How "the case system collapses" in e.g. Latin

Question

A comment on Understanding the purpose of determiners/articles/demonstratives in language suggested that case systems break down:

For unrelated reasons, the case system collapses, so that word order becomes more rigid, which means there's pressure to find another way to solve the problem which was solved by word order before. Demonstratives are a natural candidate.

Wondering what this means, what it looks like, and why it happens.

Answer 1

As jknappen says, it came down to sound shifts in Latin.

A portion of the Latin case system (first and second declension singular) looked like this:

     Fem   Masc
Nom  -a    -us
Gen  -aj   -i:
Dat  -aj   -o:
Acc  -am   -um
Abl  -a:   -o:
Voc  -a    -e

(Note that this is IPA phonemic transcription, not Latin orthography. Basically, Nom = subject, Gen = possessive, Dat = indirect object, Acc = direct object, Abl = object of a preposition, Voc = person being spoken to directly.)

In Classical times, final /m/ started to disappear, nasalizing the previous vowel (like in French). Then in Later Latin, the fourteen Classical vowels simplified down into seven, and this nasalization vanished. Now we have:

     Fem   Masc
Nom  -a    -o
Gen  -e    -i
Dat  -e    -o
Acc  -a    -o
Abl  -a    -o
Voc  -a    -e

At this point the case system is significantly less useful! In particular, the nominative and accusative look exactly the same, so word order started to take their place: the verb was placed in the middle, with the subject before it and the object after it.

Once the cases stopped being so useful, these last few vestiges faded away too: the prepositions ad "toward" and pro "for" took the place of the dative, de "down from" took the place of the genitive, and so on. And thus you end up with their descendants in modern (Tuscan) Italian:

     Fem   Masc
Nom  -a    -o

All of this simplification happened before Late Latin/Romance split apart into what would become French, Spanish, Italian, and so on. That's why Romance languages for the most part have no cases on their nouns. But this process didn't affect personal pronouns, which were already fossilized in Classical Latin and used different inflection patterns, so those retained case.

EDIT: fdb has pointed out that Old French preserves a case distinction in third-declension nouns, a type that isn't covered here. He's completely right! My bad. So the complete loss of cases in general was after the Romance split.

EDIT 2: Romanian, Dalmation, and I think maybe Sardinian do preserve some of the cases. Mea culpa.

Answer 2

What it means: the language changes in such a way that cases become indistinguishable, so that they cannot be used to express information any more.

I can't comment on the process in Latin, but the case system of North-West Semitic collapsed in two branches, Aramaic and Canaanite (Hebrew, Phoenician and some Transjordian languages), due to sound changes. The original endings are usually reconstructed as -u, -i, -a for the singular. Both Canaanite and Aramaic dropped these final short vowels. After this sound change, the cases are indistinguishable, and grammatical functions need to be indicated through other means.

These other means to indicate grammatical functions can be various. In Latin, sentence structure became more rigid so that subject and object can be distinguished. However, in Aramaic and Hebrew, sentence structure is already determined by information status and objects can be marked with an object marker (though not obligatorily so). The Aramaic object marker developed from a preposition; the origins of the Hebrew marker are unclear.

Returning to the question why it happens: sound change continuously happens (the extent to which it influences grammar varies), as a result of noise in the language acquisition process. (But this is really the topic for a different question.)

But of course, these are also slow and partially intertwined processes. Loss of short final vowels as discussed above is possible when they are less and less clearly pronounced, which happens in part because they are less and less relevant to meaning. Thus, things as object markers are already available in some contexts, get expanded, leading to a weaker need for case endings, leading to weaker articulation, leading to more usage of those object markers, etc. The process reinforces itself.

Answer 3

If I may follow up on Keelan’s remarks on Semitic: The ancient Semitic languages distinguished three cases in the singular (with suffix /u/ for the nominative, /i/ for the genitive, /a/ for the accusative) and two each in the dual (nominative /ān/, gen./acc. /ayn/) and the plural (/ūn/ and /īn/). In many languages the loss of all final short vowels means that the cases are no longer distinguished in the singular, but case distinction disappears also in the dual and plural, where only the old oblique endings survive. This shows that the loss of cases cannot be explained entirely by the loss of endings. It is more likely that the loss of the singular case endings weakened the case system to the degree that case came to be seen as redundant in the plural as well. It is as if the speakers suddenly realised that they could survive without cases.

Answer 4

The collapse of the case system in Latin was caused by sound shifts that made the case endings indistinguishable for many nouns. Already in Classical Latin, the accusative singular endings -am and -um started to loose the /m/ (first, the final vowel was nasalised, and later this nasalisation was also lost). I haven't the other details at hand, but the whole process is well-documented. By "majority rule" most words in Modern Romance languages are derived from the oblique cases and not from the nominative.