*=h2o
‘and’
- Language
- Proto-Anatolian
- Grammar
- part
- ID
- 2881
1. Individual Anatolian Languages
1.1 Hieroglyphic Luwian
/=ha/, -ha
‘and, also’1.1.1Transmission
This lexeme is attested 385 times across all regions from the 13th to the early 7th century BCE.
[A.P.]
1.1.2Forms
The above table offers a representative selection, omitting broken attestations.
[A.P.] [AH.B.]
1.1.3Graphic Features
1.1.3.1Phonological Interpretation
The phonological interpretation of this word is secure. Besides the consistent spelling found in HLuw., the CLuw. evidence supports the interpretation with singleton /h/ (see below). In HLuw., the particle regularly causes a preceding /n/ to be omitted at least in writing, cf. e.g. Melchert 2003b:182.
The particle is mostly spelt with the sign 〈ha〉 (sign *215, see Laroche 1960c:118f. for references); occurrences with 〈há〉 (*196, its syllabic value deriving from the geographical name Ḫatti) are much rarer.
[AH.B.]
1.1.4Meaning
First identified by Forrer 1932a:28-31 as /=ha/ with a general meaning ‘and’, the particle is used in different ways (see e.g. Yakubovich 2015b:6.6), as is its cognate CLuw. =ḫa (see under lemma). Always attached to a preceding orthotonic word, it can occur once, serving as a linker that conjoins words, mostly nouns, or clauses. Alternatively, it can occur doubled to /…=ha …=ha/ at phrase level.In contrast to CLuw., the HLuw. occurrences of /…=ha …=ha/ do not join only two elements in a meaning of ‘both … and …’. Instead, they appear in longer concatenations of the format /X Y=ha Z=ha/, i.e. ‘X and Y as well as Z’, and there may be further elements preceding X that belong to the same string albeit without overt linking, see e.g. KARKAMIŠ A6, §20. It should be noted that its usage as a linker is by no means mandatory, since ordinary clause linking is achieved with /a=/, a weak linker usually left untranslated, and phrase linking is mostly asyndetic. When /=ha/ is used, it seems to raise the strength of linking expressed, expanding the focus to additional elements. This can be observed e.g. in AKSARAY, §5: wa/i-ta á-mi-ia-ti (IUSTITIA)tara/i-wa/i-na-ti (DEUS)TONITRUS-hu-ti-i tara/i-ma-za-ha DEUS-ni-za za-ti BONUS-si-ia-za-ha, approx. /wa=tta ammiyadi tarrawannadi Tarhunti tanimanz=ha massaninz zadi wassiyazzahha/, meaning ‘By my loyalty, I was dear to (the storm-god) Tarhunt as well as all the gods here’ (cf. Hawkins 2000a:476). Between clauses, the same emphasis becomes evident e.g. in ALEPPO 6, §10: a-wa/i *a-pa-sa-ha OVIS(ANIMA) ka+ra/i-tu, approx. /a=wa abas=ha hawin karidu/, meaning ‘(Whether he is a country lord or a river-land lord), he, too, shall sacrifice a sheep’ (Hawkins 2011a:44f.).
The strong intra-clausal linking effect comes to the fore most noticeably in the apodoses of curse formulae such as KARKAMIŠ A11b+c, §25, calling upon the combined wrath of a number of deities to punish a perpetrator: *a-pa-ti-pa-wa/i-tá CAELUM (DEUS)TONITRUS-sa (DEUS)kar-hu-ha-sá (DEUS)ku+AVIS-pa-pa-sá-ha (MONS)a+ra/i-pu-tá-wa/i-ni-sá-ha (DEUS)TONITRUS-sa (“FLUMEN+MINUS”)sà-ku+ra/i-wa/i-ni-i-zi-ha (FLUMEN.REGIO)ha-pa-tà-si DEUS-ní-zi LIS-la/i/u-sa-tú, approx. /abatti=ppa=wa=tta tiPasassis Tarhunzas Karhuhas Kubabas=ha Arputtawannis=ha Tarhunzas Sakurawanninzi=ha habadasi massaninzi salissantu/, translating as ‘(… or who erases my name from these orthostats), him, then, (the storm-god) Tarhunt of Heaven, (the god) Karhuha, and (the goddess) Kubaba shall oppose, and also (the storm-god) Tarhunt of Mount Arputta as well as the gods of the river-land of Sakura’ (cf. Hawkins 2000a:104).
[AH.B.]
1.1.5Stem
Monomorphematic and indeclinable, this particle is enclitic in both varieties of Luwian and therefore always found attached to an orthotone lexeme. When linking clauses, it is included in the clitic chain in Wackernagel position, sharing a slot with the other connector /=ba/ ‘but’And, if extant, the connector /=ppa/ ‘(strong sequencer?), then(?)’. and thus in complementary distribution with it, provided the slot is used at all. Linking phrases, it attaches to the first word of the second (or last) phrase, which may be a noun or adjective, cf. the examples cited above.
The particle is used with relative pronouns in order to form indefinite pronouns, attaching to the inflected form, viz. /kwi-=ha/ ‘someone, anyone’, a feature commonly found in the Anatolian languages, cf. e.g. Melchert 2003b:191.
[AH.B.]
1.2 Cuneiform Luwian
=ḫa
‘and, also’1.2.1Transmission
The particle =ḫa is well attested throughout the corpus.
[D.S.]
1.2.2Forms
| indecl. | ḫa | KBo 8.130 iii 5´ | (CTH 764, MS) |
| KUB 32.13 i 9 | (CTH 771, MS) | ||
| KUB 35.101 obv.? 7’ | (CTH 765, MS) | ||
| KUB 35.101 obv.? 9’ | (CTH 765, MS) | ||
| KUB 35.107+ ii 12’ | (CTH 764, MS) | ||
| KUB 35.54 ii 44’ | (CTH 758, MS) | ||
| KUB 35.108, 24’ | (CTH 765, MS) | ||
| KUB 35.87, 6’, 7’ | (CTH 765, MS) | ||
| KUB 35.103 iii 7 | (CTH 766, MS) | ||
| KBo 7.67, 2’, 6’ | (CTH 773, MS) | ||
| KUB 35.9 ii 4’ | (CTH 757, MS?) | ||
| KUB 9.31 ii 31 | (CTH 757, NS) | ||
| KUB 35.137 obv.? 4’ | (CTH 773, NS) | ||
| KBo 7.68 ii 4’ | (CTH 764, NS) | ||
| KUB 35.39+ iii 9’ | (CTH 759, LNS) | ||
| KUB 35.133 ii 8’, 9’ | (CTH 665, LNS) | ||
| KUB 35.97, 8’ | (CTH 764, LNS) | ||
| KBo 61.237, 5’ | (CTH 770, LNS) | ||
| *ḫa* | KUB 35.133 ii 30’ | (CTH 665, LNS) | |
| ˹ḫa˺ | KUB 35.104+, 8’ | (CTH 758, MS) | |
| KUB 35.108, 25’ | (CTH 765, MS) | ||
| ḫ[a-a˚] | KUB 35.39+ iii 8’ | (CTH 759, LNS) |
All occurrences are listed in the table, which shows that the coordinating particle is less frequently attested in Cuneiform Luwian than would be expected for such a word.
Kloekhorst 2014b:594 sees a geminated variant of =ḫa, i.e. =ḫḫa, which would occur after accented short vowels. However, the examples ti-iḫ-ḫa-ti[(-) and ti-i-iḫ-ḫa-ku-w[a, which he adduces, are not probative; see the same criticism under CLuw. tī, =du.
[D.S.]
1.2.3Graphic Features
Checked against the photographs in the Konkordanz (Mainzer Fotoarchiv).
[D.S.]
1.2.4Meaning
The coordinating function of the particle =ḫa was identified early enough together with its HLuw. correspondence and through its functional analysis (see Rosenkranz 1952a:91, Otten 1953a:53, Laroche 1958b:172, Laroche 1959a:37, CLL:45): In coordinated strings, it can be added to the third element of a series, e.g. u-ra-az-<za-aš> DUTU-az ta-ti-in-zi DINGIRMEŠ-in-zi DÉ.A-aš=ḫa ‘The chief Sun-god, the paternal gods and Ea’ (KUB 9.31 ii 30-31). It is also used emphatically, then often meaning “also” or “likewise”, e.g. pa=aš pu-ú-wa [ku-w]a-ti a-aš-ta na-nu-un=ḫa=aš a-pa-ti a-aš-du ‘How he was before, also now he must be this way’ (KUB 35.103 iii 6-7). Still in its emphatic function, there are some examples where =ḫa is enclitic to a verb in the middle of a sentence, e.g. a=ti az-za-aš-da=ḫa wa-a-šu ‘He also ate well.’ (KUB 35.133 ii 8’). Although the fragmentary contexts often hinder an in-depth syntactic analysis, Giusfredi 2020b:149 considers Luw. /=ha/ to be a marker of additivity. For more information on its syntactic properties, see under the HLuw. lemma.
[D.S.]
1.3 Palaic
=(y)a
‘and, also’1.3.1Transmission
This particle is well attested across the Palaic corpus.
[D.S.]
1.3.2Forms
| indecl. | ia | KUB 35.165 rev. 22’ | (CTH 751, OS or MS?) |
| IBoT 2.35 obv. 8 | (CTH 751, MS) | ||
| KBo 19.152 iv! 17’ | (CTH 751, MS) | ||
| IBoT 2.36, 4 | (CTH 751, NS) | ||
| KBo 64.179 r.col. 7’ | (CTH 751, NS) | ||
| [i]a | KUB 35.165 obv. 3 | (CTH 751, OS or MS) | |
| i-ia | KUB 32.16, 4’ | (CTH 752, OS or MS) | |
| KUB 32.18 iv 10’ | (CTH 752, OS or MS) | ||
| ˚a-a˚ | KUB 35.165 rev. 6’ | (CTH 751, OS or MS?) | |
| KBo 53.301, 5’ | (CTH 751, NS) | ||
| ˚a | KUB 35.165 rev. 23’ | (CTH 751, OS or MS?) | |
| KUB 35.168, 5’ | (CTH 752, OS) | ||
| KUB 32.17 + KUB 35.167, 21’ | (CTH 751, MS?) | ||
| KBo 37.150 r.col. 2’ | (CTH 753, NS) | ||
| KUB 48.69, 5’ | (CTH 754, NS) | ||
| KUB 48.69, 7’ | (CTH 754, NS) | ||
| a | KUB 35.18, 12’ | (CTH 752, OS or MS) |
The clitic =ya (or =iya) occurs after a vowel, while the variant =a occurs after a consonant, though without the geminating effect; see under Reconstruction.
[D.S.]
1.3.3Graphic Features
Checked against the photographs in the Konkordanz (Mainzer Fotoarchiv).
[D.S.]
1.3.4Meaning
The enclitic particle =(y)a ‘and’ can coordinate two elements within a sentence, e.g. tabarnai MUNUStawanannai=(y)a ‘for the Tabarna and the Tawananna’. It can also be used after a verb at the end of a clause, in order to mark some kind of emphasis on the verb, in which case it is translatable as ‘also’ (Sasseville 2021a:536 with n. 8). Moreover, it can occur at the beginning of a sentence within the particle chain. In this case, its exact function has so far remained unknown; cf. Carruba 1970a:49 with further literature.
[D.S.]
1.4 Hittite
=(y)a
‘and, also’1.4.1Literature / Comments
HEG I-K:336-338, Hoffner & Melchert 2008a:399-401
2. Reconstruction
2.1Reconstruction ‘and’
The Hittite particle =ya ‘and’, which is spelled =a after consonants with a geminating effect on the consonant, has been equated with CLuw. =ḫa ‘and’ by Watkins 1975a:375f. (and Melchert 1988b:224 n. 17, EDHIL:378f.). An important piece of evidence is the comparison of the Hittite universal quantifier kuišša ‘each’ with Luw. kuišḫa ‘each’ (followed by Melchert 1984a:164f.). The Palaic evidence has always been controversial because of the apparent presence of both the particles =ya and =ḫa. However, the existence of the latter in Palaic, which was based on a single attestation, is considered here to be no longer valid; see wašūḫ-. In contrast to that, the Palaic particle =(y)a, spelled =a after consonants, is attested many times and it is noteworthy that it has no geminating effect on the preceding consonants. Further support for the etymological equation can be adduced by the free-choice relative kuiša ‘whichever’, where the geminating effect is likewise absent.
Considering the evidence shown above, it is very difficult to keep =ya/=a and =ḫa separate (pace HED A:8-9). Melchert 1994a:130, 167 argues that the allomorph =ya cannot be original, since the glide would have been lost intervocalically. Indeed, since the intervocalic loss of -y- is a regular phonological development in Hittite, it would be even more improbable for it to have remained intact in a clitic word where the phonology is usually more vulnerable to the effects of erosion. Therefore, the pre-stage of Hittite, Palaic and Luwian should have had the particle *=(ḫ)a.
Regarding the genesis of the glide -y-, Melchert 1984a:164f. suggests that it occurred whenever the clitic was placed after a word ending with -i or -e, where the glide sound was natural. At some point, speakers would have reanalyzed the suffix -a as -ya and spread out its distribution to the missing environment, i.e. after the vowel -a. The starting point for the reanalysis can be exemplified with the following Palaic dative construction; e.g. ta-ba-ar-na-i MUNUSta-wa-na-an-na-ia (tabarnai tawanannai=a → tabarnai tawanannai=ya). Therefore, the reanalysis of the clitic can be regarded as an innovation shared by Hittite and Palaic.
Regarding the loss vs. retention of the laryngeal in the particle *=(ḫ)a within the Proto-Anatolian period, Melchert 1984a:164f. explains that the enclisis weakened the phonetic value of the laryngeal, which led to its loss after -s- creating an assimilating effect as in kuišša < *kuišḫa. On the other hand, he points out that the laryngeal can be reconstructed as *h2 or as *h3, which is a further problem (cf. below). Concerning the loss of the laryngeal in the coordinating use of *=(ḫ)a, Melchert 1984a assumes that it was lost intervocalically due to its weaker status (differently EDHIL:378f. “*=h3e with loss through *h3 in hiatus”), although the problem remains as to why exactly it was retained in Luwian, lost by assimilation in Hittite, and lost without a trace in Palaic.
Therefore, a new phonological explanation must be sought: first, the loss of the laryngeal in the enclitic could be due to erosion, but it is difficult to motivate why it would have occurred only in Hittite and Palaic but not in Luwian. A different approach would be to work with the position of the accent, considering that the Anatolian laryngeals “disappear” in post-tonic consonant clusters e.g. of the type VRHV as in Hitt. tarra- (MP) ‘to be strong’ < (PIE *térh2-o-to; cf. Ved. tárate) vs. Hitt. tarḫ-u-(mi) ‘to conquer’ and Hitt. marra- (MP) ‘to be stewed, cook until tender’ (< *mérh2-o-to), vs. Hitt. marḫā- ‘(a kind of stew)’ (< *mErh2-ó- or -éh2-).For a recent discussion of this phonological phenomenon, see Sasseville forthcoming with further literature; paper presented at the IG conference in Ljubljana. In other words, whenever the enclitic *=ḫa was used after a word ending not in a vowel but in a consonant (also after stops, either voiced or voiceless), the laryngeal would disappear. At some point, Hittite and Palaic generalized the variant without initial *ḫ-, whereas Luwian went the opposite direction.
Then, Hittite and Palaic went on to reanalyze the particle *=a as *=ya, as described by Melchert 1984a (see above). Afterwards, the glide was lost after -s in Hittite and Palaic independently. In case of the former, the sequence *-si̯- regularly leads to -šš-, whereas in the latter language the glide is lost without a trace, i.e. *-si̯- > -š-. Evidence for this Palaic sound change can be demonstrated by the fate of the genitival adjective *-o-si̯o- > -aša- (cf. Hitt. -šša- and Luw. -ašša/i-). Afterwards, this pattern is generalized to all instances, i.e. Hitt. =(y)a with gemination after a consonant vs. Pal. =(y)a without gemination after a consonant. This scenario provides a much better account of the phonological development involved for the fate of the Anatolian coordinating particle.
From an Indo-European perspective, Dunkel (LIPP II:334f.) reconstructs the particle as *=h2o, which is phonologically fine for the pre-form of Hittite, Palaic and Luwian, although an e-grade, i.e. *h2e, would also work from a purely Anatolian perspective.Contra Dunkel ibid., Lyc. B =ke cannot phonologically go back to the same particle, see PAnat. *ku̯e. Due to the historical account of the particle =(y)a defended above, the etymological connection with PIE *i̯o becomes void (pace LIPP II:384f.).
[D.S.] [O.B.]