Corpus-Paul

["David C. Hindley" <DHindley AT compuserve.com>]

On 05/03/99, ""Stephen C. Carlson" <scarlson AT mindspring.com>" wrote:

>> I am interested in this topic because a few
years ago, I attempted to classify the NT documents on the basis of
computer-analyzable stylistic criteria, in my case, the number of
shared three-word sequences between two documents.<<

>> Basically, my conclusions cohere with what are presented to be
Neumann's. I found that Eph, Col, and 2 Thess were more similar
to Paul than to any other writer in the NT corpus and that there
was little stylistic basis according to my criterion to exclude
them from the other, non-disputed Paulines (e.g. Romans). I also
found that the Pastorals and Hebrews were quite different from
the rest of the Pauline corpus.<<

Steven,

There must be caution, I think, in applying only one statistical measure to 
the determination of authorship, as authors will employ combinations of 
characteristics that individually may not be meaningful. That is why I was 
happy to see multileveled multivariate analysis performed by Neumann.

However that is, I had input the statistical data on sentence length found 
in the appendix to Morton & McLeman's _Paul, The Man and the Myth_ (1966), 
and then calculated the coefficient of correlation between books. Sentence 
length was measured in "spons", which were defined as "THE NUMBER OF GREEK 
WORDS BETWEEN FULL STOPS (.), COLONS (RAISED PERIOD), OR INTERROGATION 
SYMBOLS (;) AND CAN BE SEEN AS A TYPE OF SENTENCE." (Sorry about the caps, 
as I cut and pasted from a spreadsheet document and was not in a position 
to reformat the text). The results are as follows (and I apologize in 
advance for the length of the results):

1 CLEM  2 CLEM  0.988
1 CLEM  EXHORT  0.938
1 CLEM  RICHMAN 0.916
1 COR   JAMES   0.987
1 COR   GAL     0.975
1 COR   2 COR   0.968
1 COR   1 CLEM  0.964
1 COR   2 CLEM  0.930
1 COR   1 TIM   0.914
1 COR   HEB     0.906
1 COR   PHIL    0.892
1 COR   1 THE   0.887
1 COR   RICHMAN 0.874
1 COR   PHM     0.873
1 COR   2 TIM   0.864
1 COR   EXHORT  0.847
1 COR   COL     0.833
1 COR   2 THE   0.796
1 COR   TIT     0.789
1 COR   EPH     0.685
1 THE   HEB     0.968
1 THE   1 TIM   0.947
1 THE   2 CLEM  0.945
1 THE   EXHORT  0.935
1 THE   PHM     0.923
1 THE   1 CLEM  0.918
1 THE   2 THE   0.909
1 THE   JAMES   0.887
1 THE   RICHMAN 0.845
1 THE   2 TIM   0.832
1 THE   TIT     0.800
1 TIM   2 CLEM  0.980
1 TIM   1 CLEM  0.970
1 TIM   EXHORT  0.968
1 TIM   HEB     0.964
1 TIM   JAMES   0.950
1 TIM   2 TIM   0.940
1 TIM   TIT     0.928
1 TIM   RICHMAN 0.903
1 TIM   PHM     0.889
2 CLEM  EXHORT  0.964
2 CLEM  RICHMAN 0.896
2 COR   JAMES   0.978
2 COR   1 CLEM  0.978
2 COR   GAL     0.969
2 COR   2 CLEM  0.967
2 COR   1 TIM   0.963
2 COR   EXHORT  0.935
2 COR   RICHMAN 0.934
2 COR   HEB     0.932
2 COR   1 THE   0.926
2 COR   PHIL    0.915
2 COR   COL     0.910
2 COR   2 TIM   0.906
2 COR   PHM     0.897
2 COR   2 THE   0.883
2 COR   TIT     0.879
2 COR   EPH     0.784
2 THE   EXHORT  0.951
2 THE   1 TIM   0.916
2 THE   HEB     0.915
2 THE   2 CLEM  0.902
2 THE   PHM     0.891
2 THE   1 CLEM  0.876
2 THE   TIT     0.856
2 THE   RICHMAN 0.844
2 THE   2 TIM   0.844
2 THE   JAMES   0.824
2 TIM   TIT     0.963
2 TIM   1 CLEM  0.943
2 TIM   2 CLEM  0.938
2 TIM   JAMES   0.925
2 TIM   RICHMAN 0.919
2 TIM   EXHORT  0.916
2 TIM   HEB     0.886
2 TIM   PHM     0.794
COL     EXHORT  0.980
COL     HEB     0.963
COL     1 THE   0.959
COL     2 CLEM  0.956
COL     2 THE   0.952
COL     1 TIM   0.948
COL     1 CLEM  0.919
COL     2 TIM   0.893
COL     PHM     0.886
COL     TIT     0.878
COL     RICHMAN 0.874
COL     JAMES   0.860
EPH     EXHORT  0.932
EPH     COL     0.921
EPH     PHIL    0.899
EPH     TIT     0.896
EPH     2 THE   0.890
EPH     1 TIM   0.884
EPH     2 CLEM  0.883
EPH     2 TIM   0.879
EPH     HEB     0.876
EPH     1 CLEM  0.843
EPH     1 THE   0.827
EPH     RICHMAN 0.792
EPH     JAMES   0.749
EPH     PHM     0.692
GAL     1 CLEM  0.986
GAL     2 CLEM  0.977
GAL     JAMES   0.973
GAL     HEB     0.970
GAL     PHIL    0.961
GAL     1 TIM   0.957
GAL     1 THE   0.931
GAL     EXHORT  0.913
GAL     COL     0.905
GAL     2 TIM   0.894
GAL     PHM     0.884
GAL     2 THE   0.870
GAL     RICHMAN 0.858
GAL     TIT     0.836
GAL     EPH     0.799
HEB     2 CLEM  0.979
HEB     1 CLEM  0.957
HEB     EXHORT  0.954
HEB     JAMES   0.914
HEB     RICHMAN 0.831
JAMES   1 CLEM  0.981
JAMES   2 CLEM  0.952
JAMES   RICHMAN 0.910
JAMES   EXHORT  0.885
PHIL    HEB     0.991
PHIL    2 CLEM  0.972
PHIL    1 TIM   0.966
PHIL    1 CLEM  0.955
PHIL    COL     0.952
PHIL    EXHORT  0.949
PHIL    1 THE   0.942
PHIL    2 TIM   0.917
PHIL    JAMES   0.914
PHIL    2 THE   0.913
PHIL    PHM     0.885
PHIL    TIT     0.873
PHIL    RICHMAN 0.828
PHM     HEB     0.902
PHM     JAMES   0.870
PHM     2 CLEM  0.865
PHM     EXHORT  0.863
PHM     1 CLEM  0.853
PHM     RICHMAN 0.806
RICHMAN EXHORT  0.909
ROM     1 COR   0.996
ROM     JAMES   0.986
ROM     2 COR   0.981
ROM     GAL     0.974
ROM     1 CLEM  0.970
ROM     2 CLEM  0.938
ROM     1 TIM   0.924
ROM     HEB     0.909
ROM     1 THE   0.900
ROM     RICHMAN 0.899
ROM     PHIL    0.890
ROM     PHM     0.871
ROM     EXHORT  0.869
ROM     2 TIM   0.866
ROM     COL     0.851
ROM     2 THE   0.816
ROM     TIT     0.806
ROM     EPH     0.705
TIT     EXHORT  0.929
TIT     RICHMAN 0.919
TIT     2 CLEM  0.908
TIT     1 CLEM  0.900
TIT     JAMES   0.867
TIT     HEB     0.844
TIT     PHM     0.746

I never got around to determining the level of significance, but as you can 
see, the relationship between the sentence length distributions of these 
letters (1.00 being complete correlation) produced some interesting 
results. Hopefully they will illustrate the pitfall of using too few 
measurements.

Dave Hindley
Cleveland, OH
Dhindley AT compuserve.com