User blog comment:EmperorZelos/Language evolution/@comment-79749-20100511173617

My impression is that languages evolve in a kind of circle, going from simple to complex and back to simple, restarting the cycle. If we take Indoi-European as an example, we see that, in its earliest known (or theorized) forms, it seems to be just coming out of an isolating or agglutinative state, for, albeit being a flexional language, word endings are clearly identifiable. Even in Latin we can see perfectly the boundaries from one flexional element to other, if we "clean up" some obvious sound changes that must have occurred. As an example I could quote the plural accusative ending which in Classical Latin is simply -s added to the thematic vowel, but that we know, from older written records, to have come from a previous -ns. Compare then the singular accusative -m/n to the plural accusative -ns, we have an accusative element -m/n- and a plural element -s clearly identifiable. So we can imagine the highly inflectional I.-E. language coming from a more simple, agglutinative state, which is comparable to, for instance, Sumerian (there are, in fact, theories about Sumerian and I.E. having a common origin). Well, an agglutinant language can easily be traced to an isolating state, as words to be complete don't really need endings, using them only to attribute some specific function to a word. Now, coming back to the present, we see most I.-E. descendants working different levels of simplification. We go from conservative Lithuanian to "vanguardist" English or French, which now lack most of I.-E. declensions both in nominals and in verbs. English, for example, is very near to an isolating language, relying most on the use of auxiliary words rather than on endings to convey shades of meaning. Some languages, like Bengali, seem to be already one step further: new endings, which are actually particles that were once free words, are now used to fill the gap left by the loss of old inflections, making these it look more like an agglutinant than like a flexional language.

Chinese is another example. Old Chinese had rather complex syllable structures with lots of possible consonants at the end of words. Almost every word consisted in only one syllable. Now, modern Mandarin simplified so much its phonology, that an extremely high number of homophonous words appeared. The solution? Modern Mandarin has now a great number of two-syllable words, as it uses compounds to avoid ambiguity. There are also particles that are attached to "full" words, in what seems to be the seed for an agglutinant system.

Well, my point with all these examples is that there is not a universal tendency towards simplification or towards complexity. Languages are in constant evolution, and, if you have to extreme points, that means constantly going from one extreme to the other.