In 1993, Dieter Lührmann and P. J. Parsons published P.Oxy. 4009, a double-sided Gospel fragment (2.9 × 9 cm).Footnote ¹ The small and round letters and many ligatures of the fragment reveal an informal copyist. It has been difficult to date the document precisely, but on the basis of similar manuscripts, Lührmann and Parsons have allocated this small papyrus to the second century.Footnote ² Paul Foster has criticized such an early date and instead dates P.Oxy. 4009 to the early third century.Footnote ³

1. The Side Reconstructed by Dieter Lührmann

In their publication, Lührmann and Parsons chose to call the side written on the horizontal fibers the recto—the side which was also easier to reconstruct. The side reconstructed by Lührmann includes 21 lines with a narrow margin of 0.5 cm on the right and 0.8 cm at the bottom. Lührmann's reconstruction is based on Matt 10.16b, P.Oxy. ii.19–23 = Gos. Thom. 39b (lines 5–7), and 2 Clem. 5.2–4 (lines 7–19). The letters on the fragment are in bold:Footnote ⁴

Matt 10.16b

Ἰδοὐ ἐγὼ ἀποστέλλω ὑμᾶς ὡς πρόβατα ἐν μέσῳ λύκων· γίνεσθε οὖν φρόνιμοι ὡς οἱ ὄφεις καὶ ἀκέραιοι ὡς αἱ περιστεραί.

Translation of Lührmann's reconstruction:

Even though some lines of P.Oxy. 4009 may be reconstructed differently, Lührmann's reconstruction is convincing.Footnote ⁵ Jesus answers the question proposed by a disciple with a saying introduced with the words λέγει μοι (line 11). On the basis of a strikingly similar dialogue quoted in 2 Clem. 5.2–4, Lührmann concludes that P.Oxy. 4009 has preserved a dialogue between Jesus and Peter. This, in turn, makes it possible to identify the fragment as part of the Gospel of Peter.Footnote ⁶ This assumption is supported by the vocative form of the nomen sacrum on the recto of the fragment (line 13), since this Christological title is characteristic of the Akhmîm fragment, which has been safely identified as a part of Peter's Gospel.Footnote ⁷ However, Lührmann warns us against assuming that the author of 2 Clem. used the Gospel of Peter (cf. the Gospel mentioned in 2 Clem. 8.5) as his source.Footnote ⁸

The proposal of Lührmann has not been received unopposed; Kraus and Nicklas as well as Foster conclude that P.Oxy. 4009 cannot be reliably considered to be part of the Gospel of Peter.Footnote ⁹ However, the case for identification is stronger than they assume because the side reconstructed by Lührmann consists clearly of sayings material tied to the plot of the Gospels. The other side of the fragment also indicates the vocabulary of the Gospels.

In spite of the verbal differences, the harvest theme, and the dove–serpent-saying in P.Oxy. 4009, the train of thought is similar in both P.Oxy. 4009 and 2 Clem. 5.2–4. It reads like a developed form of the synoptic saying in Matt 10.16 par. Luke 10.3. In both texts, the saying about lambs and wolves is extended with (Peter's) concerned question about the physical threat against the disciples. Jesus' answer is based on another synoptic saying which we know from a different context.Footnote ¹⁰ The clusters in both documents are dependent on the synoptic tradition.

The reorganized and edited synoptic material in P.Oxy 4009 and in 2 Clem. 5.2–4 is closely related to the themes of persecution and martyrdom. Even though the tradition of 2 Clem., particularly in its formulation of the last sayings of the cluster, is closer to Luke than Matthew (μετὰ τὸ ἀποθανεῖν, ἔχοντα ἐξουσίαν, βαλεῖν), it is reasonable to assume that the cluster rather follows the structure of the Matthean narrative. As with several other sayings in 2 Clem., this passage also is a harmonized version of Matthean and Lukan texts.Footnote ¹¹ However, the author has bypassed Jesus' speech on the eschatological mission of the disciples (Matt 10.17–27) and thus produced an immediate connection between the sending of the disciples and the saying about false and true fear (Matt 10.28 par. Luke 12.4–5).Footnote ¹² With this arrangement of the text, the author has also discarded the immediate expectation of the end proclaimed by the Matthean Jesus (10.23). Thus the focus of the text shifts from the Matthean eschatological plan to the situation in which the apostles are expected to suffer bravely at the hands of their killers and fear God and the punishment of hell more than the brutal power of their oppressors.

In a similar vein, the text reconstructed by Lührmann refers to a situation that was more or less current in the days of the author and not somewhere in a distant eschatological future. However, here the cluster of sayings is even more closely connected with the sending out of the disciples (harvest, serpents and doves). In lines 5–8, P.Oxy. 4009 is in one way or another dependent on Matt 10.16, but has the two sayings (serpents and doves, wolves) in reverse order. Furthermore, while the saying of Matt 10.16b, both in the manuscript tradition and in quotations from the church fathers, is consistently presented in the order serpents–doves, in P. Oxy. 4009 we have the opposite order. In addition, this saying is addressed to all the disciples (γίνεσθε) elsewhere, while in P.Oxy. 4009 only one disciple is addressed (γίνου).Footnote ¹³ However, according to both P.Oxy. 4009 and 2 Clem. 5.2, all the addressed disciples will be like sheep among the wolves. Thus, lines 5–8 offer a free reformulation of the Matthean text. The version of P.Oxy. 4009 is also more developed than its parallel in 2 Clem. 2.2–5. The saying especially addressed to Peter in lines 5–7 looks like a later addition to the similar oral tradition that both texts are quoting. It is not clear how the text of P.Oxy. 4009 continues; lines 20–21 do not allow a reconstruction of a saying about the true fear which follows at the end of 2 Clem. 5.4 (ἀλλὰ φοβεῖσθε κτλ.).Footnote ¹⁴

Even though Lührmann has made a good case for P.Oxy. 4009 being a fragment of the Gospel of Peter, it is indeed impossible to be absolutely sure ‘whether the author is reworking oral, non-canonical or canonical gospel traditions (or even a combination of these)’.Footnote ¹⁵ However, if the other side of the fragment reveals that P.Oxy. 4009 is part of an extended gospel narrative, it becomes difficult to avoid the conclusion that we have here an important late second- or early third-century witness to the Gospel of Peter.

2. Reconstruction of the Enigmatic Other Side

The recto of the fragment has been notoriously difficult to reconstruct. According to Lührmann, ‘a reconstruction of the verso is excluded’,Footnote ¹⁶ and Foster thinks that ‘no secure identification is possible’.Footnote ¹⁷ In their edition of the fragments related to the Gospel of Peter, Kraus and Nicklas have given up all attempts to make sense of the other side.Footnote ¹⁸

The longest visible lines of Lührmann's reconstruction include 8–10 letters. In each line, he has filled the lacuna with 9–10 letters—excepting line 11, where he has added 13 letters. Considering that the left margin of the unknown side is about 0.3 cm broader than the right margin of the side reconstructed by Lührmann, it is reasonable to assume that the missing parts of lines have basically included 11–12 letters. As line 10 on the unknown side reveals, the copyist may have started some lines one letter closer to the edge of the margin. On the other hand, the reconstructed side consists partly of quite dense writing, including 7 or even 8 letters in the same space in which the unknown side has only 6 letters (lines 4–7, 11–14). Therefore, it is not surprising that the lines to be reconstructed here are slightly shorter than the ones Lührmann has reconstructed.

The reconstructed side reveals an interesting feature of this manuscript: the copyist regularly seems to leave a blank space after an introduction to oratio recta (lines 9, 11 and 15). This detail has some significance for the reconstruction of the recto.

As mentioned above, the left margin of the unknown side (0.8 cm) is broader than the right margin of the other side. Correspondingly, even the broadest lines of the unknown side (5–13) include merely 6–7 letters. Unfortunately, there are no apparent catchwords that would create a common context for both sides. My transcription of the unknown side runs differently at some points from that of LührmannFootnote ¹⁹ and Kraus and Nicklas:Footnote ²⁰

These lines are expected to include 17–19 letters. This means that the space for reconstruction of missing letters in lines 5–13 hardly covers more than 12 letters. Due to the vocative κ(ύρι)ε, the letters αφεισ in line 13 must be read as an address; the verb form is the active indicative present 2^nd person singular, i.e., ἀφεῖς, .Footnote ²¹ In principle, it is possible that the Lord is addressed about something he allows or does not allow to be done.Footnote ²² However, further considerations make it clear that the question at stake here is the forgiveness of sins.

In the light of line 13, we can return to line 8 where the copyist has had some difficulties with the same verb. The small hook in the first letter reveals that it is not ο but α.Footnote ²³ The last letter of the line is unclear, but it has clearly been corrected to be ω, as the upward hook reveals. Having first written ΑΦΙΕ[ΝΤΑΙ, the form that is known from Jesus' declaration of forgiveness of sins to the lame man in Mark 2.5 and Matt 9.2, the copyist later wants to write ΑΦΕΩ[ΝΤΑΙFootnote ²⁴ instead and corrects his mistake. He adds Ε upon the line between Φ and Ι, draws the messy Ω on the still visible small Ε, but leaves the iota in his former writing untouched. The final result is the text as it stands.Footnote ²⁵ Be as it may, the unknown side of P.Oxy. 4009 does not tell the story of the lame man healed in Mark 2:1–10 and parallels; line 9 points in a different direction.

The first six letters of line 9 reveal the whole story: λαιαμα is nothing other than a part of the expression πολλαὶ ἁμαρτίαι. In the whole Gospel tradition, there is only one person with ‘many sins’—the sinful woman of Luke 7.36–50. On the unknown side of P.Oxy. 4009, we have some sentences of a variant of this story.

Two words, διὰ and ὅτι, reveal in lines 7–8 a similar introduction to the answer of Jesus as in Luke 7.47a. On the basis of the data gathered thus far, it is possible to reconstruct lines 7–9 of the fragment:

It is possible to add αἱ before πολλαί in line 8, but the most likely length of the lines speaks for the shorter form. More importantly, before πολλαὶ ἁμαρτίαι, the space that must be conjectured for each line in the fragment does not allow -νται αὐτῆς αἱ πολ- (14 letters)Footnote ²⁶ or even -νται αὐτῇ αἱ πολ- (13 letters) but compels the choice of -νται αὐτῇ πολ- (11 letters) instead. It must be noted that some manuscripts of the Western text have a wording notably closer to that of the most likely reconstruction. The Codex Bezae has ἀφέωνται αὐτῇ πολλά (thus also ff² and l). Most old Latin manuscripts read remissa sunt illi peccata multa, and the Sinaitic Syriac translation offers a corresponding text (‘her many sins are forgiven her’).Footnote ²⁷ Different versions of the Diatessaron favor similar readings.Footnote ²⁸ Thus, the dative αὐτῇ is well attested in the Western text.

In spite of the scanty data in lines 2–7, it is possible to reconstruct the beginning of the fragment with the help of the corresponding verses in Luke 7.45–46. The sentences are quite differently constructed, but the words σὐ δέ in line 4 make it clear that Jesus addresses his host, most likely a Pharisee as in Luke, about his actions in lines 4–7, while in the preceding lines, he has described the actions of the sinful woman. The affinities between the Lukan text and the version of the fragment are marked in bold:

These affinities indicate that the washing of feet described by Luke in 7.44b does not come up in the first lines of the fragment. A reconstruction of lines 2–4 is possible because the actions of the sinful woman are known to us from Luke (vv. 45b and 46b). The anointing performed by the sinful woman is described in lines 1–2. The letters ψε (and some obscure remains of ν) in line 3 point to the word ἤλειψεν of the Lukan story. The horizontal line of σ indicates that a τ or π follows; therefore it is possible, following the text of Luke, to add π here. The following reconstruction of lines 1–3 recommends itself:

Some notes are necessary.

Line 1–2: In the Codex Bezae and W 079 sy, the Lukan v. 46b is preserved in a shorter form αὕτη δὲ μύρῳ ἤλειψεν, omitting τοὐς πόδας μου. It is impossible to say whether P.Oxy. 4009 supports the short (Western) form of the text, even though the formulation of the sentence in lines 1–3 hints at the possibility that the woman anointed Jesus and kissed his feet. Some scholars have indeed argued that the shorter text is original.Footnote ²⁹ Be that as it may, the logic of the anointing has caused problems for the copyists of the Lukan story. Some later Latin manuscripts have—for symmetry's sake—made Jesus say that the host did not anoint his feet.Footnote ³⁰

Line 2: It is possible to read either διέλειπεν or διέλιπεν.Footnote ³¹ The former reading fits better the assumed length of the lines.

Line 3: In Luke 7.45b, the old Latin manuscript e offers a striking parallel to the wording in the fragment: non intermisit pedes meos osculando.

Lines 4–7 form the next unit. In line 5, the verb form παρέσχες (active indicative aorist 2^nd person singular) indicates that, in line 4, Jesus mentions something that his host did not offer him or provide for him. The context makes it clear that Jesus speaks about oil. It is interesting to see that the author of this text is not satisfied with the Lukan formulation of Jesus' words. The host is not expected to have anointed Jesus' head, but he should have provided oil for him that he could do it himself. The letters θοντιμ in line 6 make it clear that lines 5–6 included the words οὐδὲ εἰσελ]θόντι μ[οι, using the verb εἰσέρχομαι that is used in Luke 7.44–45. In the beginning of line 7 the letters κασ indicate the verb form—again in active indicative aorist 2^nd person singular—ἐδώκας. The Lukan parallel and the spacing of the fragmentary lines reveal that these words are about the kiss (φίλημα) that Jesus did not receive from his host. These considerations lead to the following reconstruction of lines 4–7:

Lines 9–11 can also be read in the light of the Lukan parallel, which introduces here the reaction of ‘those who were at the table’ (οἱ συνανακείμενοι). The letters μενων in line 11 indicate the expression ἐκ τῶν συνανακείμενων which fits perfectly both the length of line 10 and the characterization of the guests in Luke 7.49a. The word αὐτῷ in the beginning of line 10 indicates that line 9 ended with an expression like εἶπον δὲ, introducing the reaction of the συνανακείμενοι which begins in line 11 and ends with the words ἀφεῖς, in line 13:

This reconstruction is not secure, but it has very few alternatives. An expression like εἶπον δὲ αὐτῷ ἐκ τῶν συνανακείμενων is rare, but attested in the Christian corpus of texts.Footnote ³² The introduction to the reaction heard from among the guests can hardly be more extensive because such an assumption would make it all too difficult to reconstruct lines 11–13.

The vocative κ(ύρι)ε implies that lines 11–13 include a question the guests pose to the Lord. The letters νοματ in line 12 indicate the words ἐν τῷ ὀνόματί σου since the expression ἐν τῷ ὀνόματί τοῦ θ(εο)ῦ would be too long and would make less sense here. Furthermore, the only reasonable object for ἄφεῖς is ἁμαρτίας. Since the presence of precisely these three elements—question form, the expression ἐν τῷ] ὀνόματ[ί, forgiveness (of sins)—makes it feasible to conclude that the question posed by the guests concerns Jesus' authority to forgive sins in his own name, a plausible reconstruction of the question in lines 11–13, filling precisely the space available in lines 11–12 and parallel to Luke 7.49b, runs like this:

The words διὰ τί may be replaced with πῶς, which does not as plausibly fill the space, or πῶς σύ, which is clumsier than διὰ τί. On the other hand, a question starting directly with ἐν τῷ ὀνόματί is not plausible since that would make it necessary to reconstruct a longer and more artificial introduction.Footnote ³³ The reconstructed part of line 11 διὰ τί ἐν τῷ ὀ- is shorter than other lines reconstructed here. However, as mentioned above, in lines 9, 11 and 15 of the side reconstructed by Lührmann, there are corresponding gaps between the introduction and oratio recta.

The last lines of the fragment are more than difficult to reconstruct. We cannot possibly know whether the συνανακείμενοι continue their attack or whether we should assume that the introduction to Jesus' answer begins immediately after . On the other hand, the remains of lines 15 and 16 notably limit any attempts to make sense of what follows. The letters ουθ may indicate a number of things; in any case, the horizontal line of ο ties it to the preceding letter which we do not know.Footnote ³⁴ In line 16, Kraus and Nicklas suggest that we should read here μμαι but that makes a plausible reconstruction almost impossible. It is most reasonable to assume that the letter before μαι is α; at least the letters αμα in line 9 seem to support this suggestion. If the letters in line 15 indicate the word δύναμαι, it is necessary to suppose that Jesus' answer has already begun in the preceding line. However, this decision would considerably limit the possibilities for a plausible reconstruction of lines 14 and 15. Thus, the reconstruction extends from line 1 only to the beginning of line 13:

Translation:

It is necessary to note that some uncertainties remain as regards the precise wording of the lost letters in each line. In line 9, εἶπον δὲ may have alternatives. In line 11, the words διὰ τί—or διὰ τί σύ—could be replaced with another, corresponding expression like πῶς or πῶς σύ; a reconstruction like μενων αὐτῷ· ἐν τῷ ὀ- is possible, but unlikely.