² For Mesopotamia, see Zgoll, A., “Einen Namen will ich mir machen!” Die Sehnsucht nach Unsterblichkeit im Alten Orient”, Saeculum. Jahrbuch für Universalgeschichte 54 (2003), pp. 1–11Google Scholar and Radner, K., Die Macht des Namens: Altorientalische Strategien zur Selbsterhaltung (Wiesbaden, 2006)Google Scholar.

³ Thus, unless otherwise specified, in this paper ‘the Epic’ is shorthand for ‘the Standard version of the Epic’.

⁴ The Babylonian Epic's possible survival into other scripts, languages and literary traditions (e.g. Aramaic and Greek) is a different matter, which will not concern us here.

⁵ Since it cannot be assumed that written versions are always exact witnesses to performed or otherwise oral ones, we know nothing of the Epic's oral evolution. Note however the contention of Civil, M., “Reading Gilgameš”, Aula Orientalis, 17/18 (1999–2000), p. 188Google Scholar, that a stable textual history is suggestive of “very ritualised performances in a cultual or royal court setting”, whereas the textual history of a composition with “informal performances in settings that could be called private” would tend to be more varied. (I owe this reference to S. Dalley).

⁶ Kermode, F., Shakespeare's Language (London, 2001), footnote p. 115Google Scholar.

⁷ The word ibrī ‘my friend’ is absent from line 51 on MS e (Sultantepe), but this does not affect the argument above.

⁸ This appears to have been observed only by West, M. L., The East Face of Helicon: West Asiatic Elements in Greek Poetry and Myth (Oxford, 1997), p. 338Google Scholar: “In the main body of the poem [i.e. excluding Tablet XII, which is a secondary appendage] the two men are equals, and address one another consistently as ‘my friend’.” (The switch in Tablet VIII is however not noted, nor are the addresses with personal names in earlier versions of the Epic). West's observation deserves wider notice in Assyriological literature.

⁹ In the secondarily appended Tablet XII, Enkidu addresses Gilgameš as ‘my lord’. As noted by West, Helicon, p. 338, this is due to the tablet's Sumerian origin, and reflects its non-integration into the foregoing eleven tablets.

¹⁰ The Chicago Assyrian Dictionary, vol. I–J, p. 7a, offers the following comment on the meaning of ibru ‘friend’: “The word denotes an institutionalized relationship between free persons of the same status or profession which entailed acceptance of the same code of behavior and an obligation of mutual assistance.” On the strength of this, one might wonder whether the avoidance of personal names was constitutive of the ibru-relationship. However, as we shall see below, in other versions of the Epic Gilgameš and Enkidu do address each other by name, so this explanation does not convince.

¹¹ Etana, in which certain characters also address each other as ibrī ‘my friend’, is as yet too fragmentary to yield clear patterns.

¹² L. Siddall (personal communication) observes that Gilgameš's utterance of Enkidu's name is suggestive of funerary ritual (the kispum-ceremony), in which the name of the deceased had a prominent place. (For an interpretation of the circumstances of Enkidu's death in ritual terms see Scurlock, J. A., “Soul Emplacements in Ancient Mesopotamian Funerary Rituals”, in Magic and Divination in the Ancient World, (eds) Ciraolo, and Seidel, (Leiden, 2002), pp. 1–6Google Scholar).

¹³ Also worth noting is that at OB Sch₁ obv.2′ Gilgameš speaks Enkidu's name. This occurrence is, as mentioned above, not attested in the preserved portions of the Standard Version, but as he is not directly addressing Enkidu by name, it does not, strictly speaking, contravene the pattern under discussion.

¹⁴ As observed by George, Gilgamesh, p. 354, the wording on Assyrian manuscript y₁ lines 6′–9′ is “very like the Yale tablet (OB III 141–57)”. If OB III did have Gilgameš and Enkidu consistently address each other as ibrī ‘my friend’, then this would be one respect in which Assyrian manuscript y₁ (see line 7′, cited above) differs from OB III. Whether this would mean that an ancient redactor consciously ruptured the pattern, or simply added the vocative ‘Enkidu’ without being aware that this contravened a pattern, cannot at present be ascertained.

¹⁵ Jahn, M., “Focalisation”, in The Cambridge Companion to Narrative, (ed.) Herman, D. (Cambridge, 2007), p. 94CrossRefGoogle Scholar. Jahn's chapter offers a useful survey of the issue.

¹⁶ As reported by Jahn, p. 97, this distinction was made by Genette, , Narrative Discourse (Oxford, 1980), p. 186Google Scholar. For an insightful application of the distinction between ‘who speaks’ and ‘who sees’ to Beowulf (without use of the term ‘focalisation’), see Lumiansky, R. M., “The Dramatic Audience in Beowulf”, Journal of English and Germanic Philology, 51 (1952), pp. 545–550Google Scholar.

¹⁷ To ‘see’ in this context does not just refer to the sequence of images conjured up by the narrative but can also refer to things such as the emotions which the images cause, etc.

¹⁸ Abbott, H. Porter, The Cambridge Introduction to Narrative (Cambridge, 2008), p. 73CrossRefGoogle Scholar.

¹⁹ Porter Abbott, Introduction, p. 74.

²⁰ In line 204 we follow the manuscript tradition with -šu ‘his’ (not -ša ‘her’), on the grounds that, as noted by George, Gilgamesh, p. 799, “Šamhat's observation in l. 207 might logically follow a prolonged gaze at his face”. Also, we feel that the syntax of harimtu inaṭṭala panīša ‘He watched the harlot's features’ would be odd here.

²¹ The function of line 213 may be precisely to alert us to the fact that more words were spoken than are recorded.

²² Thus also Bottéro, J., L'Epopée de Gilgamesh (Paris, 1994), p. 69Google Scholar fn. 1 and Parpola, S., “The Esoteric Meaning of the Name of Gilgamesh”, in Intellectual Life in the Ancient Near East, (ed.) Prosecký, J. (Prague, 1998), pp. 318–319Google Scholar. This interpretation admittedly contains a slight anachronism, as the Sumerian for ‘to create (a human)’ is dím rather than dù (I thank K. Zand for this observation), but by the time at which the Standard version of the Epic was redacted (whether late in the second millennium or early in the first) Sumerian had died out as a vernacular language, so this is not a real problem: the Gilgameš redactors were thinking of Akkadian names of the form ‘God Name + a form of banû’, and viewing dù in Enki-dù as equivalent to Akkadian banû (cf. Parpola p. 318 fn. 13; as noted by Parpola with reference to Galter, H. D., Der Gott Ea/Enki in der akkadischen Überlieferung: Eine Bestandsaufnahme des vorhandenen Materials (Graz, 1983)Google Scholar, the names Ea-bāni and Enki-bāni are attested; I thank J. Scurlock and R. Beal for their comments on this point).

²³ Italics in the translation signify uncertainty.

²⁴ In George's edition the line is translated “and also Enkidu himself, whose birthplace was the hills”. The issue of how to translate the line will be discussed below.

²⁵ See the comment by Wasserman, N., “Offspring of Silence, Spawn of a Fish, Son of a Gazelle . . .: Enkidu's Different Origins in the Epic of Gilgameš”, in An Experienced Scribe Who Neglects Nothing: Ancient Near Eastern Studies in Honor of Jacob Klein, (eds) Sefati, , Artzi, , Cohen, , Eichler, and Hurowitz, (Bethesda, 2005), p. 595Google Scholar: the name “Enkidu” is “first introduced in the epic . . . when the human, civilized surrounding, in the person of the harlot Šamhat approaches the nameless creature”.

²⁶ Perhaps the idea that the name was chosen by Šamhat was introduced to the Standard version in order to explain why a character called ‘Enki created him’ was in fact created by Aruru.

²⁷ Enki/Ea may have played a part in the creation of Enkidu in a Middle Babylonian version. See George, Gilgamesh, p. 289.

²⁸ See Stamm, J. J., Die akkadische Namengebung (Leipzig 1944; MVAG 44)Google Scholar, § 6.2, followed e.g. by Edzard, D. O., “Namen und Namengebung. B”, Reallexikon der Assyriologie IX (Berlin, 2001)Google Scholar, § 4, p. 109 and Kalla, G., “Namengebung und verwandschaftliche Beziehungen in der altbabylonischen Zeit”, in Altorientalische und Semitische Onomastik, (ed.) Streck, M. P. (Münster, 2002), p. 125Google Scholar.

²⁹ As N. Wasserman observes (pers. comm.), u šū appears also at I 202 (u šū īši ṭ[ēma r]apaš hasīsa ‘That one, he (now) had reason, he was wide of understanding’). The fact that in this line šū is an independent verbal subject coheres with our view that in line I 274 it is not simply an attribute of Enkidu (i.e. not ‘Enkidu himself’). (One might wonder whether the use of u šū at I 202 is a sign that the couplet I 201–202 is also focalised through Šamhat, but in the present state of knowledge this must remain highly speculative).

³⁰ Schott, A. and von Soden, W., Das Gilgamesch-Epos (Stuttgart, 1988), pp. 20–21Google Scholar.

³¹ Dalley, S., Myths from Mesopotamia: Creation, The Flood, Gilgamesh and Others (paperback edition) (Oxford, 1991), p. 55Google Scholar.

³² George, Gilgamesh, p. 549.

³³ Maul, S. M., Das Gilgamesch-Epos (München, 2008), p. 52Google Scholar.

³⁴ Röllig, W., Das Gilgamesch-Epos (Stuttgart, 2009), p. 41Google Scholar.

³⁵ Two MSS are sufficiently preserved to show that ša is absent. MS P: ù šu-ú ^den-ki-dù i-lit-ta-šú šá-du-um-ma; MS x: [. . .-k]i-dù i-lit-t[a-. . .]. As we shall argue elsewhere (and some scholars have already recognised), a priori it is not out of the question to emend Akkadian manuscripts even when they are unanimous. Here, however, emendation seems unnecessary.

³⁶ A construction after GAG § 166 (ša-less relative clause) would here produce a restrictive relative clause with the sense ‘Enkidu-whose-birth-place-is-the-mountains’, and so should be rejected. Also, it is unclear whether such a construction is possible with a name.

³⁷ The hunter (speaking to his father at I 123 and to Gilgameš at I 150) simply calls him eṭlu ‘fellow’. Ninsun does not appear to know Enkidu's name when she interprets Gilgameš's dreams as omens of Enkidu's arrival.

³⁸ Heidel, A., “A special usage of the Akkadian term šadû”, JNES, 8 (1949), pp. 233–235Google Scholar (with reference to Jensen and Poebel) argued that sometimes the word šadû has a meaning similar to ṣēru. This may well apply to the annals of Tukultī-Ninurta II (now RIMA 2 p. 174: 63), where north-west Semitic influence (cf. Hebrew śadè ‘plain, open country’) may mean that huribtu ša šadê is ‘desert (road) of the steppe’ (thus Heidel; but cf. A. K. Grayson in RIMA, ‘a desolate mountainous region’). However, though they were accepted by both dictionaries, the examples which Heidel adduces from various versions of the Epic of Gilgameš are dubious: here it is possible that a distinction is made between šadû and ṣēru. This would mean that the relevant passages draw on two sources of imagery, rendering them poetically richer.

³⁹ For references to personified mountains, see the Chicago Assyrian Dictionary, vol. Š/i, pp. 57–58 and F. Wiggermann, RlA s.v. “Mischwesen”.

⁴⁰ See Tigay, J., The Evolution of the Gilgamesh Epic (Philadelphia, 1982)Google Scholar.

⁴¹ Other indications of unity have also been suspected, but they are less sure: (a) Tigay, Evolution, p. 12 has stated that the Standard Babylonian version “is characterised by a distinctive style”. However, no examples were offered, and the notion of a distinctive style has yet to be demonstrated. Cf. the remarks of George, Gilgamesh, p. 431: “The Standard Babylonian epic is, like the older versions of the poem . . ., characterised by a lack of ornate style and very sparing use of language of the kind called ‘hymno-epic’. It is not the purpose of this introduction to write comprehensively on the literary style of the poem. Such a study would involve the examination of the many other Standard Babylonian literary texts written in the same idiom.” (b) Dalley, Myths, p. 46 has discerned strategies for promoting artistic unity in the use of formulaic phrases to introduce direct speech. However, she notes that the use of phrases to introduce direct speech are “not quite standardised throughout the epic”, testifying to “its composite nature”. (c) Dalley (ibid.) also sees such a strategy in Gilgameš himself being “present almost the entire time, although sometimes as the silent audience to another man's tale”, but other interpretations of this point are possible.

⁴² This seems to be an innovation vis-à-vis the early second millennium version(s), which lacked the Prologue (while no manuscript of any early second millennium version's Tablet I survives, the first line doubled as the title, and so appeared on the colophons of manuscripts of other Tablets). However, since no manuscript of the final Tablet of any of the early second millennium versions is extant, one cannot rule out the possibility that one or more of the early second millennium versions also ended by repeating words which occurred at the beginning, though the words in question would be different from those repeated in the Standard version.

⁴³ See e.g. Watkins, C., How to Kill a Dragon: Aspects of Indo-European Poetics (Oxford, 1995), p. 34Google Scholar etc.

⁴⁴ E.g. Dalley, Myths, p. 49.

⁴⁵ Cooper, J. S., “Gilgamesh Dreams of Enkidu: The Evolution and Dilution of Narrative”, in Essays on the Ancient Near East in Memory of Jacob Joel Finkelstein, (ed.) de Jong Ellis, M. (Hamden, Conn., 1977), pp. 39–44Google Scholar.

⁴⁶ For example, (a) Tigay, Evolution, p. 244 holds that while “hostility to Ishtar” has been abandoned in the Standard version “to the extent that her claim to the Inanna temple is recognised”, hostility towards the goddess “is still reflected atavistically in the Bull of Heaven episode”. Thus, for Tigay, the epic's multilayered history caused the Standard version to be inconsistent in its portrayal of Ištar. (b) See Dalley's remark on the epic's “composite nature” sub (b) in fn. 41 above. (c) Cooper, “Gilgamesh Dreams of Enkidu”, p. 40 holds that “clumsiness” of conflation leads to a situation in the Standard version where “we are asked to imagine Gilgamesh making love to an axe in his mother's presence!”. (It is, however, deeply doubtful whether râmu ‘to love’ must refer to sexual intercourse (though of course it can, as at I 300), for at I 241 Šamhat tells Enkidu that gilgameš šamaš irâmšu-ma ‘Gilgameš is loved by Šamaš’).

⁴⁷ See e.g. the insightful discussion by Walls, N., Desire, Discord and Death: Approaches to Ancient Near Eastern Myth (Boston, MA, 2001), p. 31Google Scholar on the switch from rehû ‘to inseminate’ to rutammu ‘to make love together’ in Tablet I: “The poet's choice of the Dt-stem conjugation in its reciprocal sense emphasizes the mutuality of the actions and the equality of both characters, male and female, as verbal subjects. Enkidu is no longer the rutting bull who abruptly mounts the passive female; Enkidu now knows eroticism, perhaps even affection (urta”amū, ‘they made love’ [I 300, restored after OB II 46, MW]), in contrast to his original beastly mating (irhi, ‘he inseminated’ [I 194, MW]).” (One could additionally connect the use of the verb râmu at I 300 to its appearance in Gilgameš's dreams in the preceding lines.)

⁴⁸ In the lines which follow VII 184, the inhabitants of the underworld are several times referred to with feminine plural morphemes on all extant manuscripts. The usage is odd. We doubt that the feminine plurale tantum nišū ‘people’ should be understood to underlie the passage, for generic statements about groups of people tend to gravitate from feminine plural to masculine plural rather than the other way round. A solution might be to presume that the underlying noun is the feminine plural eklētu, which should then be understood to mean ‘shades, ghosts’ rather than ‘shadows, dark places’. The semantic shift per se is plausible (cf. English ‘shades’), but we can adduce no further attestations of eklētu in this meaning.

⁴⁹ Should we suppose that Gilgameš was conscious of the echo? If so, it would be an utterance of bitter self-irony.

⁵⁰ Gilgameš does not yet know that Ūta-napišti has another opportunity up his sleeve (though this too will result in failure), so from his own point of view he has completely failed in his quest for eternal life.

⁵¹ Where MS T offers [. . . (a)]-a-i-ka-a (our ayyiká above), MS J has a-a-ka-ni. This is problematic, as -āni seems not to be used as an adverbial ending in Standard Babylonian – indeed it is rare already in Old Babylonian. (The discovery of -āni as an Old Babylonian adverbial ending was made by Farber, W., Studies Kraus (Leiden, 1982), pp. 37–47Google Scholar. For more attestations see W. R. Mayer, Or 64 (1995) pp. 163–164 and N. Wasserman, Style and Form, 116 with fn. 105). One could emend to a-a-ka-ni-iš, but this is not a good solution: although -āniš is morphologically acceptable (even though it would be semantically and syntactically odd), ayyikāniš would be hapax. Thus we suppose that -ni on MS J originated with a misidentification of -a, so that the Vorlage ran a-a-ka-a. (Cf. George's comment to line 134, p. 889).

⁵² The orthography and morphology of MS T seem to conform sufficiently closely to second millennium traits that uṣ-ṣab-bi-tu ₄ ek-ke-mu should be taken as plural (uṣṣabbitū ekkēmū) rather than as a vowel-indifferent spelling of the singular (uṣṣabbit ekkēmu). MS J's [uṣ-ṣab-bi-t]a ek-ke-mu (singular, uṣṣabbit ekkēmu) is however a legitimate (and indeed possibly superior) variant.

⁵³ Cooper, “Gilgamesh Dreams of Enkidu”, p. 41.

⁵⁴ Cooper, J. S., “Symmetry and Repetition in Akkadian Narrative”, JAOS, 97 (1977), p. 510Google Scholar.

⁵⁵ George, Gilgamesh, pp. 32–33: “The [early second millennium] poem that told of the glorious feats and heroic exploits of the mightiest king of old was recast [in the Standard version] as a sombre meditation on the doom of man. In the course of these changes much of the vitality of the original poetry may have been dissipated, but the poem became a vehicle for more than entertainment.”