HERMAS, THE SHEPHERD OF, title of an early Christian paraenetic apocalypse (Lat. title: Pastor [Hermae], Gk.: Poimēn “Shepherd”), composed in Greek by a certain Hermas, who presents himself as an emancipated slave and then a Roman businessman. According to the Canon Muratori he was the brother of the Roman bishop Pius (ca. 142-55, cf. Hilhorst, 1988, p. 683; Staats, 1993, p. 103). The most likely date for the Shepherd’s origin is ca. 140 C.E. (Harnack, 1958, pp. 257-59). The text at first enjoyed great popularity and almost canonical authority in Alexandria and in the western parts of the Roman Empire. It was attached to the texts of the New Testament in the Codex Sinaiticus. It fell gradually into disuse from the fourth century on, but never disappeared completely.

The Shepherd is conventionally divided into three parts: five visions, twelve mandates, and ten similitudes (Vielhauer, 1966, pp. 445-46). They have in common a didactic and admonitory purpose: to direct their better readers, and hearers to repentance and the resuscitation of the virtues of the primitive church. The moral decline of the contemporary church is indeed the author’s major concern, and in order to rectify this deplorable state of affairs he recommends (authorized by a letter from heaven) a second absolution of sins to be granted after the baptismal absolution. This was meant to contribute to the perfection of the church as a necessary prerequisite for the completion of world history and the end of the world (Vielhauer, 1966, p. 445).

The Shepherd, first composed in Greek (not completely preserved), was soon repeatedly translated into Latin (complete text). Most recent editions: Whittaker, 1956; Joly, 1958 (with French translation); detailed German translation and commentary: Dibelius, 1923, pp. 415-639. There are also translations into Coptic (Akhmimic and Sahidic) and Ethiopian (Vielhauer, 1966, p. 445).

What makes the Shepherd a subject of Iranological studies is the existence of a Manichean adaptation in Middle Persian language, attested by the Middle Persian Turfan fragment M 97, first identified and edited by F. W. K. Müller: edition, with text photo (Müller, 1905, pp. 1077-83); re-edition (Salemann, 1908, pp. 34-35); revised text with commentary (Boyce, 1975, pp. 178-79); revised edition (Burtea, 2002, pp. 47-67); text photo (Sundermann, 1996, pl. 63).

In what follows I will provide an English translation of the Middle Persian fragment M 97, largely based on Burtea’s German text. The paragraphs correspond to those in M. Boyce’s re-edition in her Reader, 1975.

The Middle Persian Text of the Shepherd. s1 . . . tower (hʾmʾbyr) . . . They . . . those women take [rock]s . . . to(?) the dark mountain. And he [appoint]s(?) six overseers [and] another team of many men who build the tower. And all the pieces of rock which went in to the gate, (they) were placed (nysyhyst) in the tower, and those which did not go into, they were damaged and taken back to the place where they had been taken from. And he explains thus: . . . altogether, all . . . is. . . . was not in the dark mountain.

§2 And the second m[ountain], the empty one, that are those hypocrites (dysmwyʾn, Gr. hypokritai) and teachers of lie who do not bear the fruit of righteousness, just as nothing is (to be found) on their mountain. But they are [not] ([!] so to be restored; cf. Joly, 1958, pp. 332-35) becoming apostates (ʾbystʾgʾn), and if it is so that they will do penance (ʾwhrʾnd), then their sin will be forgiven.

§3 The third mountain which was full of thorns and thistles (tšk, Gr. triboloi), that are the rich ones who in goods and desirable things . . .

§4 [The fourth mountaiŋthey speak of God,] but they do not have [him in] their hearts (read: bycysʾn [pd d](yl) ny ʾst, cf. Gr. epi tēn kardian de mē ekhontes). And their word is living, but in (their) work they are dead. And they stand in two minds (pd dw mnwhmyd ʿystnd, Gr. dipsykhoi).

§5 The fifth mountain, the high one, on which there are many sappy plants, that [are] those who do not accept knowledge and wisdom and who are in themselves quarrelsome and self-complacent (ʿstwrmʾn, Gr. heautois areskontes). And they desire to know and recognize everything. And . . .

§6 [The sixth mountaiŋdea]d. Those they are who are sinful and slanderers, and they shrieked against each other the shrieking of “Say, say, I say, I say.”

§7 The seventh mountain on which many plants and greeŋ[grow] and many creatures and quadrupeds graze, that are those who are always skilful and of good judgment and “soul-gathering” (rwʾncyn). And never is there greediness in them. And about the “children of God” (less likely: of gods, cf. Gr. epi tois doulois tou theou) they are happy and fortunate.

§8 [The eighth] mountain . . .


The Middle Persian text compared with the Greek Shepherd. Already F. W. K. Müller recognized that the Middle Persian text is an abbreviated rendering of the ninth similitude of the Shepherd, which, strictly speaking, does not belong to the text in its original state. It must be an early addition which extensively repeats the contents of the third vision (the building of the tower) and adds the simile of the twelve mountains (Vielhauer, 1966, p. 447). The Middle Persian version in its preserved state begins with the building of the tower (§1), followed by the interpretation of the mountain images. The explanation of mountains one to seven is partly preserved (§§1-7). This arrangement largely follows the Shepherd: description of the twelve mountains of Arcadia and of a central rock (Joly, 1958, pp. 288-93), repetition of the tower vision (Joly, pp. 292-99), interpretation of the tower vision (Joly, pp. 316-29) and of the mountains (Joly, pp. 328-51).

The Middle Persian version and the Greek text of the Shepherd correspond to each other as follows:

MP §1 = Joly, pp. 292-99. 328-31, 332-33

MP §2 = Joly, pp. 332-35

MP §3 = Joly, pp. 334-37

MP §4 = Joly, pp. 336-37

MP §5 = Joly, pp. 336-39

MP §6 = Joly, pp. 338-41

MP §7 = Joly, pp. 340-43

MP §8 = Joly, pp. 342-43


Therefore the Middle Persian text can be rather reliably restored. The only uncertain point is the original length of fragment M 97, since its upper part is torn off, so that the extent of the lacunas at the end of §§1, 3, 5 cannot be determined. In general, one can say, however, that the Middle Persian text seems to have been constructed as a conventional Manichean parable text. Burtea has analyzed a number of terms and images of the Shepherd which are reproduced in the Manichean text, as well as particularly Manichean concepts in the Manichean text (p. 63).

The mountains are, in the Manichean interpretation, twelve groups of men, sinners (mountains 2-6) and righteous ones (mountain 7). Since the tower represents the Christian church in the Shepherd, there is no reason not to assume that it also serves as the symbol of the Manichean church in M 97. The men of mountains 2 to 6 may then symbolize those rocks which are not suitable for the erection of the tower, while the people of mountain 7 are the fitting rocks.

The importance of the Shepherd as a piece of Manichean literature. The question why the Manicheans adopted and interpreted in their way the ninth similitude of the Shepherd (if not even more) suggests itself. The answer of Luigi Cirillo, who has also worked on the Christian text of the Shepherd, is: “The most important reason for which Hermas’ work was read by the Manicheans must have been the doctrine of repentance (metanoia)” (1992, pp. 190-91; cf. 1991, p. 51). Further reasons are, according to Cirillo, the comparability of the Shepherd with the Manichean Nous or (Mani’s) Twin-Spirit (1992, pp. 191-94; 1991, pp. 51-52) and lastly the connection familiar to both the Shepherd and the Manicheans, of revelation, obedience, and salvation (1992, pp. 194-96; 1991, p. 52). Of these three reasons the last one is important in so far as it points to the preservation, noted also in some other cases (Staats, 1993, pp. 102, 106), of certain Jewish and Jewish-Christian traditions in the Shepherd which may have secured the text a place in the Elkhasaite literature and so may have made it accessible to Mani himself. The second reason is somewhat doubtful in so far as we cannot be sure that the Manicheans had a clear idea if what or who the “Shepherd” was. They had a wrong idea, according to Middle Persian M 788, which lists (h)yrmʾ šwbʾn, i.e., “Hermas the Shepherd” (instead of šwbʾn ʿy hyrmʾ), as one of the past apostles of truth (W. B. Henning, “The Murder of the Magi,” JRAS, 1944, p. 142, n. 1). Convincing, however, is Cirillo’s stress on the idea of repentance. In the Manichean text it is explicitly mentioned in the passage on the second mountain and perhaps is applicable to other groups of sinners (other mountains), too. Through repentance sinners could become “fitting rocks” for the building of the church. Thus one can say that many of the people of the mountains are qualified for representing the Manichean church in its manifoldness.

Possible after-effects. Cirillo points out that the Shepherd proclaims a single (renewed) absolution for the believers and that a Manichean Sogdian text, M 139, allows for only one forgiveness of sins (W. B. Henning, Ein manichäisches Bet- und Beichtbuch, APAW 1936, Phil.-hist. K1., Berlin, 1937, pp. 49-50) and threatens lapsed believers with the denial of absolution (Cirillo, 1991, p. 51; 1992, p. 191). But how is this undeniable statement compatible with the weekly confessional practice of the Manicheans and the detailed confessional texts of their eastern communities? Cirillo explains the idea of a unique absolution as an influence of the Shepherd on the Manichean tradition (1991, p. 51; not repeated in 1992, p. 191).

Another case of influence of the Shepherd on Manichean literature has been convincingly discussed by Burtea (2002, p. 55): Verse VIIa,1 of the Parthian Angad rōšnān says: “They will become the bricks (hyštyg) (which are) spoilt and smashed, which are not fit to go up to the keepers of the Building.”



M. Boyce, A reader in Manichean Middle Persian and Parthian, Leiden, 1975, pp. 178-79.

B. Burtea, “Interpretatio manichaica am Beispiel des mittelpersischen Turfan-Fragments M 97,” Annals of the Sergiu Al-George Institute 4-5, Bucharest, 1995-96 [2002], pp. 47-67.

L. Cirillo, “Le Pasteur d’Hermas dans la tradition manichéenne: à propos du fragment M 97 en pehlvi,” Manichaica selecta. Studies presented to Professor Julien Ries on the occasion of his seventieth birthday, ed. A. Van Tongerloo and S. Giversen, Lovanii, 1991, pp. 49-52.

Idem, “ ‘Hermae Pastor’ and ‘Revelatio Manichaica’. Some remarks,” Studia Manichaica. II. Internationaler Kongreβ zum Manichaeismus, ed. by G. Wießner and H.-J. Klimkeit, Wiesbaden, 1992, pp. 189-97.

M. Dibelius, Der Hirt des Hermas, Handbuch zum Neuen Testament, Tübingen, 1923, Ergänzungsband, pp. 415-639.

A. Harnack, Geschichte der altchristlichen Literatur bis Eusebius II, 1, Leipzig, 1958.

A. Hilhorst, “Hermas,” Reallexikon für Antike und Christentum, Lieferung 108/109, Stuttgart, 1988, cols. 682-701.

R. Joly, ed., Hermas Le Pasteur. Introduction, texte critique, traduction et notes, Paris, 1958.

F. W. K. Müller, “Eine Hermas-Stelle in manichäischer Version,” SPAW 51, 1905, pp. 1077-83.

C. Salemann, “Manichaeische Studien I,” Mémoires de l’Académie Impériale des Science de St.-Pétersbourg, 1908.

R. Staats, “Hermas,” Theologische Realenzyklopädie 15, Berlin, 1993, pp. 100-108.

W. Sundermann, Iranian Manichaean Turfan texts in early publications (1904-1934), Photo Edition, London, 1996.

P. Vielhauer, “Apokalyptik des Urchristentums,” in E. Hennecke, Neutestamentliche Apokryphen in deutscher Übersetzung II, ed. W. Schneemelcher, Berlin, 1966, pp. 428-83.

M. Whittaker, ed., Die Apotolischen. Väter II. Der Hirt des Hermas, Berlin, 1956, 2nd ed., Berlin 1967.

(I thank Desmond Durkin-Meisterernst and Christiane Reck for valuable help and advice.)

(Werner Sundermann)

Originally Published: December 15, 2003

Last Updated: March 22, 2012

This article is available in print.
Vol. XII, Fasc. 3, pp. 232-234