When and how did French become a non-null-subject language?

Question

First of all, what does "null-subject" mean? Taken from the Wikipedia page for "Null-subject languages":

[…] a null-subject language is a language whose grammar permits an independent clause to lack an explicit subject. Such a clause is then said to have a null subject.

Therefore, according to this, a non-null-subject language is a language that must have a subject. This can be expressed explicitly or through a pronoun and dummy or not¹ doesn't matter.

Most Romance languages, for example when telling the weather (the sentence is "It rains") work without an explicit subject (1-5), but French is a (notable) exception and it doesn't allow sentences without expressing the subject (6), although in colloquial/oral usage, such sentences are allowed (7, 8):

1. Piove. (Italian)
2. Llueve. (Spanish)
3. Plou. (Catalan)
4. Chove. (Portuguese)
5. Plouǎ. (Romanian)
6. Il pleut. (French)
7. (Je) m'en fiche. — (I) don't care.
8. L'amour, (je) connais pas. — Love? (I) don't know.

_{Examples 7,8 are taken from Grevisse, Le Bon Usage, §234d.}

I'm not aware of other Romance languages having the same behavior. I chose to treat the most relevant ones, but if you're aware of them, feel free to share them in your answer. Although I must say I'm more interested in French, if those have some deep relation to the evolution of the French language, then they might acquire importance.

I researched a bit before asking and I found an article titled "On Null subjects and related phenomena" by Asya Pereltsvaig, which treats about this phenomenon, comparing English and Italian. Apparently it is only slightly related to this question, but it does mention something that can be interesting:

[...] «Not only that, but historically, French used to be a well-behaved Romance language like Italian, Spanish, Romanian and Latin, allowing null-subjects (or missing subjects). What also becomes clear from the historical record is that French switched in all the relevant respects — at the same time!»

This article is hinting (well, more than hinting) to the fact that French used to work like other Romance languages, such as Spanish and Italian, but at a certain point in time, it changed.

If this is true, I can't prove it because I have nothing about the old phases of French evolution, so what I'm asking is: when and how did French switched and became a language that doesn't allow null-subject in independent clauses?

---

^{1: A dummy pronoun is used when there isn't an actual subject but it's nevertheless syntactically required by a given language. The usual example is "it rains": It's not that "something" rains, but English sentences can't stand without an expressed subject.}

Answer 1

A classical grammar book indicates a rough timeline and presents the phonetic deletion of verbal endings as the main cause of this change.

« Au Moyen Âge, le pronom sujet faisait ordinairement défaut, parce que les terminaisons verbales, étant encore sonores, indiquaient suffisamment les personnes grammaticales : Que ferai donc ? (Eneas, 8729). — Ma chiere amie, que avez ? (Béroul, Tristan, 3175) — Où est ? (Fiore et Blancheflor, 676) — Feras ? (Pathelin, 1390) — Peu à peu le pronom sujet s'est imposé dans la conjugaison. Il était encore assez souvent omis au XVIe s. : Direz, chantant mes vers, en vous esmerveillant / [...] (Ronsard, t. XVII, p. 265) — Au XVIIe s., cette omission était un archaïsme, quoiqu'elle fût fréquente encore dans les réponses : Et le vais voir tantôt (Mol., Ét., V, 8). — Leur ai dit la langueur (La F., F., VIII, 3). — Non ferai, de par tous les diables (Mol., Av., V, 3) »

(from Grevisse, Le Bon Usage, a (quite expensive) book that cannot be too recommended for any advanced student [or speaker] of French, §667; the English-speaking calls it a "prescriptive grammar", but that's absurd, as the title already shows).

Rough translation, without the quotes: “During the Middle Ages, the subject pronoun was usually lacking, because the verbal endings, still pronounced, were sufficient to show the grammatical person. Bit by bit, the pronoun subject made his way in the conjugation. It was still often omitted in the 16th century. In the 17th century, this omission was already an archaism, still frequent yet in answers.” (Of course, Mol. is Molière and La F. is La Fontaine)

Answer 2

Wikipedia mentions the superstrate/substrate influence of Germanic languages as a possible reason.

The syntax shows the systematic presence of a subject pronoun in front of the verb, as in the Germanic languages: je vois, tu vois, il voit, while the subject pronoun is optional – function of the parameter pro-drop – in the other Romance languages (as in veo, ves, ve).

[...]

French is noticeably different from most other Romance languages. Some of the changes have been attributed to substrate influence—i.e. to carry-over effects from Gaulish (Celtic) or superstrate—influence from Frankish (Germanic).

The same article also mentions that the subject became mandatory around 1500 (this is unsourced, though)

Another possible reason maybe has something to do that due to sound changes, in spoken form the verb conjugation is not so obvious. Normally only the nous and vous forms sound different. If you can't tell the conjugated verbs apart, I guess the subject becomes necessary.

See also the paper: From Old French to the Theory of Pro-Drop (thanks Alex B. for the link!)

Answer 3

Kaiser (2009) is an analysis of Brazilian Portuguese (BP) considering whether it is in the process of becoming a non-null subject language. Kaiser compares developments in BP with historical changes in French as it became a non-null subject language through the course of the Middle Ages and reports on studies of French bible translations that show this occurring over the period from 1170 to around 1600.

The changes which Kaiser reports as being thought to be connected with the loss of the null subject property in French are:

1. decrease in verbal inflection
   Middle French lost its rich verbal agreement morphology in favour of a system where most inflectional endings were reduced to schwa (lost in Modern French). This has led to Modern French having no agreement marking in the present singular (except for être and avoir).

2. decrease of verb second effects

3. emergence of subject clitic pronouns
   This process occurred simultaneous to the loss of null subjects.

4. emergence of lexical expletive pronouns
   It has been claimed that there is a strong correlation between null subject and a lack of expletive subject pronouns; this is observed in the development of French, where lexical expletive pronouns arise at the same time as the null subject property is lost.

While there remains some debate as to whether French is truly non-null subject (based around analyses of its two sets of pronouns) it is generally accepted that it is. The only other Romance language that is accepted as non-null subject is Swiss Romansh. The author concludes that BP is possibly losing its null subject property but this process is incomplete and it currently retains this property.

Answer 4

There are several reasons for that, most of them were already mentioned.

1) Homophony of conjugated verb forms, mostly in present tense. Therefore, to identify the subject, people needed to express it by pronoun. Nowadays, some linguist even postulate that the pronoun became part of the predicate.

2) French is the first Romance language to have appeared as distinct from Latin. French is different from other Romance languages in many ways and the most distinct from Latin, so it has many particular characteristics.

3) The influence of German and other Germanic languages. French is the northernmost Romance language and neither German nor English are pro-drop languages. The influence of German could have been seen (mostly in the past) in many other phenomena like e.g. the word order: Hier je fus.. > Hier fus je.. (like Gestern bin ich..) or up till now in phonetics: no other Romance language has a [y] (ü) sound

Answer 5

I believe that some linguists interpret contemporary informal French as a null-subject language, mainly because the phonemic stature of French subject pronouns has diminished over time. A paper by Elly van Gelderen (The Subject Cycle: Linguistic Change and Cognitive Principles, [2007]) perhaps suggests as much. If this is the case, and as the source language of French (namely, Latin) was a null-subject language, is an unorthodox analysis of intermediate forms of French at all possible: to wit, French has never been a non-null-subject language, as English is? I in no way have the expertise to answer this question, but I do not believe that standard prescriptivist accounts of French grammar are indications in themselves of the non-null-subject state of the language. Possibly only the historical record can determine the validity of this unorthodox assertion. That is, examples of subject-less clauses within the historical corpus of unstylized, informal French might affirm the thesis, and examples of unstylized clauses wherein a free morpheme intervenes between subject pronoun and verb phrase (full syntactic configurability assumed) might disprove the thesis.