn. The list of potential ancestors for 1842 starts with 89.913% for L23 as its first potential ancestor and is less than 0.6 percentage points lower for 218. In this situation, the coherence values cannot contribute to the assessment of the origin of a in 1842.

Based on the revised data of phase 4, the list of ancestors, agreement and priority values have changed in detail, but the outcome remains basically the same. If you set “Conn” to “All”, 1509 will appear as the ninth potential ancestor of 218. The list of relatives of 218 will show 1509 as ninth potential ancestor with an agreement rate of 90.76. Now we see that there appear four potential ancestors with more than the MT/P agreement rate of 93.06. The one additional potential ancestor is 398 which in phase 3 was listed among the potential descendants. The gap between the fourth and the fifth potential ancestor, 1735, still is about 1.7 percentage points. Now it takes four more steps to reach 1509 with ranking number 9, but this still is the first relative within the same attestation. Again, the agreement rate is higher than average, but the closely related witnesses are obviously found in the first four positions, and TP of the independent emergence of variant a in 218 is confirmed.


4. Impasses according to TP

1) Luke typically connects sentences with τέ, although δέ is more common. Variation may reflect secondary adaptation to common usage or to Luke’s preference.

2) 03, known for its tendency to omit dispensable words,[4] is often the most prominent witness for omissions of phraseologically common redundant words.

3) TP often cannot provide a decisive argument when an aorist subjunctive is required by classical usage, even though the general linguistic development favors the future tense, especially if variation may be due to vowel interchange (cf. 4:16/6, 4:21/20, 7:7/12, 8:31/22).

4) The article is increasingly used with proper names even in places where it is not referring to a person being mentioned in the preceding passage (cf. 2:14/4-6). The general linguistic development supports the use of the article, which has become mandatory in modern Greek.[5]

5) θεός and κύριος are frequently used like proper names, particularly under the influence of the LXX.[6]


Unless otherwise stated, Accordance 11 (OakTree Software 2015) <> was used on the tagged Rahlfs LXX and/or NA28 editions to calculate the number of occurrences of words or phrases.



ADNTCE – Amsterdam Database of New Testament Conjectural Emendations

Con – Connectivity

GC – Genealogical Coherence

GenQ – Genealogical Queries (see Bibliography)

IR – Initial Reading

MT – Majority Text

NA – Nestle-Aland

NT – New Testament

R – Rule

TP – Transcriptional Probability

TFD – Textual Flow Diagram



Accordance 11 (OakTree Software 2015) <>

Barrett – Charles K. Barrett: A Critical and Exegetical Commentary on the Acts of the Apostles (2 vols.), ICC, Edinburgh: T&T Clark 1994, 1998.

BDAG – Walter Bauer: A Greek-English Lexicon of the New Testament and Other Early Christian Literature, ed. W.F. Arndt, F.W. Gingrich; 3rd ed. rev. by F.W. Danker. Chicago: Univ. of Chicago Press 2000.

BDF – F. Blass and A. Debrunner: A Greek Grammar of the New Testament and Other Early Christian Literature, ed. R.W. Funk. Chicago: Univ. of Chigaco Press 1961.

BDR – F. Blass/A. Debrunner/F. Rehkopf: Grammatik des neutestamentlichen Griechisch. Göttingen: Vandenhoeck und Ruprecht 17. Aufl. 1990.

Conzelmann – Hans Conzelmann: Die Apostelgeschichte, Handbuch zum Neuen Testament 7, Tübingen: Mohr 21972.

GenQ – Genealogical Queries – Acts <>

GenQ2 – Genealogical Queries – Acts (Phase 2), accessible from GenQ

GenQ3 – Genealogical Queries – Acts (Phase 3) accessible from GenQ

GenQ4 – Genealogical Queries – Acts (Phase 4) accessible from GenQ

Haenchen – Ernst Haenchen: Die Apostelgeschichte. KEK, Göttingen: Vandenhoeck und Ruprecht, 12. Aufl. 1959.

Hort – Notes On Select readings, in: B.F. Westcott/F.J.A. Hort: The New Testament in the Original Greek, Bd. II, Appendix I, Cambridge/London 1882.

Jervell – Jacob Jervell: Die Apostelgeschichte. KEK, Göttingen: Vandenhoeck und Ruprecht 1998.

Lake and Cadbury – The Beginnings of Christianity, Part I, The Acts of the Apostles, ed. F.J. Foakes Jackson and Kirsopp Lake; vol. IV, English Translation and Commentary, by K. Lake and H.J. Cadbury, London: Macmillan 1933.

LSJ – H.G. Liddell/R. Scott: A Greek-English Lexicon, rev. and augmented by H.St. Jones, Oxford: Clarendon 91940 (repr. 1985).

Metzger – Bruce M. Metzger: A Textual Commentary of the Greek New Testament, Stuttgart: Deutsche Bibelgesellschaft 21994 (2nd printing 1998).

Mink 2004 – Gerd Mink: Problems of a Highly Contaminated Tradition: The New Testament, in: P. van Reenen, A. den Hollander, M. van Mulken (eds.), Studies in Stemmatology II, Amsterdam/Philadelphia: Benjamins 2004, p. 13-85.

Mink 2009 – Gerd Mink: The Coherence-Based Genealogical Method – CBGM: Introductory Presentation, 2009, online <>.

Moulton III: Nigel Turner, Syntax, in: James H. Moulton (ed.), A Grammar of New Testament Greek, T&T Clark: Edinburgh 1963.

NA26-28 – Nestle-Aland, Novum Testamentum Graece, 26th edition 1979 by K. Aland, M. Black, C.M. Martini, B.M. Metzger, A. Wikgren, 27th/28th edition 1993/2012 by B. Aland, K. Aland, J. Karavidopoulos, C.M. Martini, B.M. Metzger; Stuttgart: Deutsche Bibelgesellschaft.

NET Bible – The NET Bible, New English Translation, Biblical Studies Press 1996-2005 [Version 3.4 used in the Accordance 11 environment].

Nicklas/Tilly: Tobias Nicklas and Michael Tilly (eds.): The Book of Acts as Church History. (BZNW 120) Berlin/New York: de Gruyter 2003.

Omanson – Roger L. Omanson: A Textual Guide to the Greek New Testament. An Adaptation of Bruce M. Metzger’s Textual Commentary for the Needs of Translators, Stuttgart: Deutsche Bibelgesellschaft 2006.

Pervo – Richard I. Pervo: Acts: A Commentary. Hermeneia, Minneapolis: Fortress 2009.

Rahlfs – Alfred Rahlfs (ed.): Psalmi cum Odis, Septuaginta vol. X, Göttingen: Vandenhoeck/Ruprecht 1967.

Ropes – The Beginnings of Christianity, Part I, The Acts of the Apostles, ed. F.J. Foakes Jackson and Kirsopp Lake; vol. III The Text of Acts, by James Hardy Ropes, London: Macmillan 1926.

Tischendorf – Novum Testamentum Graece, Editio octava critica maior. Leipzig 1869-1872.

Strutwolf 2014 – Holger Strutwolf: Urtext oder frühe Korruption? Einige Beispiele aus der Apostelgeschichte, in: Texts and Traditions. Essays in Honour of J. Keith Elliott, eds. Peter Doble und Jeffrey Kloha, Leiden/Boston 2014, 255 - 280.

UBS3-5 – The Greek New Testament, 3rd edition 1975 by K. Aland, M. Black, C.M. Martini, B.M. Metzger, A. Wikgren, Stuttgart: United Bible Societies; 4th/5th edition 1992/2014 by B. Aland, K. Aland, J. Karavidopoulos, C.M. Martini, B.M. Metzger; Stuttgart: Deutsche Bibelgesellschaft/United Bible Societies.

Wendt, KEK 1888 – Hans H. Wendt: Handbuch über die Apostelgeschichte, Kritisch Exegetischer Kommentar über das Neue Testament 3, ed. H.A.W. Meyer, Göttingen: Vandenhoeck/Ruprecht 1888.

Wevers – John W. Wevers (ed.): Genesis, Septuaginta vol. I, Göttingen: Vandenhoeck/Ruprecht 1974.

Ziegler – Joseph Ziegler (ed.): Isaias, Septuginta vol. XIV, Göttingen: Vandenhoeck/Ruprecht 1939.




[1] Based on the introduction printed in ECM III.3, p. 1-4.

[2] These guidelines were first published in “CBGM Background” p. 3, referring to cases discussed in “Examples”. The examples are now linked to the new interface (phase 4).

[3] It is methodologically important to distinguish between manuscripts and the texts they contain. To say it in the words of Gerd Mink, “The CBGM deals with texts, not with manuscripts. The text is the witness.” (Mink 2009, 38; cf. 33-38 and Mink 2004, 29). Accordingly, “early witnesses” are not necessarily preserved in early manuscripts. They are states of texts relatively close to the beginning of the transmission, the Ausgangstext. Manuscripts from the second millennium are sometimes witnesses to an earlier text.

[4] See e.g. Hort, Introduction, 234-236, who in spite of his preference for 03 concedes that “it is on the whole safer for the present to allow for a proneness on the part of the scribe of B to drop petty words not evidently required by the sense” (236); cf. Carlo Martini in his introduction to Novum Testamentum e codice Vaticano Graeco 1209, Vatican City 1968, p. xxi-xxii.

[5] Cf. BDR/BDF 260.

[6] Cf. BDR/BDF 254.