*Hep-
‘to join, attach’
- Language
- Proto-Anatolian
- Grammar
- verb
- ID
- 1580
1. Individual Anatolian Languages
1.1 Hieroglyphic Luwian
*/hab-/
‘to join, attach’1.1.1Derivatives
1.2 Cuneiform Luwian
*ḫap(p)-
‘to join, attach’1.2.1Derivatives
1.3 Lycian A
*xb-
‘to join, attach’1.3.1Derivatives
1.4 Palaic
*ḫap(p)-
‘to join, attach’1.4.1Derivatives
2. Reconstruction
2.1Reconstruction ‘to join, attach’
All Anatolian cognates listed below allow for the reconstruction of a Proto-Anatolian Root *Hep-, in which the laryngeal must have colored the root vowel to *a or *o because of Hittite ḫap- (< *Hép-(/*Hp-´)); see Oettinger 1976c:124 on the reconstruction of a stative verb. Sasseville 2021a:220f. points out that there are two allomorphs of the root found in Anatolian, i.e. one with the lenition of the root-final labial as in CLuw. 
ḫapanzuwalatar/n- and Lyc. xba-(ti) and one without CLuw. ḫappiš- n. and Hitt. ḫapp-(mi). This implies the existence of a root verb of the ḫi-conjugation *Hób-ei / *Həpp-énti (perhaps attested in Hitt. 2sg. pres. ḫapti), which follows the pattern of the Hittite type -āC1i / -aC1C1anzi (cf. Melchert 2012c and Kloekhorst 2014c). To take the strong instead of the weak stem of such an ablauting verb as basis for further derivatives is not typical, but not unparalleled either; cf. šāḫ-(ḫi) (/*šaḫḫ-) ‘to clog, stuff’ → šaḫessar/n- n. ‘fortification’. Thus, CLuw. ḫap(a)i-(di) would represent an iterative formation with the suffix *-éi̯e/o-, whose lenited /b/ would have been analogically taken from the root verb of the ḫi-conjugation, unless one simply derives it from the strong stem of *Hób-ei / *Həpp-énti → *Hob-éi̯e/o-.
This root verb of the ḫi-conjugation is also probably the base of the participial formation *ḫap-ant-, which is found in several Luwian derivatives. The assibilation found in *ḫapanzu- ‘loyal, obedient’ (→ HLuw. /habanzulawa(/i)-/, /habanzanuwa-(i)/ and CLuw. ḫapazuwalatar/n-) implies the presence of a glide *i̯ (thus *-nt-i̯o-; Melchert 2003d) or perhaps even *u̯ (thus *-nt-u̯o-). Accepting the palatalizing effect of *u̯ after -nt- would be more economical than assuming a chain *-nt-i̯o-u̯o-. Cf. the palatalizing effects of *u̯ in Palaic tāzzu- ‘to take (redupl.)’ in Sasseville 2019a:25 n. 13. The suffix *-u̯o- in its “social status” function (see Rieken & Sasseville 2014a) would be appropriate for the posited semantics of *ḫapanzu-. For the reconstruction of a Proto-Anatolian neuter stem in *-is- built on this root, see PAnat. *Hep-is-.
[D.S.]
3. Proto-Indo-European
3.1Semantic Reconstruction
Since PAnat. *Hóp‑/*Hp‑´ ‘to join, attach’ is a root verb, cf. the Root etymology for details on the root’s semantics.
3.1.1Root etymology
Besides the Anatolian material only the sparsely attested verb Lat. apere ‘to tie, fasten’ with its assumed verbal adjective Lat. aptus ‘fastened, fitted, suitable’ provides possible evidence for the root in question. The meaning is clear from the evidence: PIE *h2/3ep- ‘to join (tr.)’.Differently LIV²:269: “passen, sich fügen” (intr.). Regardless whether Gk. ἅπτω ‘to fasten, bind, join’, etc. belongs here (as per van Beek apud EDG:120), the reconstructed meaning can stay the same. Cf. Willi 2018a:68 and 160 for possible explanations of Gk. ἰάπτω ‘to shoot; to harm’ and Gk. ἀπαφίσκω ‘to cheat, beguile’, respectively.
3.2Morphological Reconstruction
PAnat. *Hóp‑/*Hp‑´ ‘to join, attach’ has no direct cognates in other IE languages, which is why no PIE formation is reconstructed.
3.2.1Root etymology
There is hardly any non-Anatolian evidence for the root underlying PAnat. *Hóp‑/*Hp‑´ ‘to join, attach’. Most likely, Lat. aptus ‘fastened, fitted, suitable’ belongs here and probably its allegedly underlying verbal formation Lat. apere ‘to tie, fasten’ (cf., e.g., HED M:114). Although Gk. ἅπτω ‘to fasten, bind, join’ would semantically fit the Anatolian and Latin formations, it is phonologically problematic to connect it.For comments on the morphology and phonology of possibly related Gk. ἰάπτω ‘to shoot; to harm’ (Od.+), Gk. ἀπαφίσκω ‘to cheat, beguile’ (Od.+), cf. Willi 2018a:68, 160. Furthermore, note the phonological similarity of Gk. ἅπτω ‘to fasten, bind’ < */haPie/o‑/ and CLuw. /hab(a)i‑/ (instead of expected */happ(a)i‑/, cf. above) ‘to bind, attach to’ as well as the correspondence between the s-stems Gk. ἅψος ‘joint, limb’, which is earliest attested as nom./acc. pl. n. /haps‑ea/ < */-esa/ ‘members, limbs’ (Od.4.794 = 18.189), and CLuw. nom./acc. sg. n. /happis(=sa)/ ‘members, limbs’. The verb ἅπτω and the form ἅψεα are – except for dat. pl. ἁψῖσι (or psilotic ἀψῖσι) ‘meshs’ – the only stems of the word family attested in Homer.Therefore, the reconstruction is based only on the Anatolian root *Hep- and Lat. apere, aptus. The root in question consists of an anlauting laryngeal whose exact nature is not finally determinable: Lat. aptus is inconclusive (< *H̥pto‑), the Anatolian data points to *h2- or *h3-, and only Lat. apere – if the root vowel goes back to a full grade and is not influenced by the verbal adjective (ape/o- ← *pe/o- < *H̥p-é/ó‑) – may show a reflex of *h2‑ (if < *h2ép‑e/o‑). In view of this uncertainty, it is best to reconstruct the root as PIE *h2/3ep- ‘to join’.LIV²:269 reconstructs *h2ep‑. Furthermore, it should be noted that the affiliation of the Latin material is not unambiguously certain, cf. the relevant etymological dictionaries of Latin and, e.g., Meiser 2003a:197f.
[A.O.]
4. Transmission
Gk. ἅπτω ‘to fasten, bind, join’ (Il.+), ἅψος n. ‘juncture, joint, limb’ (Od.+), etc.
GEW I:126f., DELG:99f., EDG:120f., García Trabazo 2011a:328, Willi 2018a:68, LSJ:231, 299
Lat. apere, apiō ‘to tie, fasten’ (Enn.Ann.499, Paul.Fest.18M, 22M), Lat. aptus ‘fastened, fitted, suitable’ (Enn.+)
WH I:57, EM:39, EDLIL:47, OLD²:162, 170f.The three main etymological dictionaries connect Lat. aptus and apere ultimately with the root PIE *h1ep‑ ‘to grab, seize’ (LIV²:237), cf. WH I:57, EM:39, EDLIL:47. LIV²:237 n. 3, on the other hand, connects only Lat. aptus with the Anatolian material presented here and disconnects it from Lat. apere.