Authorship of Letters Ascribed to Paul

Return to Index

The purpose of these notes, completed for my students in India in 1973, is to indicate some types of approaches to the study of authorship of the letters (and also some bibliography) that lie behind the following working hypothesis which I have used ever since (dates are approximate):

Letters by Paul:
Galatians (ca. 52 CE)
1 Corinthians (Spring, 52 CE)
2 Corinthians (Autumn, 52 CE – this letter is composite)
Philemon (52-54 CE)
Romans (54-58 CE)
Pauline School (Deutero-Pauline, ca. CE 65-90)
1 Thessalonians
2 Thessalonians (dependent upon 1 Thessalonians and not as well attested)
Ephesians  (dependent upon Colossians, but showing knowledge of all the above letters, and thus not written until after the letters had been collected together, sometime between CE 70 and 90)
The Pastorals (Trito-Pauline, ca. CE 95-105)
2 Timothy (ca. CE 95 – B. S. Easton’s date)
1 Timothy (ca. CE 105 – B. S. Easton’s date)


        "In general in antiquity there were numerous pseudonymous writings; … the literary genus of pseudonymous writings … was very widespread … precisely in the period of Hellenism; … out of late Judaism as well as out of early Christianity there is reliable evidence for it…." (W. G. Kümmel, Introduction to the New Testament (SCM Press, London, 1966), p. 255.
Examples (drawn from Kümmel):

Jewish Epistle of Jeremiah (late 4th century BCE)
Epistle of Aristeas (written ca. 100 BCE in Alexandria)
Wisdom of Solomon (written ca. 100-50 BCE in Alexandria)
Christian: Acts 23.26-30 (purported letter of Claudius Lysia to Felix )
Didache of the 12 Apostles (ca. 100-150 CE)
3 Corinthians (included in the Acts of Paul, ca. 180 CE)

Note that 2 Thess 2.2; 3.17 indicates the possibility of ‘false’ letters in Paul’s name.


Basically there are two kinds of tests:

1. Examination of the author’s conscious choice of
1) Vocabulary and types of grammatical construction;
2) Theological models, ideas and outlook;
3) Literary structures (e.g., chiasmus, inclusio or parallel structuring)
Chiasmus: A-B-C-D-D`-C`-B`-A`
Inclusio: A-B-C-D-A (i.e., the end returns to the beginning)
Parallel: A-B-C-A-B-C-A-B-C
4) We may also include here the apparent echoing or even paralleling of OT books, including echoing or paralleling of Jewish lectionary usage.
Judgements by scholars as to how much an author could or would vary in these things with changes of time and circumstance often differ widely, so these tests are often dependent upon the subjective judgements of the scholars involved.
The results seldom command universal assent among critical scholars. Two near-exceptions are P. N. Harrison’s work on the Pastorals (1921 and 1946) and C. L. Mitton’s work on Ephesians (1937).
2. Examination of the author’s unconscious stylistic traits such as:
a. Use and distribution of kai (‘and’, ‘even’, ‘also’) and de (‘and’ or ‘but’);
b. Sentence length distribution;
c. Use and distribution of various kinds of words with which to end clauses.

Indication of authorship by unconscious stylistic traits

A partial bibliography in chronological order

Wake, W. C.
"Sentence Length Distributions of Greek Authors," Journal of the Royal Society (London), Series A, Part 3, 1957, Vol. 120, pp. 331-346.
Morton, A. Q., and McLeman, James
Christianity and the Computer, Hodder and Stoughton, London, 1964.
Morton, A. Q.
"Statistical Analysis and New Testament Problems," in The  Authorship and Integrity of the New Testament, S.P.C.K. Theological Collections, No. 4 (SPCK, London, 1965), pp. 40-60.
Morton, A. Q.
"The Integrity of the Pauline Epistles," Journal of the Manchester Statistical Society, March, 1965.
Levison, M., Morton, A. Q., and Wake, W. C.
"Some Statistical Features of the Pauline Epistles," Journal of the Royal Philosophical Society (London), July,1966.
Morton, A. Q., and McLeman
Paul, The Man and the Myth, Hodder and Stoughton, London, 1966.
Morton, A. Q.
The Authorship and Integrity of the New Testament Epistles,  Edinburgh, 1971.
Michaelson, S. D., and Morton, A. Q.
"Last Words: A Test of Authorship for Greek Writers," New Testament Studies, Vol. 18, No. 2 (January, 1972), pp. 192-208.

The following is a passage of Greek prose with the six commonest words translated, and with a hyphen (-) for each of the other words:

               And he - - - - - - - - - and the - - - the - - - - the - - - - - - -.

The following is the same text with only the six commonest words omitted:

              - - said, there was a man who had two sons - - younger said to – father, Father, give me – share of property that falls to me.

The second version is clear if inelegant, and is readily identifiable as the start of the parable of the Prodigal Son in the RSV (Luke 15.11-12). It is these fifteen rarer words that give interest and meaning to the passage; the other seven of the twenty-two words only help to provide a framework. The frequent use of these filler words means that they tend to become habitual in use, and so a good subject for stylistic studies. (Drawn from Paul, The Man and the Myth, p. 45)

The need for controls

If habitual stylistic devices are to be used as a test for authorship, then it needs to be shown (1) that they are habitual for writers of Greek prose, (2) that each writer has a distinctively characteristic pattern, (3) that that pattern remains stable over a wide range of time, and (4) the sample size (how long a passage?) for reliably determining this pattern has to be established.

To these ends A. Q. Morton and others have investigated various characteristics for: Herodotus, Thucydides, Plato, Philo, Diodorus Siculus, Lysias, Demosthenes, Isocrates, and Clement of Alexandria among others.

The results of this work show that the material must be homogeneous prose, without dialogue or commentary (such as a biblical commentary) or a long string of quotations. Morton’s basic approach is to take Galatians as being by Paul, and then seeing how well the other letters match its characteristics by means of various ‘closeness of fit’ tests, among which is the so-called "Chi-Square Test", but he also uses many other more sophisticated tests as well.

The letters he is able to check by these means are all the letters except 2 Thessalonians, Titus and Philemon, which are too short to provide reliable samples.

Sentence Length:

Group 1: Romans, 1 & 2 Corinthians, & Galatians, with anomalies in the first samples (150 sentences) of both Romans and 2 Corinthians.
Group 2 Philippians, Colossians, 1 & 2 Thessalonians, and Pastorals
Totally isolated epistles: Hebrews and Ephesians.

Kai (‘and’, ‘even’, ‘also’):

      1. Romans, 1 & 2 Corinthians, Galatians show no overall significant difference, but the first sample of Romans shows a difference, for there are two passages with a comparatively large number of two "kais" in them, Rom 2.2-11 and 5.19-25.
      2. There is a significant difference between Ephesians, Philippians and Colossians on the one hand and 1 and 2 Timothy on the other. 1 Thessalonians could belong to either group.

De (‘and’, ‘but’)

      1. Romans, 1 & 2 Corinthians, Galatians form a group, but Rom 9.16-13.5 with all its OT quotations lying at the start of sentences causes an anomaly in the third sample of Romans. The last sample of 2 Corinthians is also somewhat out of line.
      2. Hebrews, Colossians, Ephesians and Philippians are all distinct from the above and from each other.

Word-types used to end sentences

For this test ‘sentence’ is defined as that which editors of critical texts on the Greek NT end with ; (= ? in English), · (= ; in English) or . (= . in English). The word-types are distinguished in categories: (1) Nouns, (2) Aorist verbs, (3) Non-aorist verbs, (4) Other.

      1. Galatians and 1 Corinthians are homogenous and may be taken together to establish Paul’s normal usage.
      2. Romans’ structure is complex, with neither chapter 1 nor chapters 15 and 16 being ‘Pauline.’ There are anomalies from 4.12-5.9 and from 8.16-9.21, and there is evidence of a diffuse anomaly from chapters 11 to 14. Thus ‘the tests suggest that Romans is not a free composition of the Apostle’s, a conclusion often enough reached by other scholars’ (Michaelson & Morton, NTS 18/2, p. 205).
      3. 2 Corinthians: The first chapter is not Pauline and 2.15-4.13 and 8.2-23 are anomalous. ‘The conclusion is that this episle has been fabricated from Pauline material and furnished with an introductory chapter’ (Michaelson & Morton, ibid., p. 206).
      4. Ephesians, Philippians and Colossians: By this test these are clearly distinguished from Paul but not from each other.
      5. 1 & 2 Thessalonians: This test clearly distinguishes these from Paul, but
        corroborates that these two are from a single hand.
      6. 1 2 Timothy and Titus: These are clearly distinguished from Paul, and Titus appears to be significantly different from 1 & 2 Timothy
      7. Philemon is too short to deduce anything from it.
      8. Hebrews is clearly distinguished from Paul. Heb 13 is separate from Heb 1-12.

Some further arguments (drawn from A. Q. Morton, Paul, The Man and the Myth, pp. 96-97):

      1. The proportion of adjectives and adverbs is a constant: consistent for Romans, 1 & 2 Corinthians and Galatians.
      2. Cumulative-sum techniques show that Ephesians, Philippians and Colossians are written by men with quite contrasting habits in their grouping of sentences.
      3. In Romans, 1 & 2 Corinthians and Galatians there are 26 (rhetorical) questions per 1500 words. In all other epistles: 1 question per 1500 words.
      4. How OT quotations are introduced: consistent in Romans, 1 & 2 Corinthians and Galatians; different in the others.
      5. Collection for famine relief: prominent in 3 of the 4 (Rom, 1 & 2 Cor) and probably mentioned in Gal 2.10; missing in the others.

Corroborating Evidence from Tests Regarding Conscious Choices:

            1.    Vocabulary, etc.:
      1. P. N. Harrison, The Problem of the Pastorals (Oxford U. Press, 1921) and Paulines and Pastorals (Villiers Publications, London, 1964), has massively demonstrated that (1) the Pastorals’ style and vocabulary are distinct from Paul, and (2) they closely match the style and vocabulary current in the early second century CE (as in the Apostolic Fathers and the Apologists).
      2. E. J. Goodspeed, The Meaning of Ephesians (U. of Chicago Press, 1933), has shown Ephesians’ dependence on Colossians, and its knowledge and use of the other ‘Paulines’ in a way not found in any other letter. (More recently its high proportions of Semitisms [four times as frequent as in the other letters], plus its affinities with Qumran materials have been noted.)
           2.    Literary structure:
      1. "Paul’s" letters are chiastic overall:
      2. Romans 1-15 plus 16.26-27 (the floating doxology) (work of J. M. Gibbs)
        1 Corinthians (work of J. M. Gibbs, plus help of John Bradley)
        2 Corinthians’ recognized three sections by Paul:
            2 Cor 1.23-7.16 (minus interpolation of 6.14-7.1)
            2 Cor 8-9
            2 Cor 10.1-13.10 (all work of J. M. Gibbs)
        Galatians (work of J. Bligh)
        Philemon (work of J. M. Gibbs)

      3. Philippians is chiastic overall in same way (work of J. M. Gibbs)
      4. Ephesians is not chiastic overall, but has two major overlapping chiasms (1.1-4.1 and 3.20-6.22), plus eleven lesser chiasms (work of J. M. Gibbs).
      5. Colossians 1.12-3.17 is chiastic but not the letter overall (this probably indicates that the letter is composite). (work of J. M. Gibbs)
      6. 1 Thessalonians is parallel structure throughout (work of J. M. Gibbs)
      7. Hebrews is chiastic overall (work of Albert Vanhoye and also J. Bligh)
      8. 1 & 2 Timothy and Titus appear to be chiastic by subjects, but the subject divisions within the letters then appear to be parallel to one another (work of J. M. Gibbs)
      9. 1, 2 & 3 John, Jude and 2 Peter are also chiastic overall (work of J. M. Gibbs).

Conclusion: Chiastic structure overall is not an argument for Pauline authorship, but lack of it is probably an argument against it. This data thus argues against Ephesians, 1 & 2 Thessalonians, Colossians and the Pastorals as being by Paul.

            3.    Use of Scripture for Content and Structure:
      1. "Paul’s" letters appear to be generally ‘Jewish lectionary-located’:
      2. Romans appears to echo the whole lectionary year, basically in
                sequence (work of J. M. Gibbs)
        1 Corinthians is anchored to Passover and scriptures of Nisan
                (and perhaps up to Pentecost in Sivan) (work of J. M.
        2 Corinthians: at least 2 Cor 3 appears to have a Pentecost
                nexus (Gibbs)
        Galatians is anchored to Pentecost and scriptures of Sivan
        Philemon appears to echo an Abraham-Isaac-God (Gen 22) and
                Jacob-Benjamin-Joseph (Gen 42-43) setting, two
                passages read at same time in the Nisan and Tishri
                cycles. (Gibbs)

      3. Philippians 1.1-4.20 appears to echo Isaiah (34) 35-66 in
                straight sequence (work of D. P. Francis, supplemented
                by J. W. McMillan and J. M. Gibbs)
      4. Ephesians appears to have a Pentecost background (work of J.
                C. Kirby)

The very different way (a non-lectionary way) in which scripture is used in Philippians suggests it may not be by Paul.

4.    Theological models and vocabulary

There are substantial shifts from "Paul’s" use to that of some of the deutero-Paulines and the Pastorals with regard to the use of such terms as ‘circumcision’, ‘to circumcise’ (Paul: negative or neutral usage; positive in Phil 3.3; Col 2.11-12), and the Christian’s resurrection (Paul: future passive; Col 2.12; 3.1: raised with Christ in baptism, cp. Eph 2.6).


      1. Romans, 1 & 2 Corinthians, Galatians (and in all probability Philemon) are basically from Paul (with modifications/additions made in Romans and 2 Corinthians).
      2. Philippians, Colossians and Ephesians are from three separate writers of the Pauline school.
      3. 1 & 2 Thessalonians are by the same writer, but not by Paul.
      4. 1 & 2 Timothy and Titus are not by Paul, and Titus appears to stand apart from the other two Pastorals.
      5. Hebrews is in a class by itself, with Heb 13 apparently isolated from the rest of the letter.