Why does inflection always die?

[vir litterarum]

I am currently taking German, and we discussed in class today how the genitive is being replaced in spoken German in most instances by von+ dative. Why does the movement in languages such as German, Greek, and Latin always seem to be away from inflected forms towards stricter word order and preposition usage?

[Estoniacus Inoriginale]

I am not sure, but I think it is out of unrestrained habits, for better or for worse. It has been determined that whenever a language becomes simpler in terms of morphological features, its phonology will eventually get more complex. For example, Latin actually grew in the amount of vowels and consonants, while its case system and other stuff just melted away. The same with English. Some have thought that English might start to form new case-like structures, but it is not known for certain. While the complex system of verb tenses is becoming more simpler. Thusly, there may be a fluctuation between these two modes, throughout the thousands of years a language continuum may exist, that of the language maintaining morphological simplicity over phonological complexity and vice versa. However, Latin's eroding case system has been discussed before in hoc foro.

[quendidil]

It has been determined that whenever a language becomes simpler in terms of morphological features, its phonology will eventually get more complex.

I think the reverse is also true, at least for Chinese. The reconstructed phonology for Old Chinese is pretty complex, with tons of consonant clusters, different vowels althought it possibly lacked tones. The morphology on the other hand is very simple and compact, more so than in modern Mandarin.

Modern Mandarin on the other hand, has relatively simple phonology, but has become more morphologically complex and has developed some agglutinative features.

[Estoniacus Inoriginale]

Indeed, quendidil, I agree. In the case of Chinese, the complexity remanifests in tones. So the cylce in total, can be quadrupal in its nature.

[thesaurus]

Just a thought, but does it have anything to do with the dominance and effects of other languages? I'm not sure about late Latin, but perhaps German could be taking its cues from English?

[quendidil]

Just a thought, but does it have anything to do with the dominance and effects of other languages? I'm not sure about late Latin, but perhaps German could be taking its cues from English?

Influence from other languages and non-native speakers is certainly a factor to be considered, but what of the North Germanic languages? All of them except for Icelandic, Faroese and Elfdalian have about as much noun declension as English. This in spite of the fact that they've been influenced by the case-heavy Finno-Ugric languages.

Have the cases in German been artificially preserved in a manner of speaking, by the presence of a standard language? Dutch and Low German had lost most of their cases in colloquial speech by the beginning of the 20th century, are the High German colloquial dialects in a similar position?

[Didymus]

... we discussed in class today how the genitive is being replaced in spoken German in most instances by von+ dative.

Na ja, der Dativ ist dem Genitiv sein Tod.

[vir litterarum]

That's the work in particular my professor cited.

[modus.irrealis]

Inflections don't only die -- they're created too (otherwise, of course, every language would now be completely analytic). For example French (and other Romance languages) have created new inflectional futures and conditionals by reducing forms of habeo from full words to inflectional endings (cantare + habeo > chanterai, cantare + habes > chanteras, etc.) And it seems to me that that the verb inflection system of the Romance language has many more irregularities than the Latin system and therefore is less analytic.

In some cases, new inflectional forms have been created without our being aware of it because of the way we're taught to think of a language. For example, the first time I read the claim that English auxiliary verbs are inflected for negation, it seemed very strange for me, but there actually is strong evidence for the claim, and I'm now convinced it's true. Usually people think forms like "couldn't" are contractions of "could" and "not" but many of these contractions have a now completely unpredictable form (e.g. "won't", "shan't") and furthermore the contraction can occur in places where the separate words cannot (e.g. you can say "Won't you come tomorrow?" but not "Will not you come tomorrow?"), so these contractions are treated in many ways like single words and are on their way to being inflections if they're not already there.

There are even more extreme claims, like that the French verb (in the spoken language) is actually a very complicated thing that is inflected for subject, direct and indirect object, and other things, but that this is obscured by the standard orthography which reflects the spoken language of a long time ago.

So I think there is movement towards inflection as well as away from it, although in Indo-European languages, it does seem like the trend in the past few millennia has been strongly in the direction of analyticity (especially with respect to nominal inflection, but it's not as if cases have been entirely lost across the board). My own theory is that this is due to the fact that Ancient Indo-European languages do inflection in as complicated a way as possible, with e.g. the same case being realized by different endings for different nouns, case and number being mixed together in nouns, lots of suppletion. You can compare these languages to Turkish, which is also a heavily inflected language but does inflection in a very straightforward way. Personally I think it's this extra complication that motivated speakers of these languages to heavily simplify these aspects of the language.

[loqu]

My own theory is that this is due to the fact that Ancient Indo-European languages do inflection in as complicated a way as possible, with e.g. the same case being realized by different endings for different nouns, case and number being mixed together in nouns, lots of suppletion. You can compare these languages to Turkish, which is also a heavily inflected language but does inflection in a very straightforward way. Personally I think it's this extra complication that motivated speakers of these languages to heavily simplify these aspects of the language.

and this leads us to a question that is equally interesting and non-answerable: why was it like that in PIE and ancient IE languages? Where did it all come from?

[Lucus Eques]

To answer the last question above, could it be that PIE was more complex than the Finno-Uralico-Hungarian variety because, at an earlier point in time, it was as simple — but the influence of certain vowels and consonants at the ends of words created certain blendings in different forms that evolved into the 5 Latin declensions, and other varieties?

Plus, does Turkish have gender? I know Finnish and Hungarian don't — no doubt that's a BIG contributing factor to the complexity of PIE syntheticism, having three genders.

[Estoniacus Inoriginale]

I agree with Luke. Mostly, because it sounds logical, and I previously knew that Proto-Uralic has been reconstructed and therein have been found to be only 8 cases and no gender.

[annis]

Plus, does Turkish have gender? I know Finnish and Hungarian don't — no doubt that's a BIG contributing factor to the complexity of PIE syntheticism, having three genders.

Swahili has eight — but it's quite regular.

[Lucus Eques]

Plus, does Turkish have gender? I know Finnish and Hungarian don't — no doubt that's a BIG contributing factor to the complexity of PIE syntheticism, having three genders.

Swahili has eight — but it's quite regular.

Eight genders?

[calvinist]

Plus, does Turkish have gender? I know Finnish and Hungarian don't — no doubt that's a BIG contributing factor to the complexity of PIE syntheticism, having three genders.

Swahili has eight — but it's quite regular.

Eight genders?

At first glance eight genders sounds crazy, I think because we usually mix grammatical gender with natural gender (since they overlap), but if you think about gender as a purely grammatical concept, then there's really no limit to how many genders a language could have. Just as with noun declensions, verb conjugations, etc. I'm not sure what advantage there would be in having eight genders though, from a communication standpoint.

[annis]

At first glance eight genders sounds crazy, I think because we usually mix grammatical gender with natural gender (since they overlap), but if you think about gender as a purely grammatical concept, then there's really no limit to how many genders a language could have. Just as with noun declensions, verb conjugations, etc. I'm not sure what advantage there would be in having eight genders though, from a communication standpoint.

There's no need to think in terms of advantage here. Once a language comes up with a good idea, there is a very good chance of that idea spreading out into areas where it does not, at first glance, have any obvious business going. This is a normal feature of language and language change.

Once you get more than three genders one starts to speak of noun "classes" instead. Swahili has eight (give or take) and other some Bantu languages may have quite a few more. Classifier languages are quite common, actually. Classes in Swahili are a little different from gender — there is a semantic component, so you can pick different classes for the same root for related meanings — but the formal grammatical apparatus is much the same: noun-adjective agreement in class and number, etc. In the case of Swahili, the verb has to agree with the subject class, too. I'm not aware of any Indo-European language that does this, though in most Semitic languages the 2nd and 3rd person verbs are also marked for gender.

[calvinist]

There's no need to think in terms of advantage here. Once a language comes up with a good idea, there is a very good chance of that idea spreading out into areas where it does not, at first glance, have any obvious business going. This is a normal feature of language and language change.

Once you get more than three genders one starts to speak of noun "classes" instead. Swahili has eight (give or take) and other some Bantu languages may have quite a few more. Classifier languages are quite common, actually. Classes in Swahili are a little different from gender — there is a semantic component, so you can pick different classes for the same root for related meanings — but the formal grammatical apparatus is much the same: noun-adjective agreement in class and number, etc. In the case of Swahili, the verb has to agree with the subject class, too. I'm not aware of any Indo-European language that does this, though in most Semitic languages the 2nd and 3rd person verbs are also marked for gender.

Classes, yeah that makes sense. Especially when you mention the added semantic meaning. I guess it would also allow for much more freedom in the placement of noun modifiers like adjectives than in say Latin where we have three genders and the masculine/neuter share endings in most cases.

[Estoniacus Inoriginale]

Inlection dies and may re-emerge. At the same time, phonetical complexity increases and upon generation of new inflections a related phonological simplification may take place. Can this be said to be right?

[annis]

Inlection dies and may re-emerge. At the same time, phonetical complexity increases and upon generation of new inflections a related phonological simplification may take place. Can this be said to be right?

Well, it's one right scenario.

I would drop the notion of complexity from how you're thinking about this. While it is true that phonological changes are probably the main changers of grammar, you can nonetheless have pretty dramatic grammar changes which aren't motived by some phonological change muddying things up. New grammar innovations will fit into the phonology of the language at the time — whether simple or complex.

Back to the original question, for example, I doubt very much that there's some phonological change in German motivating the retreat of the genitive (unlike in ancient Greek, where the dative probably did die with the aid of some sound changes). There's a nice German word, Sprachbund, which refers to a linguistic area in which unrelated (or distantly related) languages cross-fertilize each other. For example, in many west- and central-european languages intransitivity is indicated with reflexive pronouns (je me lave, etc.). This is a highly unusual way to do this, and is common only in Europe. If I had to guess, I'd say it was a Sprachbund effect (the de of French and similar in other Romance languages) inspiring this change. In exchange, the funky German uvular /r/ is taking over parts of France and northern Italy.

[annis]

Classes, yeah that makes sense. Especially when you mention the added semantic meaning. I guess it would also allow for much more freedom in the placement of noun modifiers like adjectives than in say Latin where we have three genders and the masculine/neuter share endings in most cases.

Not all classifier languages are highly inflected. In quite a few languages the classifiers only come out to play when you modify a noun with a demonstrative pronoun or with a number (Chinese and Japanese are the classic examples of this, but many Native American languages do this as well). In other classifier languages, the classes only marked in the verb.

In fact, at the moment I cannot think of a single language that has a rich classifier system and also uses a nonconfigurational word order.

[Lucus Eques]

In exchange, the funky German uvular /r/ is taking over parts of France and northern Italy.

I thought that had originated in the Île de France and changed French before it changed German.

[Bert]

In fact, at the moment I cannot think of a single language that has a rich classifier system and also uses a nonconfigurational word order.

I can't either.
But that's because I don't know enough about languages to make a difference

[mingshey]

Plus, does Turkish have gender? I know Finnish and Hungarian don't — no doubt that's a BIG contributing factor to the complexity of PIE syntheticism, having three genders.

Swahili has eight — but it's quite regular.

Eight genders?

If I were to redefine the grammatical terms for Swahili, I'd call it "totems" rather than "genders".

[Essorant]

It is no wonder less inflections may "catch on" so well, for most people don't wish to continue overfilling their mouths with excess syllables on the ends of words when they may shave off most of such syllables and achieve saying basically the same thing, with a few adjustments.

But something such as "of" instead of a genitive in many places, is probably more just from less inflections starting to feel more "natural" to the speaker, and therefore, even such a simple inflection may become used less and less not because the genitive is not important but because the manner of using "of" now feels more "natural".

[ThomasGR]

What made me always wonder, this is not why inflections and genders die. To me it seems all natural that a language tries to get simplified and yet retain its full expressions, though in other forms, much easier to use. But why were all these complicated rules of grammar at the beginning created? It looks the older a language is, the more complicated the rules are. Today, we can live without any dual gender, so why couldn't they? Perhaps someone has an answer.

[Swth\r]

Happy new year to everyone!

Some "sine quibus non" linguistic readings about language change (Historical-evolutional linguistics)...

GRAMMATICALISATION!!!
Google ---> "grammaticalisation"
You really have to read : Hopper, Paul J., and Traugott, Elizabeth C. (1993) Grammaticalization. Cambridge University Press.

And the most usefull: Barry J. Blake: Case !!!

Östen Dahl, The Growth and Maintenance of Linguistic Complexity