1 Denomination and Identity

The Hittite documents contain expressions such as hattili, nešili, hurlili, luwili and babilili. These are examples of a typical adverbial form that probably originated from the dat./loc. of adjectives in -li-¹ and has the formal meaning of ‘in the manner of.’ The formation is widely used to refer to linguistic behavior: for example, hattili means ‘in Hattian,’ nešili, ‘in Nesic,’ hurlili, ‘in Hurrian,’ luwili ‘in Luwian’ and babilili, ‘in Babylonian.’ Hattili generally introduces texts or portions of texts written in the non-Indo-European isolate language that is now called Hattic or Hattian. The term Proto-Hattian (German, prohattisch; Italian, protocattico proto(k)hattico; French, proto-hittite) is common in the older secondary literature.² It arose from the idea that Hittites, as Indo-Europeans, could not be treated as an autochthonous Anatolian population. It was too easily believed that people speaking a non-Indo-European language that was attested only in Anatolia must have been the autochthonous population of at least the central part of modern Turkey.³ Thus ‘Hattian’ became the name used for both a language and a population. Yet historically the Hittite kingdom was known as the ‘land of Hatti.’ So, in the historical kingdom that we know as Hittite, people wrote on cuneiform clay tablets using the Indo-European language that we call Hittite as their main language but identifying their territory and institutions with a name that referred to the territory whose language should have been the non-Indo-European language Hattian (hattili). The relationship between the original Hatti and the Hatti of the Hittites is a fascinating case, not only for cultural and linguistic contact research but also for research on the value of language in politics and cultural identity. Today we call the main language of the Hittite clay tablets ‘Hittite’—using a word deriving from Hatti—because of a tradition established when modern scholars began studying the clay tablets found in Hattuša. The political name Hatti survived the fall of the kingdom. Thanks to the so-called Neo-Hittite states, it lasted into the first millennium BCE and entered the Biblical tradition.⁴ The expression originally used to refer to Hittite language was nešili, that is, ‘in the manner of (the people of the city of) Neša.’⁵ Attempts made in the first half of the 20th century to rename the Hittite language Nesite or Kanesite failed.⁶

The clearest and most concise summary of this situation was given by H.G. Güterbock in a paper published in 1957, from which the passage below is quoted.

[…] The situation becomes more complicated if cultural manifestations other than languages are taken into consideration: what is Hittite art, Hittite architecture, Hittite pottery, in short, Hittite civilization? Can such names be used at all, and to what kind of Hittites do they refer? […] If we want to reach some clarity we must strictly separate two spheres: linguistic and cultural. Since the name ‘Hittite’ has, for forty years been applied to the main language of the Boğazköy archives, we cannot easily abandon it (although there are other names for the same language […]). The speakers of this language took part in what may be called ‘Hittite civilization,’ but the latter is a mixed culture and cannot in its entirety be ascribed to a single ethnic group. Consequently, the name ‘Hittite’ must mean one thing if applied to a language, another thing if applied to a civilization (Güterbock 1957:233–234).

One of the main goals of the research has been to define the contribution of the Hittites to the Hittite civilization. This could be pursued by investigating the contribution of the other ‘ethnic groups’ that could be projected behind the linguistic denominations and the texts in languages other than Hittite that are recorded in the Hittite tablets.⁷ The Hattian contribution, based on the texts we have, was almost exclusively restricted to the cult. Thus Hattian texts are mainly concerned with cult-related performances: music and dance with songs and formulas, recitations, incantations, and narratives (myths). The Hittite word hattili is chiefly used in Hittite texts as a description or instruction for specific performances as in the following examples.

KUB 1.17 iii 48–49 = CTH 591.II.A, Klinger 1996.

(48) ^LÚALAMZU₉ ha-at-ti-li (49) ki-iš-ša-an me-ma-i

“The ^LÚALAMZU₉ recites in Hattian as follows”⁸

KUB 2.13 v 2 = CTH 591.IV.A, Klinger 1996.

(2) ^LÚNAR ^URUha-at-ti-li SÌR-RU

“The singer sings in Hattian”

Sometimes, instead of the adverbial form hattili, scribes used the adjective hattili-.

VAT 7683 iii^? 11′–12′ = CTH 591.IV.D, Klinger 1996.

[^L]^Ú.MEŠNAR ha-at-ti-li-eš (12′) [SÌ]R-RU

“The Hattian singers sing”⁹

In other instances, the word hattili is omitted. This is the case for the main bilingual tablets (CTH 725 and 726) recording the rituals that relate to building activities of the palace.

KUB 2.2+ ii 38–39 = CTH 725.A, Schuster 1974

(38) ^LÚzi-li-pu-ri-ya-tal-la-aš (39) a-pí-ya-ak-ku a-ni-ya-zi ta ke-e INIM^MEŠ me-ma-i

“The Zilipuriatalla carries out (rites) in that place, and recites these words” (i.e., “this story, these facts”)

The adverb hattili is used with the verbs mema- ‘speak (of), tell, recite’; halzai- ‘cry, call out, summon, invoke’; malt- ‘declaim, recite, vow’; SÌR-RU ‘sing’ (zamārum in Akkadian and išhamai- in Hittite), and annia- ‘perform, carry out, (magically) treat’. The focus on cult performance is evident.¹⁰ Some of the tablets offer guidelines for performances; others also include the text to be delivered. In some cases, the text and the instructions related to them may be preserved on separate tablets.¹¹ The case of CTH 591 is of particular interest. This catalog entry details multiple manuscripts of the Festival of the month (Fête du mois). Some (I.1.A in Klinger 1996) preserve the Hattian texts and others do not. According to Steitler (2014), Ms. I.1.A (KUB 1.17) may reflect a misunderstanding of Hattian recitations. Although the obverse seems to respect the correct connection between Hattian recitations and the “description of the ritual activities [because] the deity honored by a particular rite is subsequently identified in the appurtenant recitation” (2014:301), rites for the Kaneš gods are interpolated on the reverse, disrupting “the original correspondence between rites and recitations” (ibid.). Steitler concludes that even if the Hattian texts were not well understood, they constituted an “expression of the Hittites’ own cultural identity” (ibid.).

The presence of the Hattian texts probably implied the existence of a cultural context (mainly religious and ideological) toward which whatever was, or should have been, originally Hittite converged, leaving present-day researchers without a picture of a pure and original (Indo-European) Hittite contribution as distinct from the Hattian one. Klinger (1996) and Steitler (2017) emphasize the historical relevance of Hattian cults during the existence of the Hittite kingdom; thus the Hattian Kultschicht (or milieu, following Steitler’s analysis) characteristic of the older phases likely survived into later phases.

2 The Textual Documentation

2.1 Writing Habits

Hattian texts are preserved on typical documents of the Mesopotamian tradition, that is, through cuneiform writing on clay tablets kept in archives and/or libraries (in a broad sense). We have no evidence of monumental or display texts. The Hittite culture developed schools and systems of conservation and cataloging,¹² along with textual genres, formats, and layouts.¹³ Of particular importance for the Hattian documentation is the format of the bilingual tablets, on which Hattian texts are transcribed with corresponding translations in Hittite (cf. below).

The Hittites adopted a form of cuneiform from the Syro-Mesopotamian tradition, although with idiosyncrasies that diverge from the Babylonian standard (see Chapter 6). The orthographic system applied to texts in the Hattian language reveals further distinguishing characteristics. Some concern the repertoire of signs. The Hattian texts abound in signs built on the cuneiform PI, to which a smaller sign is juxtaposed—generally, but not exclusively, a vowel (e.g., PI_A). The PI sign has a pure consonant value, so the subscript signs function as an indication of the vocalization of the syllable. In Hittite the syllabic value /pi/ is never rendered with PI, for which BI is used instead; therefore, BI is also transcribed as -pé- and -pí-. The syllabic value of PI is usually /wa/. Therefore, the PI_V combinations are transcribed as -wV_V- (wa_a-, wi_i-, -wu_u-, -wú_ú- etc.). It is believed that the sign PI with subscripts is used to indicate a fricative of the labiodental type, such as [f] or [v]. The Hittite scribes used concordance tables to identify correspondences between signs of this type and simple signs; the tables were practical solutions to cases of variation in spelling resulting from divergent traditions or schools or simplifications. It is not uncommon to find variants in which signs such as BI (-pí/é-) and PI_V alternate. The orthography of Hattian in Hittite documents has been recently reviewed in two important works: Soysal 2004 and Simon 2012.

There is plenty of variation in the documents in Hattian. Many words show graphic variants. Some are the result of mistakes. However, it is important to remember that both the Hittite and Hurrian orthographies include variation. For example, in Hittite, spellings with and without consonant gemination may alternate; the same can be said for vocalic scriptio plena or the use of -Ci- and -Ce- syllabograms. Due to the nature and frequency of these spelling variants, it seems reasonable to interpret them, at least to an extent, as real alternatives, that is, instances of some sort of allography.¹⁴

Hittite documents in Hurrian, for example, diverge from the orthography of the Mittani letter. Mittani Hurrian shows unambiguous spellings in many cases in which ambiguities exist in Hittite Hurrian. For instance, we can consider syllabographic pairs such as KI and GI and KU and GU, which are unambiguously used in the Mittani letter to code vowel quality: KI is /ki/, GI is /ke/, KU is /ko/, and GU is /ku/.¹⁵ In Hittite Hurrian, there are on the contrary many cases of non-strict one-to-one correspondence. For example, according to Giorgieri and Wilhelm (1995), a Hurrian syllable with /e/ could be written using signs of the type -Ci- and -iC- (e.g., NI or IN), -Ce-, and -eC- (e.g., NE or EN); syllables with /i/ could be written only with -Ci- and -iC- sign types.¹⁶ A situation in which one sign has a unique phonemic correspondence, whereas another sign corresponds to both that sound and a different one, may be represented as a contrastive opposition between fully specified vs. underspecified features. Tentatively, the Hittite Hurrian orthographic treatment of /i/ and /e/ with CV and VC syllabograms can be represented as follows:

-Ce- and -eC-: [+front +mid], that is, only the mid-front vowel

-Ci- and -iC-: [+front -low], that is, any higher front vowel.

Other variations are not as coherent as these. For example, the GA, KA, and QA signs may alternate in the same text to render a velar plosive with /a/.¹⁷

Returning to Hattian, it should be evident by now that we face two theoretical pathways for considering variant spellings: comparing manuscripts to identify error patterns and comparing words and word forms to identify functional values. To be sure, we have instances of both cases, but to uncover functional values we need to consider all solutions that are typical of Hittite documents: not only signs that may alternate or seem alternate freely for the same value, but also cases of underspecification, such as the Hittite plene-vowel spellings, which should be treated as more specific than the non-plene ones, or the scriptio geminata, which should be treated as more specific than the scriptio simplex.

The Hattian situation proves to be rather obscure. A recent and major effort to determine the value of variant spellings is found in Simon 2012, to which we refer the interested reader. Even if the phonological values proposed in that study cannot be taken as definitive, it is very important to stress that scribal mistakes, misunderstandings, uncertainties, and the like are not enough to explain the situation.

A structural analogy among Hittite, Hurrian, and Hattian orthographies must be assumed; a simple transfer of the Hittite and Hittite Hurrian patterns, however, is not convincing, at least not in all cases. Both Hittite Hurrian and Hattian use the same scribal habit in providing the sign PI with a mater lectionis. As already stated, the sign PI is never used with value /pi/ or /bi/ in Hittite, Hurrian (including the Mittani letter), or Hattian. The sign PI in the Mittani letter has four values, conventionally transcribed as -wa-, -we-, -wi-, -wu- (it works like a consonantal sign and phonologically most likely codes a labiodental fricative, either /f/ or /v/). Hurrian and Hattian in the Hittite documents discriminate values by adding a subscript vocalic sign, PI_V, (i.e., -wa_a, -we_e, -wi_i, -wu_u, -wú_ú). While we notice here a more detailed system, we have to remember that those signs could alternate with signs for labial plosives (PA, BI, PU), a nonadmissible option in the Mittani letter.¹⁸

As mentioned above, the scribes had at their disposal concordances that listed such alternatives. One example survives on a tablet, which was published as KBo 37.21; the scribe drew a table on the lower edge containing the alternatives.¹⁹ See the picture of KBo 37.21 in this Fig. 9.1.

View Full Size

Figure 9.1

The table of orthographic variants in KBo 37.21

• Download Figure
• Download figure as PowerPoint slide

The signs aligned in table layout at the bottom of the tablet read as follows:

21	[waa ]	pa-a	wii	p[í-i ]
22	[wee ]	pé-e	wuu	[pu-u ]
23		vacat	wúú	[pu-ú ]

A similar, complete table is preserved in Emar (Msk 7462).²⁰ The table in KBo 37.21 is for Hattian, and that in Msk 7462 is for Hurrian. The Hattian and Hurrian texts show further composite signs of the same type but with a VC or CV syllabogram such as PI_AB, PI_BI PI_BU, i.e., -wa_ap-, -wi_pí-, or -wu_pu- (see HZL for details) that structurally conforms to the situations traced in the tables in KBo 37.21 and Msk 7462. Alternations between the signs noted in the school tables must have been available and accepted alternatives rather than uncertainties. The reason why the scribal community did not develop a more efficient system is not discernible in the surviving data. Nonetheless, we must consider that what may appear dysfunctional to us could have been perfectly functional for the scope and the objectives of the original system.

Another interesting but more obscure example of Hattian orthographic problems is the alternation between vowel plene writing and geminate writing of consonants for the same word or word form (see Soysal 2004 for details). We list just a few significant examples:

te-e-pu-ut	te-ep-pu-ut
up-pí-in	u-pí-i-in
a-ši-i-ia-ú-i	a-aš-ši-ia-ú-i
(Soysal 2004:75).

This kind of evidence prevents us from transferring the norms of Hittite orthography indiscriminately into Hattian texts.

The issues described thus far should be sufficient to suggest how problematic the study of Hattian orthography can be. Other challenges are detailed in Soysal 2004 (chapter 2).

3 The Status of Hattian in Hittite Anatolia

To examine the status of Hattian during the existence of the Hittite kingdom, we have to assume a few initial distinctions. First, we must distinguish between an ‘extinct’ and a ‘dead’ language. The former is here assumed to be a language that no longer has speakers, whether native or nonnative. The latter is here assumed to be “[o]ne that is no longer the native language of any community. Such languages may remain in use, like Latin or Sanskrit, as second or learned (e.g., as liturgical) languages” (CODL², s.v. dead language).

We can easily exclude that Hattian was an extinct language. There is abundant evidence of performers using Hattian: it may have been a very corrupted form of the original language, perhaps performed without precise knowledge of the original tongue, but it was certainly in use.

Almost all scholars in the field judge, based on the sources, that understanding of the language became quite poor by the era of the Hittite Empire;³⁴ some researchers believe that it had become a dead language by the Old Hittite phase,³⁵ but others disagree. Soysal (2004:14) finds plausible a scenario in which Hattian was lost over the years that the Hittite kingdom flourished. However, this scenario cannot be proved or disproved because it is based on state official documents produced by Hittites who did not use Hattian outside the cultic sphere. Soysal defines the Hattian preserved in the documents as a “professional language of priests and other cult functionaries, which barely developed and was memorized in trivial phrases and repeated over and over again in liturgical formulas” (2004:15).³⁶ Moreover, he considers that the language was used incorrectly in the Hittite scribal schools (2004:27). In a 1981 paper, Gerd Steiner went so far as to argue that Hattian was the true native language of the population of the core region of the (old) Hittite kingdom. Hittite (i.e., Nesic), according to Steiner, was a supraregional language, known natively (if ever) only in Kaneš and used as a language of communication that was neutral with respect to the various linguistic components of the kingdom,³⁷ which were mainly Hattian and, eventually, Luwian.³⁸

Although we can be safe in hypothesizing that Hattian as witnessed in the Hittite documents was a learned liturgical language used mainly, if not exclusively, by trained specialists, it is very hard to prove that it was no longer the native language of any community across the entire period covered by the Hittite written documentation. Considering the reports that there were songs and recitations performed in Hattian in certain places in the Hittite territory, we cannot exclude that it remained the native language of some part of the population. However, the Hittite tablets neither state this explicitly nor offer sure contexts where this situation could be inferred.³⁹

It is highly probable, however, that Hattian-speaking populations were in contact with Indo-European Anatolian speakers before the era for which documentation exists. In a 2008 study of the language communities of central Anatolia in the Old Assyrian Colony period, Goedegebuure presented a very interesting hypothesis: that a large population speaking a (proto-)Luwian dialect shifted to Hattian during or before the Old Assyrian Colony period, producing syntactic features in the latter that resulted in typological asymmetries.⁴⁰

Goedegebuure recalls that, when a large population group is speaking a low-prestige language, some people may decide or be forced to learn and use a more prestigious language. Over time, especially in crucial places such as political, administrative, or economical centers, the linguistic habits of the population converge, resulting in a language shift: the wholesale use of the prestigious language. In this scenario, the Hattian language that we know would be a language learned imperfectly by an originally non-Hattian speaking population and used and handed down as such (2008:166).

But we must stress a point that perhaps was not fully highlighted in Goedegebuure’s article: another Hattian variant, previously and perhaps also contemporaneously, must have been spoken by communities not descending from the ones that shifted. The Hattian of such speakers would not have shown the asymmetries in question. Thus the Hattian texts we read must have been produced by the ‘new’ speakers of Hattian (primarily descended from the shifting community), and the new Hattian variant—and this is the main point—must have become more prestigious than the older one.

From the historical point of view, Goedegebuure prefers a scenario in which Luwians (or proto-Luwians) rather than Palaeans or Hittites merged with the Hattian population of central Anatolia before the conquest of the kingdom of Hatti by the Hittites. Many questions remain about this reconstruction. The arguments are coherent linguistically, but the historical attestations are so scanty that to accept this scenario without reservation would be imprudent. To support the theory, Geodegebuure searched for hints that would allow us to postulate the presence of Luwians in the Hatti area during or even before the Colony Period. The role played by Luwians in the Hittite kingdom, however, is not a strong argument. Nor is the establishment of the original location of some dialects witnessed in the archives of Hattuša;⁴¹ Luwian dialects preserved in Hittite documents are unlikely to be those of the communities that shifted to Hattian—if the Luwian of the Hatto-Luwians survived, then the sociolinguistic conditions required for the shifting scenario are not fully met. The various Luwian dialects documented in the Hittite archives must have had sufficient prestige to be preserved, and the Luwian of the hieroglyphic inscriptions is certainly posterior to the shift; none of these dialects are relevant to Goedegebuure’s hypothesis. The Luwians that shifted to Hattian should not be confused with those who did not. This fact, unfortunately, makes the Hatto-Luwians rather elusive. Furthermore, while Hittite typological consistency has been thoroughly investigated, comparable research is not available for Luwian.

Returning to Hattian, the clay tablets we have are scribal copies. Although scribes preserved and handed down Hattian texts, it is unclear how well the Hattian language was known. Their handling of the language is far from perfect in the manuscripts that have been preserved; nonetheless it cannot be qualified as totally corrupted.⁴² But we should never forget that we see only the scribal witness and what survived, especially from Hattuša, might not have been the best editions.⁴³

Curiously, in KUB 28.80 (CTH 737, ‘regular’ festival of Nerik) the scribe states (iv 1′–11′) that the tablet, which is new, bears a text for a malteššar recitation that does not comply with the ancient one. In a seminal study by E. Laroche (1947a), this comment was taken to refer to errors in Hattian caused by imperfect knowledge of the language, but the matter is not so straightforward.

KUB 28.80 iv 1′–11′

1′–2′	“Tablet of the malteššar of the festival of Nerik, regular.
3′	Now (there) is a new tablet.
4′–6′	When, during the hostile years, the festival of Nerik started to be celebrated in Hakmiš,
6′–8′	the man of the Storm god and the GUDU₁₂-priest came (relocated) from Nerik,
8′–9′	and (thus) this malteššar has been established/taken for/from them
10′–11′	It does not conform to the ancient malteššar”

There is no agreement among scholars about the correct interpretation of a-pí-e-⸢da⸣-aš da-a-e-ir at line 9′. Some interpret it to mean that the recitation was ‘placed’ (prepared or undertaken) for them; others that it was ‘taken’ (recorded) from them.⁴⁴ We will further analyze this question elsewhere. Here we concentrate on a couple of considerations that can and must be drawn.

1. There is no explicit reference to linguistic competence.

2. The focus is on conformity to and compliance with some ancient tradition and/or source.

This text can be better interpreted in the light of the problem of conformity or, as Schwemer (2016) puts it, in terms of ‘quality assurance.’ Schwemer identifies three main factors for text production and conservation:

1. “the practice of centrally administering and controlling the regular performances, also outside the capital (especially, but not exclusively, cultic events that involved members of the royal family);”

2. “the ideal of preserving and restoring the correct, original tradition;”

3. “the necessity of regular, often annual, adaptation and change” (Schwemer 2016:23).

The problem that our scribe faced was more probably conformity, due to the changes that interested cultic performances. Our main question about this text should no longer be the scribe’s competence in Hattian but rather how a scribe could determine whether a source was compliant. In this specific case, we have two possible answers: either the scribe could not find a written source with which to compare the version in question or the scribe checked the present version against an ‘official’ (authorized) source, which was not explicitly mentioned. We would not take this text as evidence of poor knowledge of Hattian in either scenario.⁴⁵ Although a possible cause of the non-conformity could have been poor knowledge of Hattian, this is only one possibility among others. The culturally internal Hittite focus was on compliant performances rather than linguistic (i.e., an abstract lexical and grammatical) competence.⁴⁶

Our view is conditioned by the nature and history of the surviving documentation, which was largely the work of scribes and cult personnel who preserved texts that were selected based on state traditions. We must pay due attention to the fact that we judge specific texts—those restricted to the religious and ideological sphere—above all from their conditions of transmission. We do not have enough historical sources to judge the status of the language outside the state structure and personnel active in the territory. While we can marginalize Steiner’s opinion, we should accept that any consideration about the knowledge of the language is dependent on the more or less narrow view that our sources provide. We must also consider the state of knowledge of the texts. KUB 28.80, as previously mentioned, testifies that scribes were aware of the emergence of variant versions of models considered original or official.

Perhaps the most balanced conclusion that we can provide at the moment is that Hattian, among the circle of scribes in Hattuša but perhaps more generally among the cult personnel, became in the course of the history of the Hittite kingdom a residual sacred language that was not acquired natively but rather by explicit instruction. However, we must also accept that what was retained was less knowledge of the language than knowledge of a selection of texts. They were pieces to be performed from memory or read aloud during cult activities, but their contents were no longer easily accessible without the support of the Hittite translations that were handed down in parallel.