Glottochronology

MOESAN

Elite member
Messages
5,864
Reaction score
1,281
Points
113
Location
Brittany
Ethnic group
more celtic
Y-DNA haplogroup
R1b - L21/S145*
mtDNA haplogroup
H3c
I searched a thread already open about this but I did not found it.
So, I give here under my current amateur's opinion

Glottology : accuracy for an amateur?s eye:
Already I said I have big doubts about the value of glottology. Not because the concept is of no interest, but because the evaluation of speed of languages evolution is very uncertain.
Reasons:
- phonetically, nothing proves every language changes at the same speed; the differences can depend
- on the internal structure of the language (agglutinative or not, B.I.)
- on the population speaking it on the long time, faster or slower evolution according to the pop
- on language shift, and here, the changes can be fast and steepy, before long time different rhythms of evolution after the first ?break?
- for lexicon, it depends on external contacts, trade (loanwords), and level of culture, all of them tightly tied one together, and as time passes, these factors can change, so we cannot expect a linear evolution over time
- concerning both phonetic and lexicon, their evolution depends also on internal contacts, influenced by sedentism or
nomadism on a side, and by level of culture and by economic and political organization (exchanges, centralism, media, school, religious written language ; strong sedentism in a decentralized state crumbles quickly languages: here we find the problem of criteria concerning lexicon.
So, the very concept of calculation of evolution speed (since supposed separation) implied by the term ?chrono? is already kind of a dream. But the method of comparisons between languages of cultures having known very different histories is uneasy to settle : according to the type of organization and economy are we looking for cognates only in the ?national? official languages or also in its dialects (very often sources of richness)? Are we restricting the comparisons to dictionaries entries or do we search the words with a far drift meaning but which are evident cognates? Do we pay attention to words derived from the same root as the researched cognate(s), these last ones erased today?
All that complicates very heavily a reliable foundation of a time scale based on today distances, and according to method (I avow I have not studied them) we can have very different results.


We compare often languages spoken today but written very lastly, and others written early but today disappeared without evident ?daughter? or ?son? tongue (sometimes glottochronology uses even reconstructed words). When we compare comparable things, like the Romance languages, between them, we see differences in the speed and the way of evolution: whatever the severe injuries underwent by the Latin, we see O?l French is gone farther than its ?cousin? tongues of Spain and Portugal, themselves gone farther than Italian; nevertheless these countries have been latinized at the same time, more or less, and leaved roughly at he same time by the Roman empire, Italy apart. Without to forget that the dates proposed based upon romance tongues reconstruction don?t check to precisely the real dates historically verified.


Concerning the difficulty of calculating the distances based upon lexicon :
- more than a word for a concept or a thing or a being; ex: ?horse? more than a word in more than a language -
ex: ?horse?, Ga?lic ?each?, ?marc?, ?capal?, Welsh ?march?, ?ceffyll?, Breton ?marc?h?, ?jaw?, ?ro?se?: no cognate for the IE descendant ?each?, ?epos? in Gaulish; but if you search ?foal? or ?colt?, you find Welsh ?ebol?, Breton ?ebeul? - - Slavic ?kon(j)? for ?horse?, but ?kobila? for ?mare?, Spanish ?yegua? - ?mare? itself close to ?marc? (see ?mars(c)hal?)
ex: ?head?, Germanics ?hoofd? (Dutch), ?hoved?, ?huvud? but also ?kop? (Dutch) and ?kopf? (German) and ?holle? (Frisian), but German has also ?haupt? = ?principal?, ?head-?, with equivalent in French as for ?chef? = ?leader?, ?principal?, taken its place in the original meaning by ?t?te? < < Lat. ?testa? = ?shell?
- words disappeared, but still present in ?congealed? expressions or humoristic names:
ex: ?ear?, Ga?lic ?cluas?, Welsh ?clust?, Breton ?skouarn? (Cornish ?skovarn?), but ?hare?, Welsh ?ysgyfarnog? = Breton ?skouarneg? (?long ears? +?half deaf? by derision ) -
ex: ?tail?, Welsh ?cynffon?, Breton ?lost?, but Welsh ?llostllydan? = ?afanc?, Breton ?avank? = ?beaver? but Breton ?lost ledan? = ?broad tail? -
-mot dialectaux:
ex: ?ewe?, French ?brebis?, ?oveja?, but dialectal French ?ouaille? -
?cock?, ?rooster?, French ?coq? (< Gaulish?), but dialectal ?jau? (O?l) and ?gau?, ?gal? (Oc) < < Lat. ?gallus?
?hen?, French ?poule?, but dialectal ?j?line? (O?l), ?galina? (Oc)?
&: speaking here only of presence of cognates, not of their phonetic evolution...
When a language is no more the elite?s one, its use tends to disappear of certain parts of public life and to restrict itself to smaller spheres, it tends to disappear of writings too, and it crumbles, even for basic common words because a living tongue creates new words, or uses words with close meanings in place of others, loosing synonyms: spoken Breton doesn?t share any word with Welsh for ?to speak?, presenting komz, kaoseal ?to talk?, parlantal, prezeg ?to preach?, safari? ?to make noise?.




Here we are in front of very common words, the most of them not compounds, and basic, as required by glottochronology: even some names concerning family are different, by instance in Romance languages but not only (?brother?, ?sister?, ?uncle?, ?aunt?, ?nephew?, ?niece??) ; and I don?t speak of polysemy which acts upon basic words too, diminishing the list of reliable words spite we need rather a thousand of them to make something serious - So IMO to calculate the drift or to score the differences in lexicon we should be obliged to give diverse numbers of ?points? to basic concepts according to presence vs total absence of cognates or to their presence but with derived/drift meanings or presence as roots in derived words? And this would remain accurate with certainty only to calculate distances, not timing.
Even if I?m interested, my conclusion is that we can calculate distances but not the speed of evolution, not at a precise enough level, at least, and not in an ?universal? scale for all languages. I suppose the calculations have been based for the most upon ancient languages compared to reconstructed PIE, but the evolution of our ?modern? languages shows how unsteady is the ground here. ATW ancient languages like Greek and Latin, spite written classical ones, had already undergone visible, sometimes striking changes, as not too much time went. And what ?weight? of lexicon have we for less known languages of Antiquity?
 
The most clearly useful aspect is it's a tool to put a chronologic order about some of the loans from language to language, but for absolute chronology it's rather unreliable.
 
Yes, especially glottochronology based on Bayesian methods comes off to me as rather unreliable because it makes assumptions about languages behaving more or less like biological entities that are not exactly proven correct. Besides, we have historical examples of languages that just changed faster than others, and there are also several sociocultural phenomena that intervene in the language's development either causing faster divergence or the reverse i.e. a relative re-convergence with other languages/dialects. The rate of "mutation" is much more variable and uncertain than in living beings. There's also the problem of roots that may appear like innovation, but are actually the sole remnants of a lexical variation on the same meaning that already existed since the proto-language.

It's sometimes hard to disentangle loanwords from core vocabulary, but that can theoretically be done if you pay close attention at the sound rules that should've applied if the words came directly from the earliest proto-language (that can only be done, though, for large and very studied language families with a lot of documented attestation of the languages, like Indo-European). For me a bigger problem is that languages may be closely related to others, but develop much faster than its siblings, or rather it can be much more distantly related to other languages, but through areal features and bilingualism share some of the phonological and grammatical innovations that happened in its distant cousins (or even other language families).

All in all, I think glottochronolgy may be useful if it's not taken way too seriously, and especially if you avoid relying too much on a "biological-like" methodology to explain the phylogenetics of the language family.
 

This thread has been viewed 2552 times.

Back
Top