Proto-Romance language

(Redirected from Proto-Romance)

Proto-Romance is the comparatively reconstructed ancestor of all Romance languages. It reflects a late variety of spoken Latin prior to regional fragmentation.[1]

Reconstruction ofRomance languages
RegionRoman Empire
Erac. 3rd–4th centuries CE?




Front Central Back
Close i u
Near-close ɪ ʊ
Close-mid e o
Open-mid ɛ ɔ
Open a


The only phonemic diphthong was /au̯/.[2]


  • Vowels were lengthened in stressed open syllables.[3]
  • Vowel breaking: stressed /ɛ, ɔ/ may have yielded the incipient diphthongs [e͜ɛ, o͜ɔ] when followed by a syllable containing a close vowel.[4]
    • Whatever the precise outcome, Maiden argues that this would have been limited, at the Proto-Romance stage, to open syllables. That is, it would have applied only to instances of /ɛ/ and /ɔ/ that had been subject to stressed-open-syllable lengthening.[5]


  • Neither a distinct /ɛ/ nor /ɔ/ occurred in unstressed position on account of having merged into /e/ and /o/ respectively.[6]
  • Neither a distinct /i/ nor /u/ occurred in the second syllable of words with the structure [ˌσσˈσσ] (such as càntatóre 'singer') on account of having merged into /ɪ/ and /ʊ/ respectively.[7]


Per Burger (1955:25)
Labial Coronal Palatal Velar
non-labial labial
Nasal m mʲ n nʲ
Occlusive p pʲ b bʲ t tʲ d dʲ j k kʲ ɡ ɡʲ () (ɡʷ)
Fricative f () β βʲ s sʲ
Vibrant r rʲ
Approximant l lʲ w

/tʲ/ appears to have affricated to [t͡sʲ] and /kʲ/ was at least fronted to [c], if not also affricated to [c͡ç].[8]


The following features are reconstructed with varying degrees of certainty:

  • A prop-vowel [ɪ] was added before word-initial /sC/ clusters not already preceded by a vowel (as in /sˈtare/ [ɪsˈtaːɾe]).[9]
  • Palatalized consonants, other than /sʲ rʲ/, tended to geminate in intervocalic position, although this varied widely depending on the consonant in question.[10] For /bʲ dʲ ɡʲ/, see below.
  • The sequence /ɡn/ was likely realized as [ɣn] at first, with subsequent developments varying by region.[11][i]
  • /j/ was likely realized as [ʝ] or [ɟ], possibly with gemination in intervocalic position.[12]
  • /d/ and /ɡ/ might have been fricatives or approximants in intervocalic position.[13]
  • /s/ might have been apico-alveolar.[14]
  • /ll/ might have been retroflex.[15][ii]
  • /f/ might have been bilabial.[16]


  • /b/ and /bʲ/ did not occur intervocalically on account of having merged with /β, βʲ/ (spirantization).[17]
    • The same may also have occurred after /r/ or /l/.[18]
  • /dʲ/ and /ɡʲ/ did not occur intervocalically on account of both having merged with /j/.[19]
  • /kʷ/ and /ɡʷ/ did not occur before back vowels on account of having delabialized to /k, ɡ/.[20]


Proto-Romance nouns appear to have had three cases: a nominative, an accusative, and a combined genitive-dative.

Per Lausberg (1973:29, 32, 66–67)[iii]
NOM kápra kápras kaβállʊs kaβálli páter pátres~pátri máter mátres
ACC kaβállu kaβállos pátre pátres
GEN-DAT kápre kápris kaβállo kaβállis pátri pátris mátri mátris
Translation goat horse father mother

Several Class III nouns had inflexions that differed by syllable count or stress position.

Per Hall (1983:28)
Number SG
NOM ɔ́mo pástor sɔ́ror
ACC ɔ́mɪne pastóre soróre
GEN-DAT ɔ́mɪni pastóri soróri
Translation man pastor sister

Some nouns were pluralized with -a or -ora, having originally been neuter in Classical Latin. Their singular was treated as grammatically masculine, while their plural was treated as feminine.[21]

Per Lausberg (1973:47)
Class II III
Number SG PL SG PL
NOM brákʲu brákʲa tɛ́mpʊs tɛ́mpora
GEN-DAT brákʲo brákʲis tɛ́mpori tɛ́mporis
Translation arm time

Such nouns, due to their plurals, were often reanalyzed as collective feminine nouns.

Per Alkire & Rosen (2010:193–194)
Number SG PL SG PL
Original noun fɔ́lʲu fɔ́lʲa lɪ́ɡnu lɪ́ɡna
Fem. variant fɔ́lʲa fɔ́lʲas lɪ́ɡna lɪ́ɡnas
Translation leaf, leaves firewood



Per Lausberg (1973:108–109, 119–122)
Class I/II III
Gender M F M F
NOM bɔ́nʊs bɔ́ni bɔ́na bɔ́nas βɪ́rdɪs βɪ́rdes~βɪ́rdi βɪ́rdɪs βɪ́rdes
ACC bɔ́nu bɔ́nos βɪ́rde βɪ́rdes βɪ́rde
GEN-DAT bɔ́no bɔ́nis bɔ́ne bɔ́nis βɪ́rdi βɪ́rdis βɪ́rdi βɪ́rdis
Translation good green


Proto-Romance inherited the comparative suffix -ior from Latin, but only in a limited number of adjectives.[22][iv]

Per Lausberg (1973:129–131)
Number SG
Gender M+F N
NOM mɛ́lʲor mɛ́lʲʊs
ACC melʲóre
Translation better

Otherwise, the typical way to form a comparative seems to have been to add either plus or magis (meaning 'more') to a positive adjective.[23]


With the exception of a few fossilized forms, such as /ˈpɛssɪmʊs/ 'worst', superlatives were formed by adding an intensifying adverb or prefix (/mʊltu, bɛne, per-, tras-/ etc.) to a positive adjective. Comparative forms could also have been made superlative by adding a demonstrative adjective.[24]


Feminine singular forms shown below. In certain cases there was an opposition between 'strong' (stressed) and 'weak' (unstressed) variants.[25]

Per Lausberg (1973:§§754–755)
1P 2P 3P INT
SG mɛ́a~ma tʊ́a~ta sʊ́a~sa kúja
PL nɔ́stra βɔ́stra



Numerous variant forms appear to have existed. For the third-person genitive-dative inflexions, there appears to have been an opposition between 'strong' (stressed) and 'weak' (unstressed) variants, as also with the possessive adjectives.

Per Hall (1983:39) and De Dardel & Wüest (1993:39–43, 57)
1P 2P 3P.M 3P.F
NOM ɛ́ɡo nós βós ɪ́lle~ɪ́lli ɪ́lli ɪ́lla ɪ́llas
ACC mé~méne té~téne ɪ́llu ɪ́llos
GEN-DAT mí~mɪ́βɪ nóβɪs tí~tɪ́βɪ βóβɪs ɪlli~ɪllúi ɪllis~ɪllóru ɪlli~ɪllɛ́i ɪllis~ɪllóru


Per Elcock (1960:95–96)
Gender M+F N
NOM kʷí kɔ́d
ACC kʷɛ́n

The interrogative pronouns were the same, except that the neuter nominative-accusative form was /ˈkʷɪd/.


Proto Romance verbs belonged to three main classes, each characterized by a different thematic vowel. Their conjugations were built on three stems and involved various combinations of mood, aspect, and tense.[26]

Present indicativeEdit

Per Van Den Bussche (1985:§§2.3–2.3.2)[v]
Verb class 1P 2P 3P Translation
I kánto kantámʊs kántas kantátɪs kántat kántant sing
II.a βɪ́jo βɪdémʊs βɪ́des βɪdétɪs βɪ́det βɪ́jʊnt~βɪ́dʊnt~βɪ́dent see
II.b βɛ́ndo βɛ́ndɪmʊs βɛ́ndɪs βɛ́ndɪtɪs βɛ́ndɪt βɛ́ndʊnt~βɛ́ndent sell
III dɔ́rmo~dɔ́rmʲo dormímʊs dɔ́rmɪs dɔrmítɪs dɔ́rmɪt dɔ́rmʊnt~dɔ́rment sleep
Irregular sʊ́n sʊ́mʊs~sémʊs ɛ́s ɛ́stɪs~sétɪs~sʊ́tɪs ɛ́st sʊ́nt be
áβʲo~ájo aβémʊs áes~ás aβétɪs áet~át áu̯nt~áent~ánt have
dáo dámʊs dás dátɪs dát dáu̯nt~dáent~dánt give
βádo~βáo ímʊs[27] βáɪs~βás ítɪs[27] βáɪt~βát βáu̯nt~βáent~βánt go


As in Latin, present participles had an active sense and inflected like class III adjectives, while past participles had a passive sense and inflected like class I/II adjectives. Regular forms would have been as follows (in the accusative feminine singular):

Per Hall (1983:122–3)
Type PRES.ACT Translation PERF.PASS Translation
I amánte adoring amáta adored
II aβɛ́nte having aβúta had
III finɛ́nte finishing finíta finished

See alsoEdit


  1. ^ The ultimate outcome of /ɡn/ in most of Romance is /ɲ/. Exceptions include Balkan Romance and Dalmatian, where it yielded /mn/; Sardinian, where it yielded /nn/; and certain dialects of southern Italy, where it yielded /u̯m/ or /i̯m/.
  2. ^ For further discussion on /ll/, see Zampaulo 2019 (71–77) and Lausberg 1970 (§§494–499).
  3. ^ De Dardel & Gaeng (1992: 104) differ from Lausberg on the following points: 1) They believe that the genitive-dative case was limited to animate nouns. 2) They reconstruct a universal GEN-DAT.PL ending /-ˈoru/. 3) They reconstruct, for class I nouns, a NOM.PL /-e/, albeit in competition with /-as/ (De Dardel & Wüest 1993: 57). They are in agreement with Lausberg regarding the remaining inflexions.
  4. ^ All comparatives inflected the same way. Further examples are pɛ́jor, májor, mɪ́nor, fɔ́rtjor, and ɡɛ́ntjor; meaning 'worse, greater, lesser, stronger, nobler' (Hall 1983: 32, 120).
  5. ^ Nearly all of the below is per Van Den Bussche (1985), a critique of, and elaboration on, Hall (1983). Since the former does not discuss the inflexions of essere 'to be', those have been taken unchanged from Hall (1983: 55). Van Den Bussche leaves out the 1PL and 2PL inflexions of vadere 'to go' because there was suppletion with forms of Latin ire, as indicated more explicitly by Maiden (1995: 135).


  1. ^ Dworkin 2016:13
  2. ^ Ferguson 1976:84; Gouvert 2015:81
  3. ^ Gouvert 2015:118‒119; Loporcaro 2015; Leppänen & Alho 2018, §§5.1, 6
  4. ^ Ferguson 1976, chapter 7
  5. ^ Maiden 2016
  6. ^ Ferguson 1976:76; Gouvert 2015:78–81, 121–122
  7. ^ Lausberg 1970, §§192–196 apud Gouvert 2015:78–79
  8. ^ Lausberg 1970, §§452, 467; Gouvert 2015:86, 92; Zampaulo 2019:94
  9. ^ Lloyd 1987:148–150; Gouvert 2015:125–126
  10. ^ Lausberg 1970, §§451–478; Wilkinson 1976:11–14; Gouvert 2015:95, 111, 115
  11. ^ Lausberg 1970, §444; Chambon 2013 apud Gouvert 2015:95; Zampaulo 2019:80
  12. ^ Lausberg 1970, §§329, 471; Lloyd 1987:132; Gouvert 2015:83, 91; Zampaulo 2019:83‒84, 88
  13. ^ Lloyd 1987:141; Gouvert 2016:48
  14. ^ Lloyd 1987:80–81; Zampaulo 2019:93
  15. ^ Gouvert 2015:15
  16. ^ Lloyd 1987:80; Gouvert 2016:28
  17. ^ Lausberg 1970, §§366, 475; Gouvert 2015:86
  18. ^ Gouvert 2015:84
  19. ^ Lloyd 1987:133; Gouvert 2016:43; Zampaulo 2019:87‒88
  20. ^ Grandgent 1907, §§226, 254; Lausberg 1970, §§344, 486
  21. ^ Hall 1983:23–4, 29–30
  22. ^ Maltby 2016:340
  23. ^ Lausberg 1973:126–127; Maltby 2016:340–346
  24. ^ Lausberg 1973, §§686–687; Bauer 2016:340, 359
  25. ^ Lyons 1986:20–24
  26. ^ Hall 1983:47–50
  27. ^ a b Maiden 1995:135


  • Adams, J. N. (2013). Social Variation and the Latin Language. Cambridge University Press. ISBN 978-0521886147.
  • Alkire, Ti & Rosen, Carol (2010). Romance Languages: A Historical Introduction. New York: Cambridge University Press. ASIN B003VS0CSS.
  • Bauer, Brigitte (2016). "The development of the comparative in Latin texts". In Adams, J.N. & Vincent, Nigel (eds.). Early and Late Latin: Continuity or Change?. Cambridge University Press. pp. 313–339. doi:10.1017/CBO9781316450826.015. ISBN 9781316450826.
  • Burger, André (1955). "Phonématique et diachronie a propos de la palatalisation des consonnes romanes" [Phonemics and Diachrony of the Palatalization of Romance Consonants]. Cahiers Ferdinand de Saussure (in French). Librairie Droz (13): 19–33.
  • Chambon, Jean-Pierre (2013). "Notes sur un problème de la reconstruction phonétique et phonologique du protoroman: Le groupe */ɡn/". Bulletin de la Société de Linguistique de Paris. CVIII: 273–282.
  • De Dardel, R. & Gaeng, P. A. (1992). "La declinaison nominale du latin non classique: Essai d'une methode de synthese". Probus (in French). 4 (2): 91–125. doi:10.1515/prbs.1992.4.2.91. S2CID 171003000.
  • De Dardel, R. & Wüest, Jakob (1993). "Les systèmes casuels du protoroman: Les deux cycles de simplification". Vox Romanica (52): 25–65.
  • Dworkin, Steven N. (2016). "Do Romanists Need to Reconstruct Proto-Romance? The Case of the Dictionnaire Étymologique Roman Project" (PDF). Zeitschrift für romanische Philologie (132): 1–19. doi:10.1515/zrp-2016-0001. S2CID 163635015.
  • Elcock, W. D. (1960). The Romance Languages. London: Faber and Faber.
  • Ferguson, Thaddeus (1976). A History of the Romance Vowel Systems through Paradigmatic Reconstruction. Berlin: De Gruyter.
  • Gouvert, Xavier (2015). "Le système phonologique du protoroman: essai de reconstruction". In Buchi, Éva; Schweickard, Wolfgang (eds.). Dictionnaire Étymologique Roman. Beihefte zur Zeitschrift für romanische Philologie. Vol. 381. De Gruyter. ISBN 9783110453614.
  • Gouvert, Xavier (2016). "Du protoitalique au protoroman: deux problèmes de reconstruction phonologique". In Buchi, Éva & Schweickard, Wolfgang (eds.). Dictionnaire Étymologique Roman 2. Beihefte zur Zeitschrift für romanische Philologie. Vol. 402. De Gruyter. pp. 27–51.
  • Grandgent, C. H. (1907). An Introduction to Vulgar Latin. Boston: D.C. Heath & Co.
  • Hall, Robert Anderson (1976). Proto-Romance Phonology. New York: Elsevier.
  • Hall, Robert Anderson (1983). Proto-Romance Morphology. Philadelphia: John Benjamins.
  • Lausberg, Heinrich (1970). Lingüística románica. Vol. I: Fonética. Madrid: Gredos.
  • Lausberg, Heinrich (1973). Lingüística románica. Vol. II: Morfología. Madrid: Gredos.
  • Leppänen, V.; Alho, T. (2018). "On the mergers of Latin close-mid vowels". Transactions of the Philological Society. 116 (3): 460–483. doi:10.1111/1467-968X.12130. S2CID 150148733.
  • Loporcaro, Michele (2015). Vowel Length From Latin to Romance. Oxford University Press.
  • Lloyd, Paul M. (1987). From Latin to Spanish. Philadelphia: American Philological Society.
  • Lyons, Christopher (1986). "On the Origin of the Old French Strong-Weak Possessive Distinction". Transactions of the Philological Society. 84 (1): 1–41. doi:10.1111/j.1467-968X.1986.tb01046.x.
  • Maiden, Marten (1995). A Linguistic History of Italian. New York: Routledge.
  • Maiden, Martin (2016). "Diphthongization". In Ledgeway, Adam; Maiden, Martin (eds.). The Oxford guide to the Romance languages. Oxford University Press. pp. 647–57.
  • Maltby, Robert (2016). "Analytic and synthetic forms of the comparative and superlative from early to late Latin". In Adams, J.N.; Vincent, Nigel (eds.). Early and Late Latin: Continuity or Change?. Cambridge University Press. pp. 340–366.
  • Van Den Bussche, H. (1985). "Proto-Romance Inflectional Morphology. Review of Proto-Romance Morphology by Robert Hall". Lingua. 66 (2–3): 225–260. doi:10.1016/S0024-3841(85)90336-5.
  • Wilkinson, Hugh E. (1976). "Notes on the development of -kj-, -tj- in Spanish and Portuguese". Ronshu. 17: 19–36. CiteSeerX
  • Zampaulo, André (2019). Palatal Sound Change in the Romance languages: Diachronic and Synchronic Perspectives. Oxford Studies in Diachronic and Historical Linguistics. Vol. 38. Oxford University Press. ISBN 9780192534293.