Methodology Glottochronology

1 methodology

1.1 word list
1.2 glottochronologic constant
1.3 divergence time
1.4 results

methodology
word list

the original method presumed core vocabulary of language replaced @ constant (or constant average) rate across languages , cultures, , can therefore used measure passage of time. process makes use of list of lexical terms. lists compiled morris swadesh , assumed resistant against borrowing (originally designed in 1952 list of 200 items; however, refined 100 word list in swadesh (1955) more common among modern day linguists). core vocabulary designed encompass concepts common every human language (such personal pronouns, body parts, heavenly bodies, verbs of basic actions, numerals 1 , 2 , etc.), eliminating concepts specific particular culture or time. has been found ideal not in fact possible , meaning set may need tailored languages being compared. many alternative word lists have been compiled other linguists, using fewer meaning slots.

the percentage of cognates (words have common origin) in these word lists measured. larger percentage of cognates, more 2 languages being compared presumed have separated.

glottochronologic constant

robert lees obtained value glottochronological constant (r) of words considering known changes in 13 pairs of languages using 200 word list. obtained value of 0.805 ± 0.0176 90% confidence. 100 word list swadesh obtained value of 0.86, higher value reflecting elimination of semantically unstable words. constant may related retention rate of words by:-

l
=
2
ln
⁡
(
r
)


{\displaystyle l=2\ln(r)}

where l rate of replacement, ln logarithm base e, , r glottochronological constant

divergence time

the basic formula of glottochronology in shortest form is:-

t
=



ln
⁡
(
c
)


−
l





{\displaystyle t={\frac {\ln(c)}{-l}}}

where t = given period of time 1 stage of language another, c = proportion of wordlist items retained @ end of period, , l = rate of replacement word list.

one can therefore formulate that:

t
=
−



ln
⁡
(
c
)


2
ln
⁡
(
r
)





{\displaystyle t=-{\frac {\ln(c)}{2\ln(r)}}}

by testing historically verifiable cases have knowledge of t through non-linguistic data (e. g. approximate distance classical latin modern romance languages), swadesh arrived @ empirical value of approximately 0.14 l (meaning rate of replacement constitutes around 14 words 100-wordlist per millennium).

results

glottochronology found work in case of indo-european, accounting 87% of variance. postulated work hamito-semitic (fleming 1973), chinese (munro 1978) , amerind (stark 1973; baumhoff , olmsted 1963). latter, correlations have been obtained radiocarbon dating , blood groups archaeology.

note approach of gray , atkinson, say, has nothing glottochronology .