Skip to main content
Loading Explore guides…
Romance Language Grammar Evolution · Acedra
Home
Explore
Languages
Romance Language Grammar Evolution
Sign in to save
Chat
If YouTube asks you to sign in, open the video on YouTube.
Watch on YouTube
Reload player
Transcript
66 segments
Copy
Hide video
0:00
Hey guys, what’s up. Several months ago, we talked about the evolution of sounds from Classical Latin to the modern Romance languages. So naturally, just like I did with the Germanic languages, now I gotta explain how grammar evolved in the Romance languages. Just to run through the organization again, I’ll go through each broad grammatical category in the same order I do in the language overviews, starting with the sentence structure. Each section will start
0:19
with how it was in Latin, and be followed by what happened afterwards in the descendant languages. I won’t talk about changes that are solely sound-based, unless it led to a particularly interesting grammatical development, as all the sound shifts were covered in the prequel. Speaking of the Romance sound shifts video, if you haven’t watched it yet, go do that now, or you’ll be missing a lot of context for this video. Back? Awesome. And just like in that video,
0:38
the proto-language will be Classical Latin, which I explained my reasoning for last time, although there was almost no one who objected to that, which was a pleasant surprise. Oh ya, and speaking of controversial decisions, I promise, I had reasons to use the supposedly “wrong” flag of Catalonia, which you can pause to read if you want. Long story short, I thought that both flags had explicit political connotations- either you support Catalan independence, or you don’t,
0:59
so I was like “go big or go home in pleasing the Catalans”. But now it seems more like the senyera is a neutral representation of Catalonia, especially considering how the yellow-and-red stripe pattern is present on all of the flags of Catalan-speaking regions. But to make it fair, this time I’ll use the flag of the Balearic Islands to represent Catalan, and I’d use the senyera in an eventual Catalan language overview. So with all that out of the way, let’s go!
1:19
So let’s start with the sentence structure of the proto-language, Classical Latin… ya, it’s completely fluid. Like, we’re talking fluid to the point where linguists can’t even agree on if there was a default word order, much less what that order was. I can tell you that Latin pretty much only used prepositions, the adpositions that go before the noun phrase, and this will come up later, but that’s about all I can tell you for certain. Subjects, verbs,
1:39
and objects were ordered in such a way that the topic, the information already present in the conversation, would go first, while the focus, the new information, follows, regardless of those words’ roles in the sentence. That being said, there appears to have been a general tendency for SOV sentences in Latin- the main disagreements are over whether this is enough to constitute a default word order. The verb itself could also move around based on many factors. As well,
2:01
adjectives could go on either side of the noun, and genitive (or “possessive”) nouns could go on either side of the main noun. So how did this whole mess end up getting organized in modern Romance languages? Well, it was probably a result of both intermixing with other European languages in the early AD years, and the fact that Latin always used prepositions pulling the languages in that direction to comply with Greenberg Universals, but the whole branch has now
2:20
shifted to SVO word order. (To quickly explain the Greenberg Universals thing, basically, languages with prepositions are more likely to be SVO due to the sentence flowing more logically in our brains that way than vice versa, so I speculate that Latin’s use of prepositions may have helped pull the syntax towards SVO in conjunction with all the European migration.) Anyway, for example,
2:39
there was the Latin sentence Clāuēs suās in turre capra dēseruit. This means “The goat left her keys in the tower”, and I organized it as OSV because I imagined it as a response to the question “Where did the goat leave her keys?”, which would make the keys the topic, and the tower the focus, while the verb is neutral. But then all the Romance languages evolved to SVO, so in Spanish, this sentence is La cabra dejó sus llaves en la torre, and in Romanian,
2:59
it’s Capra a lăsat cheile ei în turnul. All those extra small words and suffixes will get explained later, so don’t worry about that yet. But speaking of small words, Romance languages today aren’t always SVO. When the verbal object is a pronoun, it will go before the verb, making those sentences SOV. So if you replace “keys” with a pronoun, the Spanish sentence becomes La cabra las dejó en la torre, and the Romanian sentence becomes Capra le a lăsat în turnul. That being said,
3:23
Portuguese is more flexible with this rule in its entirety, including the SOV business. But aside from the basic word order, notice how in the Spanish sentence, the possessive pronoun moved to the other side of the noun, before it. This is another development that happened as Latin’s free word order got restricted- Romanian and Sardinian put the possessive pronouns after the noun, while the rest of the Romance languages put them before the noun, although to my understanding,
3:45
very formal Italian can put the pronoun after the noun as well. As for adjective placement, that was more consistent across the branch. All Romance languages now have a system where most adjectives follow the noun, but a few short and common ones precede the noun instead- with the exception of Romanian, in which all adjectives go after the noun. Now we can get onto more solid ground with the nouns. Latin nouns had two numbers, singular and
4:06
plural, three genders, masculine, feminine, and neuter, and six cases, Nominative, Accusative, Genitive, Dative, Ablative, and Vocative, although there was a marginal seventh case, the Locative, that only existed on a few nouns- otherwise, it had gotten absorbed into the Ablative case by the time of Classical Latin. A noun’s gender could pretty easily be determined by looking at the noun, as in the nominative case, masculine nouns ended with -us, feminine nouns ended with -a,
4:29
and neuter nouns ended with -um. But there were also some other suffixes that nouns could take, like -is, -es, or just -s, which are harder to tell, they could be either masculine or feminine. By the way, this group also includes nouns ending with X- you can basically see that as a combination of C and S. There were some exceptions tho. These suffixes, for the most part, also determined how the cases looked, as they were fusional and also depended on the noun class.
4:51
Latin scholars talk about five “declensions” when it comes to lining genders up with cases: The first declension is for feminine nouns ending with -a, while the second is for masculine and neuter nouns ending with -us and -um respectively. The third declension includes the -is, -es, and -s nouns, as well as irregular neuter nouns ending with -us in the nominative. If you’re wondering what in the world these declensions share, it’s the genitive singular ending -is. Then,
5:13
the fourth declension has the famous irregularly feminine noun manus, and the fifth declension is for nouns ending with -ēs with a long vowel. So what happened with this case system? Yeahhhh they’re gone. There are no modern Romance languages with a case system even remotely close to the Latin one. The only one that still has cases of any kind is Romanian, and the declensions are severely depleted. There’s just three cases- a nominative-accusative case, a genitive-dative
5:35
case descended from the Latin genitive case which has lost a lot of its functions to the nominative-accusative, and a vocative case which, interestingly, follows patterns associated with the vocative case in Slavic languages, like the feminine suffix being -o, for example, which indicates that that case was actually loaned in by the Slavic languages, and is not descended from Latin. The Latin vocative itself was actually identical to the nominative everywhere
5:56
except in regular masculine nouns. So then the nominative-accusative case was descended from the nominative, right? No of course not, that would be too logical. In a very interesting turn of events, when the Latin case system was being massively simplified across the Romance languages, they actually kept the accusative around to become the new all-purpose form of the noun. This makes more sense when you consider that for the more irregular nouns, the rest of the cases’
6:17
roots looked more like the accusative than the nominative, so the root of the accusative was generally more common than the nominative. This is why, back in the Romance sound shifts video, I had to give you the accusative form of several nouns as the original Latin form. Like, I had to give noctem for “night” instead of nox, because noctem is the form that got inherited. This also had a massive effect on the gender system- look at the masculine and neuter nouns, particularly
6:40
their accusative forms, and notice how they’re identical. Yep, this resulted in the neuter gender getting absorbed by the masculine. Most modern Romance languages just have a two-gender system of masculine and feminine, in which the masculine generally ends with -o, from -um, and the feminine ends with -a, from -am (both of the suffixes went through denasalization). But not all of the
6:58
languages have this system. You see, we talked in the sound shifts video about how French, Catalan, and Romanian systematically dropped certain sounds off the ends of words. Since Catalan and Romanian did this to a lesser degree, with just steps 1 and 3 on this list, they now largely perceive the gender system as masculine nouns ending with a consonant, while feminine nouns end with a reduced
7:16
/a/. But in French, they did even more dropping of sounds, to the point that the system in French, at least from a purely phonological standpoint, is now masculine nouns ending with a silent consonant, while feminine nouns end with a pronounced consonant… and a lot of exceptions. Of course, for French learners, the easier thing to do would be to use the orthography and see which words end with E, but even that doesn’t always work, the real rule is “what gender was that
7:38
word in Latin?”. But some of y’all may be getting impatient with me, since I said that Romanian has two genders. Well, at first it did. The neuter merging into the masculine still happened in Romanian, but then, Romanians started changing many masculine nouns, predominantly inanimate objects, to be feminine in the plural, resulting in a new neuter gender (although a lot of people think that’s not the best name for the class). And yes, Italian did the same thing to some nouns, but
7:59
the reason we don’t talk about Italian having a neuter gender is simply number- in Italian there are only a handful of masculine-singular feminine-plural nouns, but in Romanian, that class encompasses the majority of previously masculine inanimate nouns. K Klein has a video on this, and postulates a potential reason for this development at the end, which I agree with. So then that leaves us with the plural nouns- did they do the same thing as the singular nouns and keep the
8:21
accusative case to be their new form? Yes!… but only in the west. Italian and Romanian actually kept the nominative case for the plural forms- the ones that ended with -i or -e, as the accusative plural ended with -s, hence that form in the west. You can actually draw a very messy line on the map dividing the eastern and western Romance languages that splits up the -s plural from the -i/-e
8:40
plurals (with the caveat that the Gallo-Italic languages dropped the S and now just use articles to show plurality). But also, Romanian developed some entirely new plural suffixes, so where did those come from? Well, -le is actually descended from Latin -ae, just like -e is, but for those nouns, an epenthetic [l] was added, since the words’ roots ended with vowels and speakers didn’t like that. And -uri is descended from the third declension neuter plural suffix -ora, which
9:03
then underwent reanalyzation with the feminine plural to become -ure, became -uri following many feminine nouns changing their own plurals to -i, and then became more popular among neuter nouns due to yet another round of reanalyzation where that suffix became viewed as neuter. So now let’s talk about small words- articles. As you may suspect from the example sentences earlier, Latin didn’t have articles in the same way its descendants do, known as anaphoric
9:25
articles. There were demonstrative articles tho, but the only one we need to know about today is distal ille, with its full set of case declensions minus the vocative. So how did it lead to the anaphoric articles in modern Romance languages? Well as mentioned, ille, the distal demonstrative, is the article that matters to us, because in most Romance languages, it expanded its function to become the definite articles. (The exception is Sardinian- they
9:46
developed their definite articles from forms of ipse, which meant “that person or thing itself” in Latin.) But which forms did all these languages use? It followed the same pattern as the nouns, so the western languages developed their articles from accusative illōs and illās, while the eastern languages made theirs out of nominative illī and illae. Except in the case of Sardinian, which turned ipsum and ipsam into su and sa. But here’s the weird part:
10:08
We’ve been talking about small words preceding the noun phrase to this point, right? Yes!... except Romanian. When it comes to definite articles in Romanian, they took the same detour that the North Germanic languages did, and attach their articles to the noun as a suffix- -ul and -a in the nominative-accusative singular, -i and -le in the nominative-accusative plural. But since this is Romanian, we can also talk about the genitive-dative articles: They came from the
10:30
Latin genitive case just like the nouns. So Latin illī̆us became Romanian -ului in the singular, and in the plural, Latin illōrum became Romanian -lor. But it gets better with Romanian: Remember how Romanian was one of the languages that dropped -o and -u off the ends of words? This may have been due to rebracketing with the suffixed articles on masculine nouns- the -u started being seen as part of the article rather than the noun. So enough about the definite articles-
10:52
where’d the indefinite articles come from? You probably saw this coming- it’s the word for 1, ūnus. More specifically, the accusative forms ūnum for masculine and ūnam for feminine. And then for the plural indefinite articles, Spanish, Portuguese, and Catalan actually have plural forms of this word, from ūnōs and ūnās. But French and Italian instead formed partitive articles, meaning “some of the thing or things”, by combining the preposition de in French and di
11:15
in Italian with the appropriate definite article, roughly translating to “of the”. And in Romanian, they actually use a plural form of the number 1 for the genitive-dative plural- unor, from the Latin genitive plural ūnōrum, while for the singular, they use niște, meaning “some”, from the Latin phrase nesciō quid, “I don’t know what”. Uhh, interesting. Now let’s talk about the pronouns, so here’s the table of them in Latin. At first glance,
11:36
Latin had three people, two numbers, and five cases (with the nominative also being available as the vocative case). However, be aware that what I’m showing as the third person pronouns are actually just reflexive- they refer back to the subject, provided the subject is third person, which is a very interesting difference I’ve noticed between the Romance and Slavic languages. Anyway so, ya, Latin didn’t have a solid set of third person pronouns,
11:57
which will make those developments interesting, although onscreen now is an explanation of what Latin did have that some people may call “third-person pronouns”. What Latin did have tho, was an entire set of possessive pronouns separate from the main pronouns. Here’s “my”... “your”... “our”... “y’all’s”... and the reflexive one, which could be either singular or plural. They were all inflected like adjectives, so we’ll talk about that part later. But how did all of this
12:19
turn out in the modern Romance languages? As you may expect, there was a general loss of the case system. But since these are the pronouns, they actually maintained at least a subject/object distinction, meaning that the Latin nominative stayed around! So as far as developments to the first and second person nominative pronouns, one of them was that, for the plural pronouns, Spanish and Catalan added respectively -otros and -altres to the ends of the subject pronouns,
12:40
meaning “others”. And because that’s an adjective, in Spanish that now makes the pronoun gendered, where nosotros and vosotros are masculine, while nosotras and vosotras are feminine. Note that Catalan still has grammatical gender, it just doesn’t make a difference in this particular situation. So then what came of the confusing third person? For this, we get to revisit our old friend ille from the article section, because it and its other forms also evolved into the
13:02
third person pronouns in all modern Romance languages! And since I’ve introduced that, now we can talk about object forms. Instead of four object cases, Romance languages today largely have two sets of object pronouns: Conjunctive, the verbal objects, and disjunctive, the prepositional objects. Since prepositional objects tend to be more stressed than verbal objects, and in Latin they were often not accusative, this has led to the modern verbal objects developing
13:24
from the Latin accusative case, and the modern prepositional objects developing from the Latin dative case. As for some exceptions: Catalan only does this for first person singular, the French disjunctive pronouns just developed from a more stressed form of accusative mē, and the same is true of Romanian… but just for the accusative. Romanian actually retained four distinct cases on the pronouns, not simplifying them like they did to the nouns. Speaking of that, the Romance
13:46
languages that generally lost all the cases actually did keep an accusative-dative distinction in the third person pronouns, and Romanian has a stressed and unstressed form for both of them. But also in Italian, the third person disjunctive pronouns lui, lei, and loro, displaced the subject pronouns! So now that all that groundwork is set, we can talk about honorifics. The first important
14:05
thing that happened was that, in probably the 300s AD, Latin speakers started acting like the emperor was multiple people to symbolize power, so if you were addressing him, you’d use a plural pronoun, and even weirder, if he was talking about himself, he’d say nos instead of ego. This system, at least for the second person, got expanded to any respectful context, and we now know it as
14:23
T-V distinction, named after the typical initials of the second-person singular and plural pronouns in the Indo-European languages that use it. It’s still the system today in French and partially Catalan. But… why partially? Because Catalan, along with the rest of the Romance languages minus French, started out doing regular T-V distinction, but then they combined the formal plural version of the pronoun with some noun to make it even more honorific, and since the pronoun was now
14:46
etymologically a regular noun, its verbs are conjugated for third person. In Catalan, where this form is even more formal than the other form, plus Spanish and Portuguese, the phrase is “your mercy”, while in Romanian, it’s “your lordship”. Italian in the past could have either “your lordship” or “your magnificence”, but they took the further step of replacing it with a pronoun, so… lei, “she”, since both those words in Italian are feminine. And finally for the pronouns, we got
15:09
the possessive pronouns. They pretty much behave like adjectives, so we’ll address that evolution soon. There is one interesting part tho- the reflexive pronoun suus actually expanded its role, and became the possessive pronoun for all of the third person. This includes the third person plural in Portuguese, Spanish, and Catalan, while in the non-Iberian Peninsula Romance languages I’ve studied, it’s just singular, and the plural possessive pronoun is an evolved form of illōrum.
15:32
Alright, it’s time to run through some adjective declensions. Latin adjectives had the same three genders, two numbers, and six cases as the nouns (including the vocative distinction), and all the endings were identical to those of the nouns! It’s a super convenient system, if you ask me. Also, the same declension patterns found on nouns are on adjectives too, so not all of them differentiate masculine and feminine. There were also kind of comparative and superlative suffixes for
15:53
adjectives. The comparative, which meant “more” of the adjective, used the suffix -ior, while the superlative, which could either mean “most” of the adjective or just “very”, used the suffix -issimus. Both of the suffixes were inflected for everything a normal adjective would be, of course. And Latin also had an adverb suffix- long ē. So how’d all this turn out today? Well, the adjectives did exactly the same things the nouns did- they just took the accusative and
16:15
ran with it, plus that famous east-west schism with the nominative plural, except Romanian, which also made the genitive into a combined genitive-dative case, you know the drill. The new gender rules I described for Catalan, Romanian, and especially French, also apply here. However, we do have a bit of a loss. The comparative suffix -ior is no longer productive in the Romance languages- it’s only stuck around on particularly common adjectives. The superlative
16:36
suffix -issimus is still productive, just not in Romanian, but it has been inherited by the languages further west. And finally, we have a major change on the adverbial suffix: long ē disappeared. However, the languages west of Romanian developed a new feature to replace it; they created phrases with the noun mentem, the accusative form of mēns, meaning “mind”, effectively making adverbs “of a {blank} mind”. It should be mentioned that this innovation was
16:58
already in its early stages by Classical Latin. And since that noun happens to be feminine, it’s now the feminine form of the adjective that goes into the compound adverb phrase. So now we get to the entirely novel section of verbs, and… damn they’re complicated. As is normal in the historical verb sections, prepare for me to oversimplify a lot. Latin verbs were conjugated for three people and two numbers, but no gender, a fairly pedestrian start. But
17:20
as for the wider categories, let’s first say that there were four moods: Indicative, Subjunctive, Imperative, and Infinitive. Yes, there were, in fact, six infinitives, each one representing a tense. So then what were the tenses? The tenses represented by the infinitive are present, perfect (meaning “has happened”), and future. And then they were split by voice, with some active (the subject does the verb), and some passive (the subject has the verb done to it). So ya,
17:42
two of each tense in the infinitive. But in the indicative mood active voice, there are actually more tenses: The imperfect (something that was ongoing or habitual in the past), the pluperfect (happened before something else in the past), and the future perfect (will have happened before something in the future). As for verb classes, those could largely be determined by looking at the active present infinitive form of the verb. There were four of them: The first conjugation,
18:03
as it’s known to linguists, ended with -āre, the second conjugation ended with -ēre, the third conjugation also ended with -ere, but that e was short and the stress was on the root syllable before it, and the fourth conjugation ended with -īre. So then how did all this turn out in the modern languages? Picking up with the verb classes, most languages slightly simplified the four-class system to a three-class system of -ar, -er, and -ir, or something along those lines.
18:25
French actually changed each suffix slightly to respectively -er, -ir, and -re, but here’s where the Latin classes actually matter: In French, the original Latin -ēre verbs with penultimate stress became French -ir verbs, like how Latin tenēre became French tenir. But the -ere verbs with antepenultimate stress became French -re verbs, as the penultimate /ɛ/ vowel got reduced out of existence, like how Latin prēndere became French prendre. French also, along with Italian,
18:48
developed plenty of irregular verbs with regard to classes, but I’m not gonna talk about all that here. But Romanian did a lot with verb classes. First off, the -re endings from the Latin infinitive are now used in Romanian as the verbal noun- in the actual infinitives, they’re basically just cut off. Therefore, Romanian infinitives have the basic tripartite split of -a, -e, and -i, but they also added a new ending, -î, mostly from Slavic loanwords,
19:10
but also at least one Latin verb, horrēre, evolved into Romanian as urî. So now about conjugations. The good news is that all the passive conjugations just… disappeared. Now to form a passive verb, you have to put the passive participle after the copula verb, so like “It is done”. In Latin, that’s fit, but in French, that’s C’est fait. As for the tenses, a lot of them will go missing as well, although this is more language-dependent. The future perfect, for one,
19:33
is gone, while the pluperfect is gone everywhere except Portuguese, and now I need to introduce the “have” perfect. Basically, a lot of languages in Europe, including all Romance languages, developed a new way to express the perfect tenses- they use a conjugated form of the verb meaning “have”, normally descended from Latin habēre, followed by the passive participle of the action verb.
19:52
The “have” verb can be conjugated for any tense, so to indicate the pluperfect, a Romance language other than Portuguese would put that verb in the imperfect tense. This relates to an interesting thing that happened with the Latin perfect tense, which describes things that happened in a defined time in the past. It survived intact in Spanish and Portuguese, where it’s now known as the preterite. But in French, Italian, and most of Romanian, the present perfect with “have”
20:14
fully displaced it- that’s now how you’d say that something happened once. Meanwhile in Catalan, they developed something called the periphrastic past, which is where they conjugate the verb anar, meaning “go”, in the present tense, followed by the action verb in the infinitive. And ya, this makes very little intuitive sense, especially considering how a lot of other languages use that exact construction optionally to express the near future, and it’s the main way to express the
20:35
future in Romanian. Speaking of, the verb habēre was also used to create a new future tense outside of Romanian, replacing the old one from Latin. In this form, the languages took the infinitive verb, and added the appropriate present tense conjugation of their verb descended from habēre as a suffix. But there’s more: Those languages can also put the habēre suffix into the imperfect, and when you do that, it creates an entirely new mood- the conditional,
20:59
which expresses the verb hypothetically happening under certain conditions. So now let’s keep talking about infinitive verbs, because Portuguese made two fascinating innovations with them. First, they added personal suffixes to the end of infinitive verbs in some situations, creating the personal infinitive, which is usable as a substitute for the subjunctive mood in some places. However, then they actually took those same innovated conjugations and made
21:20
them subjunctive, creating the subjunctive future tense, which didn’t exist in Latin. But now let’s take a moment of silence for the other tenses the Romance languages have lost. The one other big change to address with all verbs is that French stopped doing pronoun dropping because, due to dropping too many sounds off of words, they now need the subject pronoun to tell who’s doing the verb. So now a bit towards the end here, something else that primarily affects
21:40
French: negation. To negate a verb in Latin, you’d put the negative particle nōn before the verb, and this has largely been inherited by the modern Romance languages. But French had to be different. The particle evolved to ne, and still goes before the verb, but you also have to put what’s called a “negative adverbial” after the verb. The most common negative adverbial is pas, which means “step”, so you’re basically saying “not a step” when you negate a verb. And because of this
22:02
construction, you can even drop ne in more casual speech. So the final thing to talk about with Romance verbs is the copulas, the verbs meaning “be”. And I say “verbs” plural because Spanish, Portuguese, and Catalan have two. The original Latin copula verb was esse, and those three languages now use it for “be” in a more permanent manner. For the more temporary manner of “be”, they repurposed Latin stāre, which meant “stand”. That verb is also used as somewhat of a copula in
22:25
Italian, but specifically for the purpose of the compound progressive aspect. Oh ya, so that’s a thing in Italian plus the Iberian languages. I don’t wanna talk too much about it tho, it’s been a long verb section. The interesting thing with these two verbs is that the French infinitive actually combined them into être. So what about the other conjugations? Well, they’re pretty much intact from Latin. There are various forms beginning with E, S, or F, but unfortunately,
22:45
it wouldn’t make sense to talk about that here, as that’s an issue of evolution from PIE. NVNC DĪCAM IN LATĪNŌ MODO QVĀSDAM SENTENTIĀS. QVAMQVAM FACILE PAVLVM PLVS POSSVM DĪCERE- PRŌ HĀC PICTV̄RĀ MOVENTE FĒLĪX SVM OB QVANTVM SCRĪPTVM EX LINGVĀ ORĪGINĀLĪ HABĒMVS, EXIMIĒ PRŌ VERBĪS PRŌ SPECIFICĪS THEMATIBVS. NVMQVAM POTVERIT PICTV̄RA MOVĒNS DĒ ĒVOLV̄TIŌNE GERMĀNICĀ. SED LATĪNVM NŌN STVDVERAM VĒRĒ, QVOD POTERIT PROBLĒMA FORE QVIA INSICVT GERMĀNICVM,
23:15
AGNŌSCŌ VT MVLTĪ VIDENTĒS MEĪ LATĪNVM STVDVERANT, ET MĒ IV̄DICANT. So that’s it, thanks for watching, and I’ll see y’all next time.
Here to help you learn
Sign in to keep a private conversation with your course tutor.
Ask me anything about the material…
Sign in to chat