The Texts

Definitions of Hermes Trismegistus to Asclepius

The Definitions of Hermes Trismegistus to Asclepius are a short book of sayings, and Jean-Pierre Mahé, who edited them in 1982, argued that they are older than the treatises of the Corpus Hermeticum that made Hermes Trismegistus famous. The claim is serious and disputed, and few readers can weigh it for a plain reason: no public-domain English translation exists. The internet quotes what it can copy freely; the Definitions never entered that economy.

What do we know about the Definitions for certain?

Everything firm about the Definitions comes from medieval copies and modern editions, and all of it can be dated: Armenian manuscripts from the 13th century on, Greek excerpts in a manuscript of the late 13th or early 14th century, and Mahé’s editions of 1982 and 2019. The card gives each key fact, the evidence for it and its status.

Checked on 2 November 2026, each source at the link in the card or in the list of sources below.

Are the Definitions the oldest Hermetic text?

Only on Jean-Pierre Mahé’s reconstruction, and even then only among the philosophical Hermetica: no copy of the Definitions older than the 13th century survives, and their early date rests on parallels with other texts. In 1976 Mahé argued, from the number of its parallels with other Hermetic texts, that the Armenian faithfully renders a Greek work at least as old as the other Hermetica, and noted that several of its definitions appear woven into treatises of the Corpus Hermeticum. In his edition of 1982 and in his article for the Encyclopedia of Religion (1987) he went further: saying 9.4 is the source of Corpus Hermeticum I.18, a treatise he placed before the expulsion of Egypt’s Jews after the revolt of 115-117 because it echoes Jewish liturgy, so the Definitions date from the 1st century CE at the latest and could be older.

Reviewing the 1982 volume, Daniel A. Bertrand (1985) judged the text’s antiquity far from assured, despite a comparison by Gilles Quispel that Mahé invokes, and suspected that Mahé, as the scholar who had in effect discovered the Armenian text, had overrated its importance. Even Mahé’s early date would not make the Definitions the oldest writing under Hermes’ name: in the same encyclopedia article he reports André-Jean Festugière’s view that Hermetic astrology began in the 3rd or 2nd century BCE. The evidence supports something narrower: the Definitions may preserve one of the earliest layers of philosophical Hermetic teaching, and that is how this page treats them.

What is in the Definitions?

The Definitions are ten chapters of short statements, sentences rather than dialogues, in which Hermes gives Asclepius compressed teaching on God, the cosmos, the mind and the human being: the same concerns that the longer dialogues unfold. Do not confuse them with Corpus Hermeticum XVI, the Definitions of Asclepius to King Ammon: there Asclepius writes to a king; here Hermes teaches Asclepius. Mahé (1982) built a theory of Hermetic writing on them: the basic unit is the sentence, the Greek gnômê, and sentences set side by side, chained, commented, turned into myths or placed in prayers grew into the treatises of the Corpus and the Asclepius; he traced the form back to Egyptian instructions handed down by a “father” to a “son”. If he is right, the tradition literally begins as a list of definitions, an idea worth holding next to the grand visions of nous elsewhere in the corpus. The Armenian text also carries an eleventh section that Mahé identifies as an interpolation drawn from Nemesius (c. 390 CE).

Why does a Greek text survive only in Armenian?

The Definitions survive because Armenian translators rendered them, as they rendered many Greek philosophical works, and Armenian scribes went on copying the translation; the Greek original is lost except for excerpts. The version is very literal, in the manner of the so-called Hellenizing school. Mahé dates it to the second half of the 6th century (1987), while Sen Arevshatyan placed it between the late 450s and the 480s (1973, as quoted by Aelita Dolukhanyan in 2016). The oldest of the six manuscripts used for the first edition was copied before 1282 by the scribe Mkhitar Ayrivanetsi at the monastery of Geghard, in a miscellany that also holds Porphyry’s Introduction, the Art of Grammar of Dionysius Thrax and the Definitions of Philosophy of David the Invincible.

The text was first printed in Yerevan in 1956, in Hakob Manandyan’s edition with a Russian translation by S. Arevshatyan. Heinrich Dörrie reviewed it in 1957; two French translations followed in 1976, by Mahé and by M. G. de Durand; and Mahé’s critical edition of 1982, based on the most conservative branch of the manuscripts, differs markedly from Manandyan’s text.

The Greek came back by another road. In Oxford, in Bodleian Library MS Clarke 11, a manuscript of the late 13th or early 14th century, Joseph Paramelle recognised a Hermetic anthology that includes excerpts of the Definitions; he and Mahé published them in 1990-1991, and Mahé (2020) notes that the manuscript is older than every complete copy of the Corpus Hermeticum. The Armenian also helps to read the Greek of other Hermetica: in 1981 Mahé used one definition to read a sentence of Stobaean excerpt 19.1 differently from Festugière’s reconstruction. The same compressed form survives in Greek in another excerpt kept by Stobaeus (SH 11), where Hermes sums up his teaching to Tat in forty-eight short headings; Agata Sowińska (Vox Patrum, 2022), following Mahé, reads them beside these definitions (see The Stobaean Fragments).

Are the Definitions part of the Corpus Hermeticum or of Nag Hammadi?

No: the Definitions are a separate work that shares sayings with the Corpus Hermeticum, and they were not found at Nag Hammadi. Mahé published them in 1982 in the Québec series Bibliothèque copte de Nag Hammadi, in the same volume as the Coptic Perfect Discourse of Codex VI (the Coptic version of Asclepius 21-29), and in 2019 he set them again beside the Hermetic texts of Codex VI. Bertrand (1985) pointed out that the Definitions are foreign to the Nag Hammadi library; what joins them to it is Mahé’s argument about Hermetic sayings, not the find.

How can you read the Definitions today?

In English there is one published translation, by Mahé himself, in Clement Salaman, Dorine van Oyen, William D. Wharton and Jean-Pierre Mahé, The Way of Hermes (1999), which pairs a readable Corpus Hermeticum with Mahé’s translation of the Definitions. Its treatises carry the usual numbers, written “Book 1”, “Book 4” and so on, so its Book 4 is CH IV, not the Book 4 of Everard’s Divine Pymander (CH X); the concordance of Corpus Hermeticum numberings tells the systems apart. The Way of Hermes is in copyright: there is no legitimate free copy, and we don’t link pirated ones. That single fact explains the text’s obscurity better than any conspiracy: one of the tradition’s earliest voices is simply behind a bookshop counter. Readers of French have Mahé’s Budé volume of 2019, with the Armenian text, the Greek parallels and notes. Either way, the Definitions are the sharpest test of whether a reader wants Hermetic history or Hermetic branding.

Key facts

Greek original	Lost except for excerpts; the 1st century CE at the latest according to Mahé (1982), a dating that is disputed
Armenian version	Literal; 6th century (Mahé) or 5th (Arevshatyan); six manuscripts in Yerevan, 13th-16th century
Greek excerpts	Oxford, Bodleian Library, MS Clarke 11 (late 13th or early 14th century), published 1990-1991
First printed	Yerevan, 1956 (Manandyan, with a Russian translation)
Editions	Mahé, Hermès en Haute-Égypte II (1982); Budé, Hermès Trismégiste V (2019)
English translation	Mahé, in The Way of Hermes (1999)
Form	Sayings (definitions) in ten chapters, plus an interpolated eleventh section

Fair questions

Are the Definitions of Hermes the oldest Hermetic text?

Not as an established fact. Jean-Pierre Mahé (1982, 1987) dated the collection to the 1st century CE at the latest and counted such books of sayings among the oldest philosophical Hermetica, while Daniel A. Bertrand (1985) judged its antiquity far from assured. By André-Jean Festugière’s dating, astrological writings under Hermes’ name began earlier, in the 3rd or 2nd century BCE.

Are the Definitions really older than the Corpus Hermeticum?

Possibly. Mahé argued that saying 9.4 is the source of Corpus Hermeticum I.18, a treatise he dated before the expulsion of Egypt’s Jews after the revolt of 115-117, so the Definitions would be earlier. The argument rests on parallels, not on old copies: the oldest Armenian manuscript was written before 1282, and the Greek excerpts are in a manuscript of the late 13th or early 14th century.

Why does a Greek text survive in Armenian?

Armenian translators rendered many Greek philosophical works, this one very literally; the Greek original was then largely lost, so the Armenian copies, six of them in the Matenadaran in Yerevan, became the text’s lifeboat. Mahé dates the translation to the 6th century, Sen Arevshatyan to the 5th.

Is there a free English translation?

No. The only published English translation is Mahé’s, in The Way of Hermes (1999), which is in copyright, and no legitimate free copy exists. That absence, not suppression, is why quote sites and popular esoteric books almost never mention the Definitions. Readers of French can use Mahé’s Budé edition of 2019.

Are the Definitions part of the Corpus Hermeticum or the Nag Hammadi codices?

Neither. Corpus Hermeticum XVI, the Definitions of Asclepius to King Ammon, is a different text, and the Definitions were not found at Nag Hammadi; Mahé edited them in 1982 in the same volume as the Coptic Perfect Discourse of Codex VI.

Sources & further reading

• Jean-Pierre Mahé, Hermès en Haute-Égypte, 2 vols. (Québec: Presses de l’Université Laval, 1978-1982); the Definitions are in vol. II, Le fragment du Discours parfait et les Définitions hermétiques arméniennes (NH VI, 8.8a) (1982), Bibliothèque copte de Nag Hammadi, Textes 7.
• Jean-Pierre Mahé (ed. and tr.), Hermès Trismégiste, vol. V, Paralipomènes: Codex VI de Nag Hammadi, Codex Clarkianus 11 Oxoniensis, Définitions hermétiques, Divers, Collection des Universités de France 546 (Paris: Les Belles Lettres, 2019), publisher’s page.
• Clement Salaman, Dorine van Oyen and William D. Wharton (the Corpus Hermeticum) and Jean-Pierre Mahé (the Definitions), The Way of Hermes (London: Duckworth, 1999; Rochester, Vermont: Inner Traditions, 2000), catalogue record.
• Jean-Pierre Mahé, “Les Définitions d’Hermès Trismégiste à Asclépius (traduction de l’arménien)”, Revue des sciences religieuses 50 (1976), pp. 193-214, Persée.
• M. G. de Durand, “Un traité hermétique conservé en arménien”, Revue de l’histoire des religions 190 (1976), pp. 55-72, Persée.
• Jean-Pierre Mahé, “Stobaei Hermetica XIX, 1 et les définitions hermétiques arméniennes”, Revue des Études Grecques 94 (1981), pp. 523-525, Persée.
• Daniel A. Bertrand, review of Mahé, Hermès en Haute-Égypte II, Revue d’histoire et de philosophie religieuses 65 (1985), p. 327, Persée.
• Jean-Pierre Mahé, “Hermes Trismegistos”, tr. Paul C. Duggan, in Encyclopedia of Religion (1987; 2nd ed., 2005), Encyclopedia.com.
• Joseph Paramelle and Jean-Pierre Mahé, “Nouveaux parallèles grecs aux Définitions hermétiques arméniennes”, Revue des Études Arméniennes 22 (1990-1991), pp. 115-134.
• Joseph Paramelle and Jean-Pierre Mahé, “Extraits hermétiques inédits dans un manuscrit d’Oxford”, Revue des Études Grecques 104 (1991), pp. 109-139, Persée.
• IRHT, Pinakes: Oxford, Bodleian Library, Clarke 11 (diktyon 47766), record.
• M. David Litwa, Hermetica II (Cambridge University Press, 2018), Cambridge Core; general introduction in the publisher’s excerpt.
• Jean-Pierre Mahé, “Monumentum aere perennius: Hermès trismégiste, soixante-dix ans après l’édition Nock-Festugière”, Comptes rendus de l’Académie des Inscriptions et Belles-Lettres 164 (2020), pp. 153-168, Persée.
• Aelita Dolukhanyan, “The main theses of the Armenian translation of the Definitions by Hermes Trismegistus”, Wisdom 2 (7) (2016), pp. 136-141, CyberLeninka.
• Agata Ewa Sowińska, “[Hermes Trismegistos] Stobaei Hermetica XI”, Vox Patrum 81 (2022), pp. 233-280, DOI.
• Álvaro Basols Berges, review of Jorge Cano Cuenca, Textos herméticos (Madrid: Alianza, 2024), Cuadernos de Filología Clásica. Estudios griegos e indoeuropeos 36 (2026), pp. 367-372, DOI.