Language complexity

Trawick   Mon Nov 21, 2005 3:01 pm GMT
Much in the way that evolutionists try to determine the conditions that led to human intelligence, how have linguists determined what conditions lead to a given language's complexity or simplicity?

One thing that I find quite interesting is that many primitive languages are often exceedingly more complicated than those of Western cultures. Like Navajo, for instance. Each verb in Navajo has 990 conjugations, based upon their disjunct prefix, conjunct prefix, followed by one of 11 stems which classifies the verb based upon the noun. And then, of course, each of those stems actually has 3 different forms, based upon the manner of movement of the given action. Once you've figured that out, it's pretty much smooth sailing--oh yeah, except for the fact that you have to learn how to rank the nouns in a given sentence based on their level of "animacy." And the little issue of pronunciation, what with words changing their meaning completely when you forget to aspirate, labialize, nasalize or use the correct tone for a given sound.

Is it perhaps BECAUSE some groups have so little contact with outsiders that their language is able to develop into something so forbiddingly complex?
Brennus   Mon Nov 21, 2005 11:15 pm GMT
One would expect *primitive languages to have a simple Dick and Jane or Me Tarzan-You-Jane quality to them but the truth is almost the opposite. The history of languages shows that they tend to start out complex with lots of flexional luggage and then become simpler over time, slowly shedding their flexional endings. There is evidence that Chinese had inflected forms at one time too.

Usually the phonolgy or sound system of a language also undergoes simplification thus Old English Hlaf weard (Bread Guardian) became 'Lord' in Modern English and Latin 'masculus' became macho in Spanish and mâle in French.


*Primitive (I'm not ashamed to use that word even though it's another one that political correctness is giving a hard time these days).

Re: "Is it perhaps BECAUSE some groups have so little contact with outsiders that their language is able to develop into something so forbiddingly complex? " --- Trawik

Probably yes. Languages isolated from the scientific community, Gaelic, Icelandic, Eskimo even Russian historically have remained more complex.
greg   Tue Nov 22, 2005 1:38 am GMT
Brennus : en quoi la perte des désinences ou terminaisons est-elle un signe de simplification quand, simultanément, l'agencement syntaxique obéit à des lois plus complexes et le vocabulaire disponible est multiplié par 50 voire 100 ?

La comparaison entre La <masculus> et Fr <mâle> est tout sauf un exemple de simplification, contrairement aux apparences. Le processus qui a conduit à un tel résultat est le produit de variations graphiques et phonologiques complexes sur une période d'au moins 1000 ans. D'autre part l'existence de paires telles que Fr <mâle> / Fr <maculin> (avec de surcroît spécialisation sémantique) obère sérieusement tes considération sur la "simplification" des langues.

Je ne prends pas d'exemple en anglais : d'autres s'en chargeront certainement.
greg   Tue Nov 22, 2005 1:39 am GMT
Erratum : l'existence de paires telles que Fr <mâle> / Fr <masculin>.
Gjones2   Tue Nov 22, 2005 5:38 am GMT
In the following posts I'm assuming that there's such a thing as unnecessary inflections, and that reducing them does make a language simpler. It's true, though, that sometimes other kinds of complexity arise when inflections are lost. Take the problem of how to convey the information that we convey in subject pronouns. There's no significant difference in simplicity between a language that doesn't use a subject pronoun but conveys that information with an ending on the verb, and a language that uses a subject pronoun but doesn't change the verb. 'I search' and 'search-I' [or 'searchi' -- made-up inflection] would be equally simple. Both show that searching takes place and who does the searching.

Still, there are many unnecessarily complex inflections in natural inflectional languages. Once we have 'search' in English for the present tense, we don't really need 'searches'. It would be simpler to say I, you, he, she, it, we, they search. Why have to learn 'searches' for he, she, and it? This inflectional complexity is unnecessary.

French has even more. It uses subject pronouns but also has many present tense endings. Chercher (to search): Je cherche, tu cherches, il [etc.] cherche, nous cherchons, vous cherchez, ils [etc.] cherchent. Also different verbs belong to different conjugations. Besides the -er conjugation, there's the -ir, and -re, each with many different endings, and of course there are verbs that are irregular and don't fit in a conjugation.

Compare Esperanto with them. It uses subject pronouns and has only a single verb ending for each tense. No worrying about changing the ending to fit the subject pronouns, and no worrying about different kinds of conjugations. A movement towards this kind of sparse, minimized inflection would clearly be a movement towards simplicity, and that has taken place to some degree in many languages.
Gjones2   Tue Nov 22, 2005 5:49 am GMT
In the "I, you, me, they etc (always so short words?)" discussion, most of us were pointing out how the languages that we knew tended to have short personal pronouns (and usually short words for most other things that are said frequently). Some persons, though, were able to cite some examples that didn't fit this pattern. For instance, Japanese has relatively long personal pronouns. Here's what I said about evolution there.

"I'd expect, based on Darwin's theory of natural selection, that primitive peoples who had effective words of warning (about predators, for instance) would tend to survive at a greater rate. Whether this would still be an important factor with modern languages I don't know. The survival value of a bird's alarm call is obvious. With complex human societies, though -- especially modern ones -- it's hard to isolate what contributes to survival and what doesn't." http://www.antimoon.com/forum/t1051.htm

As long as a language isn't so complex as to prevent its being learned competently by children as an ordinary part of their maturation, then differences in grammatical complexity may not have a significant effect on survival value. If I recall correctly, Navahos served as American radio operators during World War II because their language hadn't been described in academic publications read by Axis scholars. Yet despite its inflectional complexity it served to communicate in life-or-death situations in which communication had to take place quickly and effectively.
Gjones2   Tue Nov 22, 2005 6:03 am GMT
>The history of languages shows that they tend to start out complex with lots of flexional luggage and then become simpler over time, slowly shedding their flexional endings. [Brennus]

I'm accepting the assumption that many languages seem to be losing their inflections, but one thing we need to bear in mind is that recorded "history" probably only covers a relatively small part of the time that language has existed. Language almost certainly has been around for at least tens of thousands of years. Some scientists suggest that a likely time for its origin would be around two million years ago when human brain size had a significant increase. The brains of other surviving primates don't seem capable of handling language beyond a relatively primitive level. This leads writers about evolution to suggest that the parts of the brain that handle language had to arise concurrently with language itself. Language couldn't advance very far without the selection of brains that could handle it, and there would be no selective value in having those brains without the use of language. (The chicken and the egg had to develop together, so to speak.)

So the languages of "primitive" peoples now are only the languages of peoples not advanced in terms of technology and other qualities usually associated with civilization (big cities, that sort of thing). They aren't anywhere close to the "primitive" beginnings of language itself. We can't rely on history to tell us anything about those beginnings. All we can do is speculate.

Despite our not having been there when it happened, it seems to me almost certain that the original development of language had to move from simple to more complex, starting with simple signs like the other primates use or even creatures like birds use. I can understand how gradually more and more signs (words) could be added by human innovators and learned by the rest of the speech community (to the degree that their brains allowed them at that time). But if some fellow on the African veldt or in a cave somewhere suddenly started speaking like a Navaho with -- as Travis puts it -- "990 conjugations, based upon their disjunct prefix, conjunct prefix, followed by one of 11 stems which classifies the verb based upon the noun. And then, of course, each of those stems actually has 3 different forms, based upon the manner of movement of the given action", how would anybody possibly be able to understand him? :-)

That couldn't have happened. I think we can safely assume that language started simple and became more complex, so this apparent evolution from complex to simple within historical times can't have been going on for a million years or more, not even for 40,000 years.
Gjones2   Tue Nov 22, 2005 6:34 am GMT
>Is it perhaps BECAUSE some groups have so little contact with outsiders that their language is able to develop into something so forbiddingly complex? [Travis]

As you suggest, isolation may have something to do with it. (I haven't read anything about this subject recently, and I'm just speculating.) After all, what has been the key characteristic of historical times? A great increase in population compared with pre-historical times and less isolation. In pre-history nearly all human beings were part of small hunting and gathering groups that were scattered throughout the large areas needed to support this kind of subsistence. (On the other hand, hunter-gatherers and pastoral peoples were often nomadic and moved from place to place within those sparsely populated areas, so sometimes they'd meet other peoples. Also some trade among peoples took place even in stone-age times.) Agriculture produced a great increase in the population. Because of it large groups of people could live close together.

History (which arose with the invention of writing) primarily covers the time after that rise of agriculture, though not every agricultural people had developed it when writers of history encountered them. And what do we observe during historical times? In many languages we see a gradual decrease in complex inflections, a decrease that couldn't have started 40,000 or a million years ago, else we wouldn't have had such complex languages at the start of historical times (unless such changes are cyclical).

So the connection between evolution towards grammatical simplicity and the rise of large human communities seems to fit. What remains to explain, though, is why. Why should farmers or city dwellers who are used to seeing more people use fewer inflections than hunters and gatherers (assuming that they are more isolated)? What's really causing this loss of inflections? I don't know much about pidgin languages but assume that they tend to be inflectionally simple. Is that related to what we're seeing? The major languages that are getting more simple have usually had some contact and influence from other languages. I don't know, though, if the process that produces pidgin languages is sufficient to explain what seems to be happening to the languages of very large groups of people.

I haven't thought this through well enough to be confident about what took place or is taking place.
Gjones2   Tue Nov 22, 2005 7:09 am GMT
"Some scientists suggest that a likely time for its origin would be around two million years ago when human brain size had a significant increase." I should have said 'pre-human' instead of 'human' (though if they really spoke, maybe we'd count them as human). These early creatures weren't the human beings that we call 'Homo sapiens'. They were hominids within the same genus. Also the increase in brain size could be related to walking erect and using the hand for many new purposes.

Even if this extremely early date for language is mistaken, Homo sapiens has been around for hundreds of thousands of years -- physically about the same as now, and evidence of burial ceremonies and art goes back tens of thousands of years. These aren't likely to have arisen suddenly among non-speaking creatures. They probably indicate a complex human society that has traditions preserved in language.
Xatufan   Thu Nov 24, 2005 12:42 am GMT
If I had learned Navajo as my mother tongue, would it be easier to learn other languages, like Spanish or French or Romanian or whatever, because they are more simple?

I think the answer is NO: Romanian, Basque or whatever language would look really complex to me: they have other kinds of difficulty. Perhaps Navajo does not have words for modern things like "television" or "space ship", so if I had to learn Interlingua I'd have to learn lots of new concepts.
Brennus   Thu Nov 24, 2005 9:33 am GMT
Greg,

Dans la linguistique, on appelle mes exemples "les formes réduites". Un autre exemple d'une forme réduite dans l'anglais est "hussy". Aujourd'hui cela signifie "une femme adultère ou prostituée" mais à l'origine il a signifié "la ménagère" <vieux anglais hus +wif. Le français est plein des formes réduites. Un autre exemple est "On dit" (Ong dee) du latin Homo dixt; aussi "soleil" (sol-ay) du soliculus.

Les formes réduites de mots sont certainement plus simples que leur formes ancestrales. Je crois que nous pouvons y être d'accord.
Le mot "mâle" vient de la forme parlée de latins (le latin vulgaire) *masc (u) lu mais "masculin" est cela que les linguistes appellent un "mot étudié ." 'Victoria' dans l'espagnol et 'victorie' dans le roumain sont aussi les "mots étudiés ". La forme latin vulgaire aurait produit *vechoria dans l'espagnol et *viptorie dans le roumain.

In linguistics, my examples are called "reduced forms". Another example of a reduced form in English is "hussy". Today it means an "adulterous woman" or "prostitute" but originally it meant "housewife" < Old English hus +wif. French is full of reduced forms. Another example is "On dit" (Ong dee) from Latin Homo dixt; also "soleil" (sol-ay) from soliculus.

Reduced forms of words are certainly more simple than their ancestral forms. I think we can agree on that.

"Male" comes from the spoken form of Latin (Vulgar Latin) *masc(u)lu but " masculin" is what linguists call a "learned word." 'Victoria' in Spanish and 'victorie' in Romanian are also "learned words". The Vulgar Latin form would have produced *vechoria in Spanish and *viptorie in Romanian.
Brennus   Thu Nov 24, 2005 9:47 am GMT
Gjones,

You leave much to think about. In short however I would say that languages steadily improve over time. Thus Greek is an improvement over Sanskrit, Latin is an improvement over Greek and both Spanish and Modern English are improvements over Latin. It takes a certain amount of sophistication to speak a language which gets by with a minimum amount of grammar like Chinese, Vietnamese and (almost) Modern English and the Scandinavian languages (excepting Icelandic). Therefore, its no wonder that the oldest languages in the world - living and dead - tend to be very convoluted and complex.
Brennus   Thu Nov 24, 2005 9:55 am GMT
Xatufan,

All languages have some difficulties but I'm inclined to think that if you spoke a language as structurally complicated as Navajo or Basque (They may even be distantly related to each other) any of the modern European languages would be a pushover for you. In Latin America, as you probably know, the Spanish conquistadors rarely learned Indian languages; it was much easier for them to have the Indians they captured or conquered learn Spanish instead because Spanish was considerably easier to learn.
greg   Thu Nov 24, 2005 4:27 pm GMT
Brennus,

Je te propose d’examiner l’évolution de <soliculus> [solikilus] jusqu’à <soleil> [solEj] telle que retracée par Nelly Andrieux-Reix, une spécialiste reconnue de l’ancien français :

<soliculus> [solikulus] — latin parlé / vulgaire : époque royale — langue écrite : latin archaïque
<soliculus> [solikulu] — latin parlé / vulgaire : époque républicaine — langue écrite : latin classique
<soliculus> [soleklu] — latin parlé / vulgaire : Ier-IIIe siècle ap. J.-C. — langue écrite : latin impérial
<soliculus> [solekju]¹ — latin parlé / vulgaire : IIIe-Ve s. — langue écrite : bas latin
<soliculus> [soleLo] — roman commun puis gallo-roman : mi Ve s. — langue écrite : latin médiéval
<soleil> <solel> [soleL] — protofrançais puis ancien français : VIIIe-XIe s. — langue écrite : protofrançais puis ancien français
<soleil> <soleill> [soleL] / <souleil> <souleill> [suleL] — ancien français puis moyen français : XIe-XVIe s. — langue écrite : ancien français puis moyen français
<soleil> [soleL] — moyen français puis français classique : XVIe-XVIIe s. — langue écrite : moyen français puis français classique
<soleil> [solEj] — français classique puis français contemporain : XVIe-XVIIe s. — langue écrite : français classique puis français contemporain

Et encore je me suis abstenu de faire figurer d’autres formes telles que <soulaiz>, <soleuz>, <solaux>, <solez>, <soleux>, <soloy>, <soleilh>, <solleil> etc².

¹On retrouve un stade phonétique voisin de [solekju] dans It <solecchio> — <far solecchio> = “farsi schermo con la mano agli occhi per impedire di essere abbagliati dalla luce solare” = “se proyéger les yeux de la main pour éviter d’être aveuglé par la lumière du soleil” — et dans It <solicchio> — <preferisco prendere una bella boccata d'aria sotto il bel solicchio di questi giorni> = “je préfère prendre l’air sous ce beau soleil” = “je préfère prendre un bon bol d’air par cette belle journée ensoleillée”.

²Tu retrouves une variété de formes en catalan moderne — <solei> <soley> <solell> — mais la signification n’est pas “soleil” exactement.

Le moins qu’on puisse dire c’est que réduction syllabique ne rime pas avec simplification car plusieurs orthographes et prononciations se concurrencent. D’autre part on constate que la réduction de 4 à 2 syllabes est acquise bien avant l’apparition de l’ancien français : le phénomène de réduction (syllabique) auquel tu faisais allusion concerne donc la latin et le roman et non le français car, dans cette langue, le nombre de syllabes est resté identique de 850 à 2005. Si on considère l’hésitation entre [so] & [su] pendant plusieurs siècles, d’une part, et la transformation de [leL] en [lEj] au début de l’époque moderne, d’autre part, on voit bien que l’étymon se modifie sans réduction quantitative : il s’agit d’un phénomène de complexification et non de simplification.

Pour répondre à ta remarque, si l’on prend l’ancien français <solel> [soleL] pour forme ancestrale du français <soleil> [solEj], non, je ne suis pas d’accord pour dire que la forme actuelle est plus simple — ni même plus réduite d’ailleurs — que la forme ancestrale.
Quand au gallo-roman <soliculus> [soleLo] ou latin vulgaire tardif <soliculus> [solekju], ils sont certainement plus réduits que le <soliculus> [solikulus] archaïque, mais pourquoi plus simples ?

Quand à Fr <on dit>, une telle combinaison sujet/prédicat ne peut être dérivée de la combinaison correspondante La <homo dixt> pour une bonne et simple raison : les deux combinaisons n’ont pas du tout la même valeur sémantique car la syntaxe du français et du latin diffèrent totalement. En principe, le français nécessite l’utilisation de pronom personnels — tels que <on> — alors que s’en passer est la norme en latin. De plus La <homo> est un substantif et non un pronom à l’instar de Fr <on>.
Par conséquent, même s’il est vrai d’affirmer que Fr <on> dérive de La <homme> et Fr <dit> de La <dixt> (on retrouve AF <dixt> d’ailleurs), il est totalement spécieux, à mon avis, de mettre Fr <on dit> et la <homo dixt> en équivalence.

PS : les “mots étudiés” dont tu parlais sont en fait des « formes savantes ».
greg   Thu Nov 24, 2005 8:18 pm GMT
Précision :

— An (adj) <learned> [l3:nId] = Fr (adj) <savant> [savÃ], <instruit> [E~stRH], <érudit> [eRydi], <docte> [dOkt], <cultivé> [kyltive], <intellectuel> [E~telEktHEl] / [E~telEktyEl], <éminent> [eminÃ] etc

— An (pp) <learned> [l3:nd] / <learnt> [l3:nt] = Fr (pp) <appris> [apRi], <étudié> [etydje], <acquis> [aki] etc.



Ainsi An <learned terms> [l3:nId] (inkhorn terms) = Fr <mots savants> ; Fr <mots étudiés> = An <words learned> / <words learnt> / <learned words> [l3:nd] / <learnt words>.