EvoWiki is now a project of the RationalMedia Foundation.
We are moving all content to RationalWiki.
See the EvoWiki project page for details!

Linguistic evolution

From EvoWiki

(Redirected from Linguistic Evolution)
Jump to: navigation, search

The evolution of languages, like that of religions, and unlike that of living organisms, is Lamarckian, not Darwinian. Languages evolve according to acquisition of inherited characteristics from parent to child. However, despite this basic difference between linguistic and biological evolution, there are many shared points: for example, linguistic speciation, just like biological, works on the group level (the level of society).

It might at first thought seem impossible to do that, since speakers of most languages will borrow words almost indiscriminately from other languages unless discouraged by linguistic purism. But there is a core of words that is relatively rarely borrowed, and that is words for relatively commonplace items and actions and relationships. English, though having borrowed an abundance of vocabulary from Old French and Latin and elsewhere, still has a Germanic core that can be traced back to Old English.

Languages speciate when they are no longer mutually intelligible. For instance, English and German share a common ancestor (Proto-Germanic), but their speakers no longer understand one another. The shared ancestry of languages can be determined by comparing systematics, especially in their older forms.

As an example, English, German and Swedish share the following vowel correspondence:

EnglishGermanSwedish
homeHeimhem
stoneSteinsten
goatGeissget
ropeReifrep
(al)oneeinen

This correspondence stays when we examine older Germanic tongues:

Old EnglishOld High GermanOld NorseGothic
hāmheimheimrhaims
stānsteinsteinnstains
gātgeizgeitrgaits
rāpreifreipraip
āneineinnains

Such resemblances, as well as grammatical correspondences, and sharing of core vocabulary, enables us to establish English, German and Swedish as sharing a common ancestor. The same can be shown for many other language families, and one can even see ancestor-descendant relationships in those language families with long written records, like Latin and the Romance languages. For more, see Orbis Latinus ("Latin World").

In contrast, vocabulary similarities broader than the core level are no pointer to common ancestry. English and French share much vocabulary, but once we get to Old English, the French-derived words are absent. Those are, then, loanwords from French. Ultimately, English and French are indeed related, but their common ancestor is much more remote - the ancestor that Latin and Proto-Germanic share: Proto-Indo-European.

The more distant one goes, the less trivial the sound correspondences become, but with Indo-European, they are still recognizable:

EnglishOld EnglishLatinGreekSanskritRussian
daughterdohtor(filia)thugatērduhitār-docher-
twotwāduōduōdvaudva
tentiendecemdekadaçadesyat'
toothtōþdent-odont-dant-zub
thouþūtvamty
threeþreotrestreistrayastri
bebēo-fū-phū-bhū-bu-
bearber-fer-pher-bhar-?
fatherfæderpaterpatērpitar-(otyets)
footfōtped-pod-pad-(noga)
cowbov-bou-gauhkorova
houndhundcaniskuonçvan-(sobaka)

For more, see the American Heritage Dictionary. It has discussions not only of ancestral Indo-European, but also of ancestral Semitic.

This analysis can be done on grammatical features as well. All Germanic languages distinguish between "strong" and "weak" verbs. The former have vowel shifts in their past tenses, while the latter have -d- or -t- in their past tenses. Also, most Germanic languages have parallel "strong" and "weak" adjective declensions, the latter being used for something definite.

Grammar also evolves. Old English had only two finite verb tenses, a simple present and a simple past. Both of them survive into present-day English, but present-day English also has an enormous number of compound tenses -- tenses that were invented from pre-existing words and grammatical constructions. Likewise, while the Romance languages have eliminated much or all of the noun and pronoun declensions, the Romance languages have invented new constructions of various tenses that had existed in Latin.

Ancestral Homelands

One nice additional feature of historical linguistics is that one can reconstruct what some ancestral-language speakers had known about. Although this endeavor has many pitfalls, one can work out that the ancestral IE speakers had known about several domestic and wild animals:

dog, cow, sheep, goat, pig, horse, wolf, elk

but not:

cat, lion, camel, donkey, elephant

They were also familiar with snow, and their technology included wheeled vehicles, plows, and yokes, but only very limited use of metals, no use of iron, and no writing. They did compensate for the latter deficiency by having a well-developed art of epic poetry, and they worshipped several deities, including one of thunder and war, and one called "Father Sky". Here's why they had no word for "iron"; consider the various descendant-language words:

English: iron
Old English: īren
German: Eisen
Swedish: jarn
Latin: ferrum
Greek: sideros
Sanskrit: ayas
Russian: zhelezo

The Germanic ones look related, but it looks unrelated to all the others, which look unrelated to each other. However, the Latin one may be related to English "brass", and the Greek one possibly to English "silver". This suggests that words for iron were invented as this metal became widely known -- something which only happened after the ancestral IE speakers split up.

The Indo-European homeland has been much-debated, but linguistic and archeological detective work has converged on a homeland just to the north of the Black Sea -- which is well to the north of the Middle East, the likely location of the ancestral-Semitic homeland. It is sometimes easier to work out where it was not, and such clues can be found from local vocabulary, like place names and plants and animals.

In the United States, northeastern place names include

Connecticut, Massachusetts, Pawtucket, Narragansett

while southwestern ones include

El Toro, Los Angeles, San Francisco, San Jose, Santa Clara, Santa Cruz

This means that the US once had speakers of two different languages, one in the northeast and one in the southwest. Which is indeed the case: various American Indian languages in the northeast and Spanish in the southwest. And the same reasoning can be applied to the first-recorded versions of the various Indo-European languages. A classic example is Greek, which has lots of borrowed words with -nth- and -ss- in it, and also several place-names with these elements. This indicates that Greece was once home to speakers of some languages other than Greek, and there is evidence of exactly that. Greek is first written with the Linear B script, but its predecessor, Linear A, was used to write a language which apparently was not Greek.

Likewise, a Middle-Eastern homeland can be ruled out by comparing words for some fauna familiar from there or nearby:

EnglishGreekSanskrit
lionleonsimha
elephantelephasgaja

The English words are clearly borrowed from Greek, though by way of Latin and Old French -- and the Greek and Sanskrit words look very different.

Personal tools
Namespaces
Variants
Actions
RWF
Navigation
Toolbox