Syntax of the Hittite “Supine” Construction

Harry A. Hoffner Jr. † et H. Craig Melchert

Both authors were most grateful to Alice Mouton and other organizers of the IFEA conference for allowing us to participate in this volume honoring the centennial of Emmanuel Laroche, although we were unable to attend the conference itself. Professor Laroche was not only one of the giants of Hittitology, but also a pioneer in the study of other languages of ancient Anatolia, both Indo-European (Luvian and Lycian) and non-Indo-European (Hattic and Hurrian), and he generously assisted both of us early in our careers, as he did so many others. Professor Hoffner’s sudden and unexpected death on March 10, 2015, prevented him from participating in the final preparation of this joint paper. He had furnished many of the crucial examples and had seen and approved with changes an initial draft, but responsibility for the version published here necessarily rests with me—HCM.

  • 1 The modern terminology follows Kammenhuber 1955: 31-57. For older literature see HE: 142-143 and fo (...)

1As is well-known, the Hittite verbal form ending in -(u)wan, labeled now for more than half a century the “supine”, occurs only in a construction with either dai- ‘to put’ or tiya- ‘to step’ that expresses the notion ‘begin/undertake to do X’:1

(1) KBo 5.8 iii 3-5 (CTH 61, Annals of Muršili; NH)
namma=šmaš=kan ÉRIN.MEŠ išḫiḫḫun nu=mu ÉRIN.MEŠ piškewan dāir n=at=mu laḫḫi kattan paišgauwan tiyēr

“Then I imposed (a commitment for) troops on them, and they began to give me troops and began to go on campaign with me.”

2It is clear that in older Hittite the auxiliary was dai- ‘to put’ (note īššuwan daišten and piyanniwan daišten at KBo 8.42 Vo 2-3; OH/OS). However, likely due to the ambiguity of plural forms such as Pres3Pl ti(y)anzi, we find also in later Hittite use of tiya- ‘to step’ as the auxiliary. In the example cited as (1), both auxiliaries occur side by side.

  • 2 We are indebted to Hans Bork of UCLA for calling our attention to this issue by asking during class (...)
  • 3 Garrett 1990, following Watkins.
  • 4 See in extenso Garrett 1996 and for a summary GrHL: 280-281. The dividing line between the two clas (...)

3Our focus in the present discussion will be on the use or non-use of third-person enclitic subject pronouns in the supine construction.2 As first observed by Calvert Watkins and confirmed by Andrew Garrett, such subject enclitic pronouns never occur with transitive verbs in Hittite.3 In the case of intransitive verbs, their use is governed by the lexical semantics of the verb: so-called “unaccusative” verbs regularly take third-person enclitic subject pronouns, while “unergative” verbs do not.4 Since dai- is a transitive verb, it does not in its usual use as ‘to put, place’ ever take an enclitic subject pronoun. As an “unaccusative” intransitive verb, tiya- ‘to step, take up a position’ regularly does so. Readers will notice that in example (1) piškewan dāir is not accompanied by a subject pronoun, while paišgauwan tiyēr is. The question then becomes: is the use or non-use of the enclitic subject pronoun governed by the lexical semantics of the auxiliary verbs or of the verb in the supine?

4A survey of all examples known to us clearly shows that it is the latter. First of all, if the verb in the supine is transitive, there is no enclitic subject pronoun, whether the auxiliary is dai- or tiya-:

(2) VS 28.111 rev. 5-6 (CTH 530, Cult Inventory; NH)
nu=šši EZEN4 DUMU.SAL.hI.A [e]ššūwan tiēzzi

“And one undertakes to perform for him/her the festival of the daughters.”

(3) KBo 29.86 obv. 10-11 with dupl. KUB 20.16 i 10-11 (CTH 694, Festival fragment; NS) [(nu GIŠarg)]ami galgaltūri GIŠ.dINANNA.HI.A [(hazz)]iyēškewan tianzi

“They begin to play the argami, galgalturi, and Ishtar-instruments.”

  • 5 Further transitive examples occur in KUB 14.16 ii 22, KBo 4.4 iv 35 and 47, and Bo 86/299 ii 27 (‘g (...)

5Example (2) with no subject enclitic pronoun even with an unambiguous form of tiya- shows that likewise its absence with a transitive verb in the supine plus dai- (as in the first clause of example (1) above) and in the ambiguous example (3) is due to the supine, not the auxiliary. See further the second clause of example (9) below.5

6Naturally, the supine of a transitive verb can take an object enclitic pronoun, but this has nothing to do with the choice of auxiliary:

(4) KUB 32.133 i 4-7 (CTH 482, Cult Reform of Muršili; NH)
nu=za azziwita išiuliHI.A=ya kue INA É.DINGIR-LUM kattan amankatta wēr=ma=at=kan LÚ.MEŠ DUB.SA˘ R.GIŠ LÚ.M˘ EŠ É.DINGIR-LIM=ya wanuškewan dāir˘n=at mMurši-DINGIR-LIM-LUGAL.GAL tuppiyaz EGIR-pa aniyanun

“The scribes on wood and the men of the temple proceeded to begin to alter the ritual prescriptions and regulations that he (Tuthaliya) had mandated in the temple. I, Muršili, the Great King, restored them by means of a clay tablet.”

7The same remark applies to the examples with transitive ē[šš]ūwan tiyanzi in KUB 5.6 + KUB 18.54 i 23 and KUB 56.19 i 39.

8As predicted, when the supine is an unaccusative intransitive verb, an enclitic subject pronoun is required:

(5) KBo 3.67 i 8-9 with dupl. KUB 11.5 obv. 4 (CTH 19, Edict of Telipinu; OH/NS)
mān mḪantīliš«š»=a ŠU.GI [kiša(t n=aš DINGIR-L)IM-iš] kikkiššūwan dāiš

“But when Hantili became an old man, and he began to become a god.” (i.e., to die)

(6) KUB 14.8 obv. 26-28 (CTH 378.2.A, Plague Prayer of Muršili; NH)
nu LÚ.[(MEŠappa)ntan] kuin ēpper n=an maḫḫan INA KUR URUḪat[(ti)] EGIR-pa uwate[(r)] nu=kan INA ŠÀ-BI LÚ.MEŠŠU.DIB.BI.HI.A inkan kišat n=aš akk˘iš˘kewan da[]˘

“A plague broke out among the prisoners whom they took, when they brought them back to Hatti, and they began to die.”

  • 6 See Neu 1968: 181. The fact is acknowledged in Garrett 1996: 91, but with an incorrect meaning ‘to (...)

9Both kiš- ‘to become, happen’ and akk- ‘to die’, as change-of-state verbs, are well established to be unaccusative in Hittite and thus require an enclitic subject pronoun, despite the fact that the auxiliary here is dai- ‘to put’. Likewise, as a verb expressing emotion, duške- ‘to rejoice’ is unaccusative and requires an enclitic subject pronoun also in the supine:6

(7) KBo 26.65 iv 15-16 (CTH 345.3A, Song of Ullikummi; pre-NH/NS)
dTašmišuš []tamašta n=aš=za duškiškewan daiš

“Tašmišu heard, and he began to rejoice.”

10Likewise in KUB 33.112 + KUB 33.114 + KUB 36.2 ii 3. So also it is the verb of directed motion pai- ‘go’ that demands an enclitic subject pronoun in the second clause of example (1), not the auxiliary tiya-.

11In a few cases where we have no finite examples of intransitive verbs with pronominal subjects the supine construction is diagnostic:

(8) KUB 12.44 ii 27-28 (CTH 392, Ritual of Anna; MH/NS)
mān SAR.GEŠTIN kuiš UL miyēškezzi [… k]iššan aniyami n=aš miškewan dāi

“If some vineyard is not growing, I treat [it] as follows, and it will begin to grow.”

(9) KBo 3.1 i 21-22 (CTH 19, Edict of Telipinu; OH/NS)
a[ašš]=a=šmaš=šan [(t)]aštašeškeuwan dāir (22) nu ēšar=š«um»mit ēššuwan tiyēr

“They also began to conspire (whisper) against their lords and began to shed their blood.”

  • 7 Garrett 1996: 94. This example is also cited in GrHL: 281, note 16, as contrasting with the absence (...)

12The evidence of example (8) for mai- ‘to grow’ as an unaccusative verb has been recognized,7 but example (9) is thus far unique in showing by its absence of a subject pronoun (-e in Old Hittite or -at in a New Script copy) that taštašiya- ‘to whisper, conspire’ is unergative.

  • 8 For the concept and the label see Garrett 1996: 98-100.

13As expected, “detransitives”, that is, transitive verbs that are used intransitively without a direct object,8 also behave as unergative in the supine construction:

(10) KUB 9.4+ ii 22-23 (CTH 760.I.2, MUNUSŠU.GI Ritual; NS)
nu=za namma kī ukmai ēpzi nu ukkiskewan dāi

“He again takes up this incantation and begins to recite the incantation.”

14Likewise KUB 9.34+ iii 7 and KUB 53.4 rev. 9 without enclitic subject.

15We have found only one exception to this very regular distribution, and it is surely specially conditioned. In the only occurrence in a native Hittite composition, the intransitive verb w(iy)ai- ‘to wail, weep’ behaves as an unergative, as is to be expected for a verbum dicendi (compare taštašiya- ‘to whisper’ above):

(11) KUB 30.15 + KUB 39.19 obv. 34-36 (CTH 450, Royal Funeral Rite; OH/NS)
[nu G]RÍN ZI.B[A.NA ar]a duwarniyaizzi n=at=kan dUTU-i menaḫḫ[anda …] [nu kalga]linaizzi nu weškewan [dai]

  • 9 For the tentative interpretation of kalkalinai- as ‘clang’, referring to the lugubrious sound made (...)

“One breaks the scales and […] it facing the Sun-god. It clangs(?), and one begins to wail.”9

  • 1

16Predictably, it also takes no enclitic subject pronoun in the secondary transitive use ‘to bewail’:10

(12) KUB 19.4 + KBo 19.45 i 7-8 (CTH 40, Deeds of Šuppiluliuma; NH)
nu maḫḫan ABU=YA ŠA mZannanza kunātar išt[amašta nu] mZannanzan wēškewan daiš

“When my father heard about the killing of Zannanza, he began to bewail Zannanza.”

17However, in the Hurro-Hittite translation literature the same verb behaves as an unaccusative and consistently takes an enclitic subject pronoun. One example will suffice for illustration:

(13) KUB 33.120 ii 53-54 (CTH 344, Theogony; MH/NS)
[m]ān=ši=kan ZU9.ḪI.A-uš anda iškalliyanta [dKumarbiya n]=aš weiškeuwan [dāi]š

“When his, Kumarbi’s, teeth began to be torn inside, he began to wail.”

18Likewise showing an enclitic subject pronoun are KUB 33.106 iii 4-6 (CTH 345.3A, Song of Ullikummi) and KUB 17.4:7 (CTH 364.3A, Song of Silver). While the motivation for use of the enclitic subject pronoun here is unclear, we must in view of other non-native usages in the translation literature attribute it to “translationese” and follow the evidence of the native example (11) for this verb being unergative in Hittite.

  • 11 The available space at the beginning of line 8 permits only restoration of [nu], not [na-aš]. 11 Se (...)

19Our finding that in the supine construction it is the lexical verb that determines the use or non-use of enclitic subject pronouns with intransitive verbs is of interest in confirming that dai- and tiya- have been fully reduced to the status of auxiliaries. Their behavior is thus entirely parallel to that of pai- ‘to go’ and uwa- ‘to come’ in the “serial” construction, where it has long been clear that it is the lexical verb that determines the behavior of subject enclitic pronouns as well as local particles.11 We may therefore henceforth confidently use any new examples of the supine construction in non-translation literature with intransitive verbs and expressed or unexpressed pronominal subject as diagnostic for unaccusativity in Hittite.11

Bibliographie

Garrett, A., “Hittite Enclitic Subjects and Transitive Verbs”, JCS 42, 1990, 227-242.

Garrett, A., “Wackernagel’s Law and Unaccusativity in Hittite”, in: Approaching Second: Second Position Clitics and Related Phenomena, Halpern, A. / Zwicky, A. (éds.). CSLI, Stanford, 1996, 85-133.

Kammenhuber, A., “Studien zum hethitischen Infinitivsystem IV”, MIO 3, 1955, 31-57.

Neu, E., Interpretation der hethitischen mediopassiven Verbalformen. StBoT 5, Harrassowitz, Wiesbaden, 1968.

van den Hout, T. P. J., “Studies in the Hittite Phraseological Construction. I. Its Syntactic and Semantic Properties”, in: Hittite Studies in Honor of Harry A. Hoffner Jr. on the Occasion of His 65th Birthday, Beckman, G. / Beal, R. / McMahon, G. (éds.). Eisenbrauns, Winona Lake, 2003, 177-204.

Notes

1 The modern terminology follows Kammenhuber 1955: 31-57. For older literature see HE: 142-143 and for a recent summary GrHL: 338.

2 We are indebted to Hans Bork of UCLA for calling our attention to this issue by asking during class instruction about the pattern shown by example (1).

3 Garrett 1990, following Watkins.

4 See in extenso Garrett 1996 and for a summary GrHL: 280-281. The dividing line between the two classes of verbs is notoriously fluid, but by and large the patterns observed in Hittite match those found elsewhere.

5 Further transitive examples occur in KUB 14.16 ii 22, KBo 4.4 iv 35 and 47, and Bo 86/299 ii 27 (‘give’), KUB 1.1 ii 6 and 10 (‘strike’), KUB 1.1 ii 43 (‘attack’), KUB 1.1 iv 52-53 (‘send’), KBo 11.1 obv. 37, KUB 5.6 i 17 and 23 and KUB 16.77 ii 40 (‘perform’), and KUB 16.32 ii 7 (‘sacrifice’).

6 See Neu 1968: 181. The fact is acknowledged in Garrett 1996: 91, but with an incorrect meaning ‘to please’ for the verb.

7 Garrett 1996: 94. This example is also cited in GrHL: 281, note 16, as contrasting with the absence of an enclitic subject pronoun when there is no referential subject.

8 For the concept and the label see Garrett 1996: 98-100.

9 For the tentative interpretation of kalkalinai- as ‘clang’, referring to the lugubrious sound made by the breaking of the scales see HED K: 25 and the similar passage cited there.

10

11 The available space at the beginning of line 8 permits only restoration of [nu], not [na-aš]. 11 See GrHL: 325, following van den Hout 2003. The conclusions just presented for the supine construction with dai- and tiya- raise the question of what the facts are for the competing construction with -za ēpp- plus the infinitive (see GrHL: 335 and 338 on its distribution). Based on all examples known to us (see the collection in HW² E: 64), the answer is that -za ēpp- plus the infinitive appears to be restricted to collocations with transitive verbs, a limitation not previously noted. The distinction of unaccusative versus unergative is thus irrelevant for this construction.

Auteurs

Harry A. Hoffner Jr. †

University of Chicago/

H. Craig Melchert

University of California, Los Angeles