The origin of the thematic verbs

The origin of the thematic verbs Miguel Carrasquer Vidal, May 2017 Over the past few years (see here, here, here, here, here and here), I have been developing a theory about the origin of the PIE verbal system. To briefly summarize: 1. PIE inherited two separate categories of verbs: hi-verbs and mi-verbs. 2. The hi-conjugation used reduplication to mark the subject of intransitive verbs and the object of transitive ones. 3. In the hi-conjugation, the first vowel of the verb (root or reduplication) could be lengthened, creating a distinction between the A-s ste aorist a d the P-s ste perfe t . If we split the resulting paradigm into transitive, intransitive/singular object and transitive/plural object, this gives 6 verbal categories: A1 intransiti e > redupli ated prese t , A transitive > root aorist, A3 transitive plural > redupli ated aorist ; P1 intransitive > perfect, P2 transitive > olō- er s , P transitive plural > Narte er s . 4. Even before the split-up of Anatolian, the A-system largely acquired mi-conjugation endings, with only a few traces of the older hi-endings remaining (the k-aorists < 1sg. *-h2a after a laryngeal, the Vedi pre ati e sg. e di g -s, the 5:1 Ablaut of the root aorist, and the Vedic 3pl. ending -ur). 5. Hittite horizontally merged categories A1 and A2 with the mi-conjugation root presents, and merged P1 and P2 as the hi-conjugation. Reduplication was given up in the plural of A1 and P1. The wholly reduplicated categories A3 and P3 survived as lexically reduplicated verbs (mior hi-conjugation) or as reduplicated variants of unreduplicated verbs, in free (?) variation. 6. Tocharian, at least in the preterite, vertically merged categories A2 with P2 (preterite I), and A ith P preterite II . Categor A redupli ated prese ts had ee lost [to thematization: pär- < *bher-e-ti] in Pre-Tocharian, see below, so did not merge with P1 (the perfect), which merged instead with the s-aorist (preterite III). 7. In Germanic, the intransitive forms (A1 redupli ated prese t , P1 perfect) still show lack of reduplication in the singular, and reduplication in the plural. For A1, we can adduce the Goth. weak preterite sg. -da vs. pl. -dedun; for P1, we have for instance sg. C1aC2- ~ pl. C1eC1C2- > C1ēC2- in the strong verbs of classes IV and V. An additional argument in favour of an unreduplicated singular in category A1 is the fact that a 3sg. form *bhére (and not *bhi-bhére-) must underlie the thematized form *bhéreti (Ved. bhárati besides bibhárti ~ bíbharti). Outside Indo-Iranian, the reduplicated present was almost completely lost to thematization (apart from a few isolated forms such as ί o αι and τίθη ι). It is not surprising that a 3sg. form like *bhére, once it was provided with a mi-conjugation ending *bhéret(i), came to be interpreted as a thematic form. This cannot, however, explain the rise of the thematic conjugation itself, with its peculiarities such as the alternation of o ~ e in the thematic vowel, or its anomalous 1sg. present form -ō, because none of this follows merely from an ending*-e. Clearly, category A1 became thematized based on some pre-existing thematic model. Let s take a look at the paradig s of the hi-conjugation as I reconstruct them for the hypothetical verb *CeRC. Highlighted are the 3sg. forms with a quasi-thematic ending *-e or *-ó: A-system intr./tr.sg. 1 CéRC-h2a 2 CéRC-th2a 3 CéRC-e / ĆRC-s tr.sg. 1 CéRC-me 2 CéRC-e 3 C‘̥C-ér / C‘̥ C-́r tr.pl. Cí-C‘̥C-h2a Cí-C‘̥C-th2a Cí-C‘̥C-e / Cí-C‘̥ C-s intr./tr.pl. Ci-C‘̥C-mé Ci-C‘̥C-é Cí-C‘̥C-r̥ middle (Ci-)C‘̥ C-h2-ó(Ci-)C‘̥ C-th2-ó(Ci-)C‘̥ C-ópl. (Ci-)C‘̥ C- dhʷ-ó(Ci-)C‘̥ C-dhʷ-ó(Ci-)CRC-r-ó- P-system intr/tr.sg. 1 CóRC-h2a 2 CóRC-th2a 3 CóRC-e / ĆRC-s tr.sg. 1 CóRC-me 2 CóRC-e 3 CéRC-r̥ / CéRC-r̥s tr.pl. Ć-(C)RC-h2a Ć-(C)RC-th2a Ć-(C)RC-e / Ć-(C)RC-s intr./tr.pl. Ce-(:/C)RC-mé Ce-(:/C)RC-é Ć-(C)RC-r̥ middle CéRC-h2-aCéRC-th2-aCéRC-epl. CéRC- dhʷ-eCéRC-dhʷ-eCéRC-r-e- Reduplicated forms of the P-middle would have regularly merged with those of the A-middle (because unstressed **ī merges with *i), so are omitted. I e look at the pure the ati categories enumerated in LIV, we find five categories: 1c *tuk-é1d *ḱéi-e1i *si-sd-é1n *bhér-e1o *tud-é- schwundstufiger Wurzelstativ vollstufiger Wurzelstativ i-redupliziertes thematisches Präsens Präsens mit vollstufiger Wurzel und Suffix e/o Präsens mit schwundstufiger Wurzel und Suffix e/o Of course, including the two statives is cheating a bit, because they are not real thematics, but rather middle forms (lacking a 3rd person *-t), but I believe they are central to an understanding of the rise of the thematic paradigms. I other ords, I suggesti g that LIV class 1c (A middles) leads to present classes 1o (including thematic aorists) and 1i. And that LIV class 1d (P middles) leads to present class 1n (including subjunctives). Schematically: A P 1c *tuk-é- → 1d *ḱéi-e- → 1o *tud-é-, 1i *si-sd-é1n *bhér-e- All that is required is a metathesis of the constituent elements of the middle forms. Instead of having the root followed by the hi-e di gs, follo ed a the ati o el , a d fi all by the middle additio al e di gs *-m, *-r, *-Ø, *-dh, *-j), the order becomes: root + thematic vowel + hi-endings + additional e di gs: 1 2 3 1 2 3 A-middle them. C‘̥ C-ó-h2C‘̥ C-ó-th2C‘̥ C-óC‘̥ C-ó- edhʷC‘̥ C-ó-dhʷC‘̥ C-ó-r- P-middle them. CéRC-a-h2CéRC-e-th2CéRC-eCéRC-o-medhʷCéRC-o-dhʷCéRC-o-r- It is difficult to tell whether this came about by a reshaping of the paradigm after the 3sg. form, rei terpreted as a e te ded the ati stem, or whether this was always simply another possible ordering of the elements, but the resulting thematic middles are attested pretty much everywhere, and make an excellent starting point for the development of the PIE thematic forms. In Bozzone 2010, we find the following passage: Greek has several isolated future formations that match the formal description of a Type I subjunctive; historically subjunctives, these formations are normally labeled as futures in the standard grammars. The forms are: βέο αι I ill li e , ἔ ο αι I ill eat , ἔσο αι I ill be , πίο αι I ill dri k ithout full grade of the root , έο αι I ill retur , χέο αι I ill pour (the latter two also attested as presents). Very interestingly, all of these forms take middle endings, and resemble the full-grade middle inflection of κ ῖται < PIE *ḱéi-[t]o[i]. This strongly suggests that the bhárati-verbs and subjunctives can formally be derived from the etathesized middle forms of the P-system (the ἔ ο αι-type), by replacing the middle endings with active mi-endings: 1 2 3 1 2 3 CéRC-a-h2CéRC-e-th2CéRC-eCéRC-o- edhʷCéRC-o-dhʷCéRC-o-r- CéRC-ā; CéRC-o-m(i) CéRC-e-s(i) CéRC-e-t(i) CéRC-o-me(si) CéRC-e-te CéRC-o-nt(i) The tudáti-verbs and thematic aorists (as well as, with reduplication, the *si-sd-é verbs) can similarly be derived from the etathesized middle forms of the A-system: 1 2 3 1 2 3 C‘̥C-ó-h2C‘̥C-ó-th2C‘̥C-óC‘̥C-ó- edhʷC‘̥C-ó-dhʷC‘̥C-ó-r- C‘̥C-́; C‘̥ C-ó-m(i) C‘̥C-é-s(i) C‘̥C-é-t(i) C‘̥C-ó-me(si) C‘̥C-é-te C‘̥C-ó-nt(i) The Ablaut of the thematic vowel in the tudáti-verbs must be analogical after that of the bhárativerbs. In the unmodified thematic middles, Hittite1, Tocharian2 and Gothic still show some forms (all 1 2 Thematic middles, all of them of the structure CaRRaClass III middle subjunctives with thematic vowel -e- (ToA -a-), all of the structure CäR- or CäC-. of them in the case of Gothic) with the original o-grade of the thematic vowel throughout. On the other hand, the 1sg. ending -́ of the tudáti-class may have been transferred to the bhárati-class3. Since the tudáti-verbs are derived from aorist middles, this immediately explains why thematic aorists such as *wid-é- or *dhugh-s-é always have a zero-grade root with stressed thematic vowel, i.e. belong to the tudáti-type. The use of tudáti-forms as presents (LIV 1n) is secondary. At least in Slavic, it is clear that tudáti-presents are derived from thematic aorists, by application of the tezziprinciple. That the distinction between thematic aorist indicatives and subjunctives is also secondary is shown in Vedic, where the thematic aorist subjunctive is the only subjunctive with zero grade: root aorist é-aorist s-aorist sé-aorist indicative CéRCC‘̥C-éĆRC-sC‘̥C-sé- subjunctive CéRC-e/oC‘̥C-́/́CéRC-s-e/o-- This must derive from earlier: indicative subjunctive root aorist CéRCC‘̥C-és-aorist ĆRC-sC‘̥C-séIf my theory on the origin of the middle as a verbal form with dative agreement is correct, then the fact that the subjunctive originates in the middle can be explained by invoking modal functions of the dative, such as the ethical dative4. The developments would have been: 1. Specialization of etathesized (thematic) middles *h1éd-e-, *tud-é- as middle (aorist) subjunctives, perhaps origi all fro ethi al dati e -constructions encoded in the middle verb (dative agreement). 2. Creation of active subjunctives by replacing the middle endings with active mi-endings (ἔσο αι ~ erō). 3. Creation of thematic active presents from A1 ( redupli ated prese ts ) on the model of the subjunctive (based on the 3sg. *bhére → *bhéreti), thus creating new presents for aoristic verbs. Likewise, some thematic (s-)aorist subjunctives start to double as indicative thematic aorists (*wid-é-). 4. Formation of secondary subjunctives Bozzo e s T pe II su ju ti es , such as the double thematic *bhére-e-ti, etc., and creation of tudáti-presents to provide presents to thematic aorists. The alternative explanation for 1sg. present *-ō, from *-o- ʷi > *-o-wi [= Luw. -ā i] > *-ōu [= TochB -eu, -au] > *-ō, remains valid, however. Even though phonetically *-ō can come from thematic aorist middle *-ó-h2, active aorist thematics would only use 1sg. *-óm. 4 Parodied i “hakespeare s The taming of the shrew (Act 1, Scene 2): 3 PETRUCHIO Villain, I say, knock me here soundly. GRUMIO Knock you here, sir! why, sir, what am I, sir, that I should knock you here, sir? PETRUCHIO Villain, I say, knock me at this gate The lack of pure thematics in Hittite suggests that the Anatolian split-off must have taken place somewhere between stages (2) and (3) in the above list of developments. Anatolian subsequently lost the subjunctive, and did not participate in the rise of thematic presents in stage 3, because there was no model for the transformation *bhére -> *bhéreti. These verbs (A1) instead merged with the root aorists (A2) and the root presents (M) (mi-conjugation). However, the former existence of a subjunctive in Anatolian is proven by the e/o-alternation in the thematic vowel of the *-je- and *-sḱeverbs, which must have its origins in the P-system subjunctives (middle and active). References Bozzone, Chiara, 2010, The PIE Subjunctive: Function and Development, in: H. Craig Melchert (ed.), The IndoEuropean Verb: Proceedings of the Conference of the Society for Indo-European Studies, 7-18, Wiesbaden.