Matthew's Gospel 

among the Egyptians: 

An Introduction to 

an Early Manuscript

This paper was presented in the earliest months  of the author's initial research into Codex Schoyen (Feb-Apr 2007), and poses a number of questions which are now not as perplexing as intimated herein.  This webpage will be updated in the coming months to reflect the author's current research.

James M. Leonard

16 May 2007 

After senior text critic Eldon J. Epp’s painful assessment that his discipline has been in an unproductive interlude for the last half of the century, a tempest of sorts has been stirred up in recent years within its own ranks, especially in regard to the gospels.  The issue is over how to reconstruct the text prior to 180 C.E. when earlier manuscripts are lacking, and how stable the tradition was prior to that period.[2]  Surveying a wide array of texts from the period, 2nd century text specialist William L. Petersen controversially concluded,


To be brutally frank, we know next to nothing about the shape of the “autograph” gospels; indeed it is questionable if one can even speak of such a thing.  This leads to the inescapable conclusion that the text in our critical editions today is actually a text which dates from no earlier that [sic] about 180 CE at the earliest.  Our critical editions do not present us with the text that was current in 150, 120 or 100—much less in 80 CE (“The Genesis of the Gospel,” p. 62).


Thus, in a reversal of textual optimism, a number of recent works have emphasized a radical discontinuity between our modern, printed editions and whatever text might be reconstructed with data from early patristic citations, non-canonical gospels, traditions preserved in the versions and gospel harmonies, with an assumption that the primitive text evolved during this period into their present shape and larger size.  Within this stormy context, a certain Coptic manuscript now known as Codex Schøyen has come to light. 




Emergence of Codex Schøyen


The manuscript was purchased in 1999, with few details divulged.[3]  Papyrological expert Hans-Martin Schenke, who in 1981 published the fifth century Codex Scheide of Matthew in the same dialect, analyzed the present manuscript and published its editio princeps.  Schenke noted that the Schøyen experts analyzing the text found it to be a substantial papyrus codex of Matthew’s Gospel, written by the first half of the fourth century, in a hitherto unknown variety of the Middle Egyptian dialect.


If the early fourth century date can be substantiated, the emergence of Codex Schøyen would have automatic significance for textual criticism.  In fact, Codex Schøyen may well be the oldest witness to eleven chapters of Matthew, and a many number of verses elsewhere, too.[4]   Such an early date for a manuscript with 39 extant pages naturally prompts hope that further progress in understanding the second century text can be achieved.



Claims of Uniqueness


What is so controversial about this text, however, is not its dating, which has otherwise gone unchallenged, but its unusual readings.  When it is carefully compared with the mainstream of the tradition, one realizes that its text is different from any other known Matthew text.  Tjitze Baarda, despite his considerable criticism against Schenke’s more radical conclusions, concedes nevertheless that, “The Schøjen [sic] codex presents us with a most intriguing version of Matthew,” one which has an “enigmatic text” (review article, Novum Testamentum 46, p. 306).  Regardless of final conclusions, then, Codex Schøyen is unique among the manuscripts.


This enigmatic quality begs the question as to whether Codex Schøyen might be a fourth century vestige of 2nd century textual instability, or if it can be explained in terms of 3rd or 4th century dynamics or other phenomena. Consequently, the importance of Codex Schøyen is increased at least a little more than that afforded automatically to it by virtue of its antiquity.[5] 


Before moving on to a brief survey of the sorts of textual phenomena found in Codex Schøyen’s Sermon on the Mount, the place of Codex Schøyen within the larger Coptic versional tradition should be broached, along with a brief discussion of Coptic peculiarities relevant for textual criticism.



Codex Schøyen among the Coptic Versions


With the appearance of Codex Schøyen, certain aspects of the history of Coptic versions and early < namespace="" prefix="st1" ns="urn:schemas-microsoft-com:office:smarttags" xml="true">Egypt should be revisited.  As late as 1995, Frederik Wisse wrote that “some scholars believe that there is indirect evidence for the existence of Coptic translations before the fourth century” (134), but this assertion was made with great tentativity. With the emergence of Codex Schøyen, much has changed.  We now have two substantial manuscripts of Matthew written in Middle Egyptian, one presumably from the first half of the fourth century, and this despite the fact that Middle Egyptian has hitherto been perceived as merely a minor dialect in a provincial area of Egypt, well away from its urban centre.  This being the case, perhaps scholars now can be more assertive that portions of the New Testament were being translated into Coptic by the middle of the third century, and perhaps much earlier.[6]  In fact, the considerable physical dimensions of Codex Schøyen[7] and the extra labour invested in the multi-layering of its raw material, along with the relative quality of its script might even suggest that the codex was produced out of institutional need rather than personal use.  All of this counters prevailing views of early Coptic history, and so the appearance of Codex Schøyen necessitates a re-evaluation of these issues.[8]


The Coptic version has its own set of peculiarities which must be properly considered for its effective usage in textual criticism, and aspects of Codex Schøyen need to be factored into this process.  A version’s relative value in textual criticism is bound proportionately to its relative translational latitude:  the stricter the better, although the earliest versions had no such deliberate tendency.  Thus, before concluding that a version might support one Greek reading or another, one must ask whether the versional text must necessarily be a strict reflection of its Vorlage, or whether it might possibly reflect translational latitude.  One such example of the latter possibility comes from Codex Schøyen’s striking minimization of the Greek conjunctions kaiv, dev, and ijdouv.  Must this unusual data be interpreted, with Schenke, to reflect a different Greek Vorlage than that represented in the Standard Text, or could the translator have simply thought them superfluous in his language?[9]  Perhaps Codex Schøyen’s avoidance of certain conjugations would be better used to suggest something of the translator’s tendency away from literalism, and hence confirm the antiquity of the text of Codex Schøyen.[10] 



Sample Analysis of Texts from Schøyen’s Sermon on the Mount[11]



Sample 1:  A Case for a Unique Recension? (Matt 6:16) And when you fast, do not look gloomy like the hypocrites, for they disfigure their faces that their fasting may be seen by others. Truly, I say to you, they have received their reward.


Codex Schøyen

nt$ak #de ak$wan#@$ eknen/#cteu/$ mperer#th/ nnihupo$rit/c #et$e sauo#kemmeou$@/ peuh#a$ #je eueouwnh$ ebal hn nir#wm/ eu$n/cteu/ ha#m/n ]jw m$mac n/tn je #hauw$ euji bek/:


Schenke’s Retroversion

SuV (deV) o[tan (nhsteuvh/s), mhV givnou (w&s oi&) &pokritaiV skuqrwpoiv, (ou*) nivpotontai toV provswpon    au*tw`n (i{na fanw`sin) e*n  toi`s a*nqrw`pois nhsteuvontes: ajmhVn (levgw) uJmi`n o!ti (ajpevcousin) misqovn.



{Otan deV nhsteuvhte,    mhV givnesqe wJ" oiJ uJpokritaiV skuqrwpoiv, ajfanivzousin gaVr taV provswpa aujtw'n o{pw" fanw'sin toi'" ajnqrwvpoi" nhsteuvonte":  ajmhVn levgw uJmi'n, ajpevcousin toVn misqoVn aujtw'n.


Four items call for an explanation.  First is the emphatic personal pronoun beginning the verse, though the reconstruction is subject to some conjecture.  Second, consonant with the reconstructed emphatic pronoun, Codex Schøyen puts the admonition in the singular, as is clear from the (extant) second person prefix marker on the verb:  when you (sg.) fast.[12]  Third, Codex Schøyen reads not that the hypocrites disfigure their faces as reflected in N-A27, but rather that they do not wash their faces.  Fourth, as is typical for Codex Schøyen, it does not reflect gaVr as beginning the explanatory clause.  All four items are problematic because they reflect readings not found in the textual mainstream.[13]  An explanation for each item might be plausible in itself, but considered cumulatively, there may be at least a little difficulty in proffering an explanation as to how all four unrelated items found their way into one verse without attestation of any of those items anywhere else in the tradition.



Sample 2:  Reviser or Translator:  “This, then, is how you should ask…” (Matt 6:9) This, then, is how you should pray:  "'Our Father in heaven, hallowed be your name….


Codex Schøyen

nteihy oun ntwten pe#tet)neeti mm#af$ je peniwt ethn np/: pekre#n$ #ma$refouap


Schenke’s Retroversion

Ou@tws ou^n o$ ai*thvsete u&mei`s:  Pavter h&mw`n o& e*n toi`s ou*ranoi`s: a&giasqhvtw toV o@nomav sou


Ou{tw" ou\n proseuvcesqe uJmei'":  Pavter hJmw'n oJ ejn toi'" oujranoi'":  aJgiasqhvtw toV o[nomav sou


How does one account for Codex Schøyen’s unique reading that the prayer is introduced with the verb to ask, and not to pray?  Could this reflect the translator’s perceived translational latitude, despite proseuvcomai appearing nowhere else in his translation as to pray?  Or more radically, could the translator have assumed the task of reviser, thinking that changing the verb to to ask would improve the connection to the asking of the previous verse, as if to read, “For your Father knows what you need before you ask him, and so, this is how you ask him…”?  Or even more radically, does Codex Schøyen reflect a different recension of Matthew?



Sample 3:  Reconstructing a Partially Lost Dominical Logion?  (Matt 7:6) Do not give dogs what is holy, and do not throw your pearls before pigs, lest they trample them underfoot and turn to attack you.


The saying as we know it from canonical Matthew may be depicted with two main parts, with only the second part followed by two modifying lines: 


Do not give dogs what is holy,

and do not throw your pearls before pigs,

lest they trample them underfoot

and turn to attack you.[14]


Codex Schøyen is badly lacunose here, but Schenke suggests the following reading, relying in part on a single Bohairic manuscript for a daring Greek reconstruction:


Codex Schøyen

mper] mpetou#eb nneuhar mpertou-space for about  9 more letters$ oude petenmar#garit/c mperhitf ha neseou mpe$rtou<h>om ncecei#t/ mmaf ha neouer/t/ nceka$tou erwtn ncepeh #t/nou


Schenke’s Reconstruction

MhV dw`te toV a@gion (toi`s kusivn.  mh balevtwsan aujtoV eijs thVn koprivan).   mhdeV            toVn  margarivthn uJmw`n (mhV bavlhte aujtoVn e!mprosqen tw`n coivrwn.  mhV)  katapathsavtwsan kaiV mhV rJifqhvtwsan (aujtoVn [sic] uJpokavtw tw`n podw`n aujtw`n kaiV    mhV strafevntes) uJmi`n rJhxavtwsan (uJma`s).




MhV dw'te toV a{gion toi'" kusiVn mhdeV bavlhte touV" margarivta" uJmw'n e[mprosqen tw'n coivrwn, mhvpote katapathvsousin aujtouV" ejn toi'" posiVn aujtw'n kaiV strafevnte" rJhvxwsin uJma'".


As if out of the blue, Schenke reconstructs the logion to include a line about not throwing what is holy into the manure.  In a strict translation from the reconstructed Greek text, the verse may be depicted with two main parts, the first with one modifying line, and the second with two modifying lines: 


Do not give what is holy to (the) dogs,

            Do not let them throw it into the manure.

Nor your pearl [sg.], do not throw it before pigs,

Do not let them trample it and throw it under their feet

unless turning toward you, they attack you.


This, of course, must remain very speculative, since even relying on the other Middle Egyptian manuscript, Codex Schøyen is still badly mutilated.  The reconstruction, particularly the Greek reconstruction, is Schenke’s attempt to explain the content of missing sections which can hardly be explained by canonical Matthew.  Schenke’s reconstruction, while speculative, does have the merit of attempting to explain the origin of the unusual reading of the aforementioned Bohairic text, and adds balance to the logion itself which otherwise seems to be lacking in Canonical Matthew.



Sample 4:  Collaborating Evidence (Matt 6:25) Therefore I tell you, do not be anxious about your life, what you will eat or what you will drink, nor about your body, what you will put on. Is not life more than food, and the body more than clothing?


The textual tradition presents a difficult decision as to whether the statement, “Is not life more than food” might further be qualified by the addition of “or drink,” as reflected by the N-A27’s bracketing of h] tiv pivhte (or what you will drink) within the text itself.  In this case, Vaticanus and Sinaiticus are divided on the question, and so is much of the Coptic tradition.  Given that the preferred reading is so tenuous, perhaps support from Codex Schøyen might prompt a reversal by the N-A27 editors.  This is to suggest that Codex Schøyen may have considerable value for establishing the standard text.





A number of other interesting examples could be cited to illustrate matters such as unusual readings which might betray the scribe’s reaction to otherwise well known textual variants (6:8, 15, 33) and unusual readings which might reflect the scribe’s interest in being explanatory (6:17, 29; 7:3).  Issues such as these cannot be fully developed in this introductory paper.  The few examples mentioned, however, serve to indicate that there are grounds for further analysis. The question remains whether these phenomena actually reflect an unstable second century textual transmission, of if they can be explained within the context of a stable tradition.  Is Codex Schøyen a poster child for textual scepticism, or not?  Given the stormy context of modern textual criticism of the Gospels, the manuscript promises to foment interest in the coming years.

[1] This is a paper presented for the New Testament student seminar, University of Aberdeen, 2007, but represents only the most preliminary stages of my research.  Footnotes have not been fully checked, and fonts need to be converted to Unicode.

[2] See, for example, articles by Koester and Epp in Petersen’s Gospel Traditions in the Second Century:  Origins, Recensions, Text, and Transmission (Notre Dame:  University Press, 1989); M.-E. Boismard’s L’Evangile de Marc, sa préhistoire (Gabalda, 1994); Koester’s Ancient Christian Gospels (Philadelphia:  Trinitiy Press, 1990); Epp’s “The Multivalence of the Term ‘Original Text,’ (2001); Bart Ehrman’s Misquoting Jesus.

Opposing what he perceives as this face of “Anglo-Saxon textual criticism,” Martin Hengel writes, “"The text of the earliest Gospel 'according to Mark'...which was first Luke and the author of the Gospel of Matthew, was not demonstrably very different from the form of the text which we possess in the twenty-seventh revised edition of the Nestle/Aland 1993....  It would be completely claim that we may no longer speak of an 'original text' in the New Testament generally... [or] of an almost 'chaotic' diversity to which order was first brought in an 'orthodox' way, i.e. a violent way, by the process of canonization in the mainstream church.  Such a judgment would be far too one-sided" (The Four Gospels and the One Gospel of Jesus Christ: An Investigation of the Collection and Origin of the Canonical Gospels, 29-30).

[3] The Schøyen Collection asserts that it may have been a part of the original Chester Beatty hoard.  The deal for 15 papyrus pages of a Middle Egyptian Isaiah text which were found within the pages of Codex Schøyen did not materialize, and details of their whereabouts are said to be unknown (Coptic Papyri, vol. 1 [Oslo:  Hermes Publishing, 2001], p. 17.

[4] This comment from The Schøyen Collection assumes the traditional dating of two or three other manuscripts, notwithstanding occasional attempts to date them slightly earlier (The Schøyen Collection,  In 1995, Frederik Wisse wrote, “Little progress has been made during the past fifty years in Coptic palaeography.  Palaeographical dating is far more difficult for early Coptic MSS than for Greek ones, since there are almost no early dated Coptic documents.  This dearth is alleviated somewhat by the fact that Coptic scribal conventions in the Hellenistic period appear to follow closely developments in Greek calligraphy.  Thus dated Greek documents can be used also for Coptic palaeography.  Nonetheless great caution is needed in assigning dates to early Coptic biblical fragments, and an even greater latitude is called for than with Greek texts” (“The Coptic Versions of the New Testament” in The Text of the New Testament in Contemporary Research:  Essays on the Status Quaestionis [Grand Rapids:  Eerdmans, 1995],

[5] To heighten this importance all the more, Schenke argued that the text of Codex Schøyen is derived from a different recension of Matthew’s Gospel via a Hebrew or Aramaic Gospel of Matthew.  Schenke goes on to assert that his theory then has implications for the Synoptic Problem, that Marcan priority and the two source theory must be reconsidered.  Unfortunately, such a sensational claim has had the effect of undermining Schenke’s credibility in the academic community, with the net result that his more credible claims are dismissed out of hand.  Thus, Schenke’s claim that Codex Schøyen must precipitate the re-evaluation of the Synoptic Problem, instead of increasing the relative importance of Codex Schøyen, has the effect of diminishing it, at least on some sort of psychological level.

Despite the psychological impact of Schenke’s radical claims, Codex Schøyen must be appreciated for what it is:  a very ancient manuscript with an enigmatic text.  Moreover, it also must reflect one of the earliest attempts to translate a New Testament text into another language, and so may give some indication of the latitude an early translator might deem appropriate.  Further, it represents another piece of evidence in the effort to reconstruct the otherwise sparsely attested history of early Christian Egypt.  While one should be cautious so as not to overestimate the importance of Codex Schøyen, there is some cause for exigency in its examination.

[6] The Sahidic Crosby-Schøyen Codex  193, a substantial manuscript which contains 1 Peter and a number of Old Testament tests (with Melito), is thought to date to the third century.  Note also that a number of fourth century fragments from the Gospel of John (P. Mich. Inv. 3521), originally designated as Fayumic, have been now recognized as specifically as Middle Egyptian (Wisse, 136). 

Middle Egyptian manuscripts also include one containing a Greek-Coptic glossary to Hosea and Amos, which is probably the earliest Middle Egyptian manuscript extant, dating presumably from the end of the third century (Bruce Metzger, The Early Versions of the New Testament:  Their Origin, Transmission, and Limitations, [Oxford:  Clarendon Press, 1977], p. 117).  Note also that a number of fourth century fragments from the Gospel of John (P. Mich. Inv. 3521), originally designated as Fayumic, have been now recognized as Middle Egyptian (Wisse, 136).  A manuscript containing a Greek-Coptic glossary to Hosea and Amos is probably the earliest Middle Egyptian manuscript extant, dating presumably from the end of the third century (Bruce Metzger, The Early Versions of the New Testament:  Their Origin, Transmission, and Limitations, [Oxford:  Clarendon Press, 1977], p. 117).

[7] Its size is 18.0 cm x 22.5 cm, much larger than the other Middle Egyptian Matthew which measures 12.5 cm x 10.5 cm, of which Bruce Metzger wrote, “…because of the small size of the page, it is probable that [Codex Scheide was] not written for liturgical purposes, but for private use. 

[8] Was early Christian Egypt largely heterodox?  Further consideration of Codex Schøyen in light of Larry Hurtado’s recent discussion of the issue in his The Earliest Christian Artifacts:  Manuscripts and Christian Origins, may further erode Bauerian conclusions that Christianity there was more or less heterodox (p. 40). 

[9] To take an obvious example from the modern period, N-A27 lists the conjunction dev as occurring 29 times in the Sermon on the Mount, but it is only reflected 16 times in the NIV, which, by the same methodology, might make one conclude that the eclectic text behind the NIV was derived from an altogether different recension of Matthew.  In all fairness, this example probably does not do justice to Schenke’s considered argument, but it illustrates the need to consider translational latitude before concluding that a variant reading is actually due to a variant text. 

[10] The Coptic attestation of the Greek text is strengthened by the fact that, in terms of word order, it generally follows its Vorlage closely.  Moreover, about 15% of early Coptic texts consist of Greek loan words, typically derived directly from the Greek Vorlage.  This fortuitousness allows critics to affirm with some confidence that a Coptic reading does in fact support a particular reading in the Greek tradition.  Wisse notes that the trend in loan words is away from their usage with time.  Excluding kaiv, dev, and ijdouv (since Codex Schøyen tends to omit these) and other relevant contingencies, Codex Schøyen retained 43 of the Greek loan words, Codex Scheide retained 42, and the Sahidic retained 27.  In only one case did the Sahidic preserve the Greek loan word against the two Middle Egyptian texts. 

[11] Although fragmented in several places, the Sermon on the Mount may serve as a good sampling of the text, since it was so well known, and with much of it being paralleled in Luke, a number of interesting textual issues arise.  Schenke has proffered a reconstruction of the fragmentary lines following the fifth century Middle Egyptian Matthew, Codex Scheide.  No doubt, many more interesting readings otherwise characteristic of Codex Schøyen must simply remain unrecoverable in this portion of text.  English Bible translations are from the English Standard Version, and the Greek from N-A27 and from Schenke’s own retroversion which must be used with considerable discrimination.  Unless otherwise indicated, the reconstruction of Codex Schøyen text is cautiously and typically derived from the fifth century Middle Egyptian Matthew (Codex Scheide).

[12] Schenke’s bracketing of lacunose text seems to have several errors.

[13]  I.e., in the apparatuses of the standard editions.  One of the major criticisms Baarda levelled at Schenke was that Schenke’s search through the tradition for matching readings was rather shallow.

[14] A chiastic understanding of the saying, assuming that it is the dogs which might turn to attack, would be depicted thusly:

Do not give dogs what is holy,

and do not throw your pearls before pigs,

lest they trample them underfoot

and turn to attack you.