‘To Have’ and ‘To Hold’ in Mycenaean Hans Bork SCS 6 · Greek and Latin Linguistics Panel hsbork@ucla.edu Premise: The dativus possessivus construction is common in the Indo-European daughter languages, and is generally reconstructed as the primary expression of predicative possession in PIE, which has no reconstructible verb “to have.” Ouriously, while Tomeric Sreek preserves examples of the dativus possessivus, Mycenaean Linear N texts do not, though both varieties use ἔχω in possessive (or possessive-like) constructions. I propose that the absence of the dativus possessivus in Mycenaean could be the result of the discourse-frame of the actual Linear N texts, which are largely catalogues of state-owned property. · Mycenaean Possessive Syntax . Oross-linguistically, possession is marked in a grammar by means of either verbal (predicative, VP- bound) or nominal (attributive, NP-bound) structures. Examples from Alphabetic Sreek: . . Attributive: adnominal genitive (NP + Sen.NP); adnominal dative/dativus sympatheticus (NP + Pat.NP); denominal relational adjectives (e.g., in -ιος); possessive pronominal adjectives (e.g., ἐμός). .. Predicative: predicate genitive (Nom.NP + εἶναι + Sen.NP); predicate dative (Nom.NP + εἶναι + Pat.NP); transitive “have” verb (Nom.NP + ἔχω + Acc.NP). .. Note that the range of relationships marked as “possessive” is exceptionally broad. Of. the semantic differences covered by the same possessive construction in English: John’s car vs. John’s parents. Such ambiguities are variously handled in different languages. • This account is consistent with the typological overviews found in, among others, Nauer 2000: 156–61, Pixon 2010: 262– 312, and Stassen 200E: 3–6E. Ror a discussion of the wide semantic field encompassed by possession in Sreek, see Schwyzer and Pebrunner 1E50: 11Cff. on the “Pertinentiv” genitive. Ror discussion of possessive semantics generally, see Teine 1EEC, an influential study of the metaphorical background of grammatically-marked possession. . These same structures and the distinction between attributive and predicative marking are broadly attested in the Mycenaean Sreek of the Linear N tablets: . . Adnominal genitive: PY Eb 416 — te-o-jo do-e-ra (thehoio doelā), “servant of the god.” . . Relational adjective: PY Er 312 — wa-na-ka-te-ro te-me-no (wanakteron temenos), “the wanax’s precinct.” . . Predicate genitive: KN Ai 63 — pe-se-ro-jo e-e-si (pselloio ehensi), “they are [the people] of Psellos...” . . ‘Have’ verb: PY Ep C05 — o-na-to e-ke (onāton hekhei), “she holds the lease...” • NN— unless otherwise noted, citations and translations of Mycenaean forms are per Documents2 , but updated to reflect current theory, as e.g. recommended by Puhoux in Oh. E of Puhoux and Pavies 200D. • The background of possessive adjectives in Mycenaean is somewhat problematic. The -tero- suffix as a possessive marker is unique to the adjective wanaktero-; denominal possessive adjectives are otherwise formed with a suffix -i-jo (-ios) though this also sometimes appears in place of the “Stoffadjektiv” -e-jo (-eios). Ror more on these, see Tajnal 1EE4 and Killen 1ED3. . The Linear N texts do not attest the dativus sympatheticus, possessive pronouns, or the dativus possessivus. The absence of these attributive structures can be explained as a consequence of the verbal frame, which is exclusively in the 3rd person. . . Only a few potential personal pronoun forms occur in Linear N, all 3rd person: -mi (=min, acc. sg.), PY Ep C04 and PY Na C0; pe-i (spheis, dat. pl.), PY Na E26 and PY Na 3E5. All other suspected/attested pronouns are from the demonstrative or relative systems. . . Evidence for the dativus sympatheticus is ambiguous; dative NPs are common, but are usually assumed to be governed by an unexpressed verb “to give” or “to assign [to].” The status of the construction is unclear, but its existence seems doubtful. • Ror discussion of Mycenaean pronouns, see Documents2 p.DC, Nartoněk 2003: 343–45, and Tooker 1ED0: 60. See Schwyzer and Pebrunner 1E50: 14C–D for the dativus sympatheticus and its distribution in Alphabetic Sreek. . The absence of the dativus possessivus is much more difficult to explain, since all of the “ingredients” for the construction are otherwise extant in Mycenaean: . . Pative NPs are common, and the dative encoded multiple semantic roles in Mycenaean. Of. PY On 40, a tablet of 14 lines, each of which begins with a place-name in the dative and also includes a dative personal name governed by pa-ro (paro, “at the hands of”). . . Examples of the verb “to be” are attested: e.g., e-e-si (ehensi, “they are...”), in KN Ai 63 and KN Sd 4422. Of. too the compound a-pe-e-si (apehensi, “they are absent”) in PY An 614. . . Oomplex syntax was clearly licensed in the pragmatic frame of the Linear N texts: cf. PY Ta C11, which includes a subordinate clause introduced by o-te (hote, “when”); also PY Eq 213, which includes a main verb wi-de (wide, “saw”), and two apparent circumstantial participles. . . Predicative possession in general seems licensed: cf. §1.2.3 and §1.2.4 above, and note that forms of e-ke (hekhei) are incredibly common. E.g., PY Ep C04, an D-line text with 6 instances of e-ke, one of e-ko-si (hekhonsi, “they hold”), and one infinitive form e-ke-e (hekhehen, “to hold”) in indirect statement. • Ror overviews of the dative in Mycenaean, see Nartoněk 2003: 443, Tooker 1ED0, Waanders 200D, and more broadly, Tajnal 1EE5, none of which mention the dativus possessivus. Examples are sourced from Documents2 and the Diccionario Micénico. · Possessive Typology and PIE . The category of predicative possession can be further subdivided according to the syntactic relation- ship established between a possessive predicate and its arguments: . . Be-possessive constructions: • Srammatically realized by Subj-NP + Intransitive-Existential Verb + Oblique-NP. • They encode the cognitive metaphor .. Y AL A A Y . Have-possessive constructions: • Srammatically realized by Subj-NP + Transitive Verb + Obj/Patient-NP. have and hold · • Tend to occur with verbs of “holding” or “grasping,” which encode the metaphor L . .. . Y AL The two systems are not incompatible, but in languages where they co-occur, they tend not to be used with the same domains. Oompare English Jim has a car vs. ? Jim has a foot: body parts are inalienable, and thus semantically clash with the temporary semantics of the “have” construction. See the references listed under §1.1.3 for discussion. Typologically, PIE was a “be”-possessive language, as there is no common root for the word “have” in the IE daughter languages: . . E.g., Latin habeō H *gh eHb- (or *ǵh eHb-) ‘to grasp’, also Umbrian habe; Sreek ἔχω H *seǵh - ‘to control’, but Skt. sáhate ‘overcome’; English have/Serman haben H *keh2 p- ‘to take’, but also Sreek κάπτω ‘gulp’, and Latin capiō ‘seize’. . . The radical meanings of the roots that became “have” verbs in the IE daughter languages generally have to do with “grasping” or “holding”: this suggests that the semantic metaphor Y AL L was productive in PIE, but that it wasn’t lexically fixed. • Roots are per Rix and Kümmel 2001, and are cited as head-forms. The fundamental discussion of the be/have split in IE is Meillet 1E23, followed closely by Nenveniste 1E66. Note however that much of Meillet and Nenveniste’s reasoning has been subsequently disputed or typologically disproved, as discussed in e.g. Stassen 200E: D–10. See also Naldi and Ouzzolin 2005 for a detailed analysis of the various PIE “have” roots and their Einzelsprachlich reflexes. . The dativus possessivus is reconstructible for the proto-language: predicate dative constructions are found in most major IE subfamilies, and are composed using the dative case (or its syncretic analogue) + inherited *h1 es- (or a suppletive variant, such as *bh uH- or *ǵenh1 -) + a Nominative NP. • Ror examples and discussion of the predicate dative in the IE daughter languages, as well as additional bibliography, see Nauer 2000: 161–1D0, with the proviso that I do not necessarily endorse her reconstruction of an alienable/inalienable distinction in PIE. Note too that the question of why IE speakers would have used the dative case (rather than, say, the locative) to express a locational meaning in the predicate possessive formula remains unclear. Nortone 2010: 63ff. and Luraghi 2003: 51ff. posit that the dative case had an underlying “locatival” meaning, and this presumably licensed the locational metaphor behind the dativus possessivus. · Alphabetic Greek Possessives Re-examined . In Tomer, both a “be”-possessive (the dativus possessivus, often cited as the “ ἔστι μοι” construction) and a “have”-possessive (using ἔχω) are attested and productive. E.g.: . . Dativus possessivus: Il. 1.300, τῶν δ’ ἄλλων ἅ μοί ἐστι θοῇ παρὰ νηῒ μελαίνῃ, “...but of those other things that I have by [my] quick, black ship...” .. . “Have”-posessive: Il. 1.225, οἰνοβαρές, κυνὸς ὄμματ᾽ ἔχων, “...wine-sop, having the eyes of a dog...” Kulneff-Eriksson 1EEE is a comprehensive, diachronic study of the ἔχω and ἔστι μοι constructions in Sreek. The section on Tomer (pp.65–CD) draws on 5 books from each epic (Il: 1, 3, 4, 20, 24; Od: 1, E, 10, 11, 12), and yields the following distributional data: have and hold · . . Total number of possessive tokens: 1EC. . . . Iliad: 54 ἔστι μοι tokens; 45 possessive ἔχω tokens I 55% to 45%. . . Odyssey: 36 ἔστι μοι tokens; 62 possessive ἔχω tokens I 3C% to 63%. . . Kulneff-Eriksson suggests that the different distributions of the predicate dative in the two epics reflects a larger, observable decline of the predicate dative in Sreek, as well as a corresponding increase in ἔχω constructions (30–3D). Such a decline was no doubt exacerbated by the steady diachronic retraction and eventual loss of the dative case in spoken Sreek. In a series of followup studies, Nenvenuto and Pompeo have found that the diachronic decline of the dativus possessivus was not part of a more general decline in Nom.NP + εἶναι + Oblique-NP structures (Nenvenuto and Pompeo 2012, Nenvenuto and Pompeo 2015): . . In Tomer, they find a nearly 12:1 ratio of predicate dative vs. predicate genitive usage. Towever, in the Olassical era overall instances of the predicate dative decline sharply, while net instances of the predicate genitive remain relatively flat (Nenvenuto and Pompeo 2012: D1). . . They also claim that the predicate genitive diachronically expressed a narrower band of semantic meanings than the predicate dative did and unlike the predicate dative, was restricted largely to marking permanent affiliation and kinship relations (Nenvenuto and Pompeo 2015: 2E). · Mycenaean Possessives Re-examined . Apparent “have-” predicates with ἔχω are exceptionally common in Mycenaean. Kulneff-Eriksson iden- tifies 21C firm ἔχω tokens in her discussion of Mycenaean, and identifies at least 50 possible additional tokens (Kulneff-Eriksson 1EEE: 50–52): . . The most common form of ἔχω in Mycenaean is 3sg. e-ke (hekhei), but the form e-ke-qe (hekhei-kw e?) is nearly as abundant; the exact nature and function of the -qe variant is unclear, however, since the exact nature of the -qe morpheme itself is disputed. Kulneff-Eriksson does not distinguish between the two variants in her counts. .. Kulneff-Eriksson rightly does not include in her counts forms of ἔχω that have been emended or restored, although these forms are sometimes cited by others. .. Kulneff-Eriksson also excludes any form of ἔχω that does not have an overt argument; e.g., KN S D20.1 e-ko-si (hekhonsi, “they have”), which lacks an overt subject. .. Note too that her results do not include data from tablets and scholarly resources published after 1EED, nor does she include possible compound forms. • A great deal has been written on the -qe morpheme. Ror further discussion on, citations of, and bibliography relating to the element, see Nartoněk 2003: 314, EE1, Tooker 1ED0: 140, Kulneff-Eriksson 1EEE: 51, and Documents2 p.5C6 (ad loc). See Kulneff-Eriksson 1EEE: 1C5–CC for the list of Linear N texts that her corpus comprises. . The abundance of apparent “have”-possessives in Mycenaean is counter to the diachronic trends observed in Tomer and PIE. Nut the distribution of ἔχω tokens in the Linear N texts is uneven: have and hold · . . Of the 200+ instances, an overwhelming majority (J 1D0) occur in the PY E-series “landholder” texts (specifically, PY series Ea, Eb, Ed, En, Eo, Ep, Eq, Es). Of the remainder, 16 tokens occur in the PY Jn-series metals tablets, 5 in the PY Na-series commodity texts, and half a dozen or so other tokens are found in miscellaneous other texts. . . . There are only 4 secure total tokens in all of the KN tablets. Note though that the KN E-series is quite small, and that there is no KN J-series at all. In the E-series texts, transitive ἔχω always governs the terms o-na-to (onāton, “lease/piece of land”), ko-to-na (ktoinā, “plot of land”), or much more rarely, ki-ti-me-na (ktimenā “private land”). . . Examples: PY Ea CD2, o-na-to e-ke; PY En C4.2, ko-to-na e-ko-si; PY Er DD0.1, ki-ti-me-no e-ke. . . The distribution and use of ἔχω is understandable if we assign it a meaning closer to its radical meaning of “control/manage”. I.e., it does not mark predicative possession as such, and so did not semantically compete with the predicate dative. • On the meaning of e-ke in the E-series texts, note Documents2 p.243: “As usual on these tablets, ekhei/ekhonsi [sic] implies ‘tenancy’, not ownership.” This meaning was adumbrated in Meillet 1E23: 10 for Tomeric Sreek; moreover, see KulneffEriksson 1EEE: 63–64, who concludes that ἔχω never indicates “ownership” as such in Linear N. Ror the root semantics of the verb, cf. Meier-Nrügger 1EC6, who posits that the radical meaning of PIE *seǵh - was “einen im Kampf überwältigen, durch Kampf über etwas Meister werden”, and thus the shift to “have, possess” is specifically a development of later Sreek. . The Mycenaean economy is generally characterized as being of the “redistributive” or “palace” type, whereby labour and goods were distributed via a centralized authority rather than through a market system. This fact fundamentally shaped the possessive syntax of the Linear N texts. . . The discourse frame of the Linear N texts determined their patterns of possessive syntax: forms of ἔχω indicate only “managed” resources, and the dativus possessivus was systematically avoided by scribes based on the assumption that all the goods and properties named in the texts were inherently owned. Explicitly marking this ownership would have been otiose. • The amount scholarship on the Mycenaean socio-economic system is vast, and deeply specialized. A comprehensive introduction is Killen 200D. See in particular Killen 200D: 165–C1, Kulneff-Eriksson 1EEE: 52–60, and Uchitel 2005 for discussion of the Pylos E-series “landholding” texts. There has been some resistance to the idea that Aegean Nronze Age economies were necessarily “redistributive.” See Nakassis, Parkinson, and Salaty 2011 and Parkinson, Nakassis, and Salaty 2013 for a revisionist approach, though note that the model proposed therein could still generate the discourse-frame that I posit in this paper. have and hold · Bibliography Naldi, Philip, and Pierluigi Ouzzolin. 2005. “Oonsiderazioni etimologiche, areali e tipologiche dei verbi di “avere” nelle lingue indoeuropee.” In Latin et langues romanes: études de linguistique ofertes à József Herman à l’occasion de son 80ème anniversaire. Edited by Sándor Kiss, Luca Mondin, and Siampaolo Salvi, 2C–36. Tübingen: Niemeyer. Nartoněk, Antonín. 2003. Handbuch des mykenischen Griechisch. Teidelberg: Winter. Nauer, Nrigitte. 2000. Archaic Syntax in Indo-European: the spread of transitivity in Latin and French. Nerlin: de Sruyter. Nenveniste, Émile. 1E66. “‘Être’ et ‘avoir’ dans leurs fonctions linguistiques.” In Problèmes de linguistique générale I. 1DC–20C. Paris: Sallimard. Nenvenuto, Maria Oarmela, and Rlavia Pompeo. 2012. “Expressions of predicative possession in Ancient Sreek: ‘εἶναι plus dative’ and ‘εἶναι plus genitive’ constructions.” ΑΙΩΝ: Annali del Dipartimento di Studi Letterari, Linguistici, e Comparati Sezione linguistica, no. 1. . 2015. “Verbal Semantics in Ancient Sreek Possessive Oonstructions with eînai.” Journal of Greek Linguistics 15 (1): 3–33. Nortone, Pietro. 2010. Greek Prepositions from Antiquity to the Present. Oxford: Oxford University Press. Pixon, Robert M. W. 2010. Basic Linguistic Theory: Grammatical Topics. Vol. 2. Nasic Linguistic Theory. Oxford; New York: Oxford University Press. Puhoux, Yves, and Anna Morpurgo Pavies, eds. 200D. A Companion to Linear B: Mycenaean Greek Texts and Their World. Vol. 1. Louvain-la-Neuve: Peeters. Tajnal, Ivo. 1EE4. “Pie frühgriechische Rlexion der Stoffadjektive und deren ererbte Srundlagen.” In Früh-, Mittel-, Spätindogermanisch. Akten der IX. Fachtagung der Indogermanischen Gesellschaft, CC–10E. Wiesbaden. . 1EE5. Studien zum mykenischen Kasussystem. Vol. C. Untersuchungen zur indogermanischen Sprachund Kulturwissenschaft. Nerlin: de Sruyter. Teine, Nernd. 1EEC. Possession: Cognitive Sources, Forces, and Grammaticalization. Oambridge: Oambridge University Press. Tooker, James T. 1ED0. Linear B: An Introduction. London: Nristol Olassical Press. Killen, John T. 1ED3. “Mycenaean Possessive Adjectives in -e-jo.” Transactions of the Philological Society D1 (1): 66–EE. . 200D. “Mycenaean Economy.” In A Companion to Linear B: Mycenaean Greek Texts and Their World, edited by Yves Puhoux and Anna Morpurgo Pavies, 1:15E–200. Louvain-la-Neuve: Peeters. Kulneff-Eriksson, Karin. 1EEE. “On ‘have’ in ancient Sreek: an investigation on ἔχω and the construction εἶναι with a dative as expressions for ‘have’.” PhP diss., Lund University. Luraghi, Silvia. 2003. On the Meaning of Prepositions and Cases: the expression of semantic roles in Ancient Greek. Amsterdam; Philadelphia: J. Nenjamins Publishing. Meier-Nrügger, Michael. 1EC6. “Ἔχω und seine Nedeutungen im Sriechischen.” Museum Helveticum 33:1D0– 1D1. Meillet, Antoine. 1E23. “Le développment du verb ‘avoir’.” In Αντιδωρον: Festschrift Jacob Wackernagel, E–13. Söttingen: Vandenhoeck und Ruprecht. have and hold · Nakassis, Pimitri, William A. Parkinson, and Michael L. Salaty. 2011. “Redistribution in Aegean Palatial Societies. Redistributive Economies from a Theoretical and Oross-Oultural Perspective.” American Journal of Archaeology 115 (2): 1CC–1D4. Parkinson, William A., Pimitri Nakassis, and Michael L. Salaty. 2013. “Orafts, Specialists, and Markets in Mycenaean Sreece. Introduction.” American Journal of Archaeology 11C (3): 413–422. Rix, Telmut, and Martin Kümmel. 2001. LIV · Lexicon der indogermanischen Verben. 2nd edition. Wiesbaden: Ludwig Reichert. Schwyzer, Eduard, and Albert Pebrunner. 1E50. Griechische Grammatik. Band II: Syntax und syntaktische Stilistik. München: Neck. Stassen, Leon. 200E. Predicative Possession. Oxford: Oxford University Press. Uchitel, Alexander. 2005. “Land-Tenure in Mycenaean Sreece and the Tittite Empire: Linear N Land-Surveys from Pylos and Middle Tittite Land-Ponations.” Journal of the Economic and Social History of the Orient 4D (4): 4C3–4D6. Ventris, Michael, and John Ohadwick. 1EC3. Documents in Mycenaean Greek. 2nd edition. Oambridge: Oambridge University Press. Waanders, R. M. J. 200D. “The Syntax and Oo-occurrence of Oase Runctions in Mycenaean Texts.” In Colloquium Romanum. Atti del XII Colloquio Internazionale di Micenologia, Roma, 20-25 febbraio 2006, edited by A. Sacconi, M. Pel Rreo, L. Sodart, and M. Negri, CE5–D06. Pisa: Rabrizio Serra. have and hold ·