2020.05.06, updated 2020.05.15 | By H. Craig Melchert
§1. My principal aim in what follows is to consider the possible modalities by which a Mycenaean-Hittite diplomatic correspondence might have been carried out. Since the very notion of such a correspondence is controversial, I will begin by briefly reviewing what I take to be established facts or well-founded hypotheses about the issue and what I find more speculative or totally unfounded.
§2. I now regard as established that Ahhiyawa of the Hittite texts refers to a Mycenaean Greek kingdom not located in Asia Minor. Those who wish to wait for the proverbial “smoking gun” may do so, but the circumstantial evidence is now overwhelming. The alternative hypothesis of Hajnal (2003:40–42) of Ahhiyawa as a small city state of Cilicia is not credible. Hittite references show that Ahhiyawa was a formidable power influential in far western Asia Minor. I leave to others the problem of determining just which Mycenaean kingdom (or kingdoms) should be identified with the Ahhiyawa of the Hittite texts.
§3. The author (in the sense of the sender of the message) of the cuneiform letter KUB 26.91 found in Hattuša was a king of Ahhiyawa: see the independent arguments of Gurney (2002:135), citing already Kammenhuber (1981, pers. comm.), and of Starke cited by Latacz (2004:243–244), but now compare Weeden 2019:223–225! As per Starke, we may also be sure that this letter is not isolated, but forms part of an extended correspondence.
§4. KUB 26.91 is written in the standard Boğazköy ductus of the Neo-Hittite period: the tablet was thus inscribed by a scribe of Hattuša or trained in Hattuša. The Hittite chancellery did not typically make multiple copies of letters (see van den Hout 2002:864 and also 872–873 for a notable exception). The extant tablet is thus a contemporary copy, either sent from elsewhere and received in Hattuša or written there based on a message sent in some other form.
§5. After hearing the full argumentation in oral presentation, I now find fully persuasive the analysis by Starke (already cited by Latacz 2004:244) that ḫamakta in KUB 26.91 Ro 9 is used in the sense of “betrothed” or “married” and refers to a previous (peran) dynastic marriage between Ahhiyawa and Aššuwa, a key to understanding the text as a whole.
§6. On the other hand, Starke’s claim that the sequence ka-ga-mu-na-aš-za-kán of KUB 26.91 Ro 8 is to be read as *kat(a)mun=as=za=kan with a Hittite accusative singular form of the name Kadmos, king of Thebes, is quite impossible. The absence of the determinative for a personal name (a single vertical stroke) is not a serious obstacle, and the following A-BA A-BA A-B[I] ‘forefather’ in fact virtually demands that the form be a masculine personal name. As per Starke, there are also parallels to justify the emendation of the sign GA to TA(!). However, as already pointed out by Katz (2005:424) in his review of Latacz (2004), the analysis of an accusative singular *kat(a)mun plus enclitic subject -aš cannot be correct, since transitive verbs never occur with enclitic subject pronouns in Hittite (as shown by Garrett 1990, following Watkins). One therefore cannot accept an analysis that depends on an emendation that in turn produces an ungrammatical sequence.
§7. It is important to stress that all evidence of the extant text of KUB 26.91 argues that the author (that is, the composer of the text) was a Hittite native speaker (whose language contained Luvianisms, expected in Neo-Hittite). Contrary to the claims of Starke, there are no non-native usages in the text available to us. The only features that even call for comment involve the treatment of clitics.
§8. In Ro 5 we find the sentence [x]-ra-a-an-ni MU.KAM-ti=mu ŠEŠ=YA ḫa-at-r[a-a-e/iš] ‘In [ ] year my brother wro[te] to me’. The position of the clitic -mu ‘to me’ is in no way ungrammatical. As already seen by Sommer (1932:271), it shows merely that the preceding phrase is a close syntagm and thus may count as a single accentual unit. One may compare [MU-ti] mēni=ma=šši INA ITU.12.KAM… ‘in the course of the year over the twelve months…’ (KBo 14.142 i 42, NH/NS). See Güterbock and Hoffner (1980–1989:233) for the restoration and the remark “The position of =ma=ši after mēni proves that witti mēni is a single accentual unit.” One may also compare clause-initial takšan šarran=ma cited below.
§9. Other assured examples of such “phrasal stress” do argue that this feature was limited to set phrases, not freely formed syntagms (see Melchert 1998). This already casts doubt on the restoration by Starke of the opening phrase as [ku]rānni MU.KAM-ti ‘in the year of separation’. This analysis must assume that an expected construction with adnominal genitive *kurānnaš witti has been replaced with so-called “partitive apposition.” This assumption is quite impossible, since all examples of “partitive apposition” in Hittite, as expected, involve a whole-part relationship: see, for example, the genuine example MU-ti mēni ‘in the course of the year’ above. Since “separation” is patently not a whole of which “year” is a part, an apposition *kurānni MU.KAM-ti would never have occurred in Hittite.
§10. The repetition of the enclitic reflexive particle in Ro 8 in nu=za kagamunaš=za=kan may be a simple scribal error. The -za-kán appears to be written over an erasure. The scribe, having begun the clause with nu=za, belatedly realized he had forgotten the sentential particle -kan, which should have appeared after the -za. He thus began again and added both clitics to the nominal subject (in itself also perfectly grammatical), but failed to go back and delete the original nu=za. For a similar error one may compare […] ŪL daḫḫun [n]=at=za takšan šarran=za daḫḫun takšan šarran=ma ANA m.dSIN-dU EGIR-pa peḫḫun ‘I did not take [it all]. I took half of it for myself and I gave half back to Arma-Tarhunta’. (KUB 21.17 ii 1–4). Here the scribe, having begun with ‘and it for myself’, then changed to a construction with contrastive clause-initial #takšan šarran…#takšan šarran=ma and accordingly added -za to the first instance, producing a mixed construction with redundant -za.
§11. In summary, then, we find in Hattuša a Hittite-language version of a letter from a king of Ahhiyawa to the Hittite king, responding to a letter sent to him by the latter, written in standard Boğazköy ductus and so far as the extant text is concerned in quite idiomatic Hittite of the Neo-Hittite period. How are we to imagine that this correspondence was carried out?
§12. The closest available model we have in trying to address this question is that of the Egyptian-Hittite correspondence (for which see globally Edel 1994). This exchange generally employed Akkadian, but there are some letters attested in Hittite (Edel 1994:1:214–233 and 2:320–355). The ductus and language again match those of Hattuša. As per Edel (1994:2:320), Hittite versions of letters sent to Egypt may be copies of drafts translated and sent in Akkadian. Hittite versions of letters from Egypt must have some other source and motivation.
§13. One possibility is that the Egyptian pharaoh had Hattuša-trained scribe(s) in his employ and for special reasons in exceptional instances had letters to the Hittite king and queen composed and sent in Hittite. The “Arzawa” letters in Hittite exchanged between the pharaoh Amenophis III and Tarhuntaradu, king of Arzawa, show the presence of such scribes in Egypt. I would argue, however, that the situations here are quite different. The well-established international use of Akkadian as an Ancient Near Eastern diplomatic language and the implication of accommodation by the pharaoh in a matter of national prestige make such a practice vis-à-vis the Hittite king very unlikely. I know of no evidence for reciprocal use of Egyptian by the Hittite king. If we assume with Starke (1981:22624 and 23043) that the letter VBoT 2 represents a letter from Tarhuntaradu to Amenophis III, then we know that the use of Hittite by the pharaoh in writing to the king of Arzawa was in response to a direct urgent request from the latter (VBoT 2:25). The pharaoh’s use of Hittite was thus a matter of practical necessity, surely due to the incapacity of Tarhuntaradu to correspond in Akkadian. In any case, the use of Hittite would have a demeaning, not complimentary effect for the king of Arzawa (his exchange with the pharaoh is not in Luvian).
§14. I therefore find it more likely that letters received from Egypt in Akkadian as usual were translated by Hittite scribes for purposes of drafting replies. This hypothesis is consistent with the remark of Edel (1994:2:320) that all such extant letters are closely tied in content with letters sent to Egypt. It is unlikely that outgoing letters on such affairs of state were drafted without careful review of the previous correspondence in both directions, which would have been kept together as a sort of dossier. The frequent backward references in the extant letters confirm such consultation. Hittite versions of incoming letters would have facilitated discussion of appropriate replies with the king, queen, and advisors not fluent in Akkadian.
§15. I believe that the element of concern for loss of prestige and maintenance of coequal status applies equally as well to exchanges between the Mycenaean and Hittite kings as it does to the Egyptian-Hittite correspondence. This factor makes unlikely the employment of Hattuša scribes by a Mycenaean king to compose and inscribe letters in Hittite to send to the Hittite king. Why would he make such an accommodation? If the Mycenaean ruler in question was not truly a co-equal in power with the Hittite king (as suggested by Trevor Bryce at this workshop in this volume), his resulting insecurity about his international status would make such a concession even less likely.
§16. I find it equally implausible that any Hittite king would have made the accommodation of importing and employing Linear B scribes just for the purpose of reading or composing texts in Greek. The situation here is very different from that of Akkadian, for which the Hittite king had many reasons for employing scribes fluent in the language. In contrast to the situation vis-à-vis Egypt, there is also the question as to whether the Hittite kings ever had stable friendly relations with the kings of Ahhiyawa for a long enough period to have made elaborate provisions for intercommunication worthwhile.
§17. What other scenarios may we entertain? In preparing these remarks, I seriously considered the possibility that the actual messages were conveyed between the respective capitals orally by messengers, the written text being dictated to scribes only after receipt, perhaps as per above only for use in preparing replies. However, the verb used in KUB 26.91 by the king of Ahhiyawa is ḫatrā(i)- (Ro 5 as discussed above and also Ro 11). This verb does not mean “to write” in the sense “to inscribe” in attested Hittite, but as Gary Beckman has reminded me, it does mean “to send a written communication (about).” This is shown by among other things the Amarna letter VBoT 2, whose author insists that the addressee respond not merely with a messenger, but also via a written tablet: kūnn=a memian tuppiaz EGIR-pa ḫatrāi ‘also write back about this matter by means of a tablet’ (lines 12–13). We must therefore conclude that the Mycenaean-Hittite correspondence reflected in KUB 26.91 was written.
§18. I repeat, however, that all current evidence argues that KUB 26.91 is a Hittite translation. Where then was the translation made? One possibility is that messages were conveyed in writing to the respective frontier outposts of each kingdom in its own language and script (an accompanying oral version conveyed by messenger is not excluded). Messages were then conveyed to the other side orally under conditions of mutual security, where they were written down in the other language and sent on to the capital. This scenario implies employment at the frontier of trusted bilingual speakers. I view it as virtually certain that for military purposes each side would have found it expedient, indeed indispensable, to employ such speakers in the frontier area of Western Asia Minor that was the point of both contact and frequent dispute.
§19. However, I also find attractive an alternative as suggested at this conference by Gary Beckman: that the written correspondence was carried between the respective capitals by pairs of messengers, one from each side. Each representative would have been either a trusted official of some standing or at least would have had tangible signs that he was sent by his ruler. The presence of one person from each side would have assured passage across the frontiers in each direction and in times of high tension or even hostility would also have provided a built-in hostage. The Amarna letter VBoT 2 again provides support for this model. The writer begins by complaining that the addressee’s messenger Kalbaya suggested a marriage alliance, but that this was not on the accompanying tablet. He then insists that if the addressee is sincere about the proposal, he should send back at once both his messenger Kalbaya and that of the writer, as well as confirm it in writing.
§20. While we cannot be sure that the practice described here reflects standard practice, it makes very good sense. Each side would have written the letters in its own language and script in its own capital, under the respective kings’ oversight and with the full resources of each court. Translation took place at the receiving end, including preparation of a written version if it was desired for drafting replies or any other purpose. Regarding just what sort of controls were used for the oral translation of the received messages (to assure that they accurately reflected what was written) we can only speculate, and I forgo a rehearsal of these here.
Beckman, Gary M., Trevor R. Bryce, and Eric H. Cline. 2011. The Ahhiyawa Texts. Writings from the Ancient World 28. Atlanta.
Bryce, Trevor. 2019. “The Kingdom of Ahhiyawa: A Hittite Perspective.” Studi Micenei ed Egeo-Anatolici, n.s. 4:191–7. Forum article with responses.
Edel, Elmar. 1994. Die ägyptisch-hethitische Korrespondenz aus Boghazköi in babylonischer und hethtischer Sprache. 2 vols. Abhandlungen der Rheinisch-Westfälischen Akademie der Wissenschaften 77. Opladen.
Forrer, Emil. 1932. Aḫḫijavâ. Reallexikon der Assyriologie. 1:53–57. Berlin and Leipzig.
Garrett, Andrew. 1990. “Hittite Enclitic Subjects and Transitive Verbs.” Journal of Cuneiform Studies 42:227–242.
Gurney, Oliver. 2002. “The Authorship of the Tawagalawa Letter.” In Silva Anatolica: Anatolian Studies Presented to Maciej Popko on the Occasion of His 65th Birthday, ed. Piotr Taracha, 133–141. Warsaw.
Güterbock, Hans G., and Harry A. Hoffner. 1980–1989. The Hittite Dictionary of the Oriental Institute of the University of Chicago. Volume L–N. Chicago.
Hajnal, Ivo. 2003. Troia aus sprachwissenschaftlicher Sicht. Die Struktur einer Argumentation. Innsbruck.
Hoffner, Harry A., Jr. 2009. Letters from the Hittite Kingdom. Writings from the Ancient World 15. Atlanta.
van den Hout, Theo. 2002. “Another View of Hittite Literature.” In Anatolia Antica. Studi in Memoria di Fiorella Imparati, ed. Stefano de Martino and Franca Pecchioli Daddi, 857–878. Florence.
Katz, Joshua. 2005. Review of Latacz 2004. Journal of the American Oriental Society 125:422–425.
Latacz, Joachim. 2004. Troy and Homer: Towards a Solution of an Old Mystery. Trans. Kevin Windle and Rosh Ireland. Oxford.
Melchert, H. Craig. 1998. “Poetic Meter and Phrasal Stress in Hittite.” In Mír Curad: Studies in Honor of Calvert Watkins, ed. Jay Jasanoff, H. Craig Melchert, and Lisi Oliver, 483–494. Innsbruck.
Puhvel, Jaan. 1991. Hittite Etymological Dictionary. Volume 3, Words beginning with H. Berlin and New York.
Sommer, Ferdinand. 1932. Die Aḫḫijavā-Urkunden. Abhandlungen der Bayerischen Akademie der Wissenschaften, Philosophisch-Historische Klasse, n.f. 6. Munich.
Starke, Frank. 1981. “Zur Deutung der Arzau̯a-Briefstelle VBoT I, 25–27.” Zeitschrift für Assyriologie 71:221–231.
Weeden, Mark. 2019. “Hittite-Ahhiyawan Politics as Seen from the Tablets: A Reaction to Trevor Bryce’s Article from a Hittitological Perspective.” Studi Micenei ed Egeo-Anatolici, n.s. 4:217–27.
 Due to circumstances beyond the control of the organizer, Professor Annette Teffeteller, the proceedings of the workshop “Mycenaeans and Anatolians in the Late Bronze Age: The Ahhijawa Question,” held at Concordia University, January 4–5, 2006, in Montreal were never published. After keeping the original version of this paper online on my personal webpage ever since, I had decided finally to remove it from “papers to appear.” It seemed far too outdated to publish. I have now instead accepted the generous offer of Gregory Nagy and Keith DeStone of e-publication in Classical Inquiries, with due revisions and updates (in blue). I have done so for two reasons. First, the quite contradictory views of Hoffner (2009:290–291) and Beckman as cited in Beckman et al. (2011:138–139) regarding the modalities of Ahhiyawan-Hittite correspondence show that the topic retains its currency. Second, the challenge by Weeden (2019) to the dominant view that KUB 26.91 represents a letter from a king of Ahhiyawa also renews the debate about this fascinating but very fragmentary text, although there is no question that such correspondence did take place between the respective rulers. Although my criticisms of some of Frank Starke’s claims made at the workshop are at this point “old news,” it may do no harm to bring them to the attention of those unfamiliar with the debate of well over a decade ago.
 See for a very recent proposal Bryce 2019 and the invited responses to it.
 I would now qualify this interpretation as likely or probable, but far from assured. One must stress that the object of ḫamakta is lost in a break and need not refer to a woman at all.
 In view of the mention of LUGAL KUR A-a[š-šu-wa] in the immediately preceding Ro 7, it is also likely that Kagamuna (sic!) is the forefather of the King of Aššuwa, not of the King of Ahhiyawa. Restore then A-BA A-BA A-B[I=ŠU].
 I must insist on the fundamental methodological point that alleged non-native features based on free restorations of the text have no probative value. It is totally illicit to restore line Ro 2 in a way that places the sequence ku-e-ša-an (kue=šan with sentential clitic -šan) in the middle of a clause and then argue that this reflects a non-native usage. Sound philology requires rather the assumption that kue=šan (if correctly read) is clause-initial, and that any restoration contradicting this is false. The reading k[u]-e- is not assured (see Weeden 2019:223).
 Nor could such an apposition have been modeled on Greek usage. Thus the alleged Greek source of the use of Hittite kuer- to refer to the cutting off of land (already without probative value, being based on a free restoration) disappears entirely. No certainty is possible, but the context strongly suggests that [x]rānni MU.KAM-ti means “last year, in the preceding year” (thus already Forrer 1932:57): ‘You, my brother, wrote me [last] year’. The Hittite for this expression is unknown. The use of parā in peran parā ‘previously’ and parā ḫand(ant)ātar ‘(divine) providence’ < *‘attending to in advance’ (with Puhvel 1991:104–107) gives at least some basis for an adjective *parānni- ‘vorig, previous’. One must, however, stress that no such word for “prior” is attested in Hittite and that the standard Hittite expression for “last year” is wē/īttantanni (MU-anni). With space for only one sign and the position of the clitics, it is hard to imagine another restoration, but one should concede that this is a free restoration that is not entirely assured.
 I do follow Starke in reading [n]u-za at the start of the line, following Goetze’s autograph, against Sommer. Weeden (2019:224) reads x-a-ka-ga-mu-na-aš=za=kán, implying that the first two signs are part of the personal name, which would eliminate the problem entirely.
 I see no reason to think that current sensibilities regarding language use between heads of state did not apply among rulers of the Late Bronze Age. We know that they were quite sensitive regarding other matters of relative rank and status.
 It is also important to note that VBoT 1, the letter from Egypt, does show evidence of having been written (that is, composed) by a non-native speaker. Beside much correct usage it attests frequent second-position verbs, singular enclitic possessives with plural nouns, and the problematic words pippit and zinnuk. For the latter as Egyptian see Starke 1981. For his discussion of interference effects on the Hittite see Starke 1981:223–224. It is striking that there are no clear instances of non-native usage in VBoT 2 putatively sent by Tarhuntaradu.
 For a summary of its usage see Puhvel (1991:269–274), whose etymology implying that the verb did once mean “to incise” is plausible, but not assured.
 Such a procedure may be mentioned in the “Instructions for Lords, Princes, and Courtiers” (KUB 21.42+ iv 11–13): mānn=a LUGAL LÚSAG kuinki ANA ZAG KUR araḫzena<š> LUGAL-i uiyazi apāš=ma=kan INIM.MEŠ LUGAL waḫnuzi n=at tameda <memai> “Also if the king sends some courtier to the boundary, to the king <of> an external land, but he changes the words of the king and speaks otherwise…”. See for the text Miller 2013:290, but he wrongly ignores ZAG ‘boundary’, interpreting the first clause as “if the king sends some courtier to the king <of> a neighboring land…”. Compare rather von Schuler 1957:28: “wenn der König einen „Oberen‟ zur Landesgrenze zu einem benachbarten König schickt…”. The tentative rendering of tameda is mine.
 I must stress that I indebted to Professor Beckman for the basic idea. I am solely responsible for the elaboration of it that follows. Beckman as cited in Beckman et al. 2011:138–139 tentatively suggests rather that the respective rulers did employ in their courts bilingual scribes capable of composing and writing texts in the language of the recipient. For reasons given above I still very much doubt this.
 I am unable to affirm whether my late colleague and collaborator Harry Hoffner (2009:290–291) arrived at the same conclusion independently or was persuaded by my proposal.
Cover image: Ruins of the palace at Hattuša. Image via Flickr, shared under a CC BY-SA 2.0 license.