Glottochronology

Glottochronology ( v. Att - greek. Γλῶττα " tongue, language" and χρóνος "time" ) is the branch of lexicostatistics, dealing with temporal relationships between languages. In particular, claim the proponents of the method to calculate the time since the separation of the respective common predecessor between regarded as related languages. This is based on the assumption that the replacements would behave in a universal testing list of all languages ​​at all times so as with some for a certain period by written texts occupied examples.

• 2.1 Sociological Aspects
• 2.2 Historical and archaeological aspects
• 2.3 Linguistic Aspects
• 2.4 Mathematics and stochastic aspects

Origin and development

" The " glottochronology there is not. We distinguish between two fundamentally different assumptions about the nature of decay, as well as other additional factors in different authors to the aforementioned basic assumption addition. The results of the different approaches contradict each other largely. All calculations are based on the most careless or obsolete works zusammengescchriebenen word lists

Method of equating with the radioactive decay: " Classical glottochronology "

This direction is based on the formula of the radioactive decay. Always misunderstood again, contains this formula, that at any time all of the remaining radioactive isotopes have the same probability of decay, and hence in the same periods breaks down, the same percentage of the exponentially decreasing number of these isotopes. In the 1950s, they learned to use these laws to determine the age of radioactive material.

This inspired the American linguist Morris Swadesh, apply the method also to determine the age of languages. He set the properties of as originally suspected Wörternen his test lists which radioactive isotopes same because both yes would decrease with time. On the basis of fuzzy formulation is often overlooked that, under that law, that decreases the absolute number of decaying in successive periods original elements, and only these exponentially. In order to compare many different languages ​​as possible, designed Swadesh word lists as possible independent of culture, ie " universal" should be. The lists should also stable as possible vocabulary represent to also obtain between distantly related languages ​​still sufficient similarities. He named these lists vary, most appropriately as a universal test list, the lists were soon called Swadesh lists. " The " Swadesh list there is no way, because they Swadesh has reworked several times: starting with 200, extended to 215, ultimately reduced to 100 ( as in 1972 posthumously published). Next, there are over a dozen designs from other sites. First, calculated Lees (1953), the decay rate of 215 test words in 13 languages ​​with partially widely spaced text documents, such as old - Egyptian. 1955 checked Swadesh seven of them and compared at the same time its now changed to 100 words reduced test list.

The glottochronology soon met sharp criticism. K. Bergland and H. Vogt reported as early as 1962, after that the assumption of constant replacement rates is not tenable. J. Tischler found that were unreal separation data for the Indo-European languages ​​.

Advocate of glottochronology see the main reason in unrecognized borrowings, which was met in different ways:

• The linguist Sheila Embleton took advantage of the prior knowledge of borrowings in the Germanic languages ​​to this extrapolate quantitatively with additional algorithms, and found his way to impressive results. These results, however, range decreases only slightly over the period of the first documents. However, the complexity of their methodology and the uncertainties in the data analysis of other language families prevent today further testing.

• The ( died in 2005 ) Russian linguist Sergei Starostin Anatoljevich calculated simply with the " really important" internal improvements. He has, inter alia, placed an emphasis on the ent - Caucasian hypothesis, are problematic in the temporal references. His claim to have refuted the calculations of Bergland and Vogt, based on questionable various etymological interpretations. He grouped Albanian to Greek, but Balto -Slavic to Aryan. In the same tradition Václav Blazek (2007) advanced, not as severely impaired used word lists. S. Starostin came finally to the conclusion that instead of etymological meanings of "roots" to go out, but died in 2005 without having to pursue this approach further.

An overview of the history of research give Sh. Embleton ( 2000). and H. J. Holm (2007). Although the organizers of the conference Time Depth tried to balance, there is no professor of Indo-European studies, or comparative linguistics was an advocate of glottochronology.

All versions of this traditional approach is based on three faulty assumptions, namely, that the words of the lists as radionuclides (radioactive isotopes) (1 ) ( in the case are the words " replaced " ) all " decay " with the same probability; ( 2) only once decompose or can be replaced; ( 3) this is done in one for all about the same speed.

Method of equating it with evolutionary assumptions

Many - not at all - bioinformaticians assume a fixed rate of mutation of genes, whose numbers but - in contrast to the radioactive elements - so that does not decrease. Developed under these assumptions algorithms have been applied in recent years, also mechanistically on Swadesh and other word lists. The best known was the work of Gray & Atkinson. Despite the most modern methods and tricky modifications of the replacement rates, the result is convincing neither in time nor structurally: Time it extremely pointed in the past; structurally Albanian was erroneously grouped into Aryan, Germanic to Italic. In addition, the presentation also belies that initially gives only a non-directional graph bundle ( unrooted phylogeny ), and introduced the position of the Hittite later (Holm 2007). The team with changing investments publishes new and different pedigrees almost a year.

The variants of this " biological " approach based - in addition to the faulty data already mentioned - on two questionable assumptions, namely, that the words of the lists such genes and their alleles ( 1) with calculable probabilities " mutate " (in the case of the words " replaced " be ); ( 3) this is done in one for all about the same replacement rate, which must arise false early separations for early extinct languages ​​with many substitutions.

Basics

Sociological aspects

Language change is not based as opposed to the adoption of Glottochronologen on the effect of a perpetual motion machine, but has tangible, mostly traceable psycho- social and socio - historical reasons that are unpredictable and erratic. This statement remains true even if the effects in the above-mentioned Swadesh lists occur than in the rest ( the so-called Zipf distribution). About the findings of Bergland & Vogt also can be found for the language - historical and ethnological versed slightly more counter-examples with socio - historical reasons:

• There are, for example, languages ​​, which have long been little affected from the outside. Reasons for this are about its isolated position (so-called "conservative Saumlagen ", eg the Icelandic ) or the language influencing cultural self-confidence, for example, when the Greek against the influence of Romania.

• The range of language change further comprises a large number of intermediates and mixed languages ​​, so-called pidgins and creoles, not just today as a result of colonization and expulsion.

• Many languages ​​have become extinct even over longer or shorter periods of time, such as: the Hittite language after 1200 BC, the language of many so-called pygmy tribes due to greater external binding and economic dependence on neighboring Bantu tribes; the language of the Vedas ( Vedda ) in Sri Lanka, who have taken over the Sinhala or Tamil; Gallic after the subjugation by Caesar in present-day France and many others, all completely independent of any " rate ".

This socio- historical dependence of language change was and is emphasized by all the leading historical-comparative linguists worldwide and over again.

Historical and archaeological aspects

First of all ( see above) sparked criticism of the temporal results. Acceptable times in the rough time frame of the surveys are not evidence; the desired prehistoric results, however, are not verifiable. If necessary, the rates "adjusted", eg Starostin has changed the determined Swadesh rate of 14 % for the Indo-European languages ​​to five percent. The problem of the base times for the determination of suspected decay rates can be well seen in the section Scandinavian languages. The separation of the Icelandic can be, for example, with the settlement of Iceland in the 9th century start, but a Wegentwicklung the Norwegian begin until centuries later; especially the - depending on the author - has been essentially taken in the 300 years of Danish rule from 1536 to 1814 not " emerged as a function of time " 4-19 borrowings in the Swadeshliste but.

Linguistic Aspects

Criticism alone against the stability of certain semantic fields (see Haarmann 1990) meets in the Swadesh lists into the void, since a certain change is not disputed. Even the often (rightly ) criticized craft weaknesses of the test list does not meet the core of the method.

More important, however, the lack of linguistic quality of most Swadesh lists, freely available on the Internet, for example, the long time Dyen -list; already by Sh. Embleton 1995 objected to errors in the English part were never corrected; another twelve percent error contains the Albanian part.

In many, if not most cases, the assumed of Glottochronologen " language change " is not a change per period, but substrates remains of a previously seen before portfolio for which you have received during the acquisition of a new standard language for various reasons. It is known, for example, the maritime substrate of Germanic languages ​​(eg, mast, keel, sailing ), ie lexemes from areas where the established residents against the makers of the immigrant (here) Indo-European languages ​​had a higher pre- competence. The same applies to the art of weaving. Examples from the Swadesh list brings Aaron Dolgopolsky (2000: 401 f ), teacher of the above-mentioned S. Starostin, with the most knowledgeable in the rest of criticism from the humanities perspective.

Mathematics and stochastic aspects

Undeniably subject most languages ​​throughout its history more or less strong influences and changes. Through statistical summation itself can often grossly mismatched sums arise that are inaccurate as " rates " interpreted. When comparing many works there is a normally distributed Gaussian curve whose dimensions needs further studies.

On the assumption of a rate based on the second basic assumption of Glottochronologen, namely that languages ​​are closer related, the more common native words they have. This plausible at first sight ad hoc assumption overlooks the conditionality of three additional parameters defined ( Proportionalitätsfehler ). This works against the basic mathematical rule is violated in all glottochronologischen, namely first to analyze the stochastic distributions of the data used, in this case, the hypergeometric distribution and the broken Zipf or Pareto distribution.

Characteristic for the general assessment arranged a Indo-Europeanist the glottochronology even in the realm of science fiction.