Skip to main content Accessibility help
×
Hostname: page-component-78c5997874-4rdpn Total loading time: 0 Render date: 2024-11-10T18:53:01.800Z Has data issue: false hasContentIssue false

8 - Italic

Published online by Cambridge University Press:  15 September 2022

Thomas Olander
Affiliation:
University of Copenhagen

Summary

This chapter examines the distinctive innovations uniting Latin, the Sabellic languages and Venetic.

Type
Chapter
Information
The Indo-European Language Family
A Phylogenetic Perspective
, pp. 114 - 134
Publisher: Cambridge University Press
Print publication year: 2022
Creative Commons
Creative Common License - CCCreative Common License - BYCreative Common License - NCCreative Common License - ND
This content is Open Access and distributed under the terms of the Creative Commons Attribution licence CC-BY-NC-ND 4.0 https://creativecommons.org/cclicenses/

8.1 Introduction

The Italian peninsula before the Roman conquest was home to a large number of languages, both Indo-European and non-Indo-European.Footnote 1 Among these languages, the following have been thought to descend from a common ancestor, Proto-Italic (cf. Figure 8.1).

  1. 1. Latin, spoken in Latium in a number of slightly divergent dialects for most of which we have only scant information from inscriptions and glosses. The Latin of Praeneste, which is the findspot for the two earliest Latin inscriptions, and the Latin of Falerii are reasonably well attested in inscriptions dating from the seventh to second century BCE. The Latin of Falerii is often classified as an independent language called Faliscan, though this is not justified on linguistic grounds.Footnote 2 But towering above all is the Latin of Rome. In this language we have a small number of inscriptions from the seventh–sixth centuries BCE in a distinctively archaic form, which I call Very Old Latin. After slowing to a trickle in the fifth century, Latin inscriptions pick up again in the fourth century and are joined by literary documents in the third century. The Latin of Rome spread first to Italy, suppressing the previously existing linguistic diversity, and then to most of Western Europe, North Africa, and southeastern Europe north of the Jireček line.Footnote 3 Roman Latin survives today in its multiple descendants, the Romance languages.

  2. 2. The Sabellic languages. These languages, which form an as yet unquestioned subgroup, are

    1. a. Oscan, the language of the Samnites of central and southern Italy, who also expanded into Campania and Sicily, is represented by about 800 inscriptions dating from the mid-fourth century BCE to perhaps as late as the first century CE

    2. b. Umbrian, known chiefly from the Iguvine Tables from Gubbio (third–second century BCE; see Reference WeissWeiss 2010) and about forty smaller inscriptions, a few as early as the seventh century

    3. c. South Picene,Footnote 4 the language of fewer than thirty inscriptions from the Marche and Abruzzo dating from the sixth–third centuries BCE

    4. d. Pre-Samnite, the language of inscriptions from Campania before the Samnite conquest in the fifth century; the longest document is the Cippus of Tortora from the sixth–fifth centuriesFootnote 5

    5. e. In addition, there are a number of short texts in the dialects of the Volsci, Marsi, Paeligni, Marrucini, Vestini, and Hernici.Footnote 6 We also have a number of Sabellic loanwords in Latin (bōs ‘cow’ for expected *(w)ūs < *gʷou̯s being the most prominent of them).Footnote 7

  3. 3. Venetic, attested in more than 400 inscriptions from the northeast corner of Italy from the sixth to first centuries BCE. Some documents have been discovered in neighboring Slovenia and Austria. Not all scholars would agree that Venetic is an Italic language.

  4. 4. Sicel, the language of a small number (fewer than thirty) of pre-Greek inscriptions of eastern Sicily from the sixth to fourth centuries BCE and a number of glosses.Footnote 8 It is very difficult to determine much about this language beyond that it was Indo-European, as the form pibe ‘drink!’ = Ved. píba shows conclusively.Footnote 9 There are a few items in Sicilian Doric Greek that seem to match Latin and that are suspected of being of Sicel origin, e.g. lítra ~ Lat. lībra ‘pound’, kúbiton ~ Lat. cubitum ‘elbow’. The few inscriptions that are longer may show some Italic lexical material such as Mendolito geped ‘had’Footnote 10 with a simple perfect comparable to Osc. hipid, Grammichele dedaxed ‘made’ (?) (see Reference MachajdíkováMachajdíková 2018: 151), perhaps with a reduplicated k-extended form of the root *dʰeh1- like VOL fhe:fhaked, Osc. fefacid, or the female name Kup(a)ra, which recalls Sabellic *kupro- ‘good’. If Sicel is Italic, it would diverge from all other members in showing voiced reflexes of the PIE voiced aspirates in initial position in contrast to the f < *dʰ, * and h- < */*g̑ʰ seen in Latin, Sabellic, and Venetic.

Figure 8.1 The Italic languages

8.2 Evidence for the Italic Branch

Positing Proto-Italic as the superordinate node of Latin, Venetic, and Sabellic is not uncontroversial, though it is supported by substantial phonological and morphological evidence: the merger of *bʰ- and *dʰ- as *f-,Footnote 11 the gerundive in *-nd-, the ipf. subj. *-sē-, the ipf. *-βā- (the more probative morphological features are unattested in the fragmentary Venetic corpus). Proto-Italic was recognized as a node from the start of the serious scientific investigation of the Indo-European languages. But some scholars beginning with Reference WaldeWalde (1917), Reference MullerMuller (1926), and Reference DevotoDevoto (1929) have challenged this assumption and argued instead that Italic and Sabellic are two separate branches that have undergone a secondary process of convergence.Footnote 12

And indeed, there is no doubt that much convergence has happened between Latin and Sabellic. For example, the change of intervocalic *-z- to -r-, called rhotacism, affects both Latin and Umbrian but not Oscan and can be shown to have happened long after the initial separation of both languages. In Latin the change happened sometime in the fourth century, and the Umbrian change may have happened around the same time. Initial di̯- is eventually simplified in Latin, Umbrian, and Oscan to i̯- (except in Bantia), but again these changes happened within the historical record for Oscan and Latin at least. Both Latin and Sabellic show deletion of the final primary marker *-i in the 1sg., 2sg., 3sg., and 3pl. (Reference MeiserMeiser 1998: 74 after Reference RixRix 1996: 158). But the survival of an unapocopated final -i in tremonti in the Carmen Saliare makes it unlikely that this apocope dates to Proto-Italic times. The Carmen Saliare is old, but not that old, and the text has a specifically Latin form of the acc. pronoun tet and so could not be “Proto-Italic”. Instead, the apocope must be a diffused change.Footnote 13

But while it is easy to show some degree of phonological convergence and, of course, lexical interchange and syntactic influence within the historical period, I know of no case of a Sabellic morpheme being adopted into Latin or vice versa. We have no v-perfects in Sabellic, no -tt-perfects in Latin, no Latin infinitives in -om, no Latin athematic nom.pl. in -s, and so on. This difference between phonological, lexical, and syntactic permeability vs. morphological impermeability is not surprising: morphology is known to be more resistant to borrowing, but the absence of morphological borrowing within the attested timeframe, in a period when the Sabellic and Latin languages were in intense contact, should strengthen our confidence in the value of shared morphology for establishing the Proto-Italic subgroup.

There are a number of shared phonological developments unique to the Italic languages that cannot be shown to be the result of convergence and thus are good candidates for defining innovations of Proto-Italic. The difficult issue is deciding whether they are non-trivial. First on this list is the development of the PIE voiced aspirates. In initial position PIE * and * developed to f and *g̑ʰ/gʰ to h in Latin, Sabellic, and Venetic: *bʰuhₓ- > Lat. fu-ī ‘I was’, Osc. fu-st, Umb. fu-st ‘will be’; Transponat *dʰh1k- ‘make’ > Lat. faciō, Osc. fakiiad, Umb. façia subj.3sg., Ven. vhagsto pret.3sg.; *g̑ʰorto- ‘enclosure’ > Lat. hortus ‘garden’, Osc. húrz; Ven. hosti- < *gʰosti- ‘guest’. In medial position, on the other hand, the voiced aspirates became voiced fricatives, and the labial and dental fricatives were not merged:Footnote 14 Lat. nebula ‘cloud’ < *neβVla < *nebʰVleh2, cf. Ved. nábhas- ‘cloud’; aedēs ‘temple’ < *h1ai̯dʰ- ‘burn’, cf. Gr. αɩ҆θόμενος ‘burning’; Lat. Samnium Osc. safinim by anaptyxis from *safnim and Gr. Σαύνιον point to a Sabellic *saβnii̯o-;Footnote 15 Ven. louderobos ‘children’ dat.pl. ← *h1leu̯dʰerobʰos. The velar fricative went on to become h everywhere except Faliscan where it hardened to g: *meg̑ʰei̯ ‘me’ dat.sg. > Lat. mihī, Umb. mehe, *legʰeti > Fal. lecet /leget/ ‘lies’, cf. Goth. ligan.Footnote 16 The combination of devoicing, merger of * and * in f in initial position and voicing in medial position (whether this voicing directly continues the voicing of the voiced aspirates or is a revoicing) is a set of developments that is found in no other Indo-European branch.Footnote 17

A set of sound changes that are certainly shared by Sabellic and Latin is the absorption of a short vowel after yod in a medial syllable. This sound change led to the creation of the 3rd io-type in Latin and its analog in Sabellic, e.g. *kapi̯esi ‘you (sg.) take’ > *kapisi > capis. Cf. Osc. factud (Lu 1.9) fut.ipv.sg. ‘make’ < *fakitōd < *faki̯etōd. After a base of more than one mora, there was an epenthesis of i before the yod prior to absorption of the e: *sent-i̯esi > *sent-ii̯e-si ‘you (sg.) feel’ > *sent-īsi > sent-īs. Cf. Umb. amparitu fut.ipv.sg. ‘raise’ < *am-par-ii̯e-tōd.Footnote 18 The sound changes that produced this system appear to be quite early since they predate the resolution of syllabic sonorant consonants (see Reference Fortson, Hackstein and OpfermannFortson 2018), but, on the other hand, these sound changes appear to be distinct from similar changes in Germanic and Celtic.Footnote 19

Another interesting phonological development is the outcome of *m̥mV. This sequence first arose by the loss of a laryngeal or by Lindeman’s Law. It is also found in the ordinal and superlative suffix *-m̥mo-, which is of uncertain analysis. In Latin and Venetic the supporting vowel is o, e.g. *g̑ʰm̥mō > Lat. homō ‘man’, cf. Goth. guma, *dek̑m̥mos > Ven. dekomei loc.sg., cf. Celtib. tekametam ‘tenth’. In Sabellic the outcome appears to have originally been u. The best evidence for u is Osc. últiumam ‘last’, Palaeo-Umbrian setums (a personal name, lit. ‘seventh’) < *septm̥mos, the Pre-Samnite superlative ϝολαισυμος ‘best’ nom.pl.Footnote 20 It’s not certain what the Proto-Italic state was. We can certainly exclude *um since that would not be lowered to om in Latin, cf. tumor ‘swollen condition’, gumia ‘glutton’. It’s conceivable that *om would have been raised to um in a medial syllable in Sabellic, but there is no independent evidence for such a change. Rather than miss the generalization that the Italic languages uniquely have a rounded vowel as the reflex of prevocalic * which contrasts with its development in preconsonantal position, it may be preferable to reconstruct *ɵm with a close mid central unrounded vowel that merged with either o (in Latin and Venetic) or u in Sabellic.Footnote 21 There are a number of other phonological features that could be mentioned, but they are all problematic in one way or another.Footnote 22 The phonological innovations are admittedly not many, but they are indicative of a subgroup.

It is the shared morphological innovations, which, in my opinion and the opinion of most experts, make the existence of a Proto-Italic unavoidable. The Italic languages share a new verbal adjective, the gerundive, with the suffix -ndo- in Latin and *-nno- in Sabellic (Osc. úpsannam ‘to be constructed’, Umb. ocrer pihaner ‘to purify the city’). The origin of this form and the synchronically related gerund, not attested in Sabellic, are much debated. The original function of the form seems to have been quite similar to a middle participle as we can see in synchronically isolated cases such as Lat. secundus ‘following’ ~ sequor ‘I follow’, oriundus ‘arising’ ~ orior ‘I rise’, but the semantic development to a verbal adjective of necessity is found in both Sabellic and Latin. Whatever its origin, the gerundive has no analogs outside of Latin and Sabellic.

The imperfect subjunctive in *-sē-, e.g. Osc. fusíd = Lat. ‘foret’, Lat. es-sē-s ‘be’ ipf.subj.2sg., is another morpheme of disputed origin.Footnote 23 It does not have any comparanda outside of Italic.Footnote 24 The category is not attested in Umbrian or Venetic. But beyond the existence of an identical morpheme for this category, it is worth noting that a subjunctive system with present, imperfect, perfect, and presumably pluperfect is a uniquely Italic way of organizing the verb.

Another Italic-only verbal exponent is the imperfect indicative morpheme *-βā- (Lat. dūcēbās ‘you were leading’, Osc. fufans ‘they were’, Vest. profafa- i.e. = Lat. probābā- ‘was approving’).Footnote 25 It is generally agreed that *-βā- is a form of the root *bʰuhₓ- ‘be’ combined with the morpheme *ā < *eh2 also seen appended to the root *h1es- in the unique imperfect stem seen in 2sg. erā-s ‘you were’.Footnote 26 The combination of this extended root shape with a nominal form of the verb, probably originally an instrumental, is only found in Italic. Again, the corresponding forms, if any, in Umbrian and Venetic are unknown.

Another shared Italic innovation is the replacement of the Proto-Indo-European 2nd plural middle ending *-dʰu̯e with a form containing *-m: Lat. -minī, Sabellic * -mX inferable from fut.ipv.mid. ending Umb. -mu and Osc. -mur < Proto-Sabellic * -mōr.

Ideally, it would be preferable to derive these forms from the same proto-form. If, as most scholars believe, the Latin mid.2pl. ending continues the nom.pl. of a middle participle (< PIE *-mh1no-), the Proto-Italic form, if there was one, would have been *‑manōs with the inherited thematic nom.pl. ending. In Latin the analogy that produced the mid.ipv.3sg. was act.ipv.2pl. -te : fut.ipv.act.3sg. -tōd : mid.ipv.2pl. *-manoi̯ : fut.ipv.mid.3sg. *-manōd > -minō. That is, the acquirer got the idea that the fut.mid.ipv. was formed by hacking off the final syllable of the act.ipv.2pl. and substituting -ōd. But this is not the only way an acquirer could have conceived of the relation. Alternatively the “rule” could have been “remove all material but the initial consonant and substitute with -ōd,” i.e. -t-e : -t-ōd :: -*m-anōs : *-m-ōd. This path seems the only way to unify the Italic forms and retain the observation that both Sabellic and Latin have replaced the inherited mid.2pl. ending with a form beginning with *-m.Footnote 27

There are a few other features shared between Latin and Sabellic, such as the use of the interrogative-indefinite stem as a relative pronoun (Umb. po-i = Lat. quī). But this is a common development and occurred independently in Hittite, Thessalian, and elsewhere. Another oft-cited commonality is the creation of a distinct ablative singular form for all declensions,Footnote 28 e.g. Osc. toutad ‘community’ from an ā-stem. However, Celtiberian also created distinct ablatives for other stem types, e.g. ā-stem arekorataz (the name of a town attested on a coin, i.e. ‘from the town of A.’).Footnote 29 Thus this innovation could have happened (1) independently in Latin, Sabellic, and Celtiberian, (2) in Proto-Italic and Celtiberian, or (3) in Proto-Italo-Celtic.Footnote 30

The realm of derivational morphology, which is typically underexploited in discussions of subgrouping, also displays a number of striking shared Italic innovations. For example, the suffixes *-āsii̯o- (Umb. farariur ‘pertaining to grain’ nom.pl.m. = Lat. farrārius), and *-āli- ~ dissimilated to -āri- after a base containing an l (Umb. sorsale ‘of pig’ staflare ‘of the stall’ ~ Lat. mortālis ‘mortal’, mīlitāris ‘miltary’) are exclusively Italic.Footnote 31 Both Latin and Sabellic have specialized the conglomerate *-kelo- to form diminutives to nonthematic bases (Osc. zicolom Umb. tiçel ‘day, date’ ~ Lat diēcula). Both Latin and Sabellic have a predominantly deverbal adjectival formant *-dʰli- (Lat. amābilis ‘lovable’, Umb. purtifele ‘to be offered’).

The shared lexical material of Italic is extensive. Reference SafarewiczSafarewicz (1963) estimated that, with obvious loanwords excluded, 49 percent of the Oscan vocabulary known to him had exact matches in Latin.Footnote 32 Of course, a true doubter of Italic unity could claim that this high percentage results from borrowing. But there are several items where semantic divergences make recent borrowing unlikely. For example, Lat. aut means ‘or’ but in Osc. avt means ‘but’. Lat. enim means ‘then’, but Osc. íním means ‘and’. We may also note some interesting specializations of meaning and/or form not found outside of Italic: Latin, Sabellic, and Venetic all have the stem *diē- generalized from the Stang’s Law outcome of *di̯eu̯m as the word for ‘day’: Lat. diēs, Ven. loc. diei, Osc. zicolom, Umb. tiçel < *di̯ēkelos. Though the accusative form was obviously inherited (Ved. dyā́m, Phryg. Τιαν, Gr. Ζῆν(a)), it is only in the Italic languages that a new paradigm has been formed specifically in this meaning. Only in Latin and Sabellic does *h2eh1seh2 mean ‘altar’ (Lat. āra, Umb. asa abl.sg., Osc. aasaí loc.sg.). The Hittite cognate ḫaššāš means ‘hearth’ and Ved. ā́sa- m. means ‘ashes’. Latin, Sabellic, and Venetic have created a neo-root *dʰeh1k- ~ dʰh1k- from a k-extended stem originally at home in the active singular of the aorist (Lat. facere ‘to make’, Umb. façia, Osc. fakiiad pres.subj.3sg., Ven. vhagsto pret.3sg., vs. Gr. ἔθηκε, Phryg. addaket aor.act.3sg. vs. ἔθετο aor.mid.3sg.).Footnote 33

All in all, the phonological, morphological, and lexical innovations shared between Latin, Sabellic, and Venetic (when available) are too numerous and integrated to be the result of secondary approximation alone. At the same time, there are quite appreciable differences between the Italic languages. Given the fragmentary state of the Italic languages other than Latin, it is hard to know exactly how different Latin and Sabellic were in, for example, the second century BCE. In my opinion they were synchronically much less closely related or mutually intelligible than the old Germanic languages but more closely related than Old Irish and Middle Welsh or Lithuanian and OCS. From this we infer that there must have been quite a long period of divergence before the forms of Italic began to converge again in the historic period. Whether this inference can be made to correlate with any plausible archaeological or genetic scenario is an open question.

Two final points on the question of Proto-Italic: What would prove that the Sabellic languages and Latin do not belong to the same subgroup? Well, imagine, as a thought experiment, that a scholar claimed that Sabellic and Germanic formed a subgroup within Indo-European. This would be refuted to most people’s satisfaction by pointing out that Sabellic and Germanic share no innovations that (1) are not shared with other groups as well and (2) precede the earliest separate innovations in these two groups. If our imaginary proponent of Sabello-Germanic retorted that Proto-Sabello-Germanic should be reconstructed at a stage that could account for both sets of developments, we would respond that such a reconstructed state would be virtually identical to Nuclear Proto-Indo-European and therefore any Nuclear Indo-European language would be derivable from it.

If we return now to reality, can the proponents of the independence of Sabellic and Latin point to any innovations in either language group that make it impossible to derive the other branch from any other common ancestor than the proto-language? There are a few cases where the differing outcomes of the two groups lead to the reconstruction of the PIE state of affairs (e.g. syllabic nasals, labiovelars), but these must be weighed against the instances where this is not the case. Finally, if, on theoretical grounds, we believe that only binary branching is possible,Footnote 34 denying an Italic subgrouping immediately raises the question of what Sabellic or Latin should be grouped with instead. And when we put the question in those terms, it becomes clear that there is no other branch that could be more closely grouped with either Latin or Sabellic. And thus one would be forced to the position that, despite the evident shared innovations, Latin or Sabellic is more closely related to some language of which there is no trace.

8.3 The Internal Structure of Italic

There are actually fairly few innovations on the Latin side that can be shown to encompass all the Latin dialects and to have taken place before the onset of the historical record. In many cases the fragmentary dialects don’t preserve the trait in question. For example, the productive v-perfect formant is attested in Roman Latin (earliest epigraphical example probavet ‘approved’ from the Egadi rostra dated before 241 BCE; see Reference PragPrag 2014) and in Praenestine (cailavit ‘chiseled’) but not in Faliscan. This is presumably an accidental gap. The scant corpus of Faliscan does not preserve any alternative morpheme in the same functional slot. In some cases what would appear to be a defining innovation of Proto-Latin can be seen not to have affected all the dialects or to have been a later diffused change. Roman Latin, for example, has changed medial *β and *đ to stops b and d, and *γ to h, whereas Faliscan has f for the reflex of medial *β and *đ and apparently a stop g as the reflex of *γ (Fal. lecet ‘lies’ < *legʰeti). Thus Proto-Latin must have had fricatives *β, *đ, and *γ, and not *b, *d, and *h as Roman Latin. With these cautions in mind, we may point to the following Latin innovations, which are assumed to be Proto-Latin in the absence of evidence to the contrary.

As far as phonology is concerned, few secure innovations define the Latin node; most prehistoric phonological innovations are on the Sabellic side. One Latin innovative feature is the shortening of long vowels before final -m, which did not happen in Proto-Italic since South Picene and Oscan preserve distinct reflexes of a long vowel in the genitive plural in *-ōm and Oscan retained a long vowel in this environment at least in monosyllables (Reference Weiss, Jay H. Jasanoff, Melchert and OliverWeiss 1998; Reference ZairZair 2016: 82). There was eventually such a shortening in polysyllables in the Sabellic languages, but this is an independent change. Another phonological innovation on the Latin side is the contraction of some unlike vowel sequences after the loss of intervocalic yod. Whereas like vowels contracted already in Proto-Italic (e.g. the iterative-causative suffix *-ei̯e- > *-ee- > *-ē- Lat. ē, PSab. *ẹ̄), the sequence *āi̯ō contracts in Latin to -ō but remains uncontracted in Sabellic (Lat. vocō ‘I call’ vs. Umb. subocau ‘I invoke’ < *subu̯okāi̯ō). Likewise, the sequence *-āi̯ē- contracts to ē in Latin but remains uncontracted or perhaps diphthongizes in Sabellic (amēs ‘love’ subj.2sg. < *amāi̯ēs, vs. Osc. deivaid ‘may he swear’ < *dei̯u̯āi̯ēd).Footnote 35

There are other cases where both Latin and Sabellic have innovated in different ways from the Proto-Italic situation. For example, Latin has eliminated the nasal in the sequence *-Vns#, e.g. in the acc.pl. (VOL deivos ‘gods’ acc.pl.). The same treatment is found in Venetic (Ven. deivos ‘gods’ acc.pl.). In Sabellic, on the other hand, the development of final *-ns and *-nts is to -f (Umb. vitluf ‘calves’ acc.pl., traf ‘across’ ~ Lat. trāns < *trānts). The sequence that gave -nd- in the Latin gerund(ive) gave -nn- in the corresponding Sabellic morpheme (Osc. úpsannam ‘to be constructed’; Umb. pihaner /pianner/ ‘to be purified” = Lat. piandī).

Sabellic, on the other hand, has a number of distinctive phonological innovations. The most salient is the change of the voiced and voiceless labiovelars to labial stops, e.g. *kʷis ‘who’ > Osc. pis, Umb. pis-i = Lat. quis; *gʷih3u̯o- ‘alive’ > Osc. bivus ~ Lat. vīvī, *gʷem- ‘come’ > Umb. benust fut.perf.3sg. ~ Lat. vēnerit.Footnote 36 Venetic agrees with Latin in preserving the voiceless labiovelar and turning * into w (kve ‘and’, vivoi ‘alive’ dat.sg.). Another distinctive feature of Sabellic is across-the-board syncope of a short vowel before a final -s (Osc. húrz ‘garden’ = Lat. hortus, Osc. bantins ‘from Bantia’ < *bantīnos). In Latin and Venetic this type of syncope is more limited and occurs chiefly after r (Lat. sacer ‘sacred’ < *sakros,Footnote 37 Ven. teuters ‘public’ < *teu̯teros), but not elsewhere. In Sabellic, stops were lenited to fricatives before a dental stop, so *pt > ft, *kt > ht (Osc. scriftas ‘written’ nom.pl.f. ~ Lat. scrīptae, úhtavis ‘Octavius’). In Umbrian *ft > ht (screihtor ‘written’ nom.pl.n.). The voiced labial and dental fricatives -β- and -đ-, which occurred in medial position as the reflexes of the PIE voiced aspirates, merged as β ⟨f⟩ (Osc. mefiaí ‘in media’, Umb. rufru ‘rubrōs’). In Latin and Venetic these are kept separate, ultimately becoming stops in Latin and probably eventually in Venetic too (mediai ‘middle’ loc.sg. from *medʰi̯o/ā-, louderobos ‘children’ dat.pl. from *h1leu̯dʰeros, cf. Gr. ἐλεύθερος ‘free’). In initial syllables, syllabic nasals developed to *aM in Sabellic but to *eM in Latin (Osc. fangvam ‘tongue’ acc.sg. < *dʰn̥g̑(ʰ)u̯ā-, an- neg. < *n̥- vs. OLat. dingua < *dn̥g̑ʰu̯ā, in- < en-). But elsewhere the development is to en as in Latin (Umb. desen-duf ‘twelve’, cf. Lat. decem ‘ten’). In Venetic the outcome is -an at least in final syllables (donasan ‘they gave’ < *-sn̥t).

Turning to morphology, the innovations are more evenly distributed between Latin and Sabellic.Footnote 38 Oscan and very probably Umbrian have remade the inherited nom.sg. of n-stem nouns with -ō as the final-syllable vowel by introducing the n from the oblique stem and recharacterizing the nominative with -s. The resulting sequence gave -f, e.g. Osc. úíttiuf ‘use’ < *oi̯ti̯ōn-s vs. Lat. -iō, -iōnis. In the -eh2-stems, Sabellic retains a contrast between a reflex of *-ā < *-eh2 in the nom.sg. (Osc. víú, Umb. Turso [name of a goddess]) and *-ă in the vocative (Umb. Tursa) whereas Latin has a surprising and not satisfactorily explained -ă in both the nom. and voc.sg. In Sabellic, the proterokinetic i-stem gen.sg. ending *-ei̯s was generalized to the o-stems and consonant stems (medíkeís ‘meddix’ gen.sg.). In VOL o-stems retain -osi̯o, as in Valesiosio ‘of Valesios’ (Hom. Gr. -οιο, Ved. ‑asya, etc.) beside -ī, which eventually replaces -osi̯o, and consonant stems retain -es and -os. But note that in the case of *-ei̯s, Sabellic preserves an ending eliminated by Latin. Sabellic also gets rid of the athematic accusative singular ending *-em < *-, replacing it with thematic *-om. But, on the other hand, Sabellic retains the athematic nom.pl. *-es, which is syncopated to -s (Marruc. medix ‘medix’ nom.pl. < *medikes), whereas Latin has eliminated this ending in favor of -ēs < *-ei̯es originally from the -i-stems. Sabellic also retains the thematic nom.pl. *-ōs and the a-stem nom.pl. *-ās, which Latin has replaced with pronominal *-oi̯ > -ī and analogical -ai > -ae, just as in many other Indo-European languages.Footnote 39 Sabellic has even extended the thematic nom.pl. nominal ending to the nonpersonal pronouns (Osc. pús ‘who’ nom.pl.m. vs. Lat. quī). The neut.pl. in Sabellic has generalized the thematic ending -ā < *-eh2 to athematic forms (Umb. triiuper ‘three times’), but in Latin the generalization has gone the other way with -a < *-h2 in all paradigms.

In pronominal morphology, Latin has extended the accusatives of the singular personal pronouns by -(V)d (VOL, Fal., Praen. med) whereas Sabellic has used the particle *‑om (OUmb. míom). Sabellic retains the oblique stem formant -sm- in the anaphoric and relative-interrogative stem (SPic. esmín loc.sg., Umb. esmik dat.sg., pusme ‘to whom’), which Latin has replaced (istī eiiei, cui, quō). Sabellic has an innovative oblique stem of the anaphoric pronoun *ei̯s- created by reanalysis of the genitive plural *ei̯sōm, e.g. dat.pl. Osc. eizois, Umb. erer-unt. Sabellic has a unique proximal deictic stem *eko-/ekso-. In Oscan these stems are suppletive, with *eko- forming the nom.-acc. and *ekso- forming the oblique stem. Umb. has a unitary stem *esso- < *ekso-, which may be the older situation. The corresponding Latin proximal deictic is made from a stem *ho- (Lat. hic, Fal. hec ‘here’), which may be continued in the Umb. pronominal form erihont ‘the same’ nom.sg.m.Footnote 40

In the personal ending of the verb Latin has generalized the thematic 3rd person plural *‑ont(i) to athematic forms (sont ‘they are’; exception: opt.3pl. sient) whereas Sabellic has extended the range of the ending -ent (Osc. fiie(n)t), though -ont does survive in Pre-Samnite fυfϝοδ ‘they were’. In the primary 2nd plural SPic. has -tas (videtas ‘you see’), which must be from *-tās since *-tas would have syncopated to *-ts. But Latin has an incompatible -tis. Most probably, Proto-Italic had primary 2du. *-tas, 2pl. *-tes;Footnote 41 secondary 2du. *-, 2pl. *-te. Latin would have generalized the plural endings while Sabellic generalized the dual endings and leveled the *ā from the 2du. secondary to the primary dual ending. In 3rd person middle endings, Umbrian appears to have preserved the PIt. situation with primary -ter and -nter < *-tro and *-ntro contrasting with secondary endings in *-tor, -ntor (Umb. terkantur ‘let them see’). Oscan has generalized the primary endings, and Latin has generalized the secondary endings. Sabellic also retains t-less mid.3sg. forms (Umb. ier ‘one goes’, ferar ‘one should carry’, Osc. loufir ‘or’, lit. ‘(if) it is wished’), which Latin has eliminated without a trace. The mid.2pl. is not attested in Sabellic but is partially inferable on the basis of the deponent future ipv. ending *‑ (Umb. persnimu ‘pray’, Osc. censamur) which must have been created like the corresponding Lat. -minō on the basis of the 2pl. middle ending. This form therefore began with *m-. In the endings of the perfect system, Latin has for the most part preferred the endings originating in the PIE perfect (1sg. -ai, 2sg. *-istai, 3sg. -eit, 3pl. -ēre, all of perfect origin), but has also incorporated some originally aorist endings (3sg. -ed, 3pl. -ēr-ont, and -(er)ont ← *‑ond, cf. Fal. fifiqo(n)d). Sabellic has only aoristic endings in the forms we know: 1sg. ‑om, 3pl. -ens, Pre-Samn. -ο(ν)δ. The ending -e of the perf.2pl. form Pael. lexe ‘you (pl.) have read’ is sometimes compared to Ved. perf.2pl. -a but, given the overall aoristic provenance of the perfect endings of Sabellic, this is improbable. The ending may ultimately be from *(-s)te.

When we turn to tense, aspect, and mood, Latin has cobbled together a future tense out of (1) an original periphrastic with *bʰuhₓ-, which gives the b/f-future (Fal. carefo ‘I will lack’, Lat. carēbō ‘I will lack’), and (2) the PIE subjunctive (athematic erō ‘I will be’, thematic dūcēs ‘you will lead’). Sabellic has a probable trace of the thematic subjunctive formant ē in SPic. knúskem ‘know’ 1sg., but it is difficult to say whether this is used in subjunctive or future function. Otherwise, Sabellic has generalized the athematic s-future to all stem types. Latin once had comparable forms, but only the PIE subjunctive and optative of this type survive in the mainly OLat. faxo (fut.), faxim (subj.) type. In addition, the Latin future perfect and perfect subjunctive (fēcerō, fēcerim) employ this same morpheme after a union vowel /i/ appended to the perfect stem. Sabellic forms the future perfect with the same athematic s-morpheme added to a union vowel ‑u‑,Footnote 42 but the perfect subjunctive does not use the s-morpheme at all, instead adding -ē- to the perfect stem (Osc. tríbaraka-tt-í-ns). This ē is presumably the same as the subjunctive formant ē used in the present stem.

The stem formation of the perfect is very divergent between Latin and Sabellic.Footnote 43 While both Latin and Sabellic continue reduplicated perfects, simple perfects (perhaps thematic aorists or dereduplicated perfects), and lengthened-grade perfects, Latin has utilized the s-aorist formant much more than Sabellic, which does not have a single completely certain example. Venetic, on the other hand, seems to have generalized the s-aorist as the productive perfect formant (donasto ‘gave’ 3sg., donasan 3pl., vhagsto ‘made’ 3sg.). Latin has greatly expanded the range of the v-perfect, of which Sabellic has no trace. Sabellic has a number of innovative perfect formants mostly of quite unsettled origin. These are the -tt-perfect and the supposed -k-perfect of Oscan, the -nki̯- perfect of Umbrian, and the -ō-perfect of South Picene.Footnote 44

In nonfinite morphology, Latin and Sabellic have made different choices in grammaticalizing different case forms of different stem types as an infinitive. Latin makes use of the locative of an s-stem (dūcere ‘to lead’ < *deu̯kesi) whereas Sabellic used the accusative in -om (Osc. tríbarakavúm ‘to build’), which might originate in either a thematic or a root noun (since Sabellic replaces *-em with -om). For the medio-passive, Sabellic retains the form *-fiē or *-fiēi̯ (Umb. cehefi ‘to be taken’, Osc. sakrafír ‘to be sanctified’), which is an instrumental or dative of the same piece of nominal morphology that gave Ved. -dhyāi and Toch.B, Toch.A -tsi. Latin has perhaps redone the expected cognate *-diē as -rier to create its passive infinitive (see Reference FortsonFortson 2012; Reference Fortson, Cooper, Rau and Weiss2013).

Nominal derivational morphology is overall quite similar between Latin and Sabellic. Most suffixes found in one branch are also found in the other in more or less the same function. One notable difference is that Sabellic has no hesitation in adding the suffix ‑ii̯o- to a base in -ii̯o-. This is the origin of the Sabellic (mainly attested in Oscan) gentilics in -iís -ies (statiis < *stātii̯‑ii̯‑os derived from the praenomen staatis < *stātii̯-os). Latin has no trace of such iterative derivation and prefers formations like Lūcīlius and Mānīlius from Lūcius and Mānius or Lātīnus from Lātium. An interesting mismatch between Latin and Sabellic is shown by Lat. fānum ‘shrine’ < *fasnom vs. Osc. fíísnu nom.sg.f., where Latin continues a zero-grade of the root *dʰh1s- and Sabellic reflects a full-grade *dʰeh1s‑. Since ablaut in a -no- or -nā-stem is unlikely, it is possible that the derivational base showed ablaut or that one or the other forms may have been remade on the basis of related elements.

Defining the distinctive lexicon of Latin vs. Sabellic is challenging given the incommensurate sizes and natures of the corpora. For example, even if we combine all Sabellic languages, we still know fewer than eighty of the 200 items on the Swadesh list. It is often impossible to determine whether Latin and Sabellic agree on any particular lexical innovation. Nevertheless, something can be said about the distinctive lexical profile of Sabellic (see Reference BuckBuck 1928: 12–17). There are a small number of cases where Sabellic retains a form of a root or particle that has been completely eliminated from Latin.Footnote 45 Sabellic, like Vedic, has reflexes of both *u̯ih1ro- (Umb. ueiro acc.pl.n.) and *h2nēr (SPic. níír, Umb. nerf acc.pl. etc.) in the meaning ‘man’ while Latin has eliminated the latter.Footnote 46 The Indo-European word for ‘daughter’ is preserved in Osc. futír while Latin has replaced it with the neologism fīlia. Likewise, an old word *putlo- ‘son’ survives in Osc. puklum acc.sg., cf. Ved. putrá- in contrast to Lat. fīlius. Umbrian preserves the r/n-stem for ‘water’ (utur acc.sg., une abl.sg.) while Latin only continues the derivative unda ‘wave’Footnote 47 and has replaced the basic word with a West IE neologism *akʷā > Lat. aqua.Footnote 48 Sabellic has a word for ‘god’ *ai̯so- (Osc. aisús nom.pl.) that it shares with Venetic (aisus), to the exclusion of Latin. Interesting are the divergent prepositional/preverb forms: Osc. aa-, Umb. aha- ‘to’ (OHG uo-) with no analog in Latin; Osc. eh, Umb. ehe ‘from’ < *eg̑ʰ, which, like Lith. , preserves an s-less form of this particle whereas Latin has only ex and its further developments; and *dād ‘from’ (Osc. dat, Umb. da-), which has no analog outside of Sabellic at all. Some other Indo-European roots and words are preserved uniquely in the Sabellic branch: *ad- ‘law’ (Umb. arsmo ‘rites’, OIr. ad ‘law’, ada ‘fitting’), Umb. e-iscurent ‘seek’ fut.perf.3pl. (Ved. iccháti),Footnote 49 Osc. cadeis ‘hostility’ (OHG haz ‘hate’), Osc. mais ‘more’ (Goth. mais ‘more’), *nertero- ‘left’ (Umb. nertru, Osc. nertrak ~ ON norþr ‘north’), Osc. nessimas nom.pl.f. ‘nearest’ (OIr. nessam ‘nearest’), Umb. pir ‘fire’ (Gr. πῦρ, but possibly in Lat. pūrgāre ‘to purify’), Umb. terkantur ‘look’ subj.3pl. (Gr. δέρκομαι), Osc. touta ‘people’ (OIr. túath, Goth. þiuda, etc.). There appears to be no particular pattern to these items. Some have matches only in the Northern European IE languages (*ado-, *tou̯tā). Most have widespread cognates.

The syntax of Sabellic and Latin are very similar, but this may be partly the result of the generic similarities of epigraphical documents from Central Italy. The use of the locative case in Latin has been greatly curtailed, but the Sabellic languages continue to use it freely: Osc. eiseí tereí ‘in this territory’, comenei ‘in the assembly’.

8.4 The Relationship of Italic to the Other Branches

Italic, unsurprisingly, shares many features with other European branches of Indo-European. Reference MeilletMeillet (1922) famously recognized a “civilization of the Northwest”, which was shared by the branches that eventually became Balto-Slavic, Germanic, Italic, and Celtic.Footnote 50

One occasionally sees the claim that Italic has an especially close relationship with Germanic (most recently Reference Kuz’menkoKuz’menko 2011), but there are in my view no innovations in phonology or inflectional morphology shared exclusively by Italic and Germanic.Footnote 51 Most of the innovations listed by proponents of Italo-Germanic such as Reference HirtHirt 1897 or Reference DevotoDevoto 1936 are not correct, are attested elsewhere, or are suspect of being parallel developments.Footnote 52

That said, there are still a number of unique Italo-Germanic agreements in lexicon. Only in Germanic and Italic is the suffix *-no- added to the multiplicative adverb *du̯is ‘twice’ to create *du̯isnó- ‘double’, a proto-form continued in the Latin distributive numeral bīnī ‘two at a time’ and in Germanic *tu̯izná- (ON tvennr ‘two-fold’, OE ge-twinn ‘twin’, OHG zwirnōn ‘to twine’). But the addition of the suffix *-no- to adverbial forms is well known elsewhere in Indo-European, cf. Ved. purāṇá- ‘ancient’ ← purā́ ‘of old’. Lat. vadum ~ ON vað n. ‘ford’, OHG wat, OE wæd ‘sea’ is a perfect and exclusive match. Assuming Germanic *ga- really is cognate with Latin com-, the match between Lat. commūnis and Goth. gamains etc. is striking. An innovative word for ‘year’, Lat. annus, Osc. akeneí loc. sg., and Goth. aþna-, is added to PIE *h1ieh1, which survived in both Italic (hornus ‘this year’ < *ho-i̯Vrno-) and Germanic (Goth. jer), but this word is probably closely related to the Celtic words for ‘time’ OIr. am, Gaul. amman (see Reference StifterStifter 2017: 220–2). The words for ‘barley’, OHG gersta and Lat. hordeum, reflect two different genitival derivatives of a *gʰr̥sdo- ‘prickly plant’ (OE gorst). Italic and Germanic share not one but two exclusive words for ‘be silent’: (1) Lat. tacēre, Umb. taçez, Goth. þahan, ON þegja, OHG dagēn and (2) Lat. silēre, Goth. anasilan ‘to grow calm’. Both verbs have been compared more widely (LIV² 495 explains tacēre ~ þahan as a semantic specialization of *pteh2k- ‘to cower’; others prefer to connect these forms with OIr. tachtaid, MW tagu ‘strangle’ < ‘silence’, and silēre has been compared to OIr. silim ‘pour’ and more generally to the root *seh1-i- ‘let’), but even if these somewhat difficult comparisons are correct, the close formal and semantic match between Germanic and Italic remains.

A word needs to be said about two matches between Venetic (considered here as an Italic language) and Germanic. The Venetic accusative of ego ‘I’ is mego, and this has been compared to the Germanic accusative reflected by Goth. mik. Given the different extensions of the accusative stem of the 1sg.personal pronoun seen in Latin (OLat. mēd), Sabellic (OUmb. míom), and Venetic, it seems that Proto-Italic must have actually retained an unextended accusative * or *me.Footnote 53 This would mean that mego would have resulted from a secondary influence of Germanic, but given the fact that mego can be explained as an inner-Venetic conflation of *egō and *me, it’s preferable to leave the Germanic and Venetic forms unconnected.Footnote 54 The second Venetic-Germanic isogloss, Venetic sselboisselboi ‘for himself’ ~ Goth. silba ‘self’, OE seolf, OHG selbselbo, ON sjálfr is unquestionable and not likely to be coincidental. But it should be noted that the Venetic form occurs on one of the latest Venetic inscriptions, written in the Roman alphabet. It’s possible that this is the result of a quite recent and perhaps one-off borrowing from Germanic.

Although there have been the occasional attempts to connect Italic more closely with one or another branch of Indo-European, these are mainly of historical interest. In the early days of Indo-European, some scholars posited a Greco-Italic group, e.g. Georg Reference CurtiusCurtius (1858: 22).Footnote 55 In both Greek and Italic the voiced aspirates end up as voiceless segments at least in word-initial position, but the behavior of medial voiced aspirates is quite different.Footnote 56 Greek and Italic share an innovative nominal genitive plural for *eh2-stems *‑āsōm (Lat. -ārum, Osc. -azum, Hom. Aeol. Gr. -ᾱ́ων), but this is an introduction of the pronominal form (cf. Ved. tā́sām ‘of these’) into the nominal paradigm, which could have happened independently. An interesting agreement in derivational morphology is the extension of the stem *dek̑s(i)- ‘on the right’ by the oppositional suffix *-teros (Lat. dexter, Umb. destram, Gr. δεξιτερός) in contrast to *dek̑s(i)-u̯os (also in Gr. δεξι(ϝ)ός, Gaul. dexiva, Goth. taihswo ‘the right hand’) or *dek̑s(i)-no- (Ved. dákṣiṇa-, YAv. dašina-, Lith. dẽšinas, OCS desnъ). The antonym σκαι(ϝ)ός ~ scaevus ‘left’ is also limited to Greek and Latin. An exclusive match of derivation and meaning in a culturally important word is Lat. līber ‘free’, pl. līberī ‘children’, Ven. louderobos ‘children’ dat.pl., Gr. ἐλεύθερος ‘free’. The use of the suffix *-ero- suggests that *h1leu̯dʰeros originally meant ‘(one who is) in the people’ (*h1leu̯dʰo- > ON ljóðr ‘people’, OE lēod ‘man’, OCz. ĺud ‘people’) in opposition to those outside of the community.

Reference MartynovMartynov (1978) has indicated a number of cases where Slavic has two words for the same thing, one of which closely matches Italic. Most of these comparisons don’t hold up – the two best are OCS mesęcь ~ OCS luna ‘moon’, cf. Lat. lūna and OCS starъ ~ OCS matorъ ‘old’, cf. Lat. mātūrus – but these are not significant enough to support any theory of closer connection between Italic and Slavic, let alone Martynov’s view of a prehistoric conquest of pre-Slavs by pre-Italic speakers.

Finally, Reference Melchert and HansenMelchert (2016), following in the footsteps of Reference Puhvel and DunkelPuhvel (1994), has pointed out a few instances where Latin shares some innovative features with Anatolian, e.g. (1) HLuv. REL-ipa ‘indeed’~ Lat. quippe ‘indeed’ < *kʷid-pe, (2) Lyd. nãν, nã-m qid ‘whatever’ ~ Lat. quidnam ‘what on earth?’, (3) Hitt. kappūwe/a- ‘count’ ~ Lat. computāre ‘to reckon’. Whether these agreements are sufficient to support some secondary contact between Proto-Italic and Proto-Anatolian is uncertain.

To sum up: the similarities noticed between Italic and other Indo-European branches are predominantly lexical, and, when we compare these similarities to the ones noted between Italic and Celtic, the case for Proto-Italo-Celtic seems all the stronger (Chapter 7).

8.5 The Position of Italic

See Chapter 7.

Footnotes

1 See Reference WeissWeiss 2020: 15–18 for a survey.

2 On Faliscan see Reference BakkumBakkum 2009. There are about 360 linguistically informative Faliscan inscriptions dating from the sixth to second centuries BCE. On Praenestine, see Reference Franchi de BellisFranchi de Bellis 2005.

3 An imaginary line drawn by the historian Konstantin Jireček marking the southern extent of Latin influence in southeast Europe.

4 See Reference ZamponiZamponi 2021 for a survey of the evidence.

5 But see Reference ClacksonClackson 2015: 26–7 who questions Rix’s idea of uniting these texts and some inscriptions of Lucania as a unitary language. Reference AdiegoAdiego (2015) prefers Opic for the language of these inscriptions, which must have coexisted with Oscan for some time after the Samnite invasion of Campania.

6 There is no space to discuss the internal subgrouping of Sabellic, which is also not uncontroversial. See Reference ClacksonClackson 2015 and Reference Fortson, Klein, Joseph and FritzFortson 2017: 847–51.

9 On a kylix from Aidone. See Reference LejeuneLejeune 1990.

10 For an ingenious attempt to make sense of this text, see Reference Martzloff and Van HeemsMartzloff 2011.

11 But if, as I argued in Reference Weiss and GunkelWeiss 2018a (following Reference WaldeWalde 1906), *dʰragʰ- ‘drag’ > *dragʰ- > *tragʰ‑ > trah-ere by Limited Latin Grassmann’s Law and the change of *dr- to tr-, and if Limited Grassmann’s Law is only Latin, then there would be evidence that reflexes of *bʰ- and *dʰ- did not fall together in initial position in Italic so that the f- from both * and *dʰ- would have to be a diffused trait. The problem is that we can’t determine what the date of Limited Grassmann’s Law is, due to the lack of Sabellic evidence. It could be ordered very early in Proto-Italic before the devoicing of voiced aspirates or even theoretically in Proto-Italo-Celtic times.

12 A thorough survey of the arguments up until 1950 is provided by Diver’s unpublished dissertation of Reference Diver1953. Further important works are Jones 1951, Reference SafarewiczSafarewicz 1963, Reference Beeler, Birnbaum and PuhvelBeeler 1966, and Reference CampanileCampanile 1968.

13 For more on diffused changes, see Reference WeissWeiss 2020: 496–7. For what it is worth, Sicel appears to have an unapocopated esti.

14 Though eventually they did merge in Sabellic and Faliscan.

15 Gr. Σαύνιον, which must have been borrowed form a Sabellic form before anaptyxis, shows that the Greeks heard the fricative spelled by the Oscans with f as a voiced sound.

16 The reflex of * was still an obstruent in Umbrian at the time of medial syllable syncope because when this sound comes together with t by syncope it develops to j in the same way as k, e.g. *-u̯eg̑ʰetōd > *-u̯eγetōd > ařveitu ‘bring’ parallel to *fakitōd > feitu.

17 But note that if Sicel is Italic, and if the evidence is correctly interpreted as showing that Sicel had voiced reflexes in initial position, this isogloss would have to be interpreted differently. Either the initial PIE voiced aspirates first became voiced fricatives that were retained or became stops in Sicel and were devoiced in the rest of Italic, or the initial voiced aspirates became voiceless fricatives which were then voiced in Sicel in all positions (cf. the famous southern British English from which the standard dialect borrowed vat and vixen). Whichever interpretation is correct, the Italic developments would still be unique.

18 The sound changes that produced the 3rd io ~ 4th conjugation contrast are attested outside of verbal morphology, e.g. *dii̯eu̯ii̯o- ‘heavenly’ > *dīu̯ii̯o- (Osc. diíviaí), so it is uneconomical to set up athematic i-inflection for present stems which are functionally identical to the *i̯e-/i̯o-presents of other branches.

19 It is attractive to derive the endings of Old Irish i-stem verbs (absolute gaibid, conjunct ·gaib ‘takes’) from an immediate preform *-iti < *-i̯eti. But there are Gaulish forms that appear to show the unreduced sequence -ie- (bissiet ‘will be’), and the Sieversish distribution seen in Sabellic does not hold in Old Irish, e.g. bruinnid, ·bruinn ‘flows forth’ with an S 2 inflection after a heavy base.

20 The Oscan form humuns ‘men’ and Umb. homonus dat.pl. are ambiguous. humuns is written in an alphabet that does not distinguish u and o and Umbrian lowers u to o before m.

21 Alternatively, one could suppose that pre-vocalic syllabic * was preserved in Proto-Italic and then developed in slightly different ways in the daughter languages. In this way one loses the generalization that all Italic languages show a rounded vowel in this environment, but it is not too shocking that a prop vowel should develop to a rounded vowel before a labial. A distinct syllabic * arose in the 1sg. of the verb ‘to be’: *esmi became *esm̥ by the loss of the final primary marker *-i, and this developed to esom in VOL (Garigliano esom), but there is quite a bit of variation here. Latin itself attests sim, said to be Augustus’ favored form, and Sabellic has the same two variants *som (Osc. súm) and *sim (Pre-Samn. and SPic. sim). There are also Sabellic forms in esum (TE 4 SPic., Ps 4, 5 esum) and sum (Ps 13), but these are in alphabets that don’t distinguish u and o.

22 See Reference WeissWeiss 2020: 496–8. On the development of syllabic liquids in Italic, see Reference ZairZair 2017. Thurneysen-Havet’s Law (in the formulation of Reference VineVine 2006) and the development of *CR̥HC to CaraC are thought to be conditioned by the PIE accent and would be early, potentially of Proto-Italic date. I don’t believe that there are any secure examples of Thurneysen-Havet’s Law in Sabellic, but there is one good instance of *CR̥HC to CaraC (Umb. parfa (type of bird) < *parasā < *pŕ̥hₓseh2, see Reference HöflerHöfler 2017). This rule has a close parallel in Greek, however, and it is thus conceivable, though unlikely, that the Latin and Sabellic developments were independent.

24 Despite Campanile’s attempt (Reference Campanile1968: 59) to connect it with the Brittonic subjunctive.

25 On the last, see Reference DuprazDupraz 2010: 321.

26 There are other views, however, e.g. Reference WilliWilli 2016.

27 The alternatives are worse: (1) Latin and Sabellic both replaced the mid.2pl. with m-initial forms, which are unrelated. (2) While Latin had a reflex of *-mh1no-, Sabellic had *-mo- like East Baltic and Slavic, but there are to my knowledge no isoglosses connecting Sabellic and Balto-Slavic, and Balto-Slavic *-mo- may in any case come from *-mh1no-. (3) *-mh1no- gave *‑mo- in Sabellic, but this is excluded by the many cases of survival of -mn-, both primary and by syncope.

28 In the proto-language only o-stems made distinct abl.sg. forms. In all other stem types, the abl.sg. and the gen.sg. were identical. Another shared innovation of Sabellic and Latin is the extension of the PIE thematic instrumental to the dative-ablative (VOL -ois, Osc. -úís), but Venetic has retained the more archaic -obos in louderobos ‘children’ dat.pl.

29 See Reference VillarVillar 1995 and Reference Beltrán, Jordán, Sinner and VelazaBeltrán & Jordán 2019: 251–3. Young Avestan, independently, did the same thing, e.g. zaoθraiiat̰ ‘libation’ from an ā-stem, etc.

30 Closely related to the ablative phenomenon is the extension of originally instrumental adverbs in *‑eh1 > *-ē by -d: OLat. facilumed ‘very easily’, Osc. amprufid ‘improperly’.

31 One might add *-Footnote idʰo- if we were sure that the Sabellic place name Callifae were to be equated with Lat. callidus ‘experienced’ < *‘hardened’ or Lat. calidus ‘warm’. But I am skeptical of this because the name is attested only once at Livy 8.25 beside the much better attested Allifae and Allifae is known to have had a long i (Allīfāna Hor. S. 2.8.39, Ital. Alífe).

32 See also the listing and discussion in Reference Fortson, Klein, Joseph and FritzFortson 2017: 843–5.

33 There are also many items that are attested exclusively in Latin and Sabellic. To mention just two: *kubā- ‘lie’ (Lat. cubat, SPic. qupat) largely replacing *legʰ- except in Fal. lecet, SPic. veiiat, and Lat. lectum ‘bed’, *famelii̯ā ‘household’ (Lat. familia, Umb. fameřia) derived from *famelo- ‘slave’ (Lat. famulus).

34 See Reference HaleHale 2007: 238–9 on this point.

35 Cf. also the uncontracted form Umb. ahesnis ‘brazen’ abl.pl. < *ai̯esno-, although in this case the Latin cognate aēnus also remains – surprisingly – uncontracted.

36 The Sabellic treatment of the voiced aspirate labiovelar was to f in medial position: *h1u̯egʷʰ- > Umb. vufru ‘votive’ ~ Lat. voveō. There are no good examples of initial *gʷʰ, but it would be surprising if it was anything other than f.

37 The nom.sg. sakros is attested in VOL but it is probably an analogical restoration.

38 For a survey of Italic morphology with many references, see Reference Vine, Klein, Joseph and FritzVine 2017.

39 Latin does have instances of nom.pl. -ās, but these are probably not archaisms. See Reference WeissWeiss 2020: 252.

40 On the Sabellic demonstrative system, see Reference DuprazDupraz 2012.

41 There is little evidence for 2pl. primary *-tes outside of Italic. The ending *-tes may itself have been an analogical creation on the model of primary *-me/os.

42 See on this category most recently Reference ZairZair 2014.

43 This disagreement contrasts strikingly with the general agreement of present stem formation between Latin and Sabellic. The reasons for this contrast are presumably (1) that the historical “perfect” of Latin and Sabellic results from the parallel independent merger of the PIE perfect and aorist and (2) that denominal verbs had not yet acquired a productive perfect formant. For the development of the perfect system of Italic, see in general Reference MeiserMeiser 2003.

44 On these formants, see for the most recent proposals and a review of earlier scholarship: Reference WilliWilli 2010; Reference DuprazDupraz 2016: 340 (-nki̯-); Reference WilliWilli 2016; Reference DuprazDupraz 2016: 347 (-tt-); Reference DuprazDupraz 2018 (-k-); Reference ZairZair 2014 (‑ō-).

45 I am intentionally omitting any proposal of my own.

46 The gens Claudia, of legendary Sabine origin, introduced the praenomen and cognomen Nerō ‘strong’ into Latin.

47 Unless unda is deverbative to *u-ne-d-ti (Ved. unátti ‘flows’).

48 Oscan aapam acc.sg.fem. is usually compared with Ved. ā́p- ‘water’, but this root is *h2ep- and would not have a long ā in its paradigm. The Oscan word does not mean ‘water’ per se but most likely ‘water works’ vel sim. and can be explained as the substantivization of an inner-Italic vr̥ddhi formation *āpo- ‘of water’, but this could just as well be derived from *apā < *akʷā as from *ap- < *h2ep-.

49 Unless the root *h2ei̯s- is somehow continued in Lat. quaerere ‘to seek’ and aeruscāre ‘to beg’.

50 For more on the Indo-European of the Northwest see Chapter 7.

51 Reference PoloméPolomé 1966 refutes the supposed morphological and semantic isoglosses well, though obviously we might conduct the refutation somewhat differently today.

52 The supposed match between the -ne of Lat. superne and the -na of Goth. utana ‘from the outside’ is vitiated by the fact that Latin superne has a short final -e (cf. also dōnicum ‘until’ < *dōnVkʷom), and the suffix -nV added to preverbs is found widely in Indo-European, e.g. Hitt. istarna ‘among’. The derivation of adverbs from preverbs and pronominal bases with *-t(e)rō or *-t(e)rōd as in Lat. ultrō ‘willingly’, Osc. contrud ‘contrary to’, and Goth. innaþro ‘from within’, etc. is not exclusive (cf. Gr. προτέρω ‘further’ etc.), and the semantic match between Italic and Gothic is not good. While the Gothic forms have ablatival meaning, the same is not true of the Italic forms.

53 I’ve suggested (Reference Weiss and BeekWeiss 2018b: 351) that exactly such a form, me, is continued in Venetic me in the Isola Vicentina inscription.

54 Cf. also Hitt. ammuk.

55 The most recent proponent of this view was Reference ThibauThibau (1964).

56 The voiced aspirates were also devoiced in Romani.

References

Adiego, Ignasi-Xavier. 2015. Some remarks on the new Opic (“Pre-Samnite”) inscription of Niumsis Tanunis. Incontri Linguistici 38. 125–38.Google Scholar
Agostiniani, Luciano. 1992. Les parlers indigènes de la Sicile prégrecque. LALIES. Actes des sessions de linguistique et de literature 11. 125–57.Google Scholar
Ambrosini, Riccardo. 1984. Lingue nelle Sicilia pregreca. In Moreschini, Adriana Quattordio (ed.), Tre millenni di storia linguistica della Sicilia, 1336. Pisa: Giardini.Google Scholar
Bakkum, Gabriël C. L. M. 2009. The Latin dialect of the Ager Faliscus: 150 years of scholarship. Part 1. Amsterdam: Amsterdam University Press.Google Scholar
Beeler, Madison. 1966. The interrelationships within Italic. In Birnbaum, Henrik & Puhvel, Jaan (eds.), Ancient Indo-European dialects, 51−8. Berkeley: University of California Press.Google Scholar
Beltrán, F. & Jordán, C.. 2019. Writing and language in Celtiberia. In Sinner, Alejandro G. & Velaza, Javier (eds.), Palaeohispanic languages and epigraphies, 240303. New York: Oxford University Press.Google Scholar
Buck, Carl Darling. 1928. A grammar of Oscan and Umbrian. 2nd ed. Boston: Ginn.Google Scholar
Campanile, Enrico. 1968. Studi sulla posizione dialettale del Latino. Pisa: Pacini Mariotti.Google Scholar
Campanile, Enrico. 1969. Note sulle glosse sicule e sui rapporti linguistici fra siculo e latino. In Studia Classica et Orientalia Antonino Pagliaro oblata. Vol. 1, 293322. Rome: Herder.Google Scholar
Christol, Alain. 2005. Subjonctif latin (-) et futur indien (-sya-). Journal of Latin Linguistics 9. 2536.Google Scholar
Clackson, James. 2015. Subgrouping in the Sabellian branch of Indo-European. Transactions of the Philological Society 113. 437.Google Scholar
Curtius, Georg. 1858. Grundzüge der griechischen Etymologie. Erster Theil. Leipzig: Teubner.Google Scholar
Devoto, Giacomo. 1929. Italo-greco e italo-celtico. Archivio glottologico italiano 22/23. 200–40.Google Scholar
Devoto, Giacomo. 1936. Germanisch-Lateinisch und Germanisch-oskisch-umbrisch. In Germanen und Indogermanen: Festschrift für Herman Hirt. Vol. 2, 533–47. Heidelberg: Winter.Google Scholar
Diver, William G. 1953. The relation of Latin to Oscan-Umbrian. PhD thesis, Columbia University.Google Scholar
Dupraz, Emmanuel. 2010. Les Vestins à l’époque tardo-républicaine: Du nord-osque au latin. Mont-Saint-Aignan Cedex: Publications des universités de Rouen et du Havre.Google Scholar
Dupraz, Emmanuel. 2012. Sabellian demonstratives: Forms and functions. Leiden: Brill.Google Scholar
Dupraz, Emmanuel. 2016. Zu einigen Perfektbildungen im Sabellischen. Indogermanische Forschungen 121. 333–63.CrossRefGoogle Scholar
Dupraz, Emmanuel. 2018. Les soi-disant parfaits en -k- de l’osque: entre épigraphie, histoire des religions et linguistique comparative. Les études classiques 86. 4967.Google Scholar
Fortson, Benjamin W., IV. 2012. Latin -rier and its Indo-Iranian congeners. Indogermanische Forschungen 117. 75118.Google Scholar
Fortson, Benjamin W., IV. 2013. Pre-Italic *-dʰi̯ē (*-dhi̯eh1) versus pre-Indo-Iranian *-dʰi̯ōi: Bridging the gap. In Cooper, Adam I., Rau, Jeremy & Weiss, Michael (eds.), Multi nominis grammaticus: Studies in Classical and Indo-European linguistics in honor of Alan J. Nussbaum on the occasion of his sixty-fifth birthday, 5060. Ann Arbor, MI: Beech Stave.Google Scholar
Fortson, Benjamin W., IV. 2017. The dialectology of Italic. In Klein, Jared S., Joseph, Brian D., & Fritz, Matthias (eds.), Handbook of historical and comparative Indo-European linguistics. Vol. 2, 835–57. Berlin: De Gruyter Mouton.Google Scholar
Fortson, Benjamin W., IV. 2018. A cover-up: Latin amiciō. In Hackstein, Olav & Opfermann, Andreas (eds.), Priscis libentius et liberius novis: Indogermanische und sprachwissenschaftliche Studien. Festschrift für Gerhard Meiser zum 65. Geburtstag, 4964. Hamburg: Barr.Google Scholar
Franchi de Bellis, A., 2005. Iscrizioni prenestine su specchi e ciste. Alessandria: Edizioni dell’Orso.Google Scholar
Hale, Mark. 2007. Historical linguistics: Theory and method. Malden, MA: Blackwell.Google Scholar
Hartmann, Markus. 2018. Siculian. In Klein, Jared S., Joseph, Brian D., & Fritz, Matthias (eds.), Handbook of historical and comparative Indo-European linguistics. Vol. 3, 1854–7. Berlin: De Gruyter Mouton.Google Scholar
Hirt, Herman. 1897. Die Stellung des Germanischen im Kreise der verwandten Sprachen. Zeitschrift für deutsche Philologie 29. 289305.Google Scholar
Höfler, Stefan. 2017. Observations on the palma rule. Pallas: Revue d’études antiques 103. 1523.Google Scholar
Jasanoff, Jay H. 1991. The origin of the Latin imperfect subjunctive. Historische Sprachforschung 104. 84105.Google Scholar
Kuz’menko, Ju. K. 2011. O čem mogut svidetel’stvovat’ germano-italijskie innovacii [What can Germano-Italic innovations tell us?]. Acta Linguistica Petropolitana 7. 930.Google Scholar
Lejeune, Michel. 1990. Notes de linguistique italique XL: ‘Bois!’ disait ce Sicule; ‘je boirai’ répond ce Falisque. Revue des études latines 68. 2830.Google Scholar
Machajdíková, Barbora. 2018. Dve grécke výpožičky na sikulských nápisoch (darnakei, poterom) [Two Greek loanwords in Sicel inscriptions (darnakei, poterom)]. Hortus Graeco-Latinus Cassoviensis 2. 143–54.Google Scholar
Martynov, V. V. 1978. Balto-slavjano-italijskie izoglossy: Leksičeskaja sinonimija [Balto-Slavo-Italic isoglosses: Lexical synonymy]. Minsk: Nauka i texnika.Google Scholar
Martzloff, Vincent. 2011. Variation linguistique et exégèse paléo-italique: l’idiome sicule de Montagna di Marzo. In Van Heems, Gilles (ed.), La variation linguistique dans les langues de l’Italie préromaine, 93129. Lyon: Maison de l’Orient et de la Méditerranée.Google Scholar
Meillet, Antoine. 1922. Les dialectes indo-européens. 2nd ed. Paris: Champion.Google Scholar
Meiser, Gerhard. 1993. Uritalische Modussyntax: zur Genese des Konjunktiv Imperfekt. In Rix, H. (ed.), Oskisch-Umbrisch: Texte und Grammatik. 167–95. Wiesbaden: Reichert.Google Scholar
Meiser, Gerhard. 1998. Historische Laut- und Formenlehre der lateinischen Sprache. Darmstadt: Wissenschaftliche Buchgesellschaft.Google Scholar
Meiser, Gerhard. 2003. Veni vidi vici: Die Vorgeschichte des lateinischen Perfektsystems. Munich: Beck.Google Scholar
Melchert, H. Craig. 2016. “Western affinities” of Anatolian. In Hansen, Bjarne Simmelkjær Sandgaard et al. (eds.), Etymology and the European lexicon, 297305. Wiesbaden: Reichert.Google Scholar
Muller, Frederick. 1926. Altitalisches Wörterbuch. Göttingen: Vandenhoeck & Ruprecht.Google Scholar
Poccetti, Paolo. 2012. Language relations in Sicily. In Tribulato, Olga (ed.), Language and linguistic contact in ancient Sicily, 4994. Cambridge: Cambridge University Press.Google Scholar
Poccetti, Paolo. 2017. The documentation of Italic. In Klein, Jared S., Joseph, Brian D., & Fritz, Matthias (eds.), Handbook of historical and comparative Indo-European linguistics. Vol. 2, 733–43. Berlin: De Gruyter Mouton.Google Scholar
Polomé, Edgar. 1966. Germanisch und Italisch im Lichte der deskriptiven Sprachbetrachtung. Orbis 15. 190–9.Google Scholar
Prag, Jonathan R. W. 2014. Bronze rostra from the Egadi Islands off NW Sicily: The Latin inscriptions. Journal of Roman Archaeology 27. 3359.CrossRefGoogle Scholar
Puhvel, Jaan. 1994. West-Indo-European affinities of Anatolian. In Dunkel, George et al. (eds.), Früh-, Mittel-, Spätindogermanisch, 315–24. Wiesbaden: Reichert.Google Scholar
Rasmussen, Jens Elmegård. 1996. On the origin of the Latin imperfect subjunctive. Copenhagen Working Papers in Linguistics 4. 143–8.Google Scholar
Rix, Helmut. 1996. Review of Peter Schrijver, The reflexes of the Proto-Indo-European laryngeals in Latin, Amsterdam: Rodopi, 1991. Kratylos 41. 153–63.Google Scholar
Safarewicz, Jan. 1963. Two Italic studies. In Jan Safarewicz, Linguistic studies, 121–49. The Hague: Mouton. Originally published as “Dwa studia italskie” in Zeszyty naukowe Uniwersytetu Jagiellońskiego 60, Prace językoznawcze 5, 734.Google Scholar
Stifter, David. 2017. Varia II: The origin of time. Ériu 67. 219–26.Google Scholar
Thibau, R. 1964. Les rapports entre le latin et le grec: Essai de réhabilitation de l’hypothèse italo-greque. Leiden: Brill.CrossRefGoogle Scholar
Villar, F. 1995. A new interpretation of Celtiberian grammar. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck.Google Scholar
Vine, Brent. 2006. On “Thurneysen-Havet’s Law” in Latin and Italic. Historische Sprachforschung 119. 211–49.Google Scholar
Vine, Brent. 2017. The morphology of Italic. In Klein, Jared S., Joseph, Brian D., & Fritz, Matthias (eds.), Handbook of historical and comparative Indo-European linguistics. Vol. 2, 751804. Berlin: De Gruyter Mouton.CrossRefGoogle Scholar
Walde, Alois. 1906. Aspiratendissimilation im Latein. Indogermanische Forschungen 19. 98111.CrossRefGoogle Scholar
Walde, Alois. 1917. Über älteste sprachliche Beziehung zwischen Kelten und Italikern. Innsbruck: Wagner.Google Scholar
Weiss, Michael. 1998. On some problems of final syllables in South Picene. In Jay H. Jasanoff, H. Melchert, Craig & Oliver, Lisi (eds.). 1998. Mír curad: Studies in honor of Calvert Watkins, 703–15. Innsbruck: Institut für Sprachwissenschaft der Universität Innsbruck.Google Scholar
Weiss, Michael. 2010. Language and ritual in Sabellic Italy. Leiden: Brill.Google Scholar
Weiss, Michael. 2018a. Limited Latin Grassmann’s Law: Do we need it? In Gunkel, Dieter et al. (eds.), Vina diem celebrent: Studies in linguistics and philology in honor of Brent Vine, 438–47. Ann Arbor, MI: Beech Stave.Google Scholar
Weiss, Michael. 2018b. Veneti or Venetes? Observations on a widespread Indo-European tribal name. In Beek, Lucien van et al. (eds.), Farnah: Indo-Iranian and Indo-European studies in honor of Sasha Lubotsky, 349–58. Ann Arbor, MI: Beech Stave.Google Scholar
Weiss, Michael. 2020. Outline of the historical and comparative grammar of Latin. 2nd ed. Ann Arbor, MI: Beech Stave.Google Scholar
Willi, Andreas. 2008. Sikelismos. Sprache, Literatur und Gesellschaft im griechischen Sizilien (8.–5. Jh. v. Chr.). Basel: Schwabe.Google Scholar
Willi, Andreas. 2010. The Umbrian perfect in -nç-/-ns̀-. Transactions of the Philological Society 108. 114.CrossRefGoogle Scholar
Willi, Andreas. 2016. The Oscan perfect in -tt - . Transactions of the Philological Society 114. 7594.Google Scholar
Zair, Nicholas. 2014. The future perfect in Oscan and Umbrian and the ō-perfect in South Picene. Transactions of the Philological Society 112. 367–85.Google Scholar
Zair, Nicholas. 2016. Oscan in the Greek alphabet. Cambridge: Cambridge University Press.Google Scholar
Zair, Nicholas. 2017. The origins of -urC- for expected -orC- in Latin. Glotta 93. 255–89.Google Scholar
Zamponi, Raoul. 2021. South Picene. Abingdon: Routledge.CrossRefGoogle Scholar
Figure 0

Figure 8.1 The Italic languages

Save book to Kindle

To save this book to your Kindle, first ensure [email protected] is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

  • Italic
  • Edited by Thomas Olander, University of Copenhagen
  • Book: The Indo-European Language Family
  • Online publication: 15 September 2022
  • Chapter DOI: https://doi.org/10.1017/9781108758666.008
Available formats
×

Save book to Dropbox

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Dropbox.

  • Italic
  • Edited by Thomas Olander, University of Copenhagen
  • Book: The Indo-European Language Family
  • Online publication: 15 September 2022
  • Chapter DOI: https://doi.org/10.1017/9781108758666.008
Available formats
×

Save book to Google Drive

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Google Drive.

  • Italic
  • Edited by Thomas Olander, University of Copenhagen
  • Book: The Indo-European Language Family
  • Online publication: 15 September 2022
  • Chapter DOI: https://doi.org/10.1017/9781108758666.008
Available formats
×