Tyndale Bulletin 46.1 (1995) 43-54.
EXPLORING THE COMMON IDENTIFICATION OF THREE NEW TESTAMENT MANUSCRIPTS: p4, P64 AND P67
Philip W. Comfort
This article explores the common identity of three very early Gospel manuscripts. Some scholars have believed that p4,, p64 and p67 all came from the same codex; others have doubted. The newly proposed dating of p64 to the late first century makes this exploration all the more vital. This article examines the provenance and paleography of all three papyri in an attempt to demonstrate a common scribe. Then the article presents an argument for dating p4 to the second century.
Carsten Thiede has recently published a redating of the manuscript known as p64, which has three fragments of Matthew’s Gospel. Formerly dated ca. 200, Thiede has now dated this manuscript to the first century.1 This dating, if accurate, is extremely significant because it places a manuscript of the Gospel of Matthew within the same century it was written.
The manuscript p64 (at Magdalene College Library. Oxford: Gr. 17) has been identified as belonging to the same Matthean codex as p 67 (at Fundación San Lucas Evangelista, Barcelona: inv. no. 1). p 64 contains Matthew 26:7-8, 10, 14-15, 22-23, 31-33, and p67 preserves portions of 3:9, 15; 5:20-22, 25-28. p 64 was first published in 1953 by
Colin Roberts in an article entitled, ‘An Early Papyrus of the First Gospel’.2 p 67 was first published by P. Roca-Puig in 1957 in a booklet called Un Papiro Griego del Evangelio de San Mateo (Barcelona). After Roberts realised that p64 and p67 were two parts of the same manuscript and then confirmed this with Roca-Puig, the latter published another article in 1961 in which he gives a full presentation of the entire manuscript.3 Colin Roberts appended a note to this article explaining how he had discovered that p64 and p67 were part of the same manuscript.
Textual scholars generally acknowledge that p 64 and p 67 come from the same manuscript. For example, the two are listed together in Nestle-Aland’s Novum Testamentum Graece, the Alands’ The Text of the New Testament, and Metzger’s The Text of the New Testament. But few have recognised that Roberts also identified the Lukan manuscript p4 (in Paris at the Bibliothèque Nationale: Gr. 1120, suppl. 2: Luke 1:58-59: 1:62-2:1, 6-7; 3:8-4:2, 29-32, 34-35; 5:3-8; 5:30-6:16) as belonging to the same codex as p64 and p67.4 This identification, if accurate, is significant because it would mean that there is another manuscript that would have to be dated the same as p64/ p67.
II. The Identification of p 4 with p 64/p 67
In a 1965 article about new papyrus manuscripts of the New Testament, Kurt Aland presented the position that p4 probably belonged to the same codex as p64 and p67. The only hesitancy Aland had in affirming a complete identification is that the colour of the papyrus p64 was much lighter than that of p4. Otherwise, with respect to all other paleographic features, Aland noted that the manuscript p64/p 67 bears remarkable similarity to p4.5 Following Aland’s lead, the papyrologist J. van Haelst also identified P4 as probably belonging to the same manuscript as p 64 and p 67.6
The renowned papyrologist, Colin Roberts, who is best known for his dating of the Johannine manuscript p52 to the early second century, was both the editor of p64 and the scholar who identified p 67 as belonging to the same manuscript. He was convinced that p 4 also came from the same codex. Speaking of p 4, p 64 and p 67 he wrote:
There can in my mind be no doubt that all these fragments came from the same codex which was reused as packing for the binding of the late third century codex of Philo (= H. 695). An apparent discrepancy was that Ιησους appeared as Ic in the Paris fragments and as IH in the Oxford fragments; the correct reading in the latter, however, is Ic, as can be checked in the photograph.7
The above statement was made in The Schweich Lectures of the British Academy in 1977: ten years later in his publication, The Birth of the Codex, Roberts still affirms that p 4 and p 64 and p 67 are parts of the same Gospel codex.8 To my knowledge, he never changed his opinion. The only one that I know to have changed his mind on this matter is Kurt Aland. In 1963, he listed p 4 as separate from p 64/ p 67, but in 1965 he suggested that P4 belonged to the same codex as p 64 and p 67. But thereafter, he never refers to them as belonging to the same codex; the two (p4 and p64/p67) are always listed separately in Aland’s publications.9 But I have not found a reason for the change.
Thus, the common identity of the three papyri needs to be re-examined and reaffirmed, which I intend to do by exploring the provenance of the manuscripts and their paleographic features.
a) Provenance of p 4
A codex manuscript, containing two treatises by Philo10 and portions of Luke used for the binding, was discovered in Coptos (modern name, Qift), Egypt, on the east bank of the Nile, by Fr. Vincent Schell during his expedition to Upper Egypt in 1880. Jean Merrell described the circumstances of this find:
Schiel told me last June that, in 1891, having purchased in Luxor a codex including two treatises of Philo of Alexandria, he was fortunate to find the fragments of our biblical papyrus.
The papyrus was found at Coptos (Upper Egypt) in 1880. Since it was obviously considered at the time to be something very valuable, it was enclosed and concealed in a niche. (The hollow sound of the thick high wall at this point was noteworthy). In opening this area, one found in this secret place the two treatises of Philo of Alexandria. The entire document, in a well-known format, almost square, in 8” arabic books, was bound together in a leather cover, with a small tongue and cord, also in leather, wrapped around the cover. In the hiding place, the book must have been compressed in the space, the mortar was encrusted on the outside; the pages were tightly pressed together in a mass and, in addition, they were also fastened to each other by a quantity of small grains of sand, produced by an ancient condensation occurring in the vegetal tissue. After the forty-fourth sheet, in the form of a wad, I believe, and in order to fill the space provided by the cover, there were several
fragments of sheets stuck together, one of them containing the κατα μαθθαιον and the others having the fragments of St. Luke.11
According to Roberts, p4 ‘was used as stuffing for the binding of a codex of Philo, written in the later third century, and found in a jar which had been walled up in a house at Coptos’. Very likely, the owner of this manuscript ‘concealed it with the intention of removing it from its hiding place when danger had passed, either when Coptos was besieged and sacked by Diocletian in A.D. 292 or later  in his reign during the last and severest of the persecutions’.12
b) Provenance of p 64
Significantly, the manuscript p64 was purchased in the same city that p 4 was purchased—Luxor, Egypt. Roberts said, ‘the fragments were purchased by the Reverend Charles B. Huleatt in Luxor in 1901 and presented by him to his old College through the then Librarian, the Reverend H.A. Wilson’.13 Evidently, the Matthew fragments had been taken to a dealer in Luxor some time after the Luke fragments. Of course, there was a ten-year time span, so it does not mean they came from the same find. But it does suggest that p 4 and p 64 came from the same geographical area in Egypt.
c) Provenance of p 67
Though the editor of this manuscript, Roca-Puig, did not indicate its provenance, the key to determining it is probably found in Roberts’ words about other Matthean fragments, owned by Charles Huleatt, which were never given to the Magdalen Library. In his article in p64, Roberts said:
It is probable that there were further fragments of the same leaf since a letter from Mr Huleatt to the Librarian refers to purchases of fragments of the same manuscript in successive years, but nothing
beyond what is published [here] is now extant in the Library.14
Thus, it is possible that some of these fragments found their way to the Fundación San Lucas Evangelista in Barcelona, Spain. But we are not certain.
For the paleographic study of each of the three manuscripts, I have used the written analysis of each of the papyrologists who published the edito principes of each of the three manuscripts. This information, together with my own study of the paleographic features of the three manuscripts (I personally viewed p4) has yielded the following comparison of the similarities between p 4 and p 64 and p 67.
a) Area of Writing
p 4: 10 x 13.3 cm; p 64: 10.5 x 16.8 cm; p67: 10 x 15 cm.
The difference between p 4 and p64/ p67 can easily be explained: the scribe of p4 allowed more space between the two columns on the page—at least 3 cm. The area of writing for p 64 and p 67 was estimated differently by Roberts and Roca-Puig (respectively), but the difference is minimal. Turner estimated the area of writing to be 10.5 x 16 cm for both p 64 and p 67.
b) Page Size
p 4: 17 x 13.5; p 64: 17-18 x 12-13; p 67: 18-20 x 12-13 cm.
The page size is not difficult to determine for p4 because nearly complete leafs are extant. But the page size has to be reconstructed for p 64 and p 67 based on column widths and number of lines. Turner’s reconstruction of both p 64 and p 67 is 17-18 x 12-13 cm. Roca-Puig estimated that p 67 would have been 18-20 x 12-13 cm. Roca-Puig’s length is slightly greater because he estimated two more lines per page than is found in p 4 and p 64. But the long and the short of
it is, all three nearly have the same page dimensions: approximately 17-18 x 13 cm.
c) Double Columns on Each Page
All three manuscripts: p 4, p 64, p 67.
The double-column format is immediately evident in p4, but not in p64 and p67. The papyrologists who have analysed these two manuscripts have concluded that there is no other way for the extant text to have fitted on a codex page if it were not written in the double column format. This is an unusual feature, found only in these three early New Testament papyri. A few other New Testament papyri (p 34 and p 41) have double columns, but these are later (seventh and eighth centuries).
d) Lines per Column
p 4: 36 lines per column; p 64: 35-36 lines per column; p 67: 36-38 lines per column.
These line lengths were estimated by counting how much extant text and missing text it takes to fill out one column on the page. In the Lukan p4, this is not difficult to do because there is so much extant text. In p 64 and p 67, it is more difficult but still possible, because there is enough extant text to figure line lengths.
e) Letters per Line
p 4: ranges from 12-19, 15-17 on average; p 64: 15-17 letters per line; p67: ranges from 13-20, 15-17 on average.
p4: high-point (frequent), mid-point, base point colon (:) for new section (Luke 3:14; 6:8); p 64: one high point; p 67: several colons (:) as a kind of versification.
The text of p 4 is divided into sections according to a system also found in p75, which also recurs in some great fourth-century manuscripts (i.e. a and B). Furthermore, this manuscript exhibits three different positions for punctuation, as well as omission and quotation signs (in a system similar to that found in p 66). p 64 and p 67 also display punctuation marks which hint at a sophisticated system. Had more text
been extant, we would probably see punctuation similar to that in P4.
g) Paragraphing (marked as outdentation with horizontal bar)
p 4: at Luke 1:76; 1:80; 2:1; 3:19; 3:23; 5:36; 6:12 (most correspond with the beginning of a new paragraph); p 64: at the beginning of Matthew 5:27 (corresponding with the beginning of a new paragraph); p67: at the beginning of Matthew 26:31 (corresponding with the beginning of a new paragraph).
A significant difference between p 4 and p 64/ p 67 is that p4 is outdented two letters into the left margin, whereas p64/p67 are outdented one letter.
h) Nomina Sacra
p 4: θc = θεος, Kc = λιροπς, Ic = Ιησους, Χc = Χριστος, ΠΝΑ = πνευμα; p 64: KE = κυριε, Ic = Ιησους; p 67: none extant.
The letters (or uncials) in the three manuscripts are remarkably similar. The following consonants are shaped indentically: Β, Δ, Γ, Η, Θ, Κ, Λ, Ν, ξ, Π, ρ, χ. The following vowels are shaped identically: ι, ο, ω. A close examination of these letters reveals that each was stroked the same. This is especially noticeable in the letters Κ, Ν, Ξ, Π, ρ, ω (the μ is especially similar).
One remaining consonant, the sigma (ς), is quite similar but not always identical in the three manuscripts.15 The lower curve on the sigma in p 64 and p 67 doesn’t come around as far as does the sigma in P4. However, p 4 does have several sigmas that are shaped just the same; so I would expect that, had more text been extant in p 64 and p 67, we would also see some fully founded sigmas. The same is true for the vowel, epsilon (e). In all three manuscripts, the arch of the epsilon shows up as being fully curved on the underside in certain ligatures and not fully curved in others. However, the epsilon in p 64 is often more squared than the epsilon appearing in p 4 and p 67. A similar phenomenon is true for the alpha (a), which is both pointed and somewhat rounded (at the left extension) in all three manuscripts.
Finally, it should be noted that the tau (τ) always has a straight top bar in p 4, but not always so in p 67.
It would be easy to deduce that the several common features of these three papyri signify that the three came from the same codex. p 4 and p64 share a known place of purchase: Luxor, Egypt (which is quite near Coptos, the place where p4 was discovered). All three manuscripts have remarkably similar page dimensions, with double columns on each page, with about 36 lines per column and 15-17 letters per line. All three have similar punctuation, and all three display similar penmanship. But there are some marked differences. p 4 displays finer, thinner pen strokes, whereas p 64/p 67 exhibits bolder pen strokes. Though the difference could have been in the stylus, not with the scribe, it is noticeably distinct. Another marked distinction is the paragraphing outdentation noted above. Added to these particular distinctions is the fact that the colour and papyrological fibre of p 4 is quite distinct from p 64/p 67. Thus, I cannot confidently make an absolute identification of the three manuscripts as having belonged to the same codex.16 What I can suggest is that the same scribe produced all three of these manuscripts, perhaps copying Matthew’s Gospel some time prior to Luke’s Gospel—using a different stylus (a blunter one for Matthew than for Luke). Or it is possible that the scribe refined his style with time, such that what we see in Luke is a slightly more developed form of the biblical uncial hand (compared to a more primitive form of it in the Matthean fragments). In any event, this leads us to a reconsideration of the dating of p 4.
V. Dating p 4
The first editor to publish a partial edito principes of this manuscript was Vincent Schiel, in 1892. Unfortunately and mistakenly, he assigned a very late date to it: sixth century!17 The error of this dating is due to the inadequate knowledge that scholars in the nineteenth century had about papyrus manuscripts. The second editor to publish a full transcription of the text was J. Merell (1938), who redated this manuscript to the early fourth century.18 In 1963 Aland redated p 4 to the third century.19 And in 1979, Roberts classified p 4 as a late second-century manuscript, belonging to the same codex as p 64 and p67.20
We know the terminus date for P4 because we know about its provenance. It had been used for padding in a codex of Philo’s treatises, which was hidden in a house in Coptos to avoid being confiscated during the persecution of A.D. 292 or 303, when Coptos was besieged and sacked by Diocletian. The Philo Codex was dated to the third century by both Hunt and Kenyon.21 The owner of the Philo codex and the Gospel codex was probably a Christian22 and therefore would have valued the Gospels. He would not have used a newly-copied Gospel for stuffing the padding of Philo’s treatises:23 rather, this Gospel codex must have been well-used and well-worn. In fact, it must have been a discarded copy replaced by another codex. Thus, it is not unlikely that p 4 was made at least as early as a hundred years prior to the Philo Codex, if not earlier. As such, we are fairly certain of a second-century date.
This does not preclude an even earlier date because the codex may have been in use more than a hundred years before it was discarded. We know of several papyrus manuscripts that saw this kind of extended use. For example, p 46 was used from the early second century to the early fourth century, where it was buried with a Greek-reading Coptic monk.24 The manuscript is filled with the markings of various correctors and readers—at least sixteen of them, according to Kim.25 None of these correctors did any kind of thorough-going work, so as to be called a diorthotes; each, here and there, made some adjustments or marked a few books (Romans and Hebrews) for oral reading. This shows that the manuscript had been very well used. Thus, it is not unusual for a manuscript to be used for two hundred years.
In any event, it seems fairly certain that P4 belongs to the second century. The style of handwriting is virtually the same as that found in P. Oxyrhynchus 661, which is dated to the last part of the second century. My comparison of the two manuscripts affirms the remarkable similarity of handwriting.26 Of course, p 4 could be even earlier in the second century, but it all depends on when we think the earliest form of the hand known as the ‘biblical uncial’ began. The majority of paleographers date its beginning to the middle of the second century. But such dating is based on the earliest examples; thus, if earlier examples were found, the date would be pushed back. However, it is difficult to get earlier dates assigned to anything that looks like biblical uncial. Nonetheless, some manuscripts have been receiving earlier dates. For example, the Pauline codex p 46 has been redated by Kim to ca. A.D. 85. To this day, Kim’s early dating of P46 to the later part of the first century27 has not been challenged on paleographic grounds. And the Johannine codex p 66 has been dated ca.
125 by the papyrologist Herbert Hunger.28 Other manuscripts have been pushed back from the third century to the second century—namely, p 32 (ca. 175), p 45 (ca. 150), p 77 (ca. 150), p 87 (ca. 125), p 90 (ca. 150).29 Thus, if p 64 can be dated to the late first century or early second century, it should readily follow that p 4 is not far behind.30