Comparing Verb Conjugations in the Romance Languages
BY DEREK CZAJKA (“Languages of the World”)
Readers of this blog are doubtless aware of the notion of language families. Many of the world’s languages belong to a family of related languages. When languages are said to belong to a family, the implication is that these languages all descended from a common ancestor. Given sufficient data, it is often possible to reconstruct forms found in this common ancestor, by comparing forms found in extant descendants, and, with luck, forms attested in older languages that are now extinct. A longer explanation of this phenomenon, replete with examples, may be found, among many other places, in any basic historical linguistics textbook, any introductory linguistics textbook that contains a section on historical linguistics, or any textbook that deals at length with language families, such as the book currently in press from the principal author of this blog.
The comparative method, the technique of comparing related languages and attempting to derive possible proto forms, has proved to be an invaluable resource for those who study the historical relationships between languages. A great deal is now known about the language known to us, though presumably not to those who spoke it, as Proto-Indo-European, a language that, to our knowledge, was never written down. Similar efforts have been made, though perhaps not with quite the same amount of loving detail, with many other language families throughout the world. For all of its power, though, it is important to recognize that the comparative method has its limits. In order to demonstrate a little of both the potential and the limitations, I will examine two verbs in five Romance languages. One, the verb that means ‘to sing’, is a regular verb in all the languages shown here. The other, the verb meaning ‘to do’ or ‘to make’, has varying degrees of irregularity in the languages considered. For the sake of brevity, I will only consider the present indicative forms of these verbs. I will be looking at five Romance languages: French, Italian, Portuguese, Romanian, and Spanish. I will then include the forms of these verbs in classical Latin. Verb conjugations are courtesy of Verbix. I will provide these conjugations as they are spelled, but I will make remarks about pronunciation. The pronunciations were taken from the text-to-speech engines for the respective languages provided in the Apple iPhone and/or Google Translate. While this method of gleaning pronunciations is, perhaps, not the most scientific one, it is useful in cases where I have no direct experience with the languages in question (in the cases of Portuguese and Romanian), and is one computer-generated pronunciation made with consultation of native speakers more than linguists often have available when making linguistic comparisons.
To sing
p# French Italian Portuguese Romanian Spanish Latin
1S chante canto canto cânt canto cantō
2S chantes canti cantas cânţi cantas cantās
3S chante canta canta cântă canta cantat
1P chantons cantiamo cantamos cântăm cantamos cantā́mus
2p chantez cantate cantais? cântaţi cantáis cantā́tis
3P chantent cantano cantam cântă cantan cantant
Given the forms from the five modern languages, and asked to come up with forms from the protolanguage, you would probably come up with answers that are reasonably close to the Latin versions. There might be a few problems with vowel height in a couple forms, but these things are to be expected. However, there are a number of things that might lead you astray, and a couple things that you would miss altogether. One thing that makes the task easier is that I presented the words orthographically, and not phonetically. Had I done the latter, the task would have been a bit more difficult. In particular, French would have become all but useless to the task, since four of the six forms–all of the singular, plus the third-person plural–would resemble each other. French orthography provides some corroborating evidence for the sibilants at the end of the second-person endings and the first-person plural. It would be very difficult to account for these sibilants if they were not found in the proto-language. However, without the evidence from French orthography, a few might note the geographical closeness of Spanish and Portuguese, and draw the conclusion that something went a bit funny in Iberia.
Losing French orthography would also get rid of any chance to reconstruct the final -t in the third-person plural. As it is, none of these languages leaves behind any trace of the -t suffix in the third-person singular. In the case of the plural, one might be far more tempted by the Italian -ano, especially considering that it is, after all, pronounced. Anyone considering this option, however, would also need to consider the fact that we already have evidence that final -o in Italian also appears in Spanish and Portuguese. At the very least, a person hypothesizing a *-ano in the proto-language would do well to search for a verb in this conjugation where the stem ends in n. Another trap in Italian is the extra “i” that appears in the first-person plural. The way to avoid this one is to look at other verb conjugations, and see that -iamo seems to be a near universal suffix, even in cases where other languages have differing endings (compare, for example, Italian dormiamo and Spanish dormimos).
Another feature that is entirely absent from these forms is vowel length. In the Latin forms, long vowels are marked by macrons. All forms except the third-person forms have long vowels on the endings. However, even in a more conservativ language like Spanish, there is no difference between the vowels in cantas and in cantan that would suggest the change in vowel length in classical Latin.
To do
p# French Italian Portuguese Romanian Spanish Latin
1S fais faccio faço fac hago faciō
2S fais fai fazes faci haces facīs
3S fait fa faz face hace facit
1P faisons facciamo fazemos facem hacemos fácimus
2p faites fate fazeis? faceţi hacéis fácitis
3P font fanno fazem fac hacen fáciunt
AN obvious question that arises from the modern language forms is, just how regular was this verb in the proto-language? Italian and French give us the impression that it’s pretty irregular. Is there a consonant in the verb stem? It looks like there might be, sometimes, but only scattered through a few forms. In Spanish and Portuguese, however, the verb seems very close to a regular -er verb. It’s not quite there, but it’s close. Spanish has this slightly bizzare alternation between a voiced velar stop or fricative and a voiceless dental or alveolar fricative, but apart from that little oddity, it does what one would expect if one were familiar with the language. In Romanian, as well, the verb seems fairly ordinary. It does help us understand this rather unsettling alternation of the medial consonant in Spanish. Seems as though it may have been the result of palatalization. Observe that Romanian makes a similar alternation with /t/ in the verb ‘to sing’. Precisely how this palitalization led eventually to dental and alveolar fricatives in Spanish, French, and Portuguese is interesting, but beyond the scope of this article. In our small sample, more of the languages treat this verb as approximately regular, than very irregular, as in French and Italian. If the balance were tipped the other way, we might be tempted to attribute the more regular froms to regularization based on analogy. However, if we did this, we would have to ask why these languages opted for waht seems to be the rarer form when making the analogy. If, instead, we assume that the verb was more regular in Latin, but became irregular in French and Italian, we must come up with an explanation as to why they became so irregular. In order to answer this question, we would need more cognates, to see if this consonant disappeared in other words.
One thing to note about this verb is that the French forms shed some more light than they did for ‘to sing’. In both the third person singular and plural, we see the final -t. While in citation form, this “t” is not pronounced, it appears on occasion, especially when followed by a vowel. We also see the remnants of the intervocalic t in the second-person plural, which had vanished without a trace from ‘to sing’. Finally, French is the only language that hints at the back vowel in the third-person plural. Admittedly, it’s not a very good hint, since a front high vowel was converted to the same back close-mid vowel in the first-person singular, but it is worthy of note nonetheless.
An obvious criticism of this exercise is that I should not expect the modern Romance languages to echo classical Latin. The Roman empire was, after all, a large one, and it may be that some of the features of classical Latin, such as vowel length, or the final t in the third person, never existed at all in Gaul or Iberia or Eastern Europe. To these naysayers, I say, you’re absolutely right. However, this is one of the points of the exercise. Applying the comparative method to the Romance languages, especially if you use more than just the languages named after nation states, can tell you a lot about the historical ancestor of those languages. However, it can only take you as far as those modern languages allow. Imagine, if you will, construction Proto-Romance from only the modern Romance languages, Proto-Germanic from the extant GErmanic languages, Proto-Hellenic from modern Greek and maybe a few minority languages, and so forth. It would be much more difficult to reconstruct Proto-Indo-European. The reason we’ve been so successful at that enterprise is that we have ancient languages available, that were much more similar to each other than their modern descendants. However, this is the same reason that we have to stop comparison at Proto-Indo-European. It may be possible that Indo-European is related with other languages in its region, but it would be incredibly difficult to find them if all we have for inputs to the comparative method are outputs from a previous cycle. There is simply too much lost in transmission.
* Verbix does not list any verb forms for the second-person plural in Portuguese. This is likely because Portuguese, like many varieties of Spanish, has done away with the original forms from Latin in favor of newer pronouns. The forms shown here are from Wiktionary, and have not been confirmed in any other source.