Multilingual document

From PALaC Wiki
Jump to navigation Jump to search


documento plurilingue | document multilingue | mehrsprachiges Dokument


A multilingual document can be generically defined as a document written in more than one language (generally but not necessarily within a single epigraphic community). The notion of “document” can be interpreted in two ways: on the one hand, a multilingual document can be understood as a single material object on which one or more texts in different languages are written; on the other hand, if we understand a document as a text and not as a physical object, then texts in different languages do not need to share the same material support to constitute a multilingual document.

As suggested by Adiego (2012), multilingual documents can be classified according to the relationship between the texts in different languages written on them, which can be:

  1. corresponding to each other: they show more or less the same content, i.e. there is only one textual content, fully replicated in multiple texts written in different languages;
  2. complementary: they provide different pieces of information, i.e. there is only one text, partly written in different languages;
  3. consecutive: one or more texts in different languages are added over time to a document previously written in a given language, and all the texts are somehow related to each other;
  4. not related to each other: they are totally independent, although they are recorded on the same material support.

A further issue concerning multilingual documents with corresponding texts is what can be called the problem of the original version. In dealing with multilingual documents, it is generally assumed that there must be an original text written in a given language – allegedly the native language of the author/scribe – and its translation(s), and most of the studies are focused on finding the original text, through the analysis of interference phenomena that may reveal the translation. Such a situation does indeed occur in some examples, but it is not always the case, as emerges from several examples where mutual interference phenomena can be observed, which makes it impossible to identify an original version (one must always bear in mind that back-translations also exist, so that the alleged original text may not be the one written in the native language of the author/scribe). Indeed, as remarked by Giusfredi (2018, 117), in most cases it makes little sense to look for the model and the replica, because the multilingual document as a whole should be regarded as the original (especially if produced in an ancient multilingual area, having been conceived as such from the beginning.

Therefore, when judging the relationship between the different texts in a multilingual document, a number of other elements should be taken into account, such as the origin and native language of both the author and the scribe of the text, as well as the number of scribes involved, because a single scribe may not be always responsible for all versions on a multilingual document. Each of these aspects should be carefully evaluated on a case by case basis.


Each of the types of multilingual documents mentioned above can be found in the 2nd and 1st millennium BC Anatolia.
In the 2nd millennium BC Anatolia, multilingual documents in which the texts show a relationship of correspondence include e.g. the Akkadian-Hittite Testament of Hattušili I (CTH 6), the Hattian-Hittite myth of the moon that fell from heaven (CTH 727), and the Hurrian-Hittite “Song of Release” (CTH 789). In each of these documents, the Hittite version is on the right columns of the tablet, while the left columns contain the Akkadian, Hattian, or Hurrian text, respectively. In other cases, a different layout is found, in which corresponding lines or paragraphs in different languages alternate with each other within the same column, e.g. the Hattian-Hittite ritual for the consecration of a temple (CTH 725), the Sumerian-Akkadian hymn and prayer KBo 7.1+ (CTH 794), and the Sumerian-Akkadian incantation KUB 37.111 (CTH 801.4).

As for the 1st millennium BC, we can mention e.g. the Phoenician-Luwian bilingual inscriptions of KARATEPE and ÇİNEKÖY, the trilingual Letoon stele at Xanthos (N320), with inscriptions in Lycian, Greek, and Aramaic, and the Lydian-Aramaic bilingual of Sardis (LW 1).

As mentioned, texts in different languages that form a bilingual document do not need to be recorded on the same material support. Indeed, in the Hittite archives, there are several examples of bilingual documents in which each version is recorded on a separate monolingual tablet. Some Old Hittite political documents belong to this typology, e.g. the Annals of Hattušili I (CTH 4) and the Edict of Telipinu (CTH 19), both with Akkadian and Hittite versions, as well as all the Akkadian-Hittite international treaties. Some Hittite translations of original Hurrian and Akkadian compositions (e.g. mythological texts, prayers, and many omina series) also belong to this group.

Multilingual documents with complementary texts in different languages include several Hittite rituals with recitations in Akkadian, Hattian, Hurrian, Luwian, and Palaic. In all these texts, paragraphs in Hittite language provide the framework with the instructions for the ritual, while the use of foreign languages is restricted to the magical formulas to be uttered concurring with the ritual actions, usually introduced by a clause that declares the language (e.g. ‘the practitioner speaks as follows in Akkadian/Hattian/Hurrian/Luwian/Palaic’).

Examples from the 1st millennium BC are rarer, but we can mention the Greek-Phrygian bilingual inscription 19.1 (= 96), only consisting of a curse formula with a Greek protasis and a Phrygian apodosis, and the Assyrian-Urartian bilingual of Kevenli/Šušants (CTU A 5-44), in which the dedicatory inscription is in Assyrian, while the blessing formula is written in Urartian.

An example of multilingual document in which the different texts are consecutive may be the Greek-Carian bilingual stele of Hyllarima (C.Hy 1), in which the majority of the Greek text was probably written long after the Carian one (although some Greek lines are perhaps contemporary to it).

Finally, multilingual documents with unrelated texts in different languages may be exemplified by some Hittite Sammeltafeln, e.g. KUB 17.28, containing a Hattian incantation among other Hittite compositions, and the Lycian-Greek bilingual inscriptions TL 5 (Telmessos) and TL 115 (Limyra), in which the respective Greek texts were engraved later than the Lycian ones and show no connection with them with regard to content.

As far as multilingual documents with corresponding texts are concerned, examples mentioned above differ according to the relationship between the texts in different languages they consist of. In several cases, we are dealing with multilingual documents consisting of a “foreign” original text (in Akkadian, Hattian, or Hurrian) and its corresponding translation into the Hittite language, e.g. the Hurrian-Hittite “Song of release”, the Hattian-Hittite myth of the moon that fell from heaven, as well as Akkadian-Hittite omina series, wisdom literature, etc. In cases like these, an original text obviously exists, provided with a Hittite translation (thus becoming part of a multilingual document) only at Boğazköy, so that we should expect interference phenomena mostly in the Hittite version, being the product of a translation (while in the original text one might rather find transcription mistakes).

Quite different is the case with the Akkadian-Hittite bilingual political texts of the Old Kingdom, such as the Annals of Hattušili I and the Edict of Telipinu. In this case, we are not dealing with imported Akkadian materials provided with a Hittite translation, but with texts originally composed in the Hittite capital. Indeed, the analysis of the Akkadian versions of these documents reveals several interference phenomena showing that the texts were conceived in Hittite language (cf. e.g. pahru ibbašû in the Akkadian version of the Edict of Telipinu, calquing the corresponding taruppanteš ešer ‘were united’ in the Hittite version, and several other examples listed by Marazzi 1986). Strictly speaking, however, it does not mean that these documents were originally meant to be bilingual, nor that a Hittite written version existed from the beginning (bilingual tablets may be later copies): the "original" text, if there is any, could still be the Akkadian one, although composed and drafted by a non-Akkadian scribe.

A similar situation can be found in Akkadian-Hittite international treaties, with the difference that they were conceived from the beginning as bilingual texts (cf. Del Monte 1980 and Del Monte 1986 for interference phenomena in Akkadian-Hittite treaties). However, among these texts, also Hittite (back-)translations from Akkadian can be identified, e.g. KBo 10.12+ (Treaty between Šuppiluliuma I and Aziru of amurru, CTH 49.II), where several elements point to a Hittite translation from an Akkadian text, such as the imperfect translation of Akk. mithuṣu with Hitt. walh- (ii 31’), the Hittite unusual phrase išhiulaš lenkiyaš (iii 24’), translating Akk. ša riksi u ša māmīti, the occurrence once of the Akkadian phrase ÉRINMEŠ GIŠGIGIRMEŠ (ii 26’) instead of the Hittite one ÉRINMEŠ ANŠE.KUR.RAMEŠ (used elsewhere in the text), etc.

Turning to the 1st millennium BC, it is particularly interesting to look at the case of the Phoenician-Luwian bilingual inscriptions of KARATEPE and ÇİNEKÖY. Contrary to the previous assumption of an original Luwian text translated into Phoenician (cf. e.g. Payne 2007), Yakubovich (2015) recently argued for the priority of the Phoenician text on the Luwian one, as shown by the altered word-order and some aberrant constructions in the Luwian version, calquing Phoenician ones.

Another complex example is the Lycian-Greek-Aramaic trilingual inscription of the Letoon at Xanthos. The text in the three versions represents a decree of the Carian satrap Pigesere/Pixodaros concerning the establishment in Lycia of the cult of a Carian god, the “King of Kaunos”. In general, the Lycian and the Greek versions correspond quite well to each other regarding the content, with some phenomena of interference in the Greek text (e.g. the use of ὑός in the filiation formulas, calquing the Lycian structure with tideimi ‘child’, and several calques on the Lycian word order). The Aramaic version seems to be condensed instead, but it contains some important information missing in the other two versions, such as the date and the designation of Pixodaros as satrap of Caria and Lycia, while in the Lycian and Greek versions he is referred to only as satrap of Lycia. Conversely, the names of the officials appointed by the satrap are lacking, as well as other matters of local jurisdiction. Furthermore, some information contained in the Aramaic text seems to be recorded in a more accurate way than in Lycian and Greek, particularly the name of the deity (or deities) joint to the King of Kaunos: as suggested by Carruba (1999), the Aramaic generic knwth ‘his colleagues’ may be the appropriate translation of an original Carian designation referring to the divine circle of the main deity, appearing in the other versions as the obscure arKKazuma/Αρκεσιμα. This seems also to suggest that at least two different scribes were involved in the production of this document, one responsible for the Lycian and possibly also Greek text, one responsible for the Aramaic version and more familiar with the Carian cult concerned. Therefore, the evaluation of this trilingual inscription cannot be merely reduced to the relationship between the longer, allegedly original Lycian text, its Greek “imperfect” translation, and an Aramaic summary. It can be safely stated that the Aramaic text had a different scope, being the official document from the point of view of the Persian administration.


Adiego, I.-X. (2014), Las inscripciones plurilingües en Asia Menor: hacia una clasificación tipológica y un análisis funcional, in W. Eck, P. Funke (eds.), Öffentlichkeit – Monument – Text. XIV Congressus Internationalis Epigraphiae Graecae et Latinae. 27. – 31. Augusti MMXII. Akten (CIL, Auctarium, Series nova 4), Berlin – Boston: de Gruyter, pp. 231-269. Carruba, O. (1999), Ar/w/wazuma, Kadmos 38, pp. 50-58. Del Monte, G.F. (1980), Traduzione e interferenza nei trattati siro-hittiti, Vicino Oriente 3, pp. 103-119. Del Monte, G.F. (1986), Il trattato fra Muršili II di Ḫattuša e Niqmepaᶜ di Ugarit (Orientis Antiqui Collectio 18), Roma: Istituto per l’Oriente C.A. Nallino. Giusfredi, F. (2018), On Phoenicians in Ptolemaic Cyprus: A Note on CIS I 95, Vicino Oriente 22, pp. 111-120. Marazzi, M. (1986), Beiträge zu den akkadischen Texten aus Boğazköy in althethitischer Zeit (Biblioteca di Ricerche Linguistiche e Filologiche 18), Roma: Dipartimento di Studi Glottoantropologici, Università «La Sapienza». Payne, A. (2007), Multilingual Inscriptions and Their Audiences: Cilicia and Lycia, in S.L. Sanders (ed.), Margins of Writing, Origins of Cultures (Oriental Institute Seminars 2), Chicago: The Oriental Institute of the University of Chicago, pp. 125-140. Yakubovich, I. (2015), Phoenician and Luwian in Early Iron Age Cilicia, Anatolian Studies 65, pp. 35-53.