Lexical mass comparison

The Lexical mass comparison or English Mass Lexical Comparison and Multilateral Comparison is the American linguist Joseph Greenberg developed method, relationships must be justified by a large group of languages (also known as short of Greenberg's method ). The effectiveness of this method is controversial in comparative linguistics , although it has been very successful in the classification of African languages (cf. Niger-Congo , Afro-Asian and Nilo-Saharan ).

background

Linguistics has been interested in establishing relationships between languages since its inception as a modern science . This approach is largely alien to antiquity and the Middle Ages, as languages were viewed as immutable quantities during this period. Only the idea that languages can develop and change (discovered for example by observing the development of Romance languages from Latin ) creates space for the concept of the relationship between languages that have developed from a common predecessor language. In this respect, William Jones' speech in Calcutta in 1786 was indeed a breakthrough when he recognized through language comparison that a group of languages - which were later called Indo-European - probably descended from a common predecessor language that no longer existed: … sprung from some common source .

The historical-comparative method developed later in the course of the 19th century then attempted to establish the sound laws of a language family on the basis of the relationships recognized through word and morphology comparisons and finally to partially reconstruct its proto- language. The historical-comparative method, however, was only able to reconstruct a language state that dates back a maximum of a few millennia in the past (depending on the author between 5000 and 10000 years). In addition, many researchers believe that no reconstructions are possible.

Mass Lexical Comparison is now the attempt to penetrate into greater depths of time on the question of family relationships by simultaneously comparing many languages of a larger area (e.g. Africa, America, Australia, New Guinea and neighboring islands, Northern Eurasia). On the other hand, when using this method, no reconstructions of proto-languages are carried out or even attempted.

Greenberg's method of lexical mass comparison

In Greenberg's method of lexical mass comparison , the classification results from the comparison of words and morphemes from a very large group of languages (in the case of American Indian almost all indigenous languages of America, in the case of African languages almost all documented languages on the continent). Here are word equations established and from these (i. E. Groups of related languages) the classification derived. Ultimately, it is also the method with which the researchers of the early 19th century recognized the genetic unit and the essentially correct structure of Indo-European or Finno-Ugric , long before phonetic laws were established or proto- languages were reconstructed. The establishment of sound laws and the reconstruction of proto-languages is a second step that can confirm, refine or even refute the results of the previous classification hypothesis. Greenberg usually left this second step to others.

Greenberg's method is inductive-heuristic, while the methodology of classical comparative linguistics is apparently strictly deductive. Since David Hume and certainly since Karl Popper , we know that (except the axiomatic mathematics) any science that wants to describe the real world - whether natural or spiritual science - is only inductive and not true statements but only reasonable hypotheses lead can. As a sign of the scientific nature of such hypotheses, Popper emphasized their refutability . In this sense, the relationship between languages is always a hypothesis and can never be deductively “proven”, not even through the stringent use of historical-comparative methods, for example through the establishment of regular sound correspondences . In the construction of some macro families , however, Popper's criterion of refutability was not met, which then turned them into “unscientific” hypotheses. (For the "verifiability" of genetic connections see Greenberg's article The Concept of Proof in Genetic Linguistics from 2000, printed in Greenberg 2005.)

An example of the method

In the following table 12 terms from 12 African languages are compiled to illustrate the method of “mass comparison” using a simple example (cf. Ruhlen 1994).

African word equations

No	language	me myself	you / your	four	tongue	press	speak	drink	blow	year	eye	Sun	flesh
A.	ǃKung	mi	i	ǃnani	theri	tsam	ok'xui	k a	ču	kuri	ǀga	ǀam	ǃha
B.	Duala	at the	aŋo	no	yeme	came	ambo	nyo	pep	mbu	iso	oba	nyama
C.	Dinka	gene	yin	nguan	liep	nyač	jam	dec	koth	ruon	nyin	acol	ringo
D.	Zulu	ami	akhu	no	limi	kham	amb	phuza	phepheth	nyaka	so	langa	inyama
E.	Hausa	ni	kai	fudu	harše	matsa	fada	ša	busa	šekara	I do	rana	nama
F.	ǁGana	ke	tsa	.	dam	kxao	kxoi	kxxa	gom	kuri	ǃkai	ǀam	ǀka
G	Mbundu	ame	ku	kwala	limi	came	tana	nyw	pepe	lima	iso	kumbi	situ
H	Nandi	ane	inye	angwan	ngelyep	iny	mwa	ie	kut	keny	kong	asis	peny
I.	Nama	ti	tsa	haka	nami	tsam	kxu'i	kx'a	ǃgom	kuri	mus	ǀam	kx'o
J	Swahili	mimi	ako	nne	limi	kamu	amb	nyw	pepe	aka	čo	jua	nyama
K	Bole	.	.	fhwadi	lisi	matsu'o'i	puwo'i	sawo'i	pintu	soni	idi	futi	lo
L.	Masai	well	inyi	onguan	ngejep	.	iro	mat	kut	arin	ongu	olong	kiringo

Even without prior knowledge of the relationship between these languages (and without looking at the language names), groups of related languages in the 12 languages in the example can be recognized relatively easily, even for non-experts. First of all, the languages B, D, G and J match in almost all terms. They are Bantu languages . E and K also share some similarities (they belong to Chadian , a subunit of Afro-Asian ), as do C, H, and L (they are classified as Nilo- Saharan). A closer look at the remaining three languages A, F and I reveals a number of similarities here too (e.g. for speaking , drinking , year and sun ); they belong to the Khoisan languages (the unusual characters |! and ǁ make clicking sounds are). So you can easily filter out the language units Bantu , Chadian , Nilosaharan and Khoisan from a small sample of 12 languages and 12 terms . This is an example that is simplified in many respects (number of languages, number and choice of terms, simplified phonetic representation), but it shows the principle of the method.

The similarities become particularly clear when the related languages are grouped together. However, this also shows that no consistent similarities are to be expected in all of the words in the sample; the classification can only be clearly recognized when all the matches in a group are found. A binary comparison of language pairs would not lead to the desired classification result, since the differences are the most obvious. This is the reason Greenberg has categorically rejected binary comparisons.

African word equations grouped genetically

unit	No	language	me myself	you / your	four	tongue	press	speak	drink	blow	year	eye	Sun	flesh
Bantu	B.	Duala	at the	aŋo	no	yeme	came	ambo	nyo	pep	mbu	iso	oba	nyama
	D.	Zulu	ami	akhu	no	limi	kham	amb	phuza	phepheth	nyaka	so	langa	inyama
	G	Mbundu	ame	ku	kwala	limi	came	tana	nyw	pepe	lima	iso	kumbi	situ
	J	Swahili	mimi	ako	nne	limi	kamu	amb	nyw	pepe	aka	čo	jua	nyama
Chad.	E.	Hausa	ni	kai	fudu	harše	matsa	fada	ša	busa	šekara	I do	rana	nama
	K	Bole	.	.	fhwadi	lisi	matsu'o'i	puwo'i	sawo'i	pintu	soni	idi	futi	lo
Nilosah.	C.	Dinka	gene	yin	nguan	liep	nyač	jam	dec	koth	ruon	nyin	acol	ringo
	H	Nandi	ane	inye	angwan	ngelyep	iny	mwa	ie	kut	keny	kong	asis	peny
	L.	Masai	well	inyi	onguan	ngejep	.	iro	mat	kut	arin	ongu	olong	kiringo
Khoisan	A.	ǃKung	mi	i	ǃnani	theri	tsam	ok'xui	k'a	ču	kuri	ǀga	ǀam	ǃha
	F.	ǁGana	ke	tsa	.	dam	kxao	kxoi	kxxa	gom	kuri	ǃkai	ǀam	ǀka
	I.	Nama	ti	tsa	haka	nami	tsam	kxu'i	kx'a	ǃgom	kuri	mus	ǀam	kx'o

Easily identifiable matches are marked in bold .

Practical application: success and failure

Greenberg successfully applied the mass comparison method to his now largely accepted classification of African languages , but his classification results for the American and Indo-Pacific languages are rejected by most researchers, and it is still too early to establish success or failure for Eurasian .

The method was used by Greenberg's student Merritt Ruhlen to establish so-called global etymologies (word equations of words from globally distributed languages of different language families ) and to draw the conclusion of the monogenesis of all languages on earth. (See also Proto-World Language .)

Problems of method and criticism

There are many objections to the mass lexical comparison method. An important accusation is that Greenberg's approach does not always recognize loanwords and uses such unjustifiably as evidence of the genetic relationship of languages; Similar problems are caused by onomatopoeic terms that are distributed across all languages of the world, regardless of genetic relationships. Greenberg has explicitly defended himself against these allegations several times and made it clear that he has seen the problem of loanwords and onomatopoeic word formations and has dealt with them in concrete terms. Word equations for “stable terms”, which are usually not borrowed from one language to another , but rather belong to the common original stock of a language family, were particularly important to him . According to Aharon Dolgopolsky , the “most stable” in this sense are 23 terms

me / me - two / couple - you / you - who / what - tongue / language - name - eye - heart - tooth - no / not - finger / toe nail - louse - tear - water - dead - hand - night - blood - Horn - full - sun - ear - salt

This list (in order of “stability”) is based on the study of 140 languages from different language families in Europe and Asia. It is also widely used in the investigation of distant genetic relationships.

Another problem is the fact that the lexical mass comparison ignores documented older language levels in which the phonological form of a word may deviate from the synchronously observable and thus its expressiveness is devalued. (However, the examples given by Campbell in this regard are not used by Greenberg.)

With Mass Lexical Comparison, the Americanist Lyle Campbell "proved" that Finnish was related to Amerindian and wanted to make it clear that the method does not work. Greenberg (1989) replied that he would never apply his method to languages in completely heterogeneous areas, and that he would never compare a single language (Finnish) with another language or group of languages. If one sees Finnish in the context of its Eurasian neighbors, the relationship and affiliation of Finnish to the Finno-Ugric languages becomes immediately clear (in fact, the Finno-Ugric unit was recognized even earlier than the Indo-European ; see the article Uralic languages ). In return, Greenberg shows that the application of the rigid demands of the comparative Americanists to the Indo-European languages would show the non-relationship of Hittite (Greenberg 1989).

literature

Works by Joseph Greenberg on genetic linguistics

1949–50 Studies in African Linguistic Classification. 7 parts. Southwestern Journal of Anthropology. University of New Mexico Press, Albuquerque.
1957 Essays in Linguistics. The University of Chicago Press.
1963 The Languages of Africa. Bloomington: Indiana University Press.
1971 The Indo-Pacific Hypothesis. Current Trends in Linguistics 8. (Reprinted in Greenberg 2005.)
1987 Language in the Americas. Stanford University Press.
1989 Classification of American Indian Languages: a Reply to Campbell. Language 65. (Reprinted in Greenberg 2005.)
2000 Indo-European and Its Closest Relatives: The Eurasiatic Language Family. Volume I: Grammar. Stanford University Press.
2002 Indo-European and Its Closest Relatives: The Eurasiatic Language Family. Volume II: Lexicon. Stanford University Press.
2005 Genetic Linguistics: Essays on Theory and Method. Edited by William Croft. Oxford University Press.

About Greenberg's method

Lyle Campbell : Historical Linguistics: An Introduction. MIT Press, Cambridge (Ma.) 2004.
Lyle Campbell: American Indian Languages. Oxford University Press 1997.
Merritt Ruhlen : A Guide to the World's Languages. Arnold, London 1987.
Merritt Ruhlen: On the Origin of Languages. Studies in Linguistic Taxonomy. Stanford University Press 1994.