Keith Andrew Massey
County College of Morris, New Jersey
Among the Romance languages, the development of Romanian is perhaps the least understood (Rosetti et al. 1969; Du Nay 1996; Sala 2005). There is an interval of several hundred years in which no historical record mentions its exis- tence,1 let alone providing scholars with samples to track its evolution (Rosetti et al. 1969, 481; Sala 2005, 25).
Romanian, like all Romance languages, has its own distinct features, owing principally to its particular local influences. This includes a significant influx of Slavic borrowings. One word of apparently Slavic origin in Romanian is da = yes. It seems impossible that any populace would choose to borrow such a basic word from another language. But the history of human speech does, in fact, attest examples of such counter-intuitive developments.2
In this particular case, however, I will argue that a possible Latin origin for Romanian da has been overlooked. With a Romance etymology for such a basic word established, a Slavic source may be unnecessary.
2 The Slavic Assumption for da = yes
It is understandable that a Slavic origin of Romanian da = yes should have been assumed and then never questioned (Cihac 1879, 89; Ciorănescu 2003, 276). Afterall, the words enjoy a perfect oneness of phonology and semantics. Further evidence for the assumption has been the observation that da appears relatively late in the Romanian literary record,3 suggesting not just a borrowing, but a recent one at that.
3 Latin Affirmations and their Romance Coun- terparts
Classical Latin did not have only one lexical item for ‘yes’. Instead, speakers of Latin tended to employ a rich combination of words and expressions. Among these were sic, ita, vero, as well as merely repeating a phrase in full (Buck & Hale 1903, 137).4 In Late Latin, it appears that the word sic reached a level of predominance (Pucci & Harrington 1997, 11),5 setting the stage for the current usage of Spanish, Portuguese, Italian, and French.6 Latin sic did survive into Romanian with arguably an even greater prevalence, becoming the basic conjunction şi, ‘and’ (Ciorănescu 2003, 713).
4 Latin ita > Romanian da?
Because of their similarity in form and meaning, an etymological derivation of Romanian da from Latin ita would seem an intriguing possibility. The evolution of ita into da could have followed two routes.
4.1 ita > *ida > da
Though not represented in the orthography, Late Latin seems to have undergone a general voicing of intervocalic consonants, though the Romance languages outside the Iberian Peninsula subsequently experienced a significant degree of regression from this voicing (Cravens 1996, 65-66). As such, an intermediate form ∗ida likely existed at least in pronunciation. While Romanian experienced an almost total devoicing, sporadic preservation of voicing is attested.7 This would, however, leave the question of why Romanian would atypically preserve the voicing in this particular word.
An explanation for this atypical preservation could be found in the relative frequency of ‘yes’ in any language. The so-called «frequency effect» can result in relatively rapid evolution of lexical items which are in common use (Bybee 2001, 11). Further, given the overall economy of language, there is a tendency toward ease in pronunciation, especially in frequently produced lexical items (Zipf 1929). Voiced consonants are relatively shorter than their unvoiced counterparts,8 thus anticipating a preference for preserving voicing within frequently used words.
The loss of the initial vowel in the hypothetical ∗ida is more predictable. In the stage of Classical Latin, the word ita shifted the accent to the final syllable when the word received enclitic elements.9 This could have resulted in a move toward final accentuation overall. The loss of unaccented initial vowels in Romanian, though not universal, is regularly attested.10
Another possible route from ita to da would be to posit loss of the initial vowel prior to voicing. While not an ordinary development in Romanian, sporadic voicing of initial stops is also attested.11 The same argument regarding voicing of frequently used words would also apply to this case.
4.3 Latin ita > Romanian da possible but improbable
The evolution of Latin ita into Romanian da can be demonstrated as phono- logically possible. Nevertheless, the posited intermediate forms are unattested. Additionally, the proposed development appeals more to exceptions to rules than to regularly observed tendencies in Romanian. Without further evidence, there would be no compelling reason to question the assumed Slavic origin of Roma- nian da. What will shift this balance, however, is the demonstration that Latin ita has indeed survived in Romanian as da —in another form.
5 Romanian dacă = if
Romanian is alone among the Romance languages in discontinuing use of Latin si = if; Romanian se (să) was replaced by dacă, first in temporal clauses and eventually in all conditional clauses (Roques 1907, 825-839).
5.1 The traditional etymology of dacă
Within the earliest recorded Romanian the alternative forms deca, déca, and deaca are also in currency.12 This directed lexicographers to an assumed et- ymology of the item from Latin ∗de quod or ∗de ad quod.13 The form dacă has been explained as an evolution from what are assumed, on the basis of the proposed etymology, to be the original forms (Rosetti 1983, 121-122).
A problem with deriving dacă from ∗de quod / ∗de ad quod is that neither of these particular combinations is attested in Latin, even during the Medieval period. While it is not impossible for an otherwise unattested combination of words to have existed, an etymology citing attested forms would be preferable.
5.2 A new proposal for a Latin source of Romanian dacă
One feature of Late Latin is the replacement of ut by quod in subordinate clauses (Pucci & Harrington 1997, 38). For result clauses specifically, the compound ita quod = ‘so that’ became common. The combination ita quod is abundantly attested and enjoys a stable life within Latin throughout the Medieval period. As a bound pair, they came sometimes to be spelled as a single word, itaquod.14
Employing the same phonological shifts proposed above, itaquod would pro- duce the form dacă in Romanian.15 In addition to providing an etymology for dacă from an attested form, ita quod, like dacă, is also observed to evolve in the direction of a conditional particle. Intriguingly, at virtually the same time dacă in Romanian is beginning to supplant si, Francis Bacon, in a legal discussion, compares clauses introduced by Latin ita quod and si. He notes that, while ita quod ordinarily governs a subsequent conditional clause and si a precedent conditional clause, those categories can blur:
… these words, ita quod and si, howsoever in propriety the ita quod may seem subsequent and the si precedent, yet they both bow to the sense. [In the clause si ipse vellet habitare et residens esse:] there the word si amounts to a condition subsequent, for he could not be resident before he took the state; and so via versa may ita quod be precedent, for else it must be idle and void. (Bacon 1861, 82-83)
While the forms déca, deca and deaca are indeed early, the specific form dacă is attested in the very earliest records for Romanian, appearing in a letter of Cyrillic orthography dated 1581 (Hasdeu 1878, 29).
The forms déca, deca, and deaca could themselves have been regional varia- tions. Folk etymology assuming a link between dacă and the preposition de may have resulted in a sporadic realization of the word with these alternative pronun- ciations. At any rate, the emergence of dacă as the standard form throughout all of the Romanian dialects suggests that, in addition to as great a claim for antiquity, it always enjoyed the wider distribution.
6 The invisible da in early Romanian
The demonstration that ita quod provides a possible source for Romanian dacă puts the potential etymology of da from ita on a firmer footing. But if that is so, then da resided in the language for about two hundred years before appearing in literary record. This is not, however, an uncommon phenomenon. In the case of Romanian, there had been somewhat scanty orthographic evidence for the language prior to the period in which da allegedly «appears» as a presumptive Slavic borrowing. The earliest records for the Romanian language were, by definition, from the more educated registers. If da, even as a borrowing, were a more colloquial term, it could have been a part of the language long before it first appears in writing.16 The relatively late appearance of da does not, then, exclude the possibility that the word has a Latin origin but remained unattested in Romanian until a late date.
It is unfortunate that the development of the Romanian language is so obscured by scanty record. As a result, many etymologies will never be proven conclu- sively. In the present case, I have demonstrated that phonological developments could explain how a word for yes in Latin, ita, would produce in Romanian the form da. By the same token, Romanian dacă finds a more satisfactory etymol- ogy in the Latin compound conjunction ita quod. Even though da itself does not appear until late in recorded Romanian, the word could have been a more colloquial affirmation up to that point.
It is even possible that Latin ita, preserved as da in Romanian, eventu- ally emerged as the standard affirmation in Romanian under the influence of the Balkan milieu. In this case, the Slavic word of identical phonology and semantics does not so much replace a Romanian word as draw out a legitimate Romance word from a substandard register. It is unlikely that further epigraphical evi- dence will shed any more light on the topic. But a Latin source for this basic word is at least a plausible option.
1The oldest possible reference to the Romanian language is in the History of the Byzantine Chronicler Theophylactus Simocatta (circa 630), who reports that, during a campaign against the Avars in the Balkans in 587, a local person observing the load on an animal about to fall was heard to say «torna, torna» («turn back, turn back» (or, «it’s spilling, it’s spilling», Theophylacti Simocattae Historiae, II, 15, 6-9; Sala 2005, 20).
A later Byzantine Chronicler, Theophanes the Confessor, recounts the same incident but adds the word «brother» (fratre) to the phrase. (Theophanis Chronographia, I, Anno 6079 (587), 14–19; Saramandu 2001-2002, 234). After that point, however, there exists a lacuna of several centuries. The oldest attested reference to a document in Romanian comes from a Latin title of an oath made by Stephen the Great in 1485 in which it is reported that «Haec Inscriptio ex Valachico in Latinam versa est sed Rex Ruthenica Lingua scriptam accepta» (‘This Inscription was translated from Valachic [Romanian] into Latin, But the King has received it written in the Ruthenian language [Slavic]’); Tomescu 1968, 40). The oldest surviving modern sample of Romanian is a letter written in 1521 from Neacsu of Cimpulung to Hans Benkner, a judge in Brasov (Rosetti et al. 1969, 468).
2Indeed, the borrowing of a yes-particle is attested in Slovene ja and Czech jo from the German (Elsik & Matras 2006, 429). A language may borrow foreign words for which it already has a perfectly good item because of some nuance, such as prestige (Campbell 2004, 64-65).
3It is not attested prior to the second half of the 18th century. See Dimitrescu (1978, 345); Niculescu (1961; 1965), and Vrabie (1981, 237).
4Ita and vero were also frequently used together to form a single affirmative statement, ita vero (e.g., Plautus’ Miles Gloriosus III.ii.844)
5Thus Abelard is able to title his definitive work Sic et Non (Yes and No). 6French, of course, preserves si for responses to negative questions (Andison 1923, 85). 7E.g., Latin vitricus > Rom. vitreg (Dimitrescu 1978, 177). See also Lloyd (1987, 30). 8See LaVoie (2001, 9) and Blevins (2004, 145). 9E.g., itáque (Tucker 1965, 449).
10E.g., Latin excadere > Rom. scadea; Latin excambiare > Rom. schimba. Initial î is also ordinarily dropped by elision to other elements (Andersen 1986, 553).
11E.g., Latin crassus > Romanian gras (Dimitrescu 1978, 177); Latin ∗cavula > Rom. gaura (Ciorănescu 2003, 355).
12All of which are attested in the early years of the 17th Century (Gheţie & Mareş 1974, 89). 13See Gheţie & Mareş (1974, 88) and Cihac (1879, 32).
14E.g., «Itaquod in omni eventu valeat…» (‘So that it might be valid in any event’), from a letter written in the year 1263 by Beatrice of Savoy (Viard 1942, 132-134).
15For Latin quod > Rom. că, see Cihac (1879, 32) and Ciorănescu (2003, 130).
16A demonstration by analogy can be drawn from the case of the Hebrew language. By the first century AD, a sizable corpus of recorded language had developed. But when modern archaeology added linguistic records to this corpus, virtually every new text presented words never before seen. Despite the record of approximately 2,500 personal names in the Hebrew Bible, one new inscription consisting of simply ten names presented three previously unattested (Gibson 1971, 36-37).
Andersen, Henning (1986): Sandhi Phenomena in the Languages of Europe. New York: Mouton de Gruyter.
Andison, John Gordon (1923): The Affirmative Particles in French. Toronto: University of Toronto Press.
Bacon, Francis (1861): «The Case of Revocation of Uses.» In: James Sped- ding & Douglas Denon Heath [eds.], Literary and Professional Works of Francis Bacon. London: Longman & Co. [et al.].
Blevins, Juliette (2004): Evolutionary Phonology: The Emergence of Sound Patterns. Cambridge: Cambridge University Press.
Buck, Carl Darling; Hale, William Gardner (1903): A Latin Grammar. Boston: Ginn & Company.
Bybee, Joan L. (2001): Phonology and Language Use. Cambridge: Cambridge University Press.
Campbell, Lyle (2004): Historical Linguistics: An Introduction. Cambridge (MA): MIT Press.
Cihac, Alexandru (1879): Dictionnaire d’étymologie daco-romane, élé- ments slaves, magyars, turcs, grecs-moderne et albanais. Frankfurt: Lu- dolphe St-Goar. [Available online. URL: <http://www.archive.org/ details/dictionnairedt00cihauoft>].
Ciorănescu, Alexandru (2003): Dictionarul Etimologic Al Limbii Române. Bucharest: Saeculum.
Cravens, Thomas D. (1996): «Phonology, phonetics, and orthography in Late Latin and Romance: the evidence for early intervocalic sonorization.» In: Roger Wright [ed.], Latin and the Romance Languages in the Early Middle Ages. University Park (PA): Pennsylvania State University Press, 52–68.
Dimitrescu, Florica (1978): Istoria Limbii Române. Bucharest: Editura Didactică şi Pedagogică.
Du Nay, André (1977): The Early History of the Rumanian Language. Lake Bluff (IL): Jupiter Press.
(1996): The origins of the Rumanians: the early history of the Rumanian language. Buffalo: Matthias Corvinus.
Elsik, Viktor; Matras, Yaron (2006): Markedness and Language Change: The Romani Sample. Berlin; New York: Walter de Gruyter.
Gheţie, Ion; Mareş, Alexandru (1974): Graiurile dacoromâne în secolul al xvi-lea. Bucharest: Editura Academiei Republicii Socialiste România.
Gibson, John C.L. (1971): Textbook of Syrian Semitic Inscriptions. Volume 1: Hebrew and Moabite Inscriptions. Oxford: Clarendon Press.
Hasdeu, Bogdan Petriceicu (1878): Cuvente der Vătruni. Limba Română Vorbită între 1550-1600. Bucharest.
LaVoie, Lisa M. (2001): Consonant Strength: Phonological Patterns and Phonetic Manifestations. New York: Routledge.
Lloyd, Paul M. (1987): From Latin to Spanish. Darby (PA): American Philosophical Society.
Niculescu, Alexandru (1961): «Afirmarea prin da în limba româna.» Studii şi Cercetări Lingvistice 12 (4): 479–486.
(1965): Individualitatea limbii române între limbile romanice. Bucharest: Editura Ştiinţifică.
Pucci, Joseph Michael; Harrington, Karl Pomeroy (1997): Medieval Latin. Chicago: University of Chicago Press.
Roques, Mario (1907): «Recherches sur les conditionnelles sa*, de, dacă*, en Ancien Roumain.» Romanische Forschungen 23 (2): 825–839.
Rosetti, Alexandru (1983): Études de linguistique générale. Bucharest: Editura Univers.
Rosetti, Alexandru [et al.] (1969): Istoria limbii române. Vol. II. Bucharest: Editura Academiei Române.
Sala, Marius (2005): From Latin to Romanian: The Historical Development of Romanian in a Comparative Romance Context. Jackson (MS): Romance Monographs.
Saramandu, Nicolae (2001–2002): «Torna, torna, fratre.» Fonetică şi dia- lectologie XX-XXI:233–252. [Available online. URL: <http://www.ear.ro/ 3brevist/rv8/art14.pdf>].
Tomescu, Mircea (1968): Istoria cărţii româneşti de la începuturi pînă la 1918. Bucharest: Editura Ştiinţifică.
Tucker, R. Whitney (1965): «Accentuation before Enclitics in Latin.» Transactions and Proceedings of the American Philological Association 96:449–461.
Viard, Francisque (1942): Béatrice de Savoye. Lyon.
Vrabie, Emil. (1981): «Linguistic aspects of the question of the Romanians’ continuity in Dacia: A critical study.» Revue Roumaine de Lingvistique 26:213–262. [Review of Du Nay (1977)].
Zipf, George Kingsley (1929): «Relative Frequency as a Determinant of Phonetic Change.» Harvard Studies in Classical Philology 40:1–95.