Hostname: page-component-74656f9884-j7cmm Total loading time: 0 Render date: 2026-11-10T00:51:14.701Z Has data issue: false hasContentIssue false
  • English
  • Français

The Lord's Prayer and ΧΜΓ: Two Christian Papyrus Amulets*

Published online by Cambridge University Press:  23 December 2010

Brent Nongbri
Brent Nongbri*
Affiliation:
Macquarie University
Rights & Permissions [Opens in a new window]

Extract

The last few years have seen a renewed interest in the study of papyrological remains as a means of shedding light on early Christianity.1 To add further evidence to this discussion, I draw attention to two papyri in Yale University's Beinecke Library. In 1997, Yale acquired several papyri from Galerie Nefer in Zürich. The lot contained papyri ranging in date from the Ptolemaic era to the Medieval period. Two of these pieces (both, unfortunately, without specific provenance) appear to be Christian textual amulets.

Information

Type
ARTICLES
Information
Harvard Theological Review , Volume 104 , Issue 1 , January 2011 , pp. 59 - 68
Copyright
Copyright © President and Fellows of Harvard College 2010

The last few years have seen a renewed interest in the study of papyrological remains as a means of shedding light on early Christianity.Footnote 1 To add further evidence to this discussion, I draw attention to two papyri in Yale University's Beinecke Library. In 1997, Yale acquired several papyri from Galerie Nefer in Zürich. The lot contained papyri ranging in date from the Ptolemaic era to the Medieval period. Two of these pieces (both, unfortunately, without specific provenance) appear to be Christian textual amulets.

P.CtYBR inv. 4600: The Lord's PrayerFootnote 2

The first piece (fig. 1) is P.CtYBR inv. 4600, a papyrus copy of the Lord's Prayer that largely matches the version of the prayer in Matt 6:9–13.Footnote 3 The writing of the prayer runs with the fibers (→). The back of the papyrus (↓) is blank. At some point in the modern process of conservation, a large strip of clear tape was applied to the back of the papyrus in the lower left corner and has stained that portion of the papyrus.

Figure 1.

P.CtYBR inv. 4600. 9.1 × 15.5 cm. Unknown Provenance. Sixth or seventh century C.E. Courtesy of the Beinecke Rare Book and Manuscript Library, Yale University.

The piece measures 9.1 cm wide by 15.5 cm high. Margins are preserved on the top, left, and bottom. The left margin is roughly 1.0 cm; the top margin is the same. The bottom margin is about 1.5 cm (although a portion of the bottom of the papyrus may have been lost; see below). The height of the letters ranges between 0.4 cm and 1.0 cm.

The hand is practiced, sloping noticeably to the right, and while it shows some affinity to documentary hands, the letters are generally less crowded and better defined than those of documents. The vertical stroke of the η regularly rises above the notional lines, and the vertical strokes of the τ, ι, and κ regularly dip well below the notional lines; all the letters that dip below the notional line (save for ι) show hooks at their bases. The δ has a prominent loop, and ligatures regularly join the combinations αι and ει. The hand of P.CtYBR inv. 4600 is quite similar both in overall character and in individual letter forms to the hand of P.Vindob. G 39736, a letter of 643 C.E., although the latter is much more neatly executed.Footnote 4 A date for the Yale piece in the sixth or early seventh century thus seems likely, although precision in these matters is always difficult.

As is the case with a number of papyrus copies of the Lord's Prayer, the spelling of the Yale piece is barbarous. Plural personal pronouns are misspelled throughout the papyrus; the writer's distinctive η is replaced by υ in these pronouns, occasionally yielding nonsense readings.Footnote 5 A double line mark (=) occurs four times in the papyrus. In three of those instances (lines 5, 6, and 7), the symbol divides sense lines, but it appears misplaced in line 3.Footnote 6

The piece seems to have been folded both horizontally and vertically into a small square of roughly 2.5 cm by 2.5 cm. The right side of the papyrus is broken off at a crease, and a small portion of the bottom of the papyrus may have been lost for the same reason. Papyri with biblical texts that are folded in this way are often identified as amulets, and while the identification of P.CtYBR inv. 4600 as an amulet is not completely certain, it does seem highly likely.

Theodore de Bruyn has identified a cluster of characteristics that are common to papyrus amulets with biblical texts. In addition to evidence of small size (or evidence of folding into a size small enough to wear on the body), he lists the inclusion of petitions or esoteric words, the inclusion of staurograms and other symbols, as well as characteristics that suggest that the text was not written as a literary work, which include “the coarseness of the hand, the irregularity of the orthography, and the presence of peculiar readings.”Footnote 7 A number of factors suggest that the Yale piece should be regarded as an amulet. First, the Lord's Prayer (or portions of it) is known to occur on papyri that are undoubtedly amulets.Footnote 8 Second, although the Yale piece appears to lack any symbols or direct invocations (other than the Lord's Prayer itself) characteristic of many Christian amulets, its poor spelling and nonsense readings suggest that it was unlikely to have been a memory aid (another possible use for papyri with non-continuous biblical texts).Footnote 9 Finally, the blank back confirms that this piece was not part of a codex. De Bruyn has further proposed a fourfold classification for papyri with non-continuous biblical texts: those that are certainly amulets, those that are probably amulets, those that are possibly amulets, and those that are unlikely to be amulets. In this framework, the Yale piece would be in the second class; it is probably an amulet.

Text

NotesFootnote 10

1. πατηρ: Read πάτερ. The vocative πάτερ is frequently misspelled as πάτηρ in papyrus copies of the Lord's Prayer (e.g., P.Duk.inv. 778). Only traces of the ω and ν in ὑμῶν are readable. The ὑμῶν is the first of several misspellings of personal pronouns resulting from the substituion of υ for η. This phenomenon is quite common in Greek manuscripts generally. See Francis Thomas Gignac, A Grammar of the Greek Papyri of the Roman and Byzantine Periods (2 vols.; Milan: Istituto editoriale cisalpino, 1976–1981) 1:264 and 2:165; the opening line of the Lord's Prayer preserved in P.Princ. 2.107 also appears to read πατηρ υμων. There is just sufficient space in the obscured part of the first line for ὁ before the epsilon. This lone ε is a misspelling of ἐν that also occurs at this point in the Lord's Prayer in P.Iand. 1.6 (= P.Giss.Lit. 5.4). The absence of a definite article before οὐρανοῖϛ is, as far as I know, unparalleled in other papyri and the major witnesses to Matt 6:9. The reconstructed οὐρανοῖϛ in the lacuna could just as easily be the singular οὐρανῷ, which is the reading of the Lord's Prayer in the Didache 8:2.

3. κενηθητου: Read γενηθήτω. The substitutions of κ for γ and ου for ω are frequent in papyri (Gignac, Grammar of the Greek Papyri, 1:77 and 1:209–210). In this piece, the substitution of κ for γ also occurs in line 4. The = mark occurs three more times in the papyrus. In lines 5, 6, and 7, the sign marks a division between sense units. Here, however, the mark seems out of place; it would be expected before γενηθήτω.

4. κηϛ: Read γῆϛ. Most witnesses of Matt 6:10 place the definite article before γῆϛ, but several important manuscripts (such as Codex Sinaiticus, Codex Vaticanus, and the Freer Codex) lack the definite article (as does the copy of the prayer in BGU 3.954). Since τὸν ἄρτον begins a new sense unit, one might expect the = mark after γῆϛ, but there is no trace of the symbol.

5. συμερον: Read σήμερον. The same misspelling occurs in the Lord's Prayer preserved in P.Köln 8.336.

6. The personal pronouns in this line are again misspelled.

7. οϕητηϛ: Read ὀϕειλέταιϛ. The partial letter at the end of the line is almost certainly μ (see the μ of ὑμῶν in line 10).

8. πϊρασμον: Read πειρασμόν. In P.Duk.inv. 778, P.Köln 4.171, and P.Köln 8.336, πειρασμόν also lacks the ε. The partial letter at the end of the line is almost certainly κ (the vertical stroke is similar to that of the κ in line 10). If the restoration of [αί] is in fact correct, it would, as far as I know, be unique in the textual tradition (all other witnesses reading ἀλλά).

9. Once again, the personal pronoun is misspelled. The ο of ἀπό is written above π.

10. το: The opening characters in line 10 are ambiguous. They look similar to the τ and ο in τοῦ of line 9. Yet, in the five certain instances of τ followed by ο in this piece, in lines 2, 3, 4 (twice), and 9, the ο invariably appears to the right of the horizontal stroke of the τ. In line 10, the loop is under the horizontal stroke. If το is the correct reading, two possibilities arise. The letters could be a continuation of the last word in line 9 like [ὀϕειλή] ματα in lines 5 and 6. Or, given our writer's poor spelling, it is possible that the writer has substituted τὸ κυρίῳ for τῷ κυρίῳ. Finally, it is just possible that instead of two letters, there is a single character—the Christian symbol tau-rho. If that is the case, this would be a badly malformed effort. In any event, there is to my knowledge no extant manuscript of the Lord's Prayer that contains this concluding phrase (although a partial inscription of the Lord's Prayer found in Greece at Megara dated to the fourth or fifth century [O.Athens inv. 12227] ends with the vocative [κ]vύριε followed by the tau-rho symbol). Finally, after ὑμῶν, just before the papyrus breaks off, a horizontal stroke is visible; it could either be a letter or perhaps a tau-rho symbol.

P.CtYBR inv. 4710: A Papyrus Amulet with the Symbol ΧΜΓ

The second piece (fig. 2) is P.CtYBR inv. 4710, a narrow strip of papyrus with the symbol ΧΜΓ (chi-mu-gamma) appearing four times in a row on the front of the papyrus (↓); each group of letters has a separate supralinear stroke, and the ink fades from left to right such that the last group is difficult to make out:

The piece is neatly cut on all four edges. It is 33.2 cm long, and it has a height of 4.0 cm on the left side that gradually increases to 5.9 cm on the right side. The back of the papyrus (→) is blank except for a few smudges of ink and what appears to be a single curved line on the narrow end of the piece. The strip was folded in half left to right three times, dividing it into eight roughly equal portions marked by creases. The size of the papyrus and the pattern of folds are quite similar to magical amulets, which is what this piece almost certainly was.Footnote 11

Figure 2.

P.CtYBR inv. 4710. 33.2 × 5.9 cm Unknown provenance. Fifth or sixth century C.E. Courtesy of the Beinecke Rare Book and Manuscript Library, Yale University.

The orthography deserves some comment. The script is large (the height of the letters varies between 1.7 and 2.1 cm). The letters all show ligatures, but each instance of ΧΜΓ shows a different ligature between the Χ and the Μ (the ligatures between the Μ and the Γ are consistent).Footnote 12 The hand would not be out of place in the fifth or sixth century, but again, certainty is elusive.

Determining the meaning of the ΧΜΓ symbol has proven difficult for investigators. The bibliography on the symbol is large, but there is little consensus regarding the interpretation of the letters.Footnote 13 In the fourth century, this sequence of letters begins to appear on a wide range of media (papyri, amphora necks, door lintels, statuary, tiles, brick stamps, graffiti) throughout the ancient Mediterranean world.Footnote 14 In papyri, one generally finds ΧΜΓ in two contexts—centered above the first line of documents and letters, and, less frequently, centered above the first line of hymns, prayers, and magical spells.Footnote 15 The Yale piece is somewhat unusual among papyri in its repetition of ΧΜΓ, although the reverse sides of two oracular requests to St. Philoxenus assigned to the sixth century (P.Oxy. 16.1926 and P.Harr. 54) have ΧΜΓ repeated three times interspersed with crosses.Footnote 16

The two leading interpretations regard the symbol as either an acrostic or an isopsephism (that is, a numerical value that corresponds to another word or phrase with an equal numerical value). Those who claim that the symbol is an acrostic usually argue that it abbreviates phrases like Χ(ριστὸϛ ὁ ἐκ) Μ(αρίαϛ) Γ(εvvηθείϛ), Χ(ριστὸϛ) Μ(ιχαὴλ) Γ(αβριήλ), or Χ(ριστὸϛ) Μ(άρτυϛ) Γ(έvηται).Footnote 17 Support for these theories comes in the form of a small number of examples in which the symbol appears to be only partially abbreviated (such as ΧCΜΓ).Footnote 18 Further evidence for interpreting the symbol as an acrostic is provided by what appear to be alternate forms of the symbol, such as ΚΜΓ.Footnote 19 The main drawback to this set of interpretations is the lack of unambiguous attestation for the proposed unabbreviated phrases.

Those who see the symbol as an isopsephism point out that the Greek letters ΧΜΓ have a numerical value of 643 (Χ = 600, Μ = 40, Γ = 3), which corresponds to the numerical value of several Greek phrases. The most popular choice for resolving ΧΜΓ is the phrase θεὸϛ βοηθόϛ (9 + 5 +70 + 200 + 2 + 70 + 8 + 9 + 70 + 200 = 643).Footnote 20 As evidence that ΧΜΓ is an isopsephism, a number of scholars have noted that ΧΜΓ is often paired with θ(= 90 + 9 = 99), a well-known isopsephism that is equivalent to ἀμήv (1 + 40 + 8 + 50 = 99).Footnote 21 The evidence for resolving ΧΜΓ with the specific phrase θεὸϛ βοηθόϛ is chiefly that this epithet was current from the fourth century onwards.Footnote 22 Also, this phrase seems to have been abbreviated in a less cryptic form (ΘΒ) in a few papyri and inscriptions.Footnote 23 The weakness of this interpretation is that it cannot account for the non-standard occurrences of the symbol (ΧCΜΓ, ΚΜΓ, etc.).

There does not seem to be a great deal of hope for settling the question one way or the other. Given the evidence on both sides of the issue, it appears that the symbol had different meanings at different times and in different geographical areas.Footnote 24 What is clear, however, is that in some media, especially inscriptions, ΧΜΓ frequently occurs in apotropaic contexts. The inscriptions from Syria are illustrative. One of these inscriptions, a door lintel likely dating from the fifth century, reads + ΧΜΓ α + ω Ijη(σοῦ)ϛ βοήθι + (chi-mu-gamma, alpha-omega, Jesus help!).Footnote 25 Another similar piece from the fifth century reads ΧΜΓ Χ(ριστο)ῦ τὸ vῖκοϛ ϕεῦγε σαταvᾶ (chi-mu-gamma, Christ's is the victory. Flee, Satan!).Footnote 26 This type of usage does appear in papyri as well. In addition to P.Oslo 1.5 (see note 15 above), SB 16.12658, a piece identified by its editor as a Byzantine papyrus amulet, reads ΧΜΓ εὐτυχῶϛ ἀγαθῇ τύχῃ (chi-mu-gamma, successfully, by good fortune).Footnote 27 While the letters ΧΜΓ appear in P.CtYBR inv. 4710 with no additional text, the physical properties of the piece that allow for its identification as an amulet provide further evidence for the magical or apotropaic value of the symbol.

Footnotes

*

My thanks to Theodore de Bruyn, David G. Martinez, Michael L. Peppard, and Kevin W. Wilkinson for advice on different aspects of these papyri.

References

1 See Epp, Eldon Jay, “The Oxyrhynchus New Testament Papyri: ‘Not Without Honor Except in their Hometown’?” JBL 123 (2004) 5–55Google Scholar; reprinted in Perspectives on New Testament Textual Criticism: Collected Essays 1962–2004 (Leiden: Brill, 2005) 743–801; the papers collected in New Testament Manuscripts: Their Texts and Their World (ed. Thomas J. Kraus and Tobias Nicklas; Leiden: Brill, 2006); Choat, Malcom, Belief and Cult in Fourth-Century Papyri (Turnhout: Brepols, 2006)CrossRefGoogle Scholar; Hurtado, Larry W., The Earliest Christian Artifacts: Manuscripts and Christian Origins (Grand Rapids, Mich.: Eerdmans, 2006)Google Scholar; Luijendijk, AnneMarie, Greetings in the Lord: Early Christians in the Oxyrhynchus Papyri⃛ (Cambridge, Mass.: Harvard Divinity School, 2008);Google Scholar and, most recently, Bagnall, Roger S., Early Christian Books in Egypt (Princeton: Princeton University Press, 2009)CrossRefGoogle Scholar.

2 Abbreviations for papyrological publications follow the conventions of the Checklist of Editions of Greek, Latin, Demotic, and Coptic Papyri Ostraca and Tablets (online: http://scriptorium.lib.duke.edu/papyrus/texts/clist.html; last accessed 9 August, 2009). Abbreviations for inscriptions follow Horsley, G. H. R. and Lee, John A., “A Preliminary Checklist of Abbreviations of Greek Epigraphic Volumes,” Epigraphica: Periodico Internazionale di Epigrafia 56 (1994) 129–69Google Scholar.

3 For a general overview of the magical use of scripture in papyri, see Judge, Edwin A., “The Magical Use of Scripture in the Papyri,” in Perspectives on Language and Texts: Essays and Poems in Honor of Francis I. Andersen's Sixtieth Birthday (ed. Conrad, Edgar W. and Newing, Edward G.; Winona Lake, Ind.: Eisenbrauns, 1987) 339–49Google Scholar, to be reprinted in Jerusalem and Athens: Cultural Transformation in Late Antiquity (essays selected and edited by Alanna Nobbs; Tübingen: Mohr Siebeck, forthcoming). Verses from the Lord's Prayer appear frequently in papyri, usually together with other short biblical excerpts. Thomas J. Kraus has conveniently collected and discussed papyri (as well as parchments and wooden and clay tablets) containing all or part of the Lord's Prayer in “Manuscripts with the Lord's Prayer—They are More than Simply Witnesses to that Text Itself,” in New Testament Manuscripts: Their Texts and Their World, 227–66. Kraus's article has an ample bibliography on the topic that I will not repeat here; I simply list the items in his catalog and the dates that have been proposed for them: P.Ant. 2.54 (third cent. C.E.); P.Erl.Diosp. 1 (313–314 C.E.); P.Sch⊘yen 1.16 (late fourth cent. C.E.); O.Athens inv. 12227 (fourth or fifth cent. C.E.); P.Oxy. 60.4010 (fourth cent. C.E.); P.Princ. 2.107 (fourth–sixth cent. C.E.); P.Col. 11.293 (fifth cent. C.E.); P.Köln 4.171 (fifth cent. C.E.); P.Iand. 1.6 (= P.Giss.Lit. 5.4; fifth or sixth cent. C.E.); P.Duk.inv. 778 (late sixth cent. C.E.); P.Köln 8.336 (sixth cent. C.E.); PSI 6.719 (fourth–sixth cent. C.E.); BGU 3.954 (sixth cent. C.E.); MPER N.S. 15.184 (sixth or seventh cent. C.E.); Louvre MND 552 B (seventh cent. C.E.?); P.Bad. 4.60 (eighth cent. C.E.); and University of Chicago Goodspeed MS 125 (fourteenth cent. C.E.). I should add one note to Kraus's bibliography on Goodspeed MS 125: Another portion of this manuscript has been identified (Ms. M. 499 in the Morgan Library in New York, which contains a colophon dating to the third quarter of the fourteenth cent.). It preserves Psalms 35 and 91 along with the Letter of Abgar, Christ's reply, and the story of the Mandylion. See most recently Peers, Glenn, “Magic, the Mandylion and the Letter of Abgar: On a Greco-Arabic Amulet Roll in Chicago and New York,” in Intorno al Sacro Volto: Genova, Bisanzio e il Mediterraneo, secoli XI–XIV (ed. Calderoni Masetti, Anna Rosa, Dufour, Colette Bozzo, and Wolf, Gerhard; Venice: Marsilio, 2007) 163–74Google Scholar and by the same author, “Art and Identity in an Amulet Roll from Fourtheenth Century Trebizond,” in Religious Origins of Nations? The Christian Communities of the Middle East (ed. Bas ter Haar Romeny; Leiden: Brill, 2010) 153–78. A good deal of additional bibliography on the piece is available in the Morgan Library's records online at http://corsair.morganlibrary.org/msdescr/BBM0499.htm (last accessed 11 August 2009).

4 This letter is plate 43c in Cavallo, Guglielmo and Maehler, Herwig, Greek Bookhands of the Early Byzantine Period: A.D. 300–800 (London: Institute of Classical Studies, 1987)Google Scholar.

5 Garbled versions of the Lord's Prayer are not uncommon in the papyri. For example, the portion of the prayer preserved in P.Princ. 2.107 reads πατηρ υμων εν τηϛ ουρανηϛ αγιασθητω τω θελημα σου τον αρτον υμων των επιουσιων (which might be rendered as “Your father who is in heaven, hallowed be your will; your daily bread”).

6 The portion of the Lord's Prayer preserved in P.Iand. 1.6 (= P.Giss.Lit. 5.4) has colons dividing the sense lines. The part of the prayer in P.Köln 8.336 appears to use a combination of colons and high dots for the same purpose.

7 de Bruyn, Theodore, “Papyri, Parchments, Ostraca, and Tablets Written with Biblical Texts in Greek and Used as Amulets: A Preliminary List,” in Early Christian Manuscripts: Examples of Applied Method and Approach (ed. Kraus, Thomas J. and Nicklas, Tobias; Leiden: Brill, 2010) 145–89CrossRefGoogle Scholar, at 150. My thanks to Theodore de Bruyn for sharing this work with me before its publication.

8 See BGU 3.954, P.Princ. 2.107, P.Iand. 1.6 (= P.Giss.Lit. 5.4), P.Sch⊘yen 1.16, and PSI 6.719.

9 While peculiar readings and irregular orthography are indeed characteristic of biblical textual amulets, it should be noted that the proper execution of manuscript copying itself was thought by some to exert power against demons. Skemer writes, “In the sixth century, Cassiodorus reminded the monastic scribes of Vivarium that the act of copying scripture would spread Christ's words and thus wound the Devil, in effect destroying the demonic power that had been unleashed upon Christ during the passion.” Skemer, Don C., Binding Words: Textual Amulets in the Middle Ages (University Park, Pa.: The Pennsylvania State University Press, 2006) 84.Google Scholar The reference is to Cassiodorus, Institutiones 1.30. See the edition of Mynors, Roger A. B., Cassiodori Senatoris Institutiones (Oxford: Clarendon, 1937)Google Scholar.

10 Papyri with non-continuous portions of the New Testament have been only sporadically included in text critical discussions. See Alessandro Biondi, “Le citazione bibliche nei papiri magici cristiani greci,” SP 20 (1981) 93–127; and Pickering, Stuart R., “The Significance of Non-Continuous New Testament Textual Materials in Papyri,” in Studies in the Early Text of the Gospels and Acts: The Papers of the First Birmingham Colloquium on the Textual Criticism of the New Testament (ed. Taylor, David G. K.; Atlanta, Ga.: SBL, 1999) 121–41Google Scholar. Nevertheless, when relevant, I mention in these notes observations of text critical interest. For a good general discussion of the textual problems of the Lord's Prayer, see Parker, David C., The Living Text of the Gospels (Cambridge: Cambridge University Press, 1997) 49–74CrossRefGoogle Scholar.

11 One can compare the dimensions of this piece to PSI 6.719 (25.0 × 5.5 cm), a Christian amulet assigned to the sixth century. The writing on PSI 6.719 is, however, much smaller (a single long column with six lines of compactly spaced text, including the first line and doxology of the Lord's Prayer), and the piece seems to have been folded fewer times. P.Oxy. 73.4932, a recently published amulet containing Septuagint Psalm 72:21–23 assigned by its editor to the fifth century, likely had similar dimensions as well (it now measures 14.1 × 6 cm, but judging by the missing words of the psalm, its original length was probably roughly 28–30 cm).

12 In the first group, the right-leaning stroke (/) of the Χ curves downward and crosses the lower portion of the Μ. In the second group, the right-leaning stroke of the Χ has an elaborate bend and loop that could almost be another letter before it attaches to the upstroke of the Μ. In the third set, the right-leaning stroke of the X curves and joins directly with the Μ. The fourth set is difficult to see, but it appears that the bottom of the left-leaning stroke (\) of the Χ moves directly into the upstroke of the Μ.

13 I will not reproduce the bibliography here. The following three studies provide good overviews of the debate and more comprehensive bibliographies. For a summary of the earliest studies, see William K. Prentice, “ΧΜΓ, A Symbol of Christ,” CP 9 (1914) 410–16; for the state of affairs in the middle of the twentieth century, see Jan-Olof Tjäder, “Christ, Our Lord, Born of the Virgin Mary (ΧΜΓ and VDΝ),” Eranos 68 (1970) 148–90. The most recent in depth treatment of the topic is Llewelyn, Stephen R., “The Christian Symbol ΧΜΓ, an Acrostic or an Isopsephism?” in New Documents Illustrating Early Christianity, Volume 8: A Review of the Greek Inscriptions and Papyri Published 1984–85 (Grand Rapids, Mich.: Eerdmans, 1998) 156–68Google Scholar. See also the brief but informative notes in Choat, Belief and Cult, 114–16.

14 The first appearances of the symbol probably date to the third century; see Llewelyn, “The Christian Symbol ΧΜΓ,” 156–57 n. 3. For relatively recent published examples, see Tomaz Derda, “Inscriptions with the Formula θεοῦ χάριϛ κέρδοϛ on Late Roman Amphorae,” ZPE 94 (1992) 135–52 and Roland R. R. Smith, “The Statue Monument of Oecumenius: A New Portrait of a Late Antique Governor from Aphrodisias,” JRS 92 (2002) 134–56, esp. 137 and 150–53.

15 See, for example, P.NYU 1.8, a receipt likely from the fourth or fifth century, P.Mich. 8.519, a letter found in Karanis assigned to the fourth century, P.Oxy. 59.3986, a contract dating from 494 C.E., and P.Oxy. 16.1961, a lease dating from 487 C.E. For ΧΜΓ placed at the top of a Christian hymn assigned to the sixth century, see P.Oxy. 60.4011. ΧΜΓ also appears at the top of P.Oslo 1.5, a protective amulet assigned to the fourth or fifth century.

16 On these two oracular requests, written in the same hand and cut from the same sheet of papyrus, see Youtie, Herbert Ch., “Questions to a Christian Oracle,” ZPE 18 (1975) 253–57Google Scholar; reprinted in Youtie, Herbert Ch., Scriptiunculae Posteriores (2 vols.; Bonn: Rudolf Habelt, 1981) 1:225–29Google Scholar. BGU 3.948, a personal letter dated to the fourth or fifth century, has ΧΜΓ repeated three times above the first line of the letter.

17 Tjäder's article cited above was highly influential and convinced many that the symbol was originally an acrostic standing for Χ(ριστὸv) Μ(αρία) Γ(εvvᾷ). Alain Blanchard provided further support for that understanding in his essay “Sur quelques interprétations de ΧΜΓ,” Proceedings of the XIV International Congress of Papyrologists (London: Egypt Exploration Society, 1975) 19–24. Naphtali Lewis, however, has since soundly refuted their arguments (“Notationes Legentis,” BASP 13 [1976] 158–59); see also Llewelyn, “The Christian Symbol ΧΜΓ,” 157–59, esp. n. 5.

18 For a possible instance of ΧCΜΓ, see ISyriaPrentice 211(= IGLSyr 4.1486).

19 Georgina Robinson, for example, accepts that ΧΜΓ at first was an abbreviation for Χ(ριστὸϛ) Μ(άρτυϛ) Γ(ένηται), which she regards as the source of the variant forms Κ(ύριοϛ) Μ(άρτυϛ) Γ(έvvηται) and Θ(εὸϛ) Μ(άρτυϛ) Γ(έvηται) (“ΧΜΓ and ΘΜΓ for ΧΜΓ,” Tyche 1 [1986] 175–77). Bernard P. Grenfell and Arthur S. Hunt reported the discovery of many inscribed fragments of amphorae with ΧΜΓ at Oxyrhynchus, one of which has the letters reversed: ΓΜΧ. See their excavation report for the 1906–1907 season, reprinted in Oxyrhynchus: A City and Texts (ed. Alan K. Bowman et al; London: Egypt Exploration Society, 2007) 367.

20 Alternative isopsephic interpretations include ἅγειοϛ ὁ θεόϛ (God is holy). If the symbol is regarded as an isopsephism for θεὸϛ βοηθόϛ, then some caution should be exercised in using the label “Christian” to describe it, since this phrase does occur at least once in reference to another god. In this regard, Llewelyn refers to SEG 27.1016, an inscription from Jerusalem (undated by its editors) that reads: θεοϛ | βοη|θοϛ | Ποσι|δωv | βοηθει (God the helper, Poseidon, help!); see Llewelyn, “The Christian Symbol ΧΜΓ,” 163.

21 For examples of ΧΜΓ paired with θ at the top of documentary papyri, see SB 16.13037 (522 C.E.) and P.Lond. 2.483 (616 C.E.). P.Mich. 6.378, a list of payments and part of an archive dated to the fourth or fifth century, begins with these symbols along with a cross: + ΧΜΓ θ.

22 Llewelyn discusses many uses of the phrase that likely date from the fourth and fifth centuries (“The Christian Symbol ΧΜΓ,” 162–64).

23 For the abbreviation ΘΒ at the top of a letter assigned to the fifth century, see P.Heid. 4.333. One must allow, however, that this abbreviation could be resolved in different ways (such as θεὸϛ βοηθῶv or θεὲ βοήθει).

24 Such a conclusion is not new. Although William K. Prentice strongly argued for the acrostic interpretation, he wrote in 1922, “Of course it is entirely possible that many people who used the symbol were ignorant of its original meaning: to them it was merely a sign, which was useful on house-lintels and elsewhere to avert evil, or which was merely customary. Under such circumstances it would be natural that different interpretations were found for it, and perhaps some believed it the more potent [sic] the more meanings could be given to it.” Publications of the Princeton University Archaeological Expeditions to Syria in 1904–5 and 1909 Division III: Greek and Latin Inscriptions, Section B: Northern Syria (Leiden: Brill, 1922) 171.

25 ISyriaPrentice 224 (= IGLSyr 4.1452).

26 ISyriaPrentice 234 (= IGLSyr 4.1443). William K. Prentice discusses these inscriptions in “Magical Formulae on Lintels of the Christian Period in Syria,” AJA 10 (1906) 137–50. This phenomenon is not limited to Syria. The back of the head of a statue (likely from the early fifth century) discovered in 1904 in Aphrodisias is inscribed with: ΧΜΓ | Θ(ε)ὲ βοήθι (chi-mu-gamma, God, help!). For a recent discussion of the piece, see Smith, “The Statue Monument of Oecumenius,” 151.

27 See Pieter J. Sijpesteijn, “Wiener Melange,” ZPE 40 (1980) 91–110. See also Llewelyn, “The Christian Symbol ΧΜΓ,” 156.

Figure 0

Figure 1. P.CtYBR inv. 4600. 9.1 × 15.5 cm. Unknown Provenance. Sixth or seventh century C.E. Courtesy of the Beinecke Rare Book and Manuscript Library, Yale University.

Figure 1

Figure 2. P.CtYBR inv. 4710. 33.2 × 5.9 cm Unknown provenance. Fifth or sixth century C.E. Courtesy of the Beinecke Rare Book and Manuscript Library, Yale University.