¹ Bibliographical references in this article are to be interpreted as follows. Arch. or. = Archiv orientâlní. Benfey = Theodor Benfey, Handbuch der Sanskritsprache, 2 vols.; Leipzig, 1852-4. Bennett = Charles E. Bennett, Syntax of early Latin, 2 vols.; Boston, 1910-4. Boisacq = Émile Boisacq, Dictionnaire étymologique de la langue grecque; Heidelberg and Paris, 1907. Brugmann, Gdr. = Karl Brugmann and Berthold Delbrück, Grundriss der vergleichenden Grammatik der indogermanischen Sprachen², 5 vols. ; Strassburg, 1893-1916. Brugmann, KVG = Karl Brugmann, Abrégé de grammaire comparée des langues indo-européennes, transl, by J. Bloch, A. Cuny, and A. Ernout; Paris, 1905. Brugmann, Totalität = Karl Brugmann, Die Ausdrücke für den Begriff der Totalität in den indogermanischen Sprachen; Leipzig, 1893-4. Buck, Comp. gr. = Carl Darling Buck, Comparative grammar of Greek and Latin, 4th impr.; Chicago, 1948. Delbrück, CO = B. Delbrück, Der Gebrauch des Conjunctivs und Optativs im Sanskrit und Griechischen; Halle, 1871. Delbrück, Gdr., see Brugmann, Gdr. Ernout and Thomas = Alfred Ernout and François Thomas, Syntaxe latine; Paris, 1951. Forrer, BoTU = Emil Forrer, Die Boghazköi-T exte in Umschrift, 2 vols.; Leipzig, 1922-6. Friedrich, El. = Johannes Friedrich, Hethitisches Elementarbuch, 2 parts; Heidelberg, 1940-6. Friedrich, Vert. = Johannes Friedrich, Staatsverträge des Ḫatti-Reiches in hethitischer Sprache, 2 parts; Leipzig, 1926-30. Goetze, Tunnawi = Albrecht Goetze, The Hittite ritual of Tunnawi; New Haven, 1938. Goodwin = William Watson Goodwin, Syntax of the moods and tenses of the Greek verb, rev. ed.; Boston, 1900. Güterbock, Kumarbi = Hans Gustav Güterbock, Kumarbi; Zurich and New York, 1946. Hahn, Subj. and opt. = E. Adelaide Hahn, Subjunctive and optative: Their origin as futures; New York, 1953. Hirt, IG = Hermann Hirt, Indogermanische Grammatik, 7 vols.; Heidelberg, 1921-37. Hofmann = J. B. Hofmann, Etymologisches Wörterbuch des Griechischen; Munich, 1950. Hofmann, Lat. Gr. = Manu Leumann and Joh. Bapt. Hofmann, Stolz-Schmalz Lateinische Grammatik⁵; Munich, 1928. Hrozný, SH = Friedrich Hrozný, Die Sprache der Hethiter; Leipzig, 1917. Ose = Fritz Ose, Supinum und Infinitiv im Hethitischen; Leipzig, 1944. Otten, Mythen = Heinrich Otten, Mythen vom Gotte Kumarbi, Neue Fragmente; Berlin, 1950. Pedersen, Hitt. = Holger Pedersen, Hittitisch und die anderen indoeuropäischen Sprachen; Copenhagen, 1938. Pedersen, MS = Albrecht Götze and Holger Pedersen, Muršlis Sprachlähmung; Copenhagen, 1934. Ribbeck = Tragicorum romanorum fragmenta³, ed. by Otto Ribbeck; Leipzig, 1897. Schwyzer, Gr. Gr. = Eduard Schwyzer, Griechische Grammatik, 2 vols. (2nd vol. compi. and ed. by Albert Debrunner); Munich, 1939-50. Sommer, Bil. = Ferdinand Sommer and Adam Falkenstein, Die hethitisch-akkadische Bilingue des Ḫattušili I (Labarna II); Munich, 1938. Sturtevant, Chr. = Edgar H. Sturtevant and George Bechtel, A Hittite chrestomathy ; Philadelphia, 1935. Sturtevant, HG¹ = Edgar H. Sturtevant, A comparative grammar of the Hittite language; Philadelphia, 1933. Sturtevant, HG² = Edgar H. Sturtevant (and E. Adelaide Hahn), A comparative grammar of the Hittite language, vol. 1, rev. ed.; New Haven, 1951. Walde-Pokorny = Alois Walde, Vergleichendes Wörterbuch der indogermanischen Sprachen, rev. and ed. by Julius Pokorny, 3 vols.; Berlin and Leipzig, 1927-32. Whitney = William Dwight Whitney, Sanskrit grammar²; Cambridge, 1923. ZA = Zeitschrift für Assyriologie und verwandte Gebiete. References are to pages, unless there is indication to the contrary. Quotations from Hittite and Latin give only enough words to make clear the sense and the syntax, with no indication of omissions.

² A connection between the conjunction and the particle, though it is not universally accepted, seems to me exceedingly likely. See Lg. 18.103-4.

³ Also with Lycian me. Pedersen (MS 57-8) accepts this connection, but apparently thinks it rules out the possibility of a connection with μά; as for μά, this, he declares, ‘muss wegen ganz abweichender Verwendung fern bleiben’. But in the extended form has the meaning ‘nevertheless’, and is also used (like Thessalian μά) as a correlative for in place of thus it is unobjectionable as a semantic parallel for Hittite -ma.

⁴ Cf. also his earlier equation of them (JAOS 47.180-1); but here he was dealing not with the consonantism but with the vocalism, concerning which he expressed a view which he later retracted (Lg. 6.30 fn. 19).

⁵ The existence of an Indo-European pronominal stem in m- was demonstrated in 1938 by Sommer (Bil. 164-6) and Pedersen (Hitt. 67-71). Pedersen gives the stem as mo-, but presumably it had the two grades mo- and me-.

⁶ He had originally suggested something of the sort in Totalität 49 fn. 1 (1893-4); later apparently retracted this at least so far as went, IF 21.8 (1907) ; but finally returned to it, IF 37.160-1 (1916-7). Brugmann of course did not recognize the element -mo- as an indefinite stem; he identified it with the superlative ending (ib. 159). But, as I noted in the article cited (114), he was making a comparison more apt than he realized when he linked Sanskrit sama-, Greek όμόs, and Latin similis with Greek aíπόs and aíπὐs, Latin aequus in which he identified the first element with a demonstrative and the second with the indefinite stem kwo- (156-9).

⁷ We also agree in suggesting a possible connection with the reflexive pronoun (IF 58.265 fn. 1, Lg. 18.115-6).

⁸ As expressed in Totalität, cited 264 fn. 4. He does not mention the article in IF 37.

⁹ This of course accounts neatly for the much disputed forms Greek ɛκατόν (cf. sup., fn. 6) and Sanskrit sahasra- (265).

¹⁰ Both our papers appeared in 1942.

¹¹ Schwyzer (Gr. Gr. 2.569 fn. 3) seems willing to accept our view.

¹² So far as derivation goes, I have suggested (Lg. 18.91) a parallelism with Latin simplex.

¹³ On the commonness of -is adjectives in Hittite, cf. Sturtevant, Lg. 10.266-73.

¹⁴ Approved by Friedrich, ZA 15.244.

¹⁵ The etymology of asma is problematical. One thinks of the possibly similar formation -ašta ‘then, finally’ (cf. našta = nu + -ašta), but this is always enclitic. Can the -ma part be the element which, as I have noted above, I have already suggested (Lg. 12.115) is perhaps to be recognized in tamaiš ‘other’, and which, incidentally, is the one that Brugmann posited (as noted above, fn. 6) as the second element in the root sem- (IF 37.159) and that I have proposed to equate with the indefinite stem (Lg. 18.114)? This element occurs frequently in Indo-European in the formation of both ordinals and superlatives; cf. e.g. the word for ‘first’ in many languages, as Sanskrit prathama- and Latin primus. That it functioned in Hittite as an ordinal suffix is proved by the dative si-ip-ta-mi-ya ‘seventh’.

¹⁶ As do also the verbal adjectives (participles).

¹⁷ As instances of this very exceptional usage, we may cite KBo 4.4.2.63-4 ku-e KARAŠ.HI.A I-NA KUR ^URUNu-haš-ši hal-ki.HI.A-uš har-ni-in-ku-wa-an-zi pí-e-hu-da-an har-ta ‘the troops which he had led into Nuhassi in order to destroy the crops’, and KUB 24.5.2.17-8 IGI.HI.A-YA-wa-za-kán tu-uk ú-wa-an-na ha-aš-ši-ik-ki ‘open my eyes for seeing you’, i.e. ‘open my eyes so that I may see you’. A few other examples of the combination of case-forms of nouns with an infinitive are given by Ose, 57 and 86.

¹⁸ These are allied to the Latin gerund. See Sturtevant, Lg. 20.206-11, and HG² 73.

¹⁹ It has been customary to call the forms in -manzi/-wanzi supines; so e.g. Friedrich, El. 1.80. But to me it seems preferable to bestow the name of infinitive on both types; so e.g. Sturtevant, HG¹ 267 and HG² 148. Cf. Ose 88.

²⁰ The Latin construction must have been in origin precisely parallel to the Hittite: thus such a passage as Terence, Heaut. 585-6 iube hunc abire must have meant originally ‘drive him for the purpose of going’, with hunc as object of the main verb and abire a dative of purpose. But the Latin infinitive eventually lost its dative force, becoming purely an indeclinable verbal noun which functioned mainly as nominative and accusative (with the inflected gerund taking over the other case relations); and in this new development it acquired the ability to possess not merely an object (copias) as the Hittite infinitive occasionally did (cf. fn. 17), but even a subject, the accusative which had once served as object of the main verb. But this state of affairs could apply only to the infinitive in combination with a main verb of ordering, compelling, or allowing; the use of the infinitive with subject-accusative in combination with a main verb of saying or thinking or perceiving, in connection with which the infinitive could never have had dative force, must have had a quite different origin (see my discussion of this in TAPA 81.117-29).

²¹ When the main verb is in the passive, the word which would be the object of the infinitive in English is attached to the main verb as its subject. Thus if we wish to say ‘it was directed to sacrifice a bull to him’, we must recast this as ‘a bull was directed to him for sacrificing’, MS 1.11-2 nu-uš-ši GUD pu-u-hu-ga-ri-iš pí-i-ya-u-wa-an-zi SI × DI-at.

²² For other examples, see Brugmann, Gdr. 2.3.918 and Whitney 352.

²³ Both Brugmann (Gdr, 2.3.918) and Delbrück (KZ 18.104 and Gdr. 4.470) seem to think the position of the dative of reference—whether before the infinitive as in 5.2.9, or after it as in 10.14.12—is of some significance; but this I cannot see.

²⁴ So, rightly, Hofmann (Lat. Gr. 597), who translates ‘für das Geld, seine Beschaffung’.

²⁵ Once the shift was made, the gerundive might appear in passages where the gerund would scarcely have applied, as in Plautus, Pers. 5 fio miser quaerendo argento.

²⁶ I offer this theory and discuss it at length in TAPA 74.277-98.

²⁷ Of course the foregoing hypothesis would be proved only if we actually possessed in Latin such a passage as *decemviros legibus scribendo creavimus, but we do not ; where the noun is feminine or plural, we find no instance of a gerund, only the agreeing gerundive, as in Brugmann's decemviros legibus scribendis or in triumviri rei publicae constituendae. But in the genitive there are plenty of instances of the gerund: e.g., Plautus, Capt. 1008 lucis tuendi copiam, ib. 852 nominandi istorum copia. Sporadic vestiges are met in classical times, especially with the plural sui, perhaps because it looked like a singular, as in Caesar, Bell. Gall. 7.43.2 legatos sui pur gaudi gratia mittunt, and with the genitive plural in -ōrwn or -ārum, probably to avoid cacophonous repetition, as in Cicero, Inv. 2.2.5 exemplorum eligendi potestas. But infinitely commoner was the shift to the gerundive construction, from lucis tuendi to lucis tuendae, undoubtedly, in my opinion, brought about by the ambiguous form in the masculine or neuter, as Accius 275-6 (Ribbeck) luminis conspiciendi insolentiti.

²⁸ But there are two important differences between the gerund-gerundive construction just discussed, and other manifestations of the double dative in Latin. (1) The double dative in Latin (and indeed the use also of a dative of purpose alone without a second dative) seems more restricted than in Hittite or Vedic. It is mainly confined (a) to verbs of being, and (b) to verbs with the general notion of giving, assigning, or deeming, either active or passive. It is closely related either to the subject (with the copula and with passive verbs) or to the object (with active transitive verbs); indeed it may interchange with the predicate nominative in the former case and with the predicate accusative in the latter. (2) With a gerund the performer of the action is regularly either the subject of the verb modified by the gerund, or a generalized impersonal agent; and the accompanying dative substantive represents the recipient of the action. Similarly with the ordinary double dative construction, where we have a noun of action corresponding to a transitive verb, the dative substantive usually designates the recipient of the action, as in Plautus, Most. 922 ne quid captioni mihi sit-, but there are exceptions in which it designates the agent instead, as in Trin. 632 ut sic odio esses mihi. This exception seems to be the rule with one special type of noun of action, the verbal noun (really a dative supine) in -tuī or -suī, as in Curc. 501 bono usui estis nulli.

²⁹ Of course partitive apposition in Latin is by no means confined to the dative. See Hofmann's extremely interesting discussion and collection of examples, Syntaktische Gliederungsverschiebungen im Lateinischen infolge Erstarrung ursprünglich appositioneller Verhältnisse (IF 42.81-5). Hofmann cites numerous examples not only of the dative (84-5) but also of the nominative (85-6) and of the accusative (81-4), a few of the ablative (85) and of prepositional phrases (86), but none of the genitive; however, he suggests (85 fn. 2) that the construction lucís tuendi (discussed above, fn. 27) may exemplify this type. I formerly doubted this (see TAPA 74.282) ; but I now realize, as I have just said, that the fundamental relationship of the verbal noun to the substantive which it accompanies may be viewed as in origin one of quasi-apposition.

³⁰ Some might term mihi a dative of agent, but actually the so-called dative of agent is essentially a dative of reference.

³¹ On the use of this particular noun in partitive apposition in the dative, cf. its similar use in the accusative, illustrated below, fn. 34.

³² For discussion and examples, see Friedrich, ZA 1.175 and 2.47, Vert. 1.43-5 and 2.24, El. 1.69; Sommer, Bil. 106.

³³ It is hardly possible to render this literally. The force of the dative A-NA ^DUTU-ŠI must be ‘with reference to my Majesty’ and that of ČU-i ‘with reference to (my Majesty's) hand’, but the adverb or postposition anda adds a locative notion ‘in, within’ at least to SU-t and perhaps to A-NA ^DUTU-ČI also. (It must not be forgotten that the Hittite dative case serves as a locative as well.)

³⁴ As in Homer, Il. 7.14-6 and Plautus, Rud. 1345-6 te Venus eradicet caput atque aetatem tuam. With the latter instance of apposition of part and whole, contrast their coordination in the same play 486 Neptuno credat sese atque aetatem suam; and for the appositional use of the accusative aetatem, cf. the similar use of the dative aetati in Trin. 313, already cited above (see fn. 31).

³⁵ The constructions to which these names are given seem to me to be essentially the same, except that with a verb or a participle (e.g. Vergil, Aen. 1.561 voltum demissa) either name can be applied, and with an adjective (e.g. ib. 320 nuda genu) only that of an accusative of specification. It is significant that both usually indicate a part of the body, or, by an extension (as in Vergil, Ecl. 1.53-4 saepes apibus florem depasta), some portion of a whole.

³⁶ For examples, see Gdr. 2.3.919. Brugmann adds to his Vedic instances parallel examples from Avestan, and then goes on to cite examples from Greek (920-1) and from Bal to-Slavic (921-2).

³⁷ Rather oddly, these are the only ones that survived into the classical period.

³⁸ In Hittite, there is a supine in -wan which is regularly classed as a suffixless dative (or locative), but which I have sometimes been tempted to regard as an accusative; it is restricted to a very limited use in combination with not more than two or three auxiliary verbs, in the sense of ‘begin’.

³⁹ In that case the predicative construction that I view as the starting-point of the locution would have been merely a contributing, not a determining, factor in its development. Cf. sup., fn. 20.

⁴⁰ In classical Latin, to be sure, the gerund is limited to the use with a preposition, but in early Latin it is freely employed as a nominative-accusative; see TAPA 74.286-94.

⁴¹ Cf. the suggestion made at the close of the preceding paragraph about the possible relationship of pāśān to vettha.

⁴² In early Latin we find a possible parallelism in the use of certain verb forms which, like the Greek second aorist, combine a weak or aorist stem with the endings of the imperfect—these endings in the Latin forms being -am -ās etc., seen not only in the imperfect but also in the pluperfect. These are surely past indicatives in origin. Instances are ne fuas (Plautus, Capt. 443 and Trin. 267) and above all the very common ne attigas, ne attigatis (Plautus, Bacch. 445, Ep. 723, Most. 468, True. 276; Terence, And. 789; CIL 1².424 Numbers 499 and 500).

⁴³ Exactly parallel is the use in prohibitions in early Latin with the subjunctive of nōn or neque instead of nē or nēve. We find neque fairly often, but nōn is extremely are; Bennett (1.170) cites only two instances.

⁴⁴ The development of modal relations in affirmative expressions is similar. Hittite uses man with a so-called present (really present-future) indicative to express potentiality, and with a past indicative to express unreality. In Homeric Greek, which was certainly a transitional period in which modality was coming to be expressed by verbs rather than by particles, we find considerable fluctuations of usage; potentiality or remote futurity can be expressed by the indicative (future) with a modal particle () or by a mood (subjunctive or optative, though the subjunctive denotes more vivid rather than less vivid futurity) without a particle or, tautologically, with it; unreality can be expressed by the past indicative with a particle or, tautologically, by the optative with a particle. Eventually in Greek the use of the past indicative (plus a particle) became the sole norm in expressions of unreality, exactly as in Hittite ; but in expressions of potentiality, the standard form was the tautological one, optative plus particle. Sanskrit and Latin, on the other hand, use no corresponding affirmative modal particle, and depend purely on moods for the expression of modal relations so far as potentiality and unreality go.

⁴⁵ Cf. Sturtevant, HG¹ 251 fn. 70.

⁴⁶ The Greek and Latin imperatives are used only in a restricted sense; see f n. 47.

⁴⁷ When the iterative or durative aspect of an action needs to be stressed, for some reason both Greek and Latin show a tendency to use the imperative (with and nē respectively) . In Greek the regular use in prohibitions is the present imperative is met very aspect and the aorist subjunctive for punctual aspect; the aorist imperative is met very seldom, and the present subjunctive not at all (see Goodwin 89). In Latin the perfect subjunctive is regularly punctual and the present subjunctive is often, though not always, durative, as in the common phrases molestus ne sis (Plautus, As. 469 et al.) and ne vereare (Capt. 349 et al.); Delbrück (Gdr. 4.383) thinks this distinction is the rule, and, though Bennett is probably right in considering this generalization too sweeping for early Latin so far as we know it (1.174), it doubtless is correct, as Bennett himself agrees (ib.), as applied to the original distinction between the two tenses. As for the present imperative in prohibitions (as ne fle, ne time), this certainly does seem, precisely like the present imperative in Greek, to be confined to durative use; cf. Ernout and Thomas (197), ‘il s’ applique de préférence à une action déjà commencée, à un sentiment que Pon éprouve déjà lorsque la défense est formulée.'

⁴⁸ As just noted, fn. 47.

⁴⁹ Again cf. fn. 47.

⁵⁰ In view of the interchangeability of augmented and augmentless forms to express past time, it is not surprising that analogy should introduce an occasional augmented form in prohibitions as well; this happened in the epics and later, but is met only two or three times in Vedic (Whitney 218).

⁵¹ In the same way I believe that in Latin the past indicatives of the early type referred to in fn.42 generated the -ā- subjunctive. This was used as a present tense, and so eventually the -ā- endings came to be added to the present stem instead of to the aorist stem as was originally the case. For a more detailed discussion of the problem of the -ā- subjunctive, and of some of the other matters here referred to, see my monograph, Subj. and opt. 34-51.