Chapter 5. Oral-Formulaic Theory and Iranian Royal Inscriptions

Literary parallels between the Achaemenid inscriptions of the sixth and fifth centuries BCE, especially those of Darius I (late sixth century BCE), and the Sasanian inscriptions of the third and early fourth century CE had already been recognized in the 1930s and were attributed to the survival of Aramaic translations of the former. [1] In more recent times, however, these parallels have been discussed in light of the principles of oral literature, and modern scholarship increasingly favors an oral literary tradition as the source for the formulaic and thematic parallels between these two corpora of inscriptions.
In 1985 Prods Oktor Skjærvø published an exhaustive study of parallels between Achaemenid and Sasanian inscriptions. [2] He attributed these parallels to a “local (oral) literary tradition,” and concluded: “It does not seem to be too far-fetched an assumption that Darius, Xerxes, Šābuhr, Narseh, and Kerdīr all may have used literary language and formulae which they had so to speak learnt in school.” [3]
In 1990, in a review article, Philip Huyse analyzed these parallels in the light of Parry and Lord’s theories. [4] Huyse suggested a royal epic tradition (profane Überlieferung über die Könige) existing beside the sacred-religious tradition (sakral-religiöse Überlieferung) as a possible source for the parallels in the two groups of inscriptions (Achaemenid and Sasanian). [5] In order to demonstrate the traces of such an oral epic tradition in Iran, Huyse analyzed the inscriptions in relation to three criteria: formula, theme, and story pattern, the last two being those already described by Skjærvø. [6]
Since Huyse’s approach was exclusively diachronic, he only examined formulae (word groups) with similar counterparts in both sets of texts. [7] With respect to thematic and structural parallels, Huyse identified several common themes already described by Skjærvø [8] and deduced from the similarity in the structure of composition (story pattern) of the inscriptions of Darius I and Šābuhr I [9] the existence of an oral epic tradition responsible for parallels between the two epigraphic corpora. [10] Huyse’s important study, [11] in spite of the definition of the formula as a metrical wordgroup by Parry and Lord, did not address the question of the metrical nature of the formula, nor provide a new definition of the formula that would permit us to disregard the meter. Moreover, the orality of the Achaemenid and Sasanian inscriptions had been hitherto measured exclusively by the occurrence of formulaic parallels between them, and the investigation had therefore been mostly restricted to formulaic expressions attested in both corpora, but these in reality represent only an insignificant segment of these inscriptions. However, one of the yardsticks of oral composition is high formulaic recurrence, or formulaic density, that is, the concentrated presence of similar word clusters to express a specific theme. Measured by such a yardstick the few and scattered formulaic expressions pointed out so far may not constitute a complete proof of the orality of the inscriptions and an oral path of transmission responsible for the parallels. In order to demonstrate that the parallels between the Achaemenid and Sasanian inscriptions were indeed dependent upon an oral tradition, they ought to be investigated from a synchronic perspective and analyzed individually for formulaic density.
In several publications, [12] Skjærvø dealt with the relationship between oral traditions, the Iranian royal inscriptions, and Avestan hymns. In these articles he reiterated and redefined his basic position as regards the composition of the Iranian inscriptions in light of his studies in oral (epic) literature. With respect to the sources of Old Persian inscriptions, Skjærvø identified, beside the Near Eastern literary models, two other possible sources: (1) Indo-Iranian inherited themes and (2) Avestan themes and expressions. As for the latter, he proposed two alternative interpretations: (a) either the similarities in themes and formulae between Avestan texts and Old Persian inscriptions were due to a common Iranian heritage reaching back to the pre-Avestan period and transmitted over a period of more than a millennium, or (b) they represented quotations from the Avesta. Although Skjærvø did not exclude the possibility of a common Iranian heritage as the cause of the parallels between Old Persian inscriptions and the Avesta, he seemed to favor a direct influence of the Avesta. This conclusion was slightly different, however, from that of a prior study on royalty in early Iranian literature, [13] in which he had suggested that the parallels between the Achaemenid inscriptions and the Avesta came from an oral tradition in which they both were versed. [14] Skjærvø also offered a scenario for the composition of the royal inscriptions that might account for the presence of formulaic patterns typical of oral tradition: (1) the res gestae of the kings were composed by professional composer-performers who would cast their narratives in the patterns and mold of the oral tradition; (2) the deeds were subsequently conveyed to the king’s subjects by official heralds; and finally, (3) the accounts of the heralds would be retold by storytellers in whose compositions the narratives of the kings would be altered over and over again. [15]
In this study, an attempt will be made to establish that the similarities between Achaemenid and Sasanian inscriptions are not due to a common Iranian heritage, but to the decisive influence that the Achaemenid inscription of Darius I at Bisotun exercised on the oral (epic) traditions, which again served as a source of emulation for Sasanian epigraphy. Moreover, we shall investigate, through the paradigm of Darius’ inscription at Bisotun, whether Old Persian inscriptions under Darius I and Xerxes were original compositions by professional composer-performers, or were adaptations and recompositions of other versions of the Achaemenid inscriptions, including those in Elamite. However, before we can proceed to the discussion of the Iranian epigraphic material from the perspective of orality, we shall briefly outline some of the main tenets of the Oral-Formulaic School.

The Oral-Formulaic School

In the following we shall discuss some aspects of the Oral-Formulaic School as established by Parry and Lord, as well as the results of recent contributions in this field. [16]

The Formula

Parry worded his definition of the “formula” as “a group of words which is regularly employed under the same metrical conditions to express a given idea.” [17] This phrasing, however, presents several problems for our purposes here.
First, Parry’s conception of the formula concentrates solely on its verbal and metrical aspects and does not consider the syntactical dimension, which has increasingly come to the fore in recent years. [18] Second, this definition was based upon a single frame of reference, namely the Homeric poems, whose metrical features are determined by syllabicity and the internal structure of the poetic line. Although he and Lord were later able to apply this conception of the formula to the Serbo-Croatian tradition, which exhibits prosodic characteristics comparable to those of Homeric verse, to apply it to the manifold oral traditions of the world, this conception needs to be redefined and adapted to their individual metrical foundations. [19] Third, the vagueness of the very term “formula” poses severe problems. In the Homeric and Serbo-Croatian traditions, the “formula” is, from a synchronic point of view, framed by the meter, which delimits its length, [20] but Gregory Nagy has shown that it is the formula (the phraseological unit) that diachronically conditions, and even creates, the meter (the metrical unit), although the metrical unit can synchronically develop its own dynamic and regulate the new incoming phraseological units:
While traditional meter and traditional phraseology are formally correlated in origin, their evolution may diverge. At the start, traditional phraseology sets and then regulates quantitative patterns (= meter) by force of precedent, but the resulting fixed meter evolves dynamics of its own and becomes the regulator of any incoming nontraditional phraseology. By becoming a structure in its own right, meter may also develop independently of traditional phraseology. Thus it could even happen that newly-developed laws of meter may obliterate aspects of the very phraseology that had originally engendered them, if these aspects no longer match the meter. [21]
Matters become more complex, however, when we attempt to deal not with oral poetry, but with oral prose. How is the formula to be recognized without its metrical frame? Even though diachronically the formula generates the meter, without the meter we have no yardstick for identifying absolutely the formula from the perspective of synchrony. The solution to this problem must be sought, at least in part, in a new definition of the formula, one which de-emphasizes the meter as a conditioning factor. Such a redefinition must rely essentially on the theme as the diachronic factor that provoked the fixation of phrasing in formulae and thus triggered the creation of the meter. The definition of the formula by Nagy—as “a fixed phrase conditioned by the traditional themes of oral poetry” [22] —goes a long way to fulfill the needs of the student of prose formulae, but still does not suffice to “frame” the formula.
As the traditional procedure for examining a “text” to discover traces of orality is to determine the density of formulaic recurrence, the need for a well-defined criterion enabling us to determine a “formula” in a metrical context, or in prose, becomes evident, for in the absence of such a yardstick, the degree of orality of a “text” could be measured only roughly, if at all.

Formulaic Recurrence

Aside from the problem of definition, the use of density of formulaic expression as a measure for orality has also been seriously contested. Larry D. Benson, in an influential article entitled “The Literary Character of Anglo-Saxon Formulaic Poetry,” demonstrated that high-density formulaic phraseology does not constitute an infallible gauge for oral composition. [23] He revealed that several Old English poems, likely to have been composed in writing, showed the same formulaic density as two Old English oral compositions. Thus, the recurrence of formulae could not serve as a gauge for the orality of a given composition. [24] What is more, the formula itself, whose recurrent presence was thought to be the mark of oral literature, was deemed too vague a concept for one to base the edifice of oral composition upon it.
Benson’s findings were subsequently criticized by several scholars, [25] among them Zwettler, who by basing himself on the work of the medievalist H. J. Chaytor—dealing with the publication and circulation of poetic works in the early Middle Ages [26] —argued that poems initially composed in writing, rather than in performance, were primarily destined for an audience of listeners (audientia) and not an audience of readers, [27] as Benson had proposed. The fact that the poetic works in the early Middle Ages were propagated through oral performance for an aural audience [28] would suggest that the use of formulaic language was not a literary convention, as posited by Benson, but a requirement for satisfying the taste of the poet’s audience. [29] Essentially, Zwettler was intent on demonstrating that in the early Middle Ages even a poem composed in writing was governed by the rules of oral literature, for it was ultimately directed toward a public of listeners and hence had to accommodate its audience’s aural predilection. [30] Furthermore, he pointed out that reading aloud from a manuscript did not constitute the only form of oral delivery that the poet had at his disposal, as tacitly assumed by Benson, but was rather a rare alternative to the more common reciting and singing without a manuscript [31] and akin to the act of composition-in-performance, which is one of the cornerstones of the Oral-Formulaic Theory. [32]
In case the oral rendition of a work ensued by reading from a manuscript, the text itself could be deemed to be the performance, and not merely a mnemonic devise (or transcript) without consequence for the ongoing performance. [33] Here, too, while the written text must conform to the taste of an audience of listeners, being relieved of the burden of spontaneous, extemporaneous, composition, the poet is free to return to his manuscript and make changes. [34] Thus, albeit still composing in the formulaic fashion, by reworking the manuscript, the author could gradually introduce new, literary, elements. This seems to be precisely what Benson is alluding to when speaking about “literate composition.” [35]
Thus, recorded performances, while inevitably exhibiting formulaic techniques, once relieved of the pressures of spontaneous composition, [36] can result in “transitional texts,” that is, hybrid texts still keeping up the pretense of being traditional (formulaic) works, while also exhibiting evidence of literary composition.
Lord himself originally excluded unconditionally the possibility of “transitional texts,” [37] but later his view became more nuanced. In a first important survey of oral-formulaic research, he differentiated, in response to Benson’s article, between formulaic expressions that occur in texts known to be initially written and formulae that belong to the oral tradition, using the term “repeated phrases” for formulaic expressions that have lost their function as a tool for rapid spontaneous composition, reserving the term “formula” for the classical definition. [38]
However, in subsequent works dealing with oral literature, Lord defined his understanding of repetitions, formulae, and “transitional texts.” Acknowledging the possibility of “transitional texts” as residues of oral composition in a rising written tradition, Lord nonetheless refused to use the term “formula” for “formulaic expressions” that are found within these “transitional texts” on the grounds that in the new medium the “formulae” had lost their functionality and were maintained mainly for “aesthetic or referential reasons,” awaiting the emergence of a new style to replace them:
In the Singer of Tales I had argued against the existence of “transitional texts,” a concept that constantly haunted us … There seem to be texts that can be called either transitional or belonging to the first stage of written literature. [39]
When people began to write Anglo-Saxon verse … they continued to use the same traditional style, because there was as yet no other available. A new style was to evolve in time. [40]
And finally:
One of the changes that comes in the “transitional” stage is that gradually formulas, no longer being necessary for composition, give place to true repetitions, which are repeated for aesthetic or referential reasons rather than for ease in verse-making. [41]
In more recent decades, the discourse on orality has dramatically shifted its focus onto the zone of hybridity, of which transitional texts are a manifestation, and which has been meaningfully called the oral–written/literate continuum. [42] Indeed, within this discourse—and in spite of the great divide that hitherto had separated orality from literacy—the expansive domain of hybridity, where the dialectic interplay between oral and written traditions is in the foreground, has been meticulously explored. [43] The result is that past schematic conceptions of a linear movement from orality to increased literacy—leading to the eventual demise of oral traditions—has been questioned, [44] and manifold unexpected forms of cross-fertilization (in cultural milieus endowed with both oral and literary traditions) between oral composition and writing have been identified.
Perhaps most importantly, within this zone of hybridity, in which oral and literary traditions interacted, manifold ecologies emerged that concurrently drew upon compositional techniques pertaining to both traditions. [45] One pole was oral composition in performance, and the other was the literate author/scribe composing in writing. Hence, it is challenging, if not impossible, to determine whether a given text represents the textualization of a performance, a “non-performative” oral text, or even a literary product created through the use of techniques of oral composition. [46]
It is clear that on the oral–literate continuum, formulicity—or formulaic recurrence—may not serve as an unequivocal criterion for determining the “orality” of a text, [47] since formulaic expressions may have been generated in performance, or in writing without the (oral) reality that sustained them. In the first case, one would then deal with a (postulated) untainted “primary” oral tradition, [48] but in the second case with hybrid texts—which still ought to belong to the world of orality, whose confines have been shown to be much broader than assumed early on—exhibiting formulicity, surely because no literary style had yet arisen to replace it, so that it still formed the “very stuff of poetic articulation and of poetic thought”; [49] eventually, however, with the leisure to reconsider and rework a written text without the constraints of composition-in-performance, a new, literary, style was bound eventually to emerge, even though the aesthetic conservatism of the audience may have slowed theprocess. [50]
It is along this oral–literate continuum, where a new compositional onto-logy partook of oral traditions in variegated ways, that both Achaemenid and Sasanian inscriptions are to be placed. The two epigraphic corpora are tributary both to the written traditions of the ancient Near East, and to the oral traditions that may have been captured in “non-performative” oral texts, or were still performed.

Iranian Royal Inscriptions

How did oral literature interact with the royal Iranian inscriptions? Three aspects have to be considered when studying the orality of a work, namely its (1) composition, (2) transmission, and (3) reception. [51]
In line with these distinctions, we shall examine the oral aspect of the composition of the Achaemenid and the Sasanian inscriptions from a synchronic perspective. Subsequently, we shall explore their oral transmission from a diachronic perspective, in other words, investigate the possibility that themes and formulae found in the Achaemenid royal inscriptions could have been orally transmitted for centuries and then resurfaced in the Sasanian texts.

Synchronic Analysis: Achaemenid Inscriptions (Composition)

Regardless of whether the Bisotun inscription was the first inscription written in Old Persian, [52] or whether the Old Persian script had already been invented and used to write down minor royal inscriptions prior to Darius’ accession, [53] the fact remains that the Bisotun inscription represents the first attempt at composing in writing a longer narrative, and it is hence probable it may display traces of oral literature.
Inasmuch as Achaemenid inscriptions, as already established by H. Tadmor and J. Bickerman, owe numerous themes to the Assyro-Babylonian literary tradition, [54] they exhibit similarities with the Iranian epic tradition and the Avesta. This constatation may indicate that the composers/scribes responsible for the composition of the Bisotun inscription were versed in both the literary traditions and the languages of the ancient Near East, and ancient Iran. [55] They could therefore have been either Elamite (or Near Eastern) scribes who were well acquainted with the oral traditions of Iran, and well versed in Old Persian, or acculturated Iranians trained in the literary traditions of the ancient Near East. [56]
The simultaneous presence of literary (and oral) traditions in the Bisotun narrative draws attention to the inscription’s compositional genesis, as well as to the intellectual disposition and linguistic abilities of the agents who were responsible for its redaction. Several scenarios, which traditionally have been in circulation, or were formulated in more recent times, may account for the variegated structure of the Bisotun narrative. They may diverge over what is deemed to have been the so-called Urvorlage of the inscription, that is, whether the original “text” was composed in Old Persian, or in a different language of the empire, such as Elamite.
(1) It has long been assumed that an Old Persian oral text was the original composition [57] upon which other versions drew. This allows the surmise that King Darius dictated his res gestae directly to polyglot scribes who simultaneously translated the oral Old Persian Urtext and put it into writing in the empire’s other languages, be that Aramaic or Elamite; [58] alternatively, it has been put forward that the king’s oral dictation could have been ideographically captured through the use of Elamite, [59] or phonetically written down in Aramaic characters, [60] forming a transcript for further elaborations and amplifications, which, once thus enriched, would have then formed the Old Persian Vorlage for the other versions engraved at Bisotun. [61]
(2) A variant of the above, following Skjærvø, is to posit that the king ordered the oral narrative to be made by professional composer-performers who may have formulated the Old Persian res gestae using traditional (Iranian) oral formulae and themes. [62]
(3) It has also been suggested that in spite of the oral character of the Bisotun inscription, royal annals or records akin to Babylonian chronicles—possibly the selfsame βασιλικαὶ διφθεραί or βασιλικαὶ ἀναγραφαί reported by Ctesias [63] —written in Akkadian [64] or in Old Persian [65] could have been kept by the Achaemenid administration, and provided the framework for Bisotun’s composition.
(4) Another possibility is to assume, following the multiple stages of the development of the Bisotun reliefs and inscriptions, that the first draft of the Bisotun narrative was composed in Elamite, with other versions, notably Old Persian, being later translations thereof. [66]
In spite of their great ingenuity, aspects of the above propositions may be controversial and thus render their acceptance conditional. Each proposition is examined below.
(1) The production of an (oral) composition that would have simultaneously cast Iranian epic themes and story pattern(s) into the literary frame of Near Eastern narratives, and exhibit the type of complex narrative structure with which Bisotun is bestowed, presupposes the work of highly skilled polyglot oral composers or scribal schools, and may scarcely have emanated from the king’s repertoire. [67] Even the hypothesized rudimentary oral composition of the king, presumably captured in Aramaic characters in a transcript waiting to be amplified, is merely a Behelf to maintain the a priori assumption of the king’s participation in the composition, whose presumed prowess of oral dictation may have been partially prompted by the evocative formula θātiy Dārayavahuš xšāyaθiya “thus says the king.”
(2) As to the second assumption, we are confronted with yet another problem. If the original text was an Old Persian performance by professional composer-performers versed in the traditional themes of the Iranian epic, how is the concurrent use of themes and models pertaining to the literary traditions of the ancient Orient to be explained? Two departures are conceivable: either the composer-performers were also versed in the themes and story patterns of the Mesopotamian literary traditions, and hence could compose their oral narrative by occasionally embedding it into their thematic mold; [68] or a transcript of these Mesopotamian literary-narrative patterns, in which performer-composers had to cast this traditional oral narrative, was made available to them. [69]
(3) It is precisely this latter aspect—the possibility of a transcript, which could have provided a framework inspired by the literary traditions of Meso-potamia—that takes us to the third assumption. There is no need to presume royal annals at the Achaemenid court, but were one to acquiesce to their existence, one would still remain nescient as to the language used to record them. With the impossibility of deciding on the emergence of the Old Persian script before, or under, Darius, [70] and in view of the Persian-Elamite acculturation to which Iranian speakers avant l’empire were subject, the agents responsible for maintaining such postulated διφθεραί, or transcript, ought to have been Persian scribes, and the medium in which to cast it, Elamite, which leads us to the final assumption.
(4) Hitherto it has been tacitly assumed that the narrative of Achaemenid inscriptions was first composed in Old Persian with other versions being translations from this original. Some time ago, François Vallat’s investigation of the Elamite language of the foundation charter of Susa, [71] namely the Elamite texts DSf [72] and DSz [73] , as well as the Akkadian version of DSz, that is, DSaa, [74] showed that the Elamite texts, although following the same narrative pattern, were not servile copies of an original, but independent redactions that could differ in vocabulary, syntax, and content from one another, [75] as well as from other versions. [76] The Akkadian text DSaa exhibits so many discrepancies from DSz (and the older DSf), especially as regards the composition of the list(s) of products—and their provenance—that were carried forth to Susa to erect Darius’ palace, that it may be regarded an independent work possibly by an atelier of scribes writing Akkadian, albeit based on the same narrative pattern as the Elamite text(s). [77] The analysis of these texts thus may suggest that Achaemenid Elamite and Akkadian inscriptions were redacted independently on the basis of a common narrative pattern, possibly a transcript, by different groups of specialized scribes who in the end may have harmonized their respective redactions. [78]
It is therefore possible to conceive that the same Iranian scribes who were responsible for the attested “Persianisms” [79] in the Elamite version of the Bisotun inscription could have been directly writing in Elamite, instead of assuming that an oral Old Persian original was first composed and subsequently translated by polyglot scribes into Elamite, whence it was translated back (rückübersetzt) to Old Persian. Both the fact that the Elamite version was the first written realization of the king’s res gestae and the fact that DSf and DSz seem to have been independent texts rather than subservient translations of an Old Persian original may support this conclusion.
Still other elements may corroborate the above presumption. Recent studies by Tavernier [80] and Henkelman [81] in the formation of the Persian ethnogenesis, the languages of neo- and Achaemenid Elamite, and multilingualism in the Achaemenid empire indicate that our perception of the agents responsible for the redaction of Bisotun, as well as their linguistic and cultural identities, is ripe for a major revision. Indeed, the long process of cross-cultural interactions between Elamites and Persians on the Iranian plateau over the span of several centuries prior to the formation of the empire ought to have decisively shaped Persian identity, which, in spite of its Indo-Iranian heritage, may be reckoned as the culmination of this acculturation. [82]
As denoted in the case of the Persepolis Fortification Tablets, a majority of scribes bear Iranian names, and appear to be Iranophones, writing an Elamite that shows the impact of Old Persian morphological and syntactical imposition. [83] Old Persian agency in the restructuring of Achaemenid Elamite therefore may identify the scribes responsible for the redaction of the Elamite tablets as predominantly polyglot Iranian scribes, for whom writing Elamite may have been the ordinary extension of their linguistic and scribal competencies. [84]
That either Iranophone, or Elamite, scribes, continued creatively to perform within the medium of the Elamite language may be borne out by yet another example. Under Xerxes, we encounter on Achaemenid documents from Babylonia, the new title šar māt Parsu u māt Mādaya, “king of the land of Persia and of the land of Media,” for which no Old Persian precedent is attested. As we have shown in a more exhaustive study, [85] a brief survey of Elamite titulature may suggest two possible sources of inspiration for the Babylonian title: the well-attested title sunkik anšan šušunka, “king of Anšan and Susa,” which referred to the two political and ideological poles of Elamite royal authority, namely, Anšan and Susa; and the Middle Elamite interpretation of the same title, that is, hal-menik hatamtik āk šušenki, “(land)lord of Elam and Susa.”
Another Vorlage for this innovative titulary could have been the Babylonian title šar māt Šumeri u Akkadî, which on occasion had been adopted by Assyrian conquerors and restorers of Babylon, notably Assurbanipal, as a means to present themselves as bearers of Babylonia’s dynastic continuity. Thus it is possible that this old title, embracing Babylonia’s chief cultural and ethnic constituents and already adopted by Cyrus as part of his titulature in Babylonia, [86] may also have served as Vorlage in forging Xerxes’ new title component, which hinted at the Achaemenid empire’s two leading ethnogeneses, namely, the Persian and Mede ethnea, in the same way as the title šar māt Šumeri u Akkadî used to refer to the main ethnic constituents of Assurbanipal’s empire.
Another query, co-related to the prospect of Elamite and Babylonian titles of old impacting the creation of Xerxes’ new (Babylonian) title, remains the identity of the agents deemed responsible for this innovation. [87] Were one to assume that the model for Xerxes’ Babylonian title might have been the old titles of Elam, that is, sunkik anšan šušunka, “king of Anšan and Susa,” or hal-menik hatamtik āk šušenki, “(land)lord of Elam and Susa,” [88] then we would have to presume that Elamite scribes of the Achaemenid chancellery were responsible for this novelty, [89] which may have been then introduced as normative into Babylonian practices.
Alternatively, Babylonian scribes within the Achaemenid chancellery, or at the provincial level, possibly aware of the past usage of royal titulary in Meso-potamia, may have decided in favor of the adoption of the old Babylonian title šar māt Šumeri u Akkadî. [90]
Although the local nature of this complex operation, which consisted of forging a new Achaemenid title by casting it into the mold of (Elamite or) Babylonian titles of old, may point to the work of Babylonian scholars, [91] we may not exclude the agency of Persian authorities. Given that, we may infer that Elamite (or Iranophone) and Babylonian scribes at Persepolis under Xerxes’ rule were still adequately in command of Elamite and Akkadian [92] to give rise to this innovation; or were possibly able to supervise its creation by Babylonian scholars in Babylonia proper. [93]
Recapitulating, then, one may posit that Achaemenid Elamite ought to be regarded as an original, autochthonous vehicle of writing for Iranophone scribes, whose appropriation is further attested in the profound changes Old Persian induced to the linguistic structure of Achaemenid Elamite. Under these circumstances, with Elamite qualifying as a “Persian” medium of writing, [94] the prospect of the Bisotun narrative being originally composed in Elamite is undeniable.
In contrast, the hypothesis that an Old Persian oral composition served as the original for other versions raises serious problems that cannot at present be solved and so seems unlikely, although not impossible. Thus synchronically, there is no compelling reason to assume an original oral composition of the Bisotun inscription, although neither the impact of oral (epic) traditions, nor the influence of the Avesta, whose sway over Old Persian inscriptions has been investigated by Skjærvø, [95] may be excluded.
Although the impact of oral literature on the redaction of the Bisotun itself remains elusive, the effects thereof on the oral rendition of the Bisotun narrative, which is referred to as the handugā- in the inscription, are the more apparent. It is this oral (re-)composition of the Bisotun story for the consumption of an Iranian audience that was captured by Greek historiography, and we shall turn our attention to the investigation of this.
Beginning with paragraph 64 of the Bisotun inscription, Darius initiates a number of exhortations and admonitions that are firmly structured. They are concerned in sequence with the veracity of Darius’ res gestae , the preservation of his (physical) inscription (dipi - ) and reliefs (patikara - ), as well as the propagation of his announcement or narrative (handugā-) among the people in arms (kāra-), before reaching the content of paragraph 70, which we shall discuss below. [96]
Darius, having depicted at length how he had vanquished the nine Lügen-könige in sundry battles, and exposed the reason for their rebellion (namely, drauga- “Lie”), addresses directly his archetypal successor (tuvam kā xšāyaθiya haya aparam āhạy, “you who shall later be king”), whom he exhorts to believe the extent of his accomplishments within a single year:
θātiy Dā[raya]vahu[š] xšāyaθiya ima taya adam akunavam vašnā Ahuramazdāha [h]ama[h]ạyāyā θarda akunavam tuvam kā haya aparam imām dipi[m] patipạrsāhạy taya manā kạrtam vạrnavatām θuvām mātaya *dr[uxt]am/*dr[aug]am maniyāhạy
Darius the king says: “This which I have done, I have done by the greatness of Ahuramazdā in one and the same year. You who shall be reading this inscription believe what I have done, and do not think it is fabricated/a lie.”
DB 4.40–43
Here, the focus is on the physical inscription (dipi-) itself, which not only can be read (patipạrsa-), but also, as other passages indicate, is written (ni-paiθa-/nipišta-), and may be seen (vaina-):
θātiy Dārayavah>uš xšāyaθiya vašnā Ahura[mazdāha] utāmaiy aniyašciy vasiy astiy kạrtam ava ahạyāyā d[i]p[īyā] naiy nipištam avahạyarādiy naiy nipištam mātaya [haya] (apa)ram imām dipim patipạrsātiy avahạyā paruv θadayā[tiy taya] manā kạrtam naiš[im] ima vạrnavātaiy duruxtam maniyā[taiy]
King Darius says: “By the greatness of Ahuramazdā and mine, there is much else that has been done and which was not written down in this inscription, for that reason it was not written down, so that the one who later shall be reading this inscription, to him what I have done may not seem too much, and he may not believe it, and think it fabricated.”
DB 4.45–50
θātiy Dārayavahuš xšāyaθiya tuvam kā haya aparam imām dipim vaināhạy taya adam niyapaiθam imaivā patikarā taya [97] vikanāhạy yāvā *hutava [98] āhạy avaθādiš paribarā
King Darius says: you who later shall be seeing this inscription that I have written, or these reliefs, do not destroy (them), (but) thus preserve them as long as you can.”
DB 4.69–72
Another indication that dipi- here refers to both the content of the Bisotun, and also the phainomenon itself, that is, that which strikes the eye, is Darius’ caution not to destroy (vikan-), as well as his urging to preserve (paribara-), the physical inscription and its accompanying reliefs (patikarā):
θātiy Dārayavahuš xšāyaθiya yadiy imām dipim vaināh[ạy] imaivā patikarā naiydiš vikanāhạy utātaiy yāvā taumā [ahatiy] paribarāhạdiš Ahuramazdā θuvām dauštā biyā utātaiy tau[m]ā vasiy bi[y]ā utā dargam jīvā utā taya kunavāhạy avataiy Ahuramazdā hucāram kunautuv
“King Darius says: if you were to see this inscription or these reliefs, and were not to destroy them, and preserve them as long as you have strength, may Ahuramazdā be your friend, and may you have a large progeny, and live long. And what you shall be doing, may Ahuramazdā make that successful for you.”
DB 4.72–76
θātiy Dārayavahuš xšāyaθiya yadiy im[ā]m dipim imaivā patikarā vaināhạy vikanahạdiš utātaiy yāvā tau[m]ā ahati[y nai]ydiš paribarāhạy Ahuramazdātaiy jatā biyā utātaiy taum[ā] mā b(i)[yā] utā taya kunavāhạy avataiy Ahuramazdā nikatuv
King Darius says: if you were to see this inscription or these reliefs, and were to destroy them, and not preserve them as long as you have strength, may Ahuramazdā be your striker, and may you have no progeny. And what you shall be doing, may Ahuramazdā not make that successful for you.”
DB 4.76–80
Darius’ attention then shifts from the mainly phenomenal aspect of the inscription to its content, that is, Bisotun’s account (handugā-) and the reporting (θaha-) thereof. Clearly, the focus of this subsection is the dissemination or oral retelling of Darius’ narrative to an Iranian? audience (kāra-). He admonishes viewers against concealing (apagaudaya-) his story, and encourages its being retold with threats and promises seen before:
θātiy Dārayavah xšāyaθiya nūram θuvām vạrnavatām taya manā kạrtam avaθā kārahạyā θādiy mā apagaudaya yadiy imām handugām naiy apagaudayāhạy kārahạyā θāhạy Ahuramazdā θuvām dauštā biyā utātaiy taumā vasiy biyā utā dargam jīvā
DB 4.52–57
King Darius says: “Now let what I have done convince you. Tell it thus to the people in arms , [and] do not conceal it; if you were not to conceal this account from, and tell (it) to, the people in arms, (then) may Ahuramazdā be your friend, may you have a large progeny, and live long.”
θātiy Dārayavahuš [xšāya]θiya yadiy imām handugām apagaudayāhạy naiy θāhạy [k]āra[hạy]ā Ahuramazdātay jatā biyā utātaiy taumā mā biyā
King Darius says: “If you were to conceal this account from, and not tell (it) to, the people in arms, (then) may Ahuramazdā strike you, and may you not have any progeny.”
DB 4.57–59
After having copiously guarded against destroying his dipi- and concealing his handugā-, Darius then moves to state how he has reproduced a copy/transcript of the narrative and sent it throughout the lands for people to know his story. Since, however, this last paragraph (paragraph 70) is among the more contested and discussed passages of Old Persian epigraphy, we shall first succinctly recall the more salient areas of contention in scholarship, before offering yet another, surely inadequate, reading of paragraph 70.
In paragraph 70 Darius appears to be introducing a new concept that differs from both the dipi- and the handugā-. Because the Old Persian passage is badly damaged, the word expressing this presumed new concept has been reconstructed as *dipiciça- by Schmitt, and without alternative readings being offered in recent times, it has been generally accepted. [99] Initially, *dipiciça- was construed as “form of writing,” [100] a meaning which Schmitt subsequently gave up in favor of “copy (of the inscription); Fassung (der Inschrift),” [101] paying heed [102] to a prior study offered by Lazard, [103] and a later one by Huyse. [104] This latter had argued that *dipiciça- expressly indicated the Old Persian version of the inscription as inscribed beneath the Babylonian and the two Elamite versions on Mount Bisotun. [105]
The form *dipiciça- consists of dipi- “inscription,” and possibly the noun ciça-, which primarily means “seed; offspring” in Old Iranian, but in view of its Middle Iranian semantic range, it could also have signified “appearance” in Old Persian. [106] Thus, *dipiciça- may have indicated “likeness of (an) inscription,” possibly a “transcript; copy” of the inscription. However, *dipiciça- being a reconstructed form, it may be methodologically unsound to infer on the precise meaning of this passage by relying solely on the Old Persian version. The Elamite version of paragraph 70 is well preserved, and there we read the term tuppi-me for *dipiciça-. [107] As argued notably by Huyse, Elamite tuppi-me ought to have had a different meaning than tuppi- “inscription,” [108] which is used for (and has historically given rise to) Old Persian dipi-, since the -me suffix usually makes abstract nouns. [109] The semantic Umfeld of Elamite tuppi- (tippi-) and tuppi-me was investigated some time ago by Lecoq, [110] and recently in an exhaustive manner by Tavernier, [111] resulting in the latter’s adopting the meaning “(inscribed) text; what is written in the tuppi- (tippi-)” for tuppi-me. [112]
However, since Bisotun decisively distinguishes between dipi- and handugā- as discrete conceptions, one may argue that *dipiciça- and its Elamite counterpart tuppi-me are also bound to express a distinct concept. What this latter nuance might be ought to be inferred from the use of actions that are associated with *dipiciça- (or tuppi-me). In the following table, verbs that have been used in Bisotun in connection with dipi-, handugā-, and *dipiciça-, have been listed:
dipi- “inscription” vaina- “see”
niyapaiθa- / nipišta- “wrote; written”
patipạrsa- “read”
(naiy/mā) vikan- “(not) destroy”
paribara- “preserve”
handugā- “report” (naiy) apagaudaya- (naiy) kārahạyā θaha- “(not) conceal, (not) reveal to the kāra-
*dipiciça- “transcript?” patišam kun- “reproduce?”
ariyā utā pavastāyā utā carmā *grạfta- ah- “captured (recorded) in Aryan, as well as on clay and on parchment”
*niyapaiθiya- utā patiyafraθiya- paišiyā *mām “written and read in my [= Darius’] presence”
*frāstāya- “sent out”
The term *dipiciça- may be “captured (recorded)” on clay and parchment (pavastā- utā carman-), “written and read (loud) in the presence” of the king (niyapaiθiya- utā patiyafraθiya- paišiyā …), and be “sent out” (* frāstāya -) everywhere in the lands (vispadā antar dahạyāva). Clearly, this *dipiciça- ought to be a mobile document, and as such distinct from the inscription itself (dipi-) and its oral rendition (handugā-). As the Elamite etymon tuppi- implies, tuppi-me ought to be a written document, but one that could be duplicated from a Vorlage, and placed on different writing surfaces to be disseminated, but also one that could have required official authentication by dint of the king’s signature? (*nāmanāfa-), and genealogy? (* h uvādātam), which were reproduced thereupon; in brief: a “transcript.”
In this light, the proposal of Lazard and Huyse that we read patišam kun- / kun- patišam (in: ima *dipiciçam taya adam akunavam patišam, “this copy that I …”) as a compound, either with the primary meaning of “mettre devant,” [113] or “put opposite,” [114] is tempting. Indeed, patišam may be considered the adverbial form (in the nominative-accusative neuter singular) of a postulated *patiša-, “adverse; contrary”—which is attested both as an adjective paitiša- with the same meaning, and an adverb “towards to,” in the Younger Avesta—or an adverbial extension of patiš (patiy), made in analogy to such adverbs as dạršam, “vigorously.” [115] If so, patišam kun- could have meant, as already suggested by Herrenschmidt, “counterfeit; replicate”; not unlike pati-kara- “relief; Abbild,” which in essence indicates a “facsimile.” [116]
Thus the Old Persian version of paragraph 70 could be read as follows:
θātiy Dārayavauš xšāyaθiya vašnā Auramazdāha ima *dipi[c]i[çam] taya adam akunavam patišam ariyā uƒtā pavastāy[ā] utā carmā *grạ[ftam āha *pat]išam[c]iy *[nāmanā]fam akunavam *pa[t]iša[m *hu]vādā[tam akuna-]va[m] utā *niyapai[θ]i[ya u]tā patiyafraθiya paišiyā mā[m] pasā[va] ima *dipi[ciça]m f[r]āstāyam vi[s]padā antar dahạyā[v]a kāra *hamā[t]axšatā
DB 4.88–92
Said Darius the king: “By the greatness of Ahuramazdā this transcript/copy that I reproduced (replicated), was captured (recorded) in Aryan, as well as on clay and on parchment. I (also) replicated (thereupon) my signature, and reproduced my genealogy. It was both written and read in my presence; afterwards, this transcript/copy I sent out everywhere in the lands, so that the people in arms may apply themselves (to know it?).”
In view of the above reading, Vallat’s recent discussion of the Elamite version of the passage ima *dipiciçam taya adam akunavam patišam “this transcript/copy that I reproduced,” which has been traditionally read: u tuppime daʾe-ikki hutta, “I made differently a text/inscription,” is pertinent. [117] Indeed, Vallat interprets daʾe-ikki not as an adverbial compound signifying “differently; otherwise,” [118] but instead attaches ikki, “at; towards to,” as preverb to the verb hutta, “to do,” with the meaning “traduire.” [119] The combination of ikki and hutta has also been recently (possibly independently) adopted by Grillot-Susini, who interprets ikki-hutta as “ajouter.” [120]
Indeed, the majority of attestations for -ikki as postposition in the Elamite version of the Bisotun are directives with the meaning “at; towards to”; [121] there are, however, instances where -ikki may be understood as “against,” as the following example illustrates:
meni taššup hupipe Haraumatiš Mimana-ikki mi[tep akkape Mi]šdatta tippe dašti
Then those troops, which Mišdatta [= Vahạyazdāta] had sent away, went forth to Arachosia against Mimana [= Vivāna]
DB El. 3.23–24
In such contexts, it reflects Old Persian abiy, but semantically covers the same ground as patiš, “against.” Thus one may wonder whether ikki-hutta is merely an attempt at representing patišam kun- and may also be understood as “counterfeit; reproduce.” [122] Thus the phrase:
u tuppime daʾe ikki-hutta harria-ma appa šašša inni šari kutta halat-ukku kutta kušmeš-ukku [123]
—could be understood as:
I reproduced another copy in Aryan, which formerly was not, neither on clay, nor on parchment.
In sum, the dissemination of the Bisotun narrative may have taken place on two levels: an oral Old Persian variant (handugā-) of the Bisotun inscription was circulated, and an Old Persian written transcript (*dipiciça - ) of the selfsame inscription, recorded on clay and parchment, [124] was sent to the four quarters of the empire; as we know the Old Persian version was not the only one to be dispersed in such a way. [125] That the * dipiciça -, by being possibly proclaimed orally, that is, performed (at satrapal courts?), could have given rise to an oral tradition of its own, independent from the original handugā-, may be, in view of the dialectical nature of the oral–literate continuum, seriously envisaged. In this light, the oral rendition of the *dipiciça- may have been responsible for some of extraordinary similarities that the Bactrian Rabatak inscription of the Kušān king Kaniška exhibits with the Bisotun inscription, as in lines 3-4: [126]
οτηια ι ιωναγγο οασο οζοαστο ταδηια αριαο ωσταδο αβο ιωγο χþονο αβο ι υνδο φροαγδαζο αβο þατριαγγε þαορε
And he *issued a Greek *edict (and) then put it into Aryan. In the year one there was *proclaimed to India, to the cities of the *kṣatri-yas/*kṣatrapas. [127]
However, it was probably the oral handugā- that gave rise to the stories captured by our Greek sources, [128] but since neither its content nor that of the “other copy in Aryan” (tuppime daʾe … harria-ma) is known to us, this remains merely an informed speculation.

Synchronic Analysis: Sasanian Inscriptions (Composition)

In this section we shall investigate the possible role of oral literature in the composition of the Sasanian inscriptions.
Unlike the Bisotun inscription, which represented the first attempt at putting into writing a longer narrative in an Iranian language, and which therefore could have been impacted by techniques of oral composition, the Sasanian inscriptions, being the culmination of a long written epigraphical tradition, are prima facie not suspect of being influenced by oral traditions. It is only through comparison with the Achaemenid inscriptions—that is, from a diachronic perspective, wherein striking parallels between the two groups of texts appear—that the question of orality becomes worth considering.
Although the incentive for our examination of the Sasanian inscriptions for traces of orality originates from a diachronic perspective, we must examine them on a synchronic level as well. As we have already pointed out, orality is generally measured by the density of formulaic recurrence, and the inscriptions of the magus Kerdīr, taken as a paradigm for the Sasanian epigraphic corpus, offer numerous passage where specific themes are repeated with similar phraseology. [129] However, the repetition of word clusters alone does not constitute an irrefutable proof of formulicity, and one has to differentiate between repetitions and the formulae proper. What distinguishes these two categories, is according to Lord, referring to Parry, [130] the “nature of an expression”:
The formula “helps the poet in his verse-making.” It is primarily for that reason that it is repeated. The “repetition,” on the other hand, is a phrase repeated to call attention to a previous occurrence, for an aesthetic or other purpose. Formulas do not point to other uses of themselves; they do not recall other occurrences. [Lord’s italics] [131]
It is however, no facile task to identify a formula, or a repetition, in an inscription based on these premises. The procedure presupposes that any knowledge about the oral or literary character of a composition precedes the formulaic analysis; in other words, what the formulaic analysis seeks to establish, that is, the orality of a composition, itself constitutes a premise for determining the formulicity of a composition. [132]
Since the texts with which we deal are not known to be literary or oral compositions, it may be futile to attempt to distinguish between repeated phrases and formulae. Moreover, it is precisely this result that the formulaic analysis is supposed to establish, not vice versa. Assuming that the repetitions within the Sasanian corpus were formulaic, observations from the Old English compositional landscape could also be applied to the Sasanian inscriptions, namely, that they represent oral texts, that is, written texts exhibiting an oral (formulaic) compositional style.
Again synchronically, we have no compelling reason to assume the orality of the Sasanian inscriptions.

Diachronic Analysis

Since the synchronic analysis of the royal Iranian inscriptions has not revealed any clear sign of orality in their composition, we must now compare the two corpora for traces of orality by adopting a diachronic approach, concentrating on the orality in the transmission.
As indicated above, numerous parallels in phraseology, themes, and overall story pattern have been identified between the Achaemenid and Sasanian inscriptions; these have been attributed to oral traditions. We have also seen that there is no compelling evidence that the earliest Old Persian inscriptions were composed orally; on the contrary, there is some evidence that they were literary translations of Elamite versions, which admittedly could have incorporated some oral Iranian traditions. We may probably postulate several steps in the diffusion of the Old Persian royal inscriptions: after the actual composition, an (officially sanctioned) oral variant of the royal proclamation, precisely the handugā-, entailing themes and points of reference most befitting the intellectual horizon of an Iranian aural audience, was communicated throughout the empire by official heralds and eventually retold by professional storytellers who possibly recast them in the pattern of oral narratives, while performing at satrapal courts for the benefit of the kāra-, as the Bisotun inscription clearly indicates. This would mean that Darius’ proclamation, albeit originally composed in writing, was orally retold and transmitted, and by dint of this process, possibly canonically incorporated into the body of oral (epic) literature, [133] upon which the composers of the Sasanian inscriptions may have been able to draw more than nine centuries later.
As mentioned before, even the * dipiciça -, that is, the postulated written Old Persian “diplomatic” copy of the Bisotun inscription, may have been read aloud, that is, performed (at satrapal courts?), and in time begotten a parallel oral tradition different from the handugā-. This tradition then may have been separately, or in confluence with the handugā-, responsible for some of the themes we encounter in Kaniška’s Rabatak inscription, but also some of the themes of Sasanian epigraphy.
There are in fact three ways to explain, on a synchronic level, the formulaic diction of Sasanian inscriptions and the numerous parallels in phraseology, themes, and story patterns between the Achaemenid and Sasanian inscriptions.
It has been suggested that Sasanian inscriptions were versed in the selfsame oral traditions wherein variations of the Achaemenid res gestae were embedded. [134] This explanation may account for the formulaic diction of the Sasanian inscriptions, as well as the parallels between Achaemenid and Sasanian inscriptions. However, even if one were to presume the existence of an oral tradition in which variants of the Achaemenid royal res gestae could have survived down to the Sasanian period, in the absence of a reliable measure with which to determine the orality of Sasanian inscriptions, the idea of themes and patterns pertaining to Achaemenid inscriptions being adopted by the Sasanians via an oral tradition would further remain a matter of conjecture. Indeed, in spite of high formulaic density, lack of enjambment, “adding style,” to name only a few gauges of oral composition, the orality of the Sasanian inscriptions remains unproven.
Second, it has been posited the Sasanian inscriptions belonged to a literary tradition that had partially preserved Aramaic translations of Achaemenid inscriptions, or at least some story patterns, themes, and even expressions, which constituted the contents of royal inscriptions. [135] This scenario explains the parallels, but not the presence of formulaic diction in the Sasanian texts.
Third, we may consider a combination of the above scenarios. The early Sasanian inscriptions could have been “oral” or “transitional” texts, [136] that is, early attempts to textualize oral composition by making use of the oral formulaic style in the medium of writing, for want of a literary style still to be developed. This explanation has the advantage of elucidating the formulaic language of the Sasanian inscriptions by taking into account their literary character as well. There is no evidence of any larger epigraphic compositions before that of Šābuhr I on the Ka ʿ be-ye Zardošt. The earlier Arsacid and Sasanian inscriptions are all very succinct and do not compare in extent and narrative elaboration with Šābuhr I’s res gestae. [137] It is therefore possible that the political reassertion of the neo-Persian empire against the imperial other, Rome, could have served as an impetus for Šābuhr I to put into writing the king’s deeds, much like the re-establishment of royal power by Darius triggered the composition of the Old Persian Bisotun inscription. Thus, if the inscription of Šābuhr I represented an early attempt to produce an elaborate literary composition, one might expect to uncover therein some residue of formulaic language. As a result, the oral-formulaic technique, although superfluous, may have survived in the early written inscriptions as a convention, before the eventual rise of a literary standard.

The Inscriptions of Darius and Narseh, and Herodotus’ Account

So far, we have proposed that the oral dissemination of Darius’ res gestae could have occasioned the incorporation of an alternate variant into oral traditions, whence it affected the structure of Sasanian inscriptions. The assumption the Bisotun inscription might have influenced the oral tradition is indebted to the observation that many themes, which Darius’ res gestae shares with the epigraphic tradition of the ancient Near East, [138] are also present in the Sasanian inscriptions, which they could only have reached through the intermediary of the Achaemenid inscription. Indeed, the only link between the content of the epigraphic tradition of the ancient Near East and the oral epic tradition of the Iranians, where Mesopotamian themes and story patterns are attested, were the Achaemenid inscriptions. Moreover, there is no evidence allowing us to assume that before the composition of Darius’ res gestae, which emulated its Near Eastern precedents, there had been any attempt to introduce Mesopotamian literary themes into the oral literature of the Iranians.
However, the belief that the inscription of Bisotun affected the thematic and linguistic structure of early Sasanian inscriptions through oral traditions remains a conjecture in the absence of an irrefutable proof for the orality of the early Sasanian inscriptions. Yet, there exists a tertium comparationis that may substantiate the orality of early Sasanian inscriptions. The account of Darius’ accession to the throne in 522 BCE is preserved not only in Darius’ own inscription at Bisotun, but also in Herodotus’ Historiae. Indeed, Herodotus’ account of Darius’ accession is closely paralleled by the inscription of the Sasanian king Narseh at Paikuli—composed in the early fourth century CE—whose accession, like that of Darius I’s, followed upon a coup d’état fomented by some members of the imperial aristocracy against their newly anointed “legitimate” sovereign, Warahrān III.
Indeed, the comparison of the three accounts of DB, NPi, and Herodotus illustrates numerous convergences between Herodotus’ narration and Narseh’s res gestae: wherever Herodotus’ report diverges from the account of the Bisotun inscription, the themes of these divergent passages have close parallels in the Paikuli inscription. The most striking departures in Herodotus’ relation from Darius’ res gestae are in the number of usurpers who took possession of the throne before Darius’ coup d’état and the formation, on two occasions, of nobiliary councils, a “consultative council,” in which the coup d’état was fomented, and a “constitutional council,” that was formed after the Seven successfully seized power to discuss the future of the state. Both of these themes are also found in the inscription of King Narseh at Paikuli. Thus, we cannot avoid the conclusion that the inscription of Narseh at Paikuli is a product of oral literature, or at least has been affected by it, for it shares themes with Herodotus’ version that are absent from the official version of the royal res gestae. Were the similarities solely between the inscriptions of Bisotun and Paikuli, we may have assumed a written source, presumably an Aramaic translation of the Achaemenid inscription, as a conceivable source for the inscription of Narseh at Paikuli, but thematic parallels between it and Herodotus’ account force us to discount this path of transmission.
Consequently, what we suspected from the beginning and strove to establish through the formulaic analysis of Kerdīr’s inscriptions—namely, the oral character of the early Sasanian inscriptions—may be now substantiated by the evidence of Narseh’s res gestae. Indeed, the presence of thematic parallels between Herodotus’ account and the Paikuli inscription and the absence of the selfsame themes from the narrative of the Bisotun inscription, combined with the evidence of formulaic density in the narrative of Kerdīr’s inscriptions, make a strong case for the influence of oral traditions on the tenor and language of Sasanian inscriptions.


[ back ] 1. See Sprengling 1953:37–45.
[ back ] 2. Skjærvø 1985:593–603.
[ back ] 3. Skjærvø 1985:603.
[ back ] 4. Huyse 1990:177–184.
[ back ] 5. Huyse 1990:177.
[ back ] 6. Huyse 1990:179–183.
[ back ] 7. “Aber im vorliegenden Zusammenhang, wo es um die mögliche gemeinsame Quelle achaimen-iden- und sāsānidenzeitlicher Formeln geht, wäre es wohl wichtiger, die den achaimenidischen und sāsānidischen Inschriften gemenisamen Formeln aufzuspüren” (Huyse 1990:179).
[ back ] 8. Huyse 1990:180–181.
[ back ] 9. Huyse 1990:181–183.
[ back ] 10. Huyse 1990:183.
[ back ] 11. See also Huyse’s other significant contributions on orality and literacy in ancient Iran, Huyse 2008:140–155; and his unpublished thèse d’habilitation entitled “Histoire orale et écrite en Iran ancien entre mémoire et oubli,” which will no doubt clarify the many questions we have broached here.
[ back ] 12. Skjærvø 1999; Skjærvø 1998c; Skjærvø 1994.
[ back ] 13. Skjærvø 1998c.
[ back ] 14. Skjærvø 1998c:100: “What this means is that the authors of the Achaemenid and Sasanian inscriptions, as well as the Manichean and Pahlavi texts, were all versed in Iranian oral literary traditions, in which not only numerous literary patterns, but also formulas had been preserved virtually unchanged for almost a millennium”; and also: “On the whole, however, they [the Old Persian inscriptions] follow the literary patterns of Near Eastern royal records, but also make abundant use of formulas from the Iranian oral literary tradition that surfaces in the Avesta” (104).
[ back ] 15. Skjærvø 1998c:105.
[ back ] 16. The wealth of scholarship produced on oral literature, the Oral-Formulaic School, and the characteristics of orality is too overwhelming to be discussed in the present confines. Therefore, only a few works most pertinent to our inquiry will find mention here; for an exhaustive survey of scholarship on the “Oral-Formulaic Theory,” see Foley 1988:57–111; Foley 1985:3–70; Foley 1983:27–95, esp. 60–79; for more salient works and manifold aspects of orality, see notably Foley 2010:17–37; Foley 2004:101–120; Foley 1999:99–108; Foley 1998b:80–93; Foley 1998b:13–33; and Foley 1991. An indispensable tool for the study of oral literature is the journal Oral Tradition, now conveniently to be found online at with archives from 1986.
[ back ] 17. Lord 1960:30.
[ back ] 18. See Kiparsky 1976; also Nagy 1990.
[ back ] 19. See Finnegan 1992:88–133; also Foley 1981b:275, who in the conclusion to his comparative study of the formula in the Homeric verses, the Yugoslav epics, and the Old English poetry writes: “In summary, then, we must discard the implicit assumption that the formula is a concept fully applicable in a single form from one tradition to another. Understanding multiformity at the level of the line—for that matter, at any level—means recognizing differences as well as similarities in poetic structure. The metrical foundation of Old English poetry, so different from those of Homeric Greek and Serbo-Croatian, demands a conception of formula which de-emphasizes the roles of syllabicity and internal structure in favor of the tradition-dependent parameters of stressed position and metrical formula.”
[ back ] 20. On the use of synchrony and diachrony in the description of the formula, see Nagy 1990:29–30.
[ back ] 21. Nagy 1974:196; also Nagy 1990:30: “Granted, meter can develop a synchronic system of its own, regulating any new incoming phraseology: nevertheless, its origins are from traditional phraseology.”
[ back ] 22. Nagy 1990:29. See also Watkins 1976:110, where he defines the formula as “the verbal and grammatical device in oral literature for encoding and transmitting a given theme or interaction of themes, with the repetition or potential repetition assuring the long-term preservation of the surface structure.”
[ back ] 23. Benson 1990.
[ back ] 24. See Benson 1990:229: “One must, that is, if it is indeed true that the presence or absence of formulas is a test of oral composition, for the whole doctrine of the oral composition of Old English poetry rests on its use of formulas. But the fact is that poems which we can be sure were not orally composed use formulas as frequently and sometimes more frequently than supposedly oral compositions such as Beowulf or the poems of Cynewulf.” From the evidence of a highly formulaic Old English translation—entitled Phoenix—of a Latin original which the translation closely follows, Benson (1990:229–230) further inferred that the author could not have spontaneously composed in the traditional way while following the original: “Moreover, it [Phoenix] is also a translation … so close to its Latin source that it is almost impossible to assume that its poet worked in the traditional way of hearing a tale, meditating on it, and then simultaneously singing and composing his own version.” From this case, as well as the versified version of a composition in prose, Meters of Boethius, Benson (1990:234–235) concluded: “Only our assumptions about formulaic poetry lead us to believe that such a style is necessarily connected with oral composition, and those assumptions, drawn from other literatures and other times, do not fit the facts of the Old English period: in that age literate poets could and did write heavily formulaic verse and, as the Meters of Boethius shows, they could do so pen in hand … not because demands of the meter or the pressures of oral composition prevent the poet from pausing to select some more suitable phrase but because this phrase is suitable, is part of a poetic diction that is clearly oral in origin but that is now just clearly a literary convention.” Similarly, also Finnegan 1992:70.
[ back ] 25. See Zwettler 1978:15–19; and Davidson 1994:63–64.
[ back ] 26. See Chaytor 1945. For an exhaustive survey of orality and literacy in medieval Europe, with emphasis on the continent, see Melve 2003:146–153, for the early Middle Ages (600–900), in particular, 158–161; for the Anglo-Saxon traditions, see Ambodio 2004:1–78.
[ back ] 27. See Chaytor 1945:10, where the author states: “Very different was the case of the medieval reader. Of the few who could read, few were habitual readers; in any case, the ordinary man of our times probably sees more printed and written matter in a week than the medieval scholar saw in a year. … Two points, therefore must be emphasised at the outset. The medieval reader, with few exceptions, did not read as we do; he was in the stage of our muttering childhood learner; each word was for him a separate entity and at times a problem, which he whispered to himself when he had found the solution; this fact is a matter of interest to those who edit the writings which he produced. Further, as readers were few and hearers numerous, literature in its early days was produced very largely for public recitation; hence, it was rhetorical rather literary in character, and rules of rhetoric governed its composition.” On performance, see more recently Jensen 2009:45–54; also Foley 2005:233–263.
[ back ] 28. See Chaytor 1945:5–21, 48–137, esp. 52–53, where it is stated: “Poetry was composed to be heard, not to be read; it was intended to give pleasure to the ear, and as the regular recurrence of rhythm has been found to be the simplest means of attaining this end, medieval poetry may be defined, in its formal aspect, as language in recurrent rhythmical form.”
[ back ] 29. See Zwettler 1978:16.
[ back ] 30. Zwettler 1978:19; on the aural dimension of communication in early medieval Europe, see Melve 2010:77–81; Melve 2003:153; still essential is the study of Stock (1983).
[ back ] 31. Comparing the tedious act of reading with reciting and singing, Zwettler (1978:19) states: “The process of reading, on the other hand, involved such concentration and labor, even for the practiced readers, that one cannot believe a representative audience of any taste would have suffered being read to, when two such satisfactory and proven techniques of rendition as singing and recitation were available.”
[ back ] 32. Zwettler 1978:18.
[ back ] 33. See Nagy 1996:33–37 and 65–69, considering the transcript to be a mnemonic devise for the performance and the manuscript tantamount to the performance itself. One of the instances for which one imagines a transcript was devised would have been at the oral rendition of a composition where composer-performers who were assembled to listen to the poet’s performance would have eventually produced a transcript upon which to base their future interpretations on the subject. See also Chaytor (1945:128), who says in this regard: “The process of circulation may have proceeded as follows. A new epic roman is produced and recited by the author or by a jongleur in his service; success is immediate and the reciter obtains no small advantage in the matter of gifts, presents, invitations, and reputation. Other jongleurs see no reason why this iniquitous monopoly should be preserved; the only method of breaking it is to produce the same thing themselves. The owner is not likely to lend them his copy. Therefore, one or two of them contrive to be present at as many recitations as they can attend, and proceed to construct their own versions upon a memory basis.”
[ back ] 34. On the publication of biographies in prose, see Chaytor 1950:133: “In the case of these lengthy works, it was usual for the author to prepare a final copy, well transcribed, finely illuminated and bound, for presentation for the patron.”
[ back ] 35. Benson 1990:231: “By ‘literate composition’ I mean a work composed to the accompaniment of the pen rather than the harp, and I mean further a work composed with the leisure and forethought that literary composition allows.”
[ back ] 36. Lord (1960:54) adequately describes the pressures on the poet during the composition-in-performance: “The need for the ‘next’ line is upon [the poet] even before he utters the final syllable of a line.”
[ back ] 37. See Lord 1960:128–130.
[ back ] 38. Lord 1990a:46: “The first answer would be that one cannot have formulas outside of oral traditional verse, because it is the function of formulas to make composition easier under the necessities of rapid composition in performance, and if that necessity no longer exists, one no longer has formulas. If one discovers repeated phrases in texts known not to be oral traditional texts, then they should be called repeated phrases rather than formulas [emphasis mine]. I do not believe that this is quibbling about terms, because the distinction is functional.”
[ back ] 39. Lord 1990b:392; see also Lord 1987:313–314. For a discussion of Lord’s understanding of the concept of transitional texts, see more recently Jensen 1998:94–114.
[ back ] 40. Lord 1990b:392.
[ back ] 41. Lord 1990b:394.
[ back ] 42. See already Finnegan 1974:53; and more recently Finnegan 2005:168: “There is not just one form of oral literary realization but many different arrangements along a continuum of more or less crystallized and stable oral texts. Nor is there just one relation between the ‘performed oral’ and the ‘textual written’ or always a clear distinction between them. … Writing can interact with oral performance in many different ways: as performance score, dictated transcription, crib sheet, memory cue, hearing aid, prompt book, calligraphic representation, ceremonial memento, notes for a speech, printed version of a memorized poem, medium for scholarly exegesis, tool for helping audiences understand a performance as it develops, script for recreating or remembering a past performance—and multiple possible combinations or sequences of all of these and more.” See also the important contributions in Ranković 2010a:1; Ranković 2010b:39–71; and Foley 2010:17–37, esp. 18, where the author states: “In recent years it has become clear that oral tradition and oral-derived texts … are ubiquitous, dwarfing the canon of textual literature in size and heterogeneity, and that they support the societies that use them in myriad ways.”
[ back ] 43. Among more recent (and seminal) works, the following may find special mention: Foley 2010:17–37, where the author explores the concept of verbal agoras or marketplaces for the oral–literate continuum, whereby each of these marketplaces is mainly dominated by one medium of composition, but may be permeable to other ecologies. As an example, the so-called o[ral]Agora is the space entailing oral compositions in performance, texts written to be orally performed, and texts drawing upon oral and literary traditions. Also the important contribution of Ranković (2010b:39–71), who, basing herself upon Finnegan’s work, considers the oral–literate continuum not to be merely a movement, but in the end an inevitable progress toward a literacy that may be characterized as a highly dialectical space, in which impulses may be observed in both directions of the continuum and, more importantly, as a place where three scales or axes are conflated: “One pertaining to the media in which works of verbal art are composed, another to their socio-historical fundamental contexts, and the third to the poetics they manifest” (Ranković 2010b:43).
[ back ] 44. See Ranković 2010b:43. For early medieval Europe, see also Melve 2003:160: “The ‘Great Divide’ theory—characterized by a dichotomous conceptualisation of the oral and the written in which the introduction of literacy in a given society immediately triggers change to a modern type of mentality—has not found empirical corroboration in Carolingian society.”
[ back ] 45. See Bakker 1997:9: “Societies as a whole may be oral or literate in various degrees, and since the transition from preliterate to a literate society in which writing is institutionalized is never an abrupt one, the notions of orality and literacy, though distinct, do not exclude each other, either diachronically or synchronically.”
[ back ] 46. On the concept of a non-performative oral text, see Amodio 2004:79–208, esp. 184: “The exclusively performance-based model which has so long been applied to ancient and medieval literatures is beginning to be questioned by scholars who are constructing a radically different paradigm for those supposedly oral poetries which survive only in written form from the ancient and medieval periods, a paradigm in which textuality is neither elided nor silently (and awkwardly) passed over and in which performance is viewed as a local, tradition-specific, and tradition-dependent feature which may well be nonessential to the production of verse, rather than a sine qua non of poetic composition. What is emerging … is a non-performative oral poetics that is employed not by an oral poet in the crucible of performance but rather by a literate poet who composes pen-in-hand.”
[ back ] 47. Similarly, also Finnegan 1992:70, who summarized her criticism as follows: “A ‘formulaic’ style is not therefore inevitably a proof of ‘oral composition.’ Theorists have now to accept that since there can be both an ‘oral’ and a ‘literary’ use of formulae one cannot necessarily discriminate between ‘oral’ and ‘written’ on the basis of a ‘formulaic’ style alone. In other words, the excitement attendant on the discoveries about the ‘oral-formulaic style’ led some scholars to an extreme application of its findings. Now that a certain reaction has set in against the more extreme claims, it has become clear that while the kind of diction found in Homeric or Yugoslav epics is a suitable, even likely, style of oral composition, it is not an infallible sign of it.”
[ back ] 48. See Amodio 2004:4: “We have come to see that orality and literacy exist along a continuum whose termini, ‘primary orality’ and ‘pure literacy,’ ought to be recognized as the theoretical constructs they are and not mistaken for discoverable, accessible real-world states.”
[ back ] 49. See Amodio 2004:30, describing the poetic and literate landscape of medieval England, but whose astute observations may be profitably applied to other paradigms with a comparable oral–literate nexus: “Although compositional devices such as formulas and formulaic systems serve highly specialized functions within performative oral traditions, poets, both oral and literate, continue to use them because they are the very stuff of poetic articulation and of poetic thought. … The central affective metonymic character of oral poetics that helped ensure its survival before the advent of writing does not vanish when the ephemeral, unreadable texts of oral tradition come to be fixed in writing but continues to survive once the pen and written text replace the mouth and ear.”
[ back ] 50. See Amodio 2004:28: “The manner of poetic production changes as literacy gains ground, but the shift from producing poetry exclusively during public performance, with all its attendant pressures and demands, to producing it privately and in writing does not necessarily entail a concomitant and equally dramatic alteration in poetic articulation. New models of poetic composition develop following the rise of print culture, but the chirographic tradition that emerges and gains increasing cultural importance throughout the medieval period does not simply and immediately replace the older one. Rather, these traditions simply merge to form a partnership, one that is most readily apparent in a specialized expressive economy that enables oral poets to articulate traditional verbal art and literate poets to draw upon oral poetics as they write the oral tradition.”
[ back ] 51. For an excellent taxonomy of these categories for oral traditions, see Foley 2010:19–21.
[ back ] 52. On the Bisotun inscription, see n1 of chapter 1. For the entire Old Persian corpus, see Schmitt 2009. For a more recent French translation of the Achaemenid epigraphic corpus (including variants not composed in Old Persian), see Lecoq 1997. For Achaemenid inscriptions at Naqš-e Rostam and Persepolis, see Schmitt 2000. For royal inscriptions in Susa, see Stève 1987:54–99.
[ back ] 53. See notably Skjærvø 1999 [2002]:160: “I … consider it possible that the scribes had practiced writing Old Persian for a while before they took on the Great King’s res gestae. This would explain why the orthography and style are so homogeneous”; similarly Herrenschmidt 1990:56: “Je pencherais pour la vision suivante de l’invention du cunéiforme vp: elle aurait été faite dans sa totalité et non au petit bonheur pour graver Bisotun—, avant Darius, assurement à la demande du pouvoir royal”; already Hinz 1968:98: “Aus dem §70 müssen wir jedoch entnehmen, daß die neue ‘iranische’ Keilschrift bereits vor einiger Zeit vor ihrer Verwendung am Felsen am Behistun ersonnen und ausprobiert worden ist.”
[ back ] 54. Bickermann and Tadmor 1978:239–261; also Skjærvø 1999:14–27, 60–61.
[ back ] 55. Bickermann and Tadmor 1978:239–246.
[ back ] 56. For the literary characteristics of Assyrian royal inscriptions, see still Grayon 1981; for their ideological dimension, see Tadmor 1981.
[ back ] 57. See Harmatta 1982:86–87; and Herrenschmidt 1990:49–51.
[ back ] 58. See more recently Huyse 1999 [2002]:57–58: “The king ‘dictated’ the text (‘Proclaims Darius, the king’) in his Persian mother tongue, and two bilingual scribes wrote it down simultaneously in Elamite (on clay tablets) and in Aramaic (on parchment). Then both texts were worked out and refined in the royal chancellery to meet the standards of official administrative language. … This work done, the final text was read aloud in front of Darius and submitted for approval, and, immediately thereafter, the stone masons started engraving the text on the rock.” See still Hinz 1968:97: “Wir müssen folgern, daß die altpersische Fassung der großen Inschrift zu Behistun in Wahrheit eine Rückübersetzung ins Persische aus der elamischen Fassung darstellt, die so entstanden war, daß das persische Diktat des Dareios von Elamischen Sekretären sogleich ins Elamische übertragen wurde”; also Hinz 1973:17.
[ back ] 59. This proposition, which (in)famously, albeit brilliantly, was put forward by Greshevitch, and dubbed as “alloglottography,” merely signifies, within our present confines, that in the transposition of Old Persian oral dictation Elamite (that is, Elameograms) served to capture ideographically the Old Persian composition—a process that in its finality may be comparable to the agency of Aramaic and Arameograms in writing Middle Iranian, where the language is arguably Parthian, or Middle Persian, but written by dint of innumerable (later to be) fossilized Aramaic words (Arameograms) that are read as Iranian. See Gershevitch 1979:122–124, 138–143, and esp. 138: “The Persians, after relying for one or two centuries exclusively on Elamography, gradually switched in the fifth century from it to Aramaeography, both of which from the very first moment Persians adopted either were to them simply the only method at their disposal for committing their own language to writing, a method we may term ‘alloglottography’ and use to define as the use of one writable language for the purpose of writing another language.” For a wider discussion and critique of (Gershevitch’s) alloglottography in ancient Iran, see Rubio 2006:33–41; also Yakubovitch 2008:202–211.
[ back ] 60. See Bae 2001:5–7, 32–40; and Bae 2009:137–140, 151–155.
[ back ] 61. In particular, see Bae 2001:159–160. Bae suggests the following schema: an oral Old Persian text dictated by the king and written phonetically in Aramaic letters, entailing merely the history of revolts, constituted an Urvorlage; the latter underwent editorial changes by the addition of literary devices and further details, thus giving rise to two Vorlagen: one being Old Persian written in Aramaic script, the source of the extant versions on Mount Bisotun (including the extant Old Persian version in Old Persian script) that ensued from further elaboration by independent language-specific editorial committees; and the second being an Aramaic translation of the Old Persian Urvorlage representing a “diplomatic copy,” whence all the abridged renderings of the Bisotun both in Aramaic (that is, the Aramaic copy found in Elephantine, and the fragment from Saqqara) and Akkadian (the two stone fragments from Babylon) sent throughout the empire derived.
[ back ] 62. Skjærvø 1998c:105.
[ back ] 63. Ctesias T3 and F1b §22.5; on these records, see Lenfant 2004:xxxvi–xxxix; Stronk 2010:22–25: “For the use of royal archives regarding this period … Ctesias obviously depended on hearsay”; and Llewellyn-Jones and Robson 2010:58–65, with reserved judgment.
[ back ] 64. See Harmatta 1982:86–87: “Diese Erscheinung spricht dafür, daß die babylonische Chronik, die in ihrer Form nach eigentlich Königsannalen darstellt auf die achämenidische Hofliteratur dieser Epoche nicht ohne Einfluß geblieben war. Aus der Tatsache, daß die auf Toten und Gefangenen bezüglichen genauen Angaben gerade in der babylonischen Version erwähnt werden, läßt sich sogar folgern, daß solche annalistischen Aufzeichnungen am altpersischen Königshof zu dieser Zeit auf Babylonisch geführt wurde.”
[ back ] 65. See Herrenschmidt 1990:51.
[ back ] 66. On the genesis of the Bisotun monument, see still Dandamaev 1976:23–52 and 76–90; Hinz 1968:95–98; Lecoq 1974:85–87; Lecoq 1983; and Lecoq 1997:19–72, 31–39; for a succinct overview, see Luschey and Schmitt 1990:289–305. More recently, see Huyse 1999 [2002]; Tuplin 2005:218–227; and Huyse 2009:73–83.
[ back ] 67. Schmitt (1993:89) is skeptical about the alleged “illiteracy” of the king: “Beiläufig will ich aus Anlaß dieser Frage darauf hinweisen, daß zwar immer wieder Behauptungen aufgestellt werden über die Sprach- und Schriftkenntnisse der Perserkönige und speziell des Darius, daß Be-hauptungen aber wie die, er habe außer seiner Muttersprache Altpersisch keine Sprache be-herrscht oder sei illiterat gewesen, weitestgehend unbegründet sind.”
[ back ] 68. See similarly Harmatta 1982:87: “[So] ist die Folgerung wohl unumgänglich, daß die Rahmen der assyrischen und babylonischen literarischen Formen in den Res Gestae des Darius mit den Wendungen und Erzählungsmodellen einer altiranischen oralen epischen Dichtung ausgeführt wurden.”
[ back ] 69. On the concept of a transcript, see Nagy 1996:34–36, 67–70.
[ back ] 70. Herrenschmidt (1990:51) based on her interpretation of paragraph 70 of the Bisotun inscription (on which see later in this chapter), posits that these annals were recorded in Old Persian: “Que pouvait bien être ce texte ‘en aryen,’ qui existait (auparavant) et qui a été ‘reproduit’, sinon des annales royales? Le paragraphe 70 de Bisotun implique l’existence d’annales.”
[ back ] 71. See Vallat 1972:3–13.
[ back ] 72. For a reconstruction of the Old Persian text of DSf, see Stève 1987:64–70; for the Akkadian text, see 72–77; also the discussion of several Elamite fragments of DSf on 70–71. For the Elamite text, still Vallat 1972:8–11.
[ back ] 73. For the Old Persian text of DSz, see Stève 1987:64–70; for the Elamite text, see Vallat 1972:10–13; also the discussion of additional fragments (Dsz 02 and Dsz 03, following Stève) by Stève 1987:81–82.
[ back ] 74. For the text and commentaries, see Vallat 1986:277–283; see also Vallat 1972:6–7; and Herrenschmidt 1983:177–179.
[ back ] 75. As already established in Vallat 1972, and further investigated in Vallat 1986:281, the Elamite versions of DSf and DSz had been written one after another (with DSz being the younger document). Indeed, whereas in DSf, El. 50–51, Darius still mentioned his father Hystaspes, for whom he would implore Ahuramazdā’s protection, in DSz, El. 55–56 (and its Old Persian variant DSz 37), as well as the Akaddian text DSaa 39–40, there were no mentions of Hystaspes; which would suggest he had passed away by then. By the same token, Hystaspes’ death establishes the age of DSz and DSaa relative to the older DSf. Although the lapse of time between the redaction of the DSf foundation charts and the DSz and DSaa seems to have been brief—as shown by Vallat (1986:281): “On peut logiquement conclure que Hystaspe est décédé entre le moment où les différentes versions de DSf ont été déposées dans les fondations de l’Apadana et celui où DSz et DSaa ont été placées dans le mur de la maison du Roi. Or nous savons aujourd’hui que tout le complexe palatial a été construit d’un seul tenant. Ces deux opérations ont donc dû se succéder dans un laps de temps très court, de l’ordre de quelques mois au maximum. Il est ainsi possible de situer la date de la mort de Hystaspe aux environs de 519/518”—it is still noteworthy that the Elamite versions of DSf and DSz show a sufficient number of differences to make them not copies of one another, but independent redactions, perhaps by different scribal ateliers?
[ back ] 76. Vallat 1972:4–7.
[ back ] 77. Herrenschmidt (1983:177–179), commenting on the composition of DSaa, points out that the scribes composing the Akkadian text, as opposed to other versions of the foundation chart, made two lists: one for materials used in the construction of the palace, wherein these are listed according to their nature (e.g., precious metals; gems; variety of woods); and another one to indicate the provenance of these primary materials, that is, the countries that provided them. Whereas the categorization of the first list may have possibly derived from a bookkeeping list (“un semblable ordonnancement des matériaux peut provenir soit d’une liste de magasins royaux, soit d’un mémento du scribe” [Herrenschmidt 1983:178]), the list of purveying countries is disconnected from that of the construction materials, as it seems to enumerate countries absent from the combined lists of DSz and DSf. It thus appears that the scribes responsible for the Akkadian variant merely copied a stereotypical list of nations without checking on its congruence with that of the construction materials: “On a donc recopié une liste extérieure incompatible avec ce qui précède et, ce faisant, on ne s’est pas soucié du fait que le document recopié était ancien, qu’il ne correspondait pas aux provenances réelles des matériaux” (Herrenschmidt 1983:179). Compare also the role of scribe(s) (teppir/tuppip) who are identified in the Elamite Persepolis tablets as (Babylonians) writing on leather (Papilip kuš ukku “Babylonians [writing] on parchment).
[ back ] 78. Vallat 1972:7.
[ back ] 79. The Elamite version (as well as the Aramaic version discovered at Elephantine) contained calques of Old Persian in addition to the inevitable termini technici; on these calques, see Dandamaev 1976:78–79; see also Reiner 1960:222–227; Friedrich 1949:1–29; Skalmowski 1976:217–229; also Tuplin 2005:219–221; compare Gershevitch 1983:51–56. The presence of these calques in the Elamite and the Aramaic (Elephantine) versions has traditionally led to the assumption that they were translated from Old Persian during dictation. For an opposite and persuasive view, see Herrenschmidt 1990:52–54; similarly Bae 2001:159–160.
[ back ] 80. Tavernier 2008:59–85; Tavernier 2011:191–261.
[ back ] 81. Henkelman 2011:575–634; Henkelman 2008; Henkelman 2003:181–231.
[ back ] 82. On recent interpretations pertaining to the formation of the Persian ethnogenesis, as resulting from Elamite and Persian acculturation, see de Miroschedji 1985:292–296; de Miroschedji 1990:69–72; de Miroschedji 2003:34–36; Stronach 1997:37–39, 49; and Stronach 2000:682–685; also Henkelman 2008:41–57; Henkelman 2011:584. Tavernier (2011: esp. 243)—based on an exhaustive survey of Iranian linguistic elements and onomastica in late Neo-Elamite administrative documents (including but not limited to the Acropole texts from Susa; the (so-called) Apadena texts; and the “Neo-Elamite Letters”), dating to early to mid sixth century BCE—has further substantiated the proposition that the origins of Iranian–Elamite acculturation are, as expected, to be placed in the late Neo-Elamite period.
[ back ] 83. For examples of this impact-induced imposition (of Old Persian on Elamite), see Henkelman 2011:590–595; on contact linguistics and the phenomenon of contact-induced imposition, see Winford 2005: esp. 420–421; and Winford 2007:36.
[ back ] 84. Interestingly, a group of Persian apprentice-scribes/subordinates (puhu paršipe), who reportedly were responsible for copying tablets (tuppime sapimanpa [on the meaning of tuppi-me, see below, pp. 97–98]), is mentioned in four documents of the Persepolis Fortification tablets (PF 0871; PF 1137; NN 1485; and NN 1588). What renders their mention noteworthy—aside from the fact that the great number of rations they receive identifies them as high-ranking Persians within the Persepolis administrative hierarchy—is the likelihood of these Persian scribes writing in Elamite; this for several reasons: (1) Elamite constituted the language predominantly used to write on clay tablets (tuppime) in the Persepolis Fortification tablets, and (2) what is more, other scribes engaged in writing Aramaic on leather were distinguished by being labeled as “Babylonians” (Papilip kuš ukku “Babylonians [writing] on parchment). See Henkelman 2008:349–350n818: “The ‘Persian servants who are copying cuneiform documents’ … are a group of scribes copying Elamite tablets. That their ethnicity is spelled out by means of the rare label ‘Persian’ may logically be explained as a means to distinguish them from the often-mentioned Babylonian scribes. That these (elite) servants were labeled as ‘Persians,’ not “Elamites,’ is not surprising: only if one continues to envisage the Elamites of the highlands as a socially and culturally distinct group would Persians writing Elamite constitute a paradox. If one accepts, on the other hand, the basic notion that ‘Persian’ culture was heir to both Iranian and Elamite traditions, there is noting unusual in Persians writing Elamite.” Also Henkelman 2011:587–589; Tavernier 2008:64–65; and Lewis 1994:26–28, esp. 28: “We are dealing with Persian scribes writing Elamite.” For a diverging view, deeming the writing of the puhu paršipe to be in Old Persian, compare already Hinz 1973:22: “Darius [ließ] persische ‘Buben’ schreiben lernen—sicher nicht auf elamisch, aramäisch oder akkadisch, sondern auf persisch”; and Hinz and Weber 1972:292–293; on the identity of the puhu as apprentice-, or highly specialized, copyists, see Giovanizzo 1995:146: “Proponiamo quindi per m/fpuhu la traduzione ‘apprendista,’ ‘aiutante’”; also Lewis 1994:27.
[ back ] 85. Shayegan 2011:260–287.
[ back ] 86. On the impact of Assurbanipal’s foundational cylinder (on Babylon’s great wall, the Dur-Imgur-Enlil) as a literary model on the Cyrus Cylinder, see Kuhrt 1983:88–92; and Kuhrt 2007:169–191; also Harmatta 1974:29. On Assurbanipal’s titles as reflected on his foundation cylinder at the Dur-Imgur-Enlil: Anaku Aššurbânaplu šarru rabû šarru dannu šar kiššati šar māt Aššur šar kibrāt irbitti apil Aššuraḫiddina šarru rabû šarru dannu šar kiššati šar māt Aššur šakkanak Bābili šar māt Šumeri u Akkadî, “I am Assurbanipal, the great king, the strong king, king of the world, king of Assur, king of the four world quarters, son of Assurhaddon, the great king, the strong king, the king of the world, king of Assur, the governor of Babylon, king of the land of Sumer and Akkad.” See still Streck 1916:235–238; for comparable titulary in Assurbanipal’s historical prism inscriptions, see Piepkorn 1933:28–29 (B i.1–5); and Borger 1996:92, 205 (B §1, i.1–5); 193–194, 209 (J 1–13).
[ back ] 87. On this subject, see Shayegan 2011:287–291.
[ back ] 88. On the titles sunkik anšan šušunka and hal-menik hatamtik āk šušenki, see Shayegan 2011:283–287; Grillot 1984:185–191.
[ back ] 89. Henkelman (2011:586–587 and 594) rightly points also to the presence of a minority of Elamo-phone scribes in Persepolis, writing in correct Elamite, possibly continuing Neo-Elamite writing practices, but which as a group still represented a rarity in contrast to the Iranophone scribes.
[ back ] 90. Shayegan 2011:289–291.
[ back ] 91. See Rollinger 1993:359 and 368.
[ back ] 92. Intriguingly, Xerxes’ chancellery continued to redact minor royal inscriptions in the three languages of the empire—with their majority imitating the models established by Darius—but there were also genuine new (bi- and trilingual) compositions, such as sections of the harem (XPf) and the daiva (XPh) inscriptions. Another example in point is XPcb, whose Babylonian version was—due to ideographic writing—shorter than the space allotted to it on the southern portico of the tačara palace in Persepolis, prompting Achaemenid scribes to alter the usual Akkadian formula in order to maintain the spatial symmetry between the three versions placed side by side. See Kozuh 2003:266–270; also Herrenschmidt 1983:180. The royal chancellery’s capability to redact new texts in Babylonian and Elamite, as exemplified in XPh and XPg, as well as its ability to make impromptu changes to traditional formulae, may indicate that in principle it possessed the expertise to introduce the new title of šar māt Parsu u māt Mādaya in Babylonia, which may, or may not, have been inspired by an Elamite model.
[ back ] 93. However, the relevance of regnal titles as indicative of changes within the central authority’s ideological postures remains subject to caution; see Kessler 1984:262–263: “Die Belege zeigen, daß die Führung dieses oder jenes Titels offensichtlich mehr von bestimmten Schreibtraditionen und –Schulen abhängig war, als etwa von dem Diktat der persischen Verwaltung.”
[ back ] 94. Henkelman 2011:595: “Achaemenid Elamite should be defined against the background of the Persian ethnogenesis as a Persian thing, and as part of the Persian identity.”
[ back ] 95. The cause of similarities between Old Persian and Avestan, as suggested by Skjærvø, may be due to (1) a common Indo-Iranian heritage; (2) a common Iranian heritage; or (3) direct quotations from the Avesta—or a combination thereof. From the alternatives proposed by Skjærvø, only the second assumes the existence of an Iranian oral tradition as the source of parallels between the Achaemenid inscriptions and the Avesta. The third alternative, in contrast, holds the Avesta to be the source of those similarities, without presupposing any other oral tradition beside the Avesta itself; see Skjærvø 1999:60: “There remain a few instances of parallels for which both Indo-Iranian heritage and Near Eastern influence seem excluded. If we assume these parallels are due to common Iranian heritage we must also conclude that the themes and forms in question (1) belong to the pre-Avestan period and (2) were transmitted for over a millennium with utmost faithfulness, which is possible. Keeping in mind, however, that the Avestan calendar was present in the Achaemenid empire from the 5th century—but had not been inherited by the Old Persians, who continued to use their own calendar—by far the simpler scenario would be to assume that the parallels in the last group are due to direct influence of the Avesta itself.”
[ back ] 96. On paragraph 70 of the Bisotun inscription, see also the pertinent remarks of Tuplin (2005:224–227).
[ back ] 97. On the unusual use of mā-taya “so that not” in the present clause, see still Utas 1966:139–140; and Schmitt 1991:71n71.
[ back ] 98. For the alternative reading *daθans “being vigorous,” in lieu of *hu-tava- “able,” see Schmitt 1991:71nn71f., who reluctantly follows in this Gershevitch 1959:197–199; also more recently Schmitt 2009:84.
[ back ] 99. See the reserved judgment of Skjærvø 1999 [2002]:160.
[ back ] 100. Schmitt 1991:73; also Schmitt 1990:59: “Schriftform, -gestalt, -bild.”
[ back ] 101. Schmitt 2009:87.
[ back ] 102. Schmitt 2009:87n489: “/patišam kar/ is am ehesten mit Lazard … und Huyse … als feste Phrase, hier als ‘hinzufügen’ übersetzt, aufzufassen.”
[ back ] 103. Lazard 1976:180–184.
[ back ] 104. Huyse 1999 [2002]:45–66.
[ back ] 105. Huyse 1999 [2002]:48: “What Darius is trying to tell us in the Old Persian version, is—I think—that he had that text added on the opposite side of (which here means: beneath) the relief and the Babylonian and the (original) Elamite versions.
[ back ] 106. See Skjærvø 1999 [2002]:160.
[ back ] 107. See the excellent discussion of the Old Persian and Elamite variants by Herrenschmidt (1989:193–208).
[ back ] 108. Huyse 1999 [2002]:46–47.
[ back ] 109. See notably Grillot-Susini 2008:29.
[ back ] 110. Lecoq 1974:67–73.
[ back ] 111. Tavernier 2007b:64–67.
[ back ] 112. Tavernier assumes two different origins for Neo- and Achaemenid Elamite tuppi-, which results from the semantic merger of the Akkadian loanword ṭuppu “inscription; tablet” and tippi-, which is derived from the Elamite root tep- “fashion” (ursprungsverschieden from Akkadian): “To summarise, in the Old and Middle Elamite periods the root tep-/tip- had nothing to do with the act of writing. In the Middle Elamite period a new form entered the Elamite lexicon: tuppi(me), ‘text,’ which was an Akkadian loanword. This loanword should not be confused with the Neo- and Achaemenid Elamite homograph tuppime (pronounced /tippime/) and is only attested in this and the Neo-Elamite period, but the phonetic similarity with Elamite tep-/tip- triggered the merger of the two forms. By the Neo-Elamite period tep-/tip- had two meanings: (1) ‘to form, shape’ and (2) ‘to write.’ ”
[ back ] 113. Lazard 1976:183–184.
[ back ] 114. Huyse 1999 [2002]:47.
[ back ] 115. See Skjærvø 2009a:80, 3.3; and Skjærvø 2007a:897–898.
[ back ] 116. See Herrenschmidt 1989:198: “Si patikara signifie ‘reproduction,’ il est bien possible que la locution verbale patišam kar- signifie reproduire.”
[ back ] 117. Among others, see Grillot, Herrenschmidt, and Malbran-Labat 1993:59; Huyse 1999 [2002]:48; also Vallat 2011:280–281, for variations on the passage’s translation.
[ back ] 118. See Hinz and Koch 1987:1.247: “andersartig”; Herrenschmidt 1989:200: “autrement”; Khačikjan 1998:43: “differently”; Huyse 1999 [2002]:48: “otherwise”; Quintania 2001: “differente(mente)”; Stolper 2004:83: “differently.” Compare also Malbran-Labat 1992: “Le suffixe -ikki est, dans l’ensemble de ce texte—comme le plus souvent—caractéristique des locatifs. Je proposerais donc de comprendre daiae ‘autre’ + ikki comme une notion locative ‘ailleurs’.”
[ back ] 119. Vallat 2011:265: “Mais il exsiste une hypothèse qui n’a jamais été envisagée. On peut considerer que ikki n’est pas un suffixe de daae mais qu’il entre en composition avec huta. Le composé ikki- hutta signifie alors ‘j’ai traduit’ … L’analyse de daae est plus simple. daae signifie ‘autre, différent’ et le mot peut être utilisé adverbialement: ‘autrement, d’une autre manière.”
[ back ] 120. See Grillot-Susini 2008:27, 98: “vers-executer” > “ajouter.”
[ back ] 121. See Hinz and Koch 1987:II.747: “bei; hin … zu/auf; nach”; Khačikjan 1998:15–16; Grillot-Susini 2008:65–66; compare Grillot-Susini 1987:26 “près; après; ensuite.” For adverbial function of the suffix -ikki, see Herrenschmidt 1989:198–201; and Khačikjan 1998:43.
[ back ] 122. Herrenschmidt (1989:200–201) had already expressed this sentiment with respect to daʾe-ikki and hutta, which still remains a strong possibility: “tá-a-e-ik-ki signifierait quelque chose comme ‘autre(ment)’ et avec le verbe hutta ‘faire’: ‘j’ai fait autrement’; cette locution serait le pendant du vp. patišam kar- ‘reproduire.’ Le scribe élamite aurait donc rendu kar- par hutta ‘faire’, ce qui est normal, et patišam par taaeikki.
[ back ] 123. DB El. 4.2–5.
[ back ] 124. Now that an example of “administrative” Old Persian texts written on clay among the Elamite Persepolis Fortification Tablets has been discovered, the possibility of Old Persian transcripts of the Bisotun inscription may be envisaged; on the said unique Old Persian administrative clay tablet, see Stolper and Tavernier 2007:9–18.
[ back ] 125. On the so-called diplomatic copies in Aramaic and Akkadian sent to Babylon and Egypt, in general see Bae 2001:33–51. For a survey of the extant fragments of the Babylonian stele depicting Darius I and possibly the two Babylonian Lügenkönige along the processional way in Babylon, and containing an abridged variant of the Akkadian version of the Bisotun inscription (DB Akk.), see Seidl 1999a and Seidl 1999b. On the inscribed fragments belonging to the Babylonian stele, notably BE 3627 (= VABab. 1502), which entails rests of thirteen lines corresponding in content to lines 55–58 and 69–72 of DB Akk., see Voigtlander 1978:63–65; and Seidl 1999a:101–102 (1), 108; for the textual remains of BE 3627, see Voigtlander 1978:64–65. Other inscribed fragments of the stele are Bab. 59328, which seems to complement the right edge of fragment BE 3627, albeit with lacunae; see Seidl 1999a:102 (7), 109–110. Also Bab. 41446, which entails text corresponding to lines 91–92 and 108–109 of DB Akk.; see Seidl 1999a:102 (2); and Voigtlander 1978:66. On the remaining fragments Bab. 59245 and Bab. 59246, which presumably contain texts paralleling respectively DB Akk. 15–18 and 26–27, see Seidl 1999a:102 (3; 6), 108–109. The Babylonian stele, as reconstructed by Seidl, further substantiates the assumption of “diplomatic” variants circulating in the empire, which might have been rendered kulturspezifisch, in view of the targeted audience, by specialized scribal ateliers writing in Aramaic, Akkadian, or Old Persian? What is more, if Seidl’s reconstruction of the relief on the Babylonian stele is accurate, and merely Babylonian usurpers were depicted thereon, then presumably the iconography of the Bisotun relief was also specifically revisited on “diplomatic” copies to befit the cultural context and political reality of recipient countries. One may also think, paying heed to Seidl’s train of thought, that “diplomatic” copies may have contained both text and image designed by Darius’ chancellery for local consumption: “Man könnte sich vorstellen, daß Bildhauermodelle mit der Szene von Bisutūn, villeicht als leicht reproduzierbare Tonreliefs in alle Teile des reiches verschickt wurden, und Daß von dieser Vorlage vor Ort vergrößerte Umbildungen geschaffen wurden, die dem jeweiligen Gebiet angemessen waren … möglicherweise wurden aber auch schon die regionalen Bilder in einer zentralen königlichen Werkstatt entworfen und als Modell in die Reichsteile geschickt” (Seidl 1999a:113); also Seidl 1999b:303; and Bae 2001:43. See also Millard 1991:213, for a description of artists accompanying Assyrian royal campaigns and recording scenes that would later inform the composition of royal reliefs. Of bearing on the present discussion is also the possible invocation on the stele of divinities proper to Babylon, such as Marduk (ina ṣilli d Bēl “under the protection of Marduk”), rather than Ahuramazdā, which further exemplifies the extent to which “diplomatic” variants are tailored to the targeted cultural context; Seidl 1999a:109–110 and 1999b:299; also Bae 2001:50. For a comparison of BE 3627 with the Elephantine Aramaic variant, see Porten 1982:8–11 and 15–16, where the author remarks that the similarity between the Aramaic variant and the Babylonian fragment, in contrast to the Akkadian version of the Bisotun inscription, is not coincidental, and one ought to presume a certain interdependence between them: “Comparison of the Babylonian fragment and Aramaic text with the Bisotun text indicates that both the former are slightly variant and abridged version of the original. While the Babylonian fragment and Aramaic text are not identical, the coincidence of many of their variants indicates a certain interdependence.” For the Aramaic variant from Elephantine, see Porten 1982; for the small Aramaic fragment found in Saqqâra, see Segal 1983:85, no. 62.
[ back ] 126. On the Rabatak inscription of the Kušān king Kaniška, which dates to the first half of the second century CE, see Sims-Williams and Cribb 1995–1996:77–142; for the dating, 105–107. The text of the Rabatak inscription has been recently reread and re-edited by Sims-Williams (2004 [2008]:55–57, and 65 for commentary). See also discussion in Shayegan 2011:19–20.
[ back ] 127. Text and translation following Sims-Williams 2004 [2008]. On the meaning of *kṣatriyas/*kṣatrapas, see n60. For a tabulation of expressions used in the above cited passage of the Rabatak inscription:
ιωναγγο οασο
“Greek announcement/*edict”
οζοαστο < *uz-vāsta-
“brought out; *issued”
ωσταδο < *ava-stāta-
“placed; put into (Aryan)”
φροαγδαζο < *fra-vāxta- + *ahāz (< *ahāt?)
“was proclaimed (to India …)”
[ back ] 128. Herodotus and Xenophon observe that Persian (oral) accounts of varied historical events were told or chanted (λεύγεται καὶ ἄδεται) during their lifetimes; see Herodotus 1.95; and Xenophon, Cyropaedia 1.2.1.
[ back ] 129. For a formulaic analysis of Kerdīr’s inscriptions, as well as a complete inventory of themes used in these inscriptions, see appendices I and II. Formulae (or repetitions) that have occurred at least twice, word for word, are underlined with a continuous line [—], in contrast, formulae that are variations from other encountered word-clusters, and thus represent “varying formulae,” are underlined with a dotted line […].
[ back ] 130. Parry (1980:304) writes: “Ιt is important at this point to remember that the formula in Homer is not necessarily a repetition, just as the repetitions of tragedy are not necessarily formulas. It is the nature of an expression [italics mine] which makes of it a formula, whereas its use a second time in Homer depends largely upon the hazard which led a poet, or a group of poets, to use it more than once in two given poems of a limited length. We are taking up the problem of the Homeric formulas from the side of the repetition, but only because it is easier to recognize a formula if we find it used a second or a third time, since we can then show more easily that it is used regularly, and that it helps the poet in his verse-making [italics mine].”
[ back ] 131. Lord 1990:394.
[ back ] 132. Lord (1990:394) thus explains this situation: “The first answer would be that one cannot have formulas outside of oral traditional verse, because it is the function of formulas to make composition easier under the necessities of rapid composition in performance, and if that necessity no longer exists, one no longer has formulas. If one discovers repeated phrases in texts known not to be oral traditional texts, then they should be called repeated phrases rather than formulas. I do not believe that this is quibbling about terms, because the distinction is functional.”
[ back ] 133. See also Balcer 1994:257–264. On other oral traditions preserved by Herodotus, see Sancisi-Weerdenburg 1994:39–55.
[ back ] 134. Compare Skjærvø 1998c:100.
[ back ] 135. See the explanations of Sims-Williams (1995–96:83) on similarities between the Bactrian Rabatak inscription and DB; he seems to favor a written Aramaic path: “It is not inconceivable that both Kushans and Sasanians may have had direct knowledge of the content of Darius’ inscription, presumably through an Aramaic version such as that found at Elephantine. … The alternative is to suppose that such later echoes of the contents of the Achaemenian inscriptions are due to an oral tradition.” See also, more recently, Skjærvø 2009c:283: “It is true that copies of the Aramaic version of Darius’ Bisotun inscription on papyrus or parchment might have been circulated throughout the Achaemenid empire and that Alexander’s soldiers might have preserved a copy they found in Bactria (or Persepolis, for that matter), which Kanishka’s speech-writers then might have consulted to produce the inscription. The principal means of dissemination of Darius’ inscription to his mostly illiterate subjects, however, is likely to have been by royal heralds.”
[ back ] 136. See Gignoux 1991:30: “Les inscriptions de Kirdīr, quant à leur forme, ont un double caractère, à la fois très structuré et très répétitif … les répétitions martèlent le cours du texte. On les trouve partout, et particulièrement dans l’introduction au récit de la vision, où il semble que l’idée-force qui justifie celle-ci pourrait être dite en quelques mots. Ces deux caractères stylistiques me semblent attester le passage de l’oralité à l’écriture, de la tradition orale à des texts écrits, dont le seul support est la pierre. L’appel au lecteur réitéré dans chacune des inscriptions semble être un ‘appel au secours,’ car bien peu nombreux devaient être ceux qui étaient en mesure de lire et de comprendre ces textes.”
[ back ] 137. For a survey of Arsacid inscriptions, see Skjærvø 1995 [1998]:283–318.
[ back ] 138. For a more recent, exhaustive, inventory of common Akkadian, Elamite, and Old Persian themes as illustrated on Mesopotamian and Achaemenid inscriptions, see Skjærvø 1999:14–27.