Hans Bork
'To Have’ and ‘To Hold’ in Mycenaean
Hans Bork, UCLA
The dativus possessivus construction found in alphabetic Ancient Greek is absent in Mycenaean texts, which is surprising given the prominence of the structure in various Indo-European daughter languages, and the usual reconstruction of the dativus possessivus for PIE itself (Keydana, forthcoming). The dativus possessivus is not mentioned in important surveys of Mycenaean grammar (e.g., Hooker 1980, Bartoněk 2003), although various other possessive or relational expressions are discussed, including the adnominal genitive (PY Eb 416, te-o-jo do-e-ra), denominal adjectives (PY An 654 a-re-ku-tu-ru-wo e-te-wo-ke-re-we-i-jo), transitive ἔχω (PY Ep 705 do-er-a e-ke), and, interestingly, the predicate genitive (KN Ai 63, pe-se-ro-jo e-e-si). Kulneff-Eriksson 1999 does note the curious absence of the predicate dative in Mycenaean, but only passingly explains this as a consequence of the Mycenaean dialect’s “exceptional position.” Nevertheless, the possessive constructions that do occur in Mycenaean conform to observed cross-linguistic typological patterns of attributive (NP-bound) and predicative (VP-bound) possession-marking, and thus there is no obvious systemic motivation for the lack of the dativus possessivus. As such, its absence deserves to be explained.
The most common predicate possessive structures in Mycenaean are transitive clauses with ἔχω, and the same construction also occurs in Homeric Greek (e.g., Il. 1.225, κυνὸς ὄμματ᾽ ἔχων, “having the eyes of a dog”). Unlike the dativus possessivus, however, transitive ἔχω meaning “have” was not inherited from Indo-European, since PIE famously has no reconstructible word for “have”. Rather, the daughter languages developed “have” possessives from a handful of roots that originally denoted grasping, holding, or physical control: e.g., Gk. ἔχω < *seǵh-, ‘to control’; Lat. habeō < *gheHb-, ‘to grab’; Gm. haben < *keh2p-, ‘to hold’. Verbs formed to these roots came to indicate possession through parallel developments of the metaphor physical control equals possession, but PIE itself was typologically a “be-possession” language, wherein permanent possession was encoded by the formula Nom.NP + copula + Dat.NP. When it branched off from PIE, Greek was also a “be-possession” language, and within it ἔχω constructions would have been semantically distinct from predicate datives. However, over time the constructions came to overlap—they occur equally often in Homer—and, indeed, ἔχω eventually fully replaced the predicate dative as Greek’s standard marker of permanent possession. In contrast, the predicate genitive was not semantically competitive with the predicate dative, and therefore did not undergo a similar decline in usage. All the linguistic data suggest that the predicate dative should have been productive in Mycenaean, and indeed, the Linear B texts show that the necessary “ingredients” for the construction—i.e., dative NPs, copula verbs, and “possessible” NP entities—were individually licensed and reasonably common.
The construction’s absence can be accounted for by reconsidering the meaning of ἔχω in Mycenaean, and to assign it a meaning closer to its radical sense, namely “control” or “govern,” rather than “have.” In the Linear B texts, ἔχω would thus denote temporary possession, a semantic schema that is typologically regular for “be-possession” languages in which secondary “have-possession” predicates also occur. Evidence of this radical meaning is seen in the fact that ἔχω lexemes are not distributed evenly in the Linear B documents, but rather the verb tends to cluster in the E-series “landholding” and Jn-series “bronze” texts, both of which concern the management of state-owned property; the use of ἔχω in these documents thereby marks temporary “governance” of the named good. The dativus possessivus was never used in any Linear B documents due to the predicate frame of the texts themselves, which were produced by state-employed scribes as catalogs of state-owned property distributed in a “command economy”. Marking of permanent possession via the dominant dativus possessivus construction would have been otiose, since Mycenaean scribes necessarily assumed that all named goods were state-owned, and thus inherently “permanently possessed.” This explanation fits well with the cultural data, and squares Mycenaean possessive syntax with the diachronic trends observed in Alphabetic Greek and reconstructed PIE.