The Consequences of Language: How does one Examine Language as a Historical Product? Chapter 7 Page XXX
The Consequences of Language
Chapter 7: How does one Examine Language As A Historical Product?
This chapter presents various approaches to the study of language change. Comparative linguistics , historical reconstruction, glottal chronology, areal linguistics, language and history..
1. Historical Background
As we noted in chapter 3, the beginning of the 20th Century marked a shift in focus from diachronic synchronic linguistics. There we presented the synchronic approach to analyzing language structure. In this chapter, we outline the major lines of the historical (diachronic) investigation of language structure. We note in passing that while it was Saussure who urged the new focus on the synchronic study of language, Saussure began his career as a very successful diachronic linguist, having published a very well received treatise on the vowel system of Proto-Indo-European.
2. Historical Linguistics
Historical or diachronic linguistics examines languages change. Underlying this approach is the family tree hypothesis: that when speakers of the same language separate into two or more groups, each group will independently introduce changes into their variety of the language and the two varieties will gradually become more and more different from each other to a point where they are no longer mutually intelligible.
This view suggests that many of the world languages are historically related to a common ancestor known as the proto-language. The group of languages descendent from the protolanguages is termed a language family. One of the first language families identified by historical linguists, Indo European is shown in the diagram above. The discovery of Indo-European demonstrated that languages from different parts of the world (Europe and South Asia) had a common ancestors.[1]
Some scholars reasoned that if some of the world’s languages were related, is it not possible that all of the world’s languages can be traced to a common ancestor language, a “proto-world”? While an intriguing possibility, we have yet to develop techniques or discover evidence that support a proto-world hypothesis.[2]
3 Methods for Studying Language Relatedness
Two general methods are currently employed for the study of language family. The first, termed lexicostatistics, has to do with comparing the relatedness of the vocabulary of a group of language. The comparative reconstruction attempts to reconstruct earlier forms of the language based on the differences found among a set of related languages.
Lexicostatistics looks for the degree of relatedness in a group of languages. It is based on the concept of basic vocabulary that presumes that the more fundamental the meaning of a vocabulary item is, the more likely it will 1) be common to all languages and 2) more important, that it will be more likely to come from the parent language than to be borrowed from another language. Thus, the more basic vocabulary two languages share, the more closely they are related.
The following example comes from Joseph Greenberg (1977:99) who is credited with providing the developing the modern the classification of African languages using this lexico statistical technique. In looking at the example, it is necessary only to see similarities and differences. One does not need to know how to pronounce these words. Notice, that in a large number of the languages the word for ‘three’ is very similar (tri, dri, tre, until we come to the last four languages. This single set separates the Indo-European languages from the Altaic languages (Finish, Estonian and Hungarian). Not all cases are this straightforward and Greenberg and other historical linguists would argue that the entire set of basic vocabulary needs to be examined before determining degree of relatedness. Note that the word for ‘two’ supports the Indo-European hypothesis as well.
Basque is a language spoken in the mountains of northern Spain and southern France. Interestingly, none of the words shows an obvious similarity with those of other words in the example. While numerous hypotheses have been suggested as to what language family Basque might belong, most linguists believe that Basque is an isolate with no genetic relationship to any language family.
LANGUAGE ONE TWO THREE HEAD NOSE MOUTH TOOTH
Breton ünan dau tri penn fri genu dant
Irish öön d tri kjan srn bjal fjaklj
Welsh in dai- tri pen truin keg dant
Danish en too? tre hoodh ns mon tan?
Sweedish en tvo tre hüvud näsa mun tand
Dutch een tvee drii hooft nöös mont tant
English wn tuw ri hed nowz maw- tuw
German ajns tsvaj draj kopf naaze muant tsaan
French œ~/yn dö trwa tt ne buu dã
Italian uno/-a due tre tsta naso bokka dnte
Spanish un/-a dos tres kabesa naso boka dnte
Rumanian un doj trej kap nas gur dinte
Albanian n' du tre kok hund goja dbmp
Greek enas dhjo tris kefali miti stoma dbondi
Lithuanian vienas du tris galva nosis burna dantis
Latvian viens divi triis galva deguns mute zobs
Polish jeden dva ci glova nos usta zöb
Czech jeden dva tri glava nos usta zup
Russian adjin dva trji galava nos rot zup
Bulgarian edin dva tri galava nos usta zeb
Serbo-Croatian jedan dva tri glava nos usta zub
Finish üksi kaksi kolme pa naenä suu hammas
Estonian üks kaks kolm pea nmina suu hammas
Hungarian ed keet haarom fö orr saaj fog
Basque bat bi hirür bürü südür aho orts
European correspondences from Greenberg (1957:42)
This reflects the fact that the first 21 languages belong to the Indo-European language family, while the last four do not. Furthermore, Finish, Estonian and Hungarian share a good deal of basic vocabulary among themselves, reflecting their membership in another language family known as Altaic.
Lexicostatistics can also help to identify branches within Indo-European (above diagram) Note the similarity of the word forms of Danish, Swedish, Dutch, English, and German for mouth (mon, mun, mont, mawand mawnt) are more closely related to each other than they are to the rest of the group. This set helps to identify the branch of Indo-European known as Germanic (see above diagram) and reflects an innovation in the common ancestor of these languages called proto-Germanic.[3] In a like way, the word for ‘head’ (galva, glova, etc.) helps to identify the Slavic branch (Lithuanian, Latvian, Polish, etc). Note that the example does not lead to the identification of those languages that have descended from Latin, known as the Romance languages (French, Spanish, Italian and Rumanian). This is why a larger basic vocabulary set is required to make a more accurate determination.
Comparative Reconstruction
During the 18th and 19th centuries, historical linguists developed the method of comparative reconstruction. This methodology seeks to reconstruct earlier forms of the language family, and is based on the principle that the phonemic correspondences between languages are systematic and not haphazard. This means that if a PIE /d/ changes to /t/ in Germanic, in one word, it will do so in all other PIE words containing /d/ as well, given the same conditions..
Grimms Law / Proto Indo-European / Proto Germanicvoiceless stops
voiced stops
voiceless fricatives / *p, *t, *k
*b, *d, *g
*f, *s, *x / f, s x (>h)
p, t, k
v, d,
Note that the Germanic words for ‘two’ and ‘tooth’ begin with a /t/, whereas elsewhere it is typically a /d/. This was actually part of a larger systematic sound change, known as Grimm’s Law, where Proto Indo-European voiced consonants (*b, *d, *g)[4] became voiceless stops (/p, t, k/) in Germanic. The regularity of such laws helps to explain why Germanic words: father, hemp and heart’, correspond to the Latin -pater, the Greek - kannabis, and kardia where Grimm’s Law did not apply.
The goal of comparative reconstruction was not only to establish the relatedness between languages, but to show how these languages were related by identifying the systematic phonological changes, which they called sound laws, by which they evolved. This process also involved reconstructing principled, albeit hypothetical, characterizations of earlier forms of these languages (called protolanguages) including the presumed ancestor of all the related languages.
Proto Indo-European was the first such family to be reconstructed, owing in part to the greater familiarity the European linguists had with these languages and to the abundance of ancient written texts to assist with the reconstruction. This technique has proven itself so successful that it has now been applied to all of the world’s other language families. In our example above, this ancestral language is called Proto Indo-European. The dramatic results of this work, which was aided both by the great familiarity with these languages and the large number of ancient written texts which these historical linguistics could use to test their methods of reconstruction.
Although these reconstructions are hypothetical, that is why they are preceded with an asterisk, these scholars were able to compare their reconstructions with early written documents. Many of these documents confirmed this methodology, while others led to an improvement in the methodology. In the case of reconstructing Proto-Romance, the ancestor of Spanish, French, Italian and Rumanian, linguists discovered that their results more closely resembled vulgar Latin rather than Classical Latin. Vulgar, in this context, carries an older meaning of ‘common’ rather than the modern meaning of ‘crude’ and ‘disgusting.’ Hence Vulgar Latin is merely the Latin of ordinary, rather than the elite Romans.
In other parts of the world, especially where there were few ancient documents, linguists began with the sound-meaning comparison of basic vocabulary. Any such work dealing with such comparisons and which assumes that the more shared basic vocabulary, the more closely related the languages is called lexicostatistics. Joseph Greenberg, building on the work of Dietrich Westermann, used this technique establishing the relationships among the languages of Africa. He identified four major, independent language families: Afro-Asiatic, which includes Hausa, Amharic, Arabic and Hebrew, Niger-Congo, which includes several subfamilies including Mande and Benue-Congo (of which Swahili is a member), Nilotic and Koi-San. The following table shows the same kind of basic-vocabulary comparison used in the Indo-European languages.
Internal Reconstruction is a method of reconstructing an earlier form of the language based on the irregularities found in the current language. In chapter 3, we introduced the various allomorphs of the plural morpheme /-s/. Internal reconstruction assumes that the allomorphs, [-s, -z, and z/ arouse through the phonological processes described in that chapter and that in an earlier stage of the language there was a single invariant form {*-s}.[5]
Glottal Chronology, a technique developed by Morris Swaddesh (1950) takes lexicostatistics one step further by assuming that the rate of change of basic vocabulary was constant. This assumption enables Swaddesh to claim not only that the more basic vocabulary two languages share, the more closely related are phylogentetically. But Swaddesh could also claim how long ago they separated from a common ancestor. While not all linguists accept the assumptions of Swaddesh’s hypothesis, the theory has been shown to produce results remarkably consistent with other historical information (Eheret 19xx).
Most linguists will agree that while lexicostatistics and glottal chronology are very useful in establishing relationships, especially for languages that have been separated for a long time. The best way of showing this relationship is through comparative reconstruction. However, it should be noted that comparative reconstruction requires a much richer knowledge of the languages involved, far beyond the minimal knowledge of the languages’ basic vocabulary and for this reason, glottal chronology and lexicostatistics are often used initially.
4. Areal Linguistics
The lexico-statistical comparison of languages often reveals another very interesting phenomenon, that languages spoken in the same geographic area often exhibit similar properties, syntactical, lexical and phonological. In my own work comparing the basic structural features Southwestern Mande, I found several such areal phenomena.
Different forms for the word ‘mother’ in Mande and Non-Mande lgs.Prototype / Mande / Non-Mande
(n)ye / Mende nja/ye Bandi nje/ye
Loko nje/ye / Gola yee
(n)de / Kono nde Dan de
Lorma dde Wen de
Kpelle nee Mwe ne
Mano le / Guro ne Kru di
Grebo de Bassa ne
ba / Bamana ba Vai ba
(n)du / Kuwaa nu Guere dou
Kissi nduaa Bele da
Koyo n Godie da
Dida n
(n)ga / Susu nga / Aizi kk
The adjacent table, shows the words for ‘mother’ in both Mande (column 2) and Non-Mande (colum 3) languages of Sierra Leone and Liberia. Each row represents a similar phonetic form. Mande shows four very different forms for ‘mother.’ This reflects the fact that there has been a good deal of innovation in the Mande languages. Furthermore, the fact that Mande shares many of these forms with non-Mande languages suggest that these groups have been interacting. A one explanation for why two languages would share the same word for ‘mother’ is intermarriage. If Mende[6] men frequently married Gola woman, their children are likely to refer to their mother in her language with the result that the Gola word enters the Mende language.
Note how different this table is from those given earlier in which similar lexical signs were seen to be an indication of a common ancestor. Greenberg noted that language families can be identified by mass comparisons, but he did not rule out other explanations, like borrowing, to explain, similar forms. The above example is such a case and represents what we call an areal phenomenon, a given linguistic feature found commonly in a geographic area and which cuts across language membership. Other such examples of this kind of phenomena involve phonological changes such as the change of /f/ > /h/[7] and grammatical phenomena such as the use of wa ‘to come’ to mark the future show an areal distribution. More widely in Africa, we see additional areal phenomena such as the distribution of serial verbs (the use of several verbs in a row), ideophones (special linguistic forms to dramatize verbal meanings), the use of the word pass to mark comparative structures (He is bigger than his brother/He passes his brother with respect to goodness) and the presence of the labiovelar consonants /kp/ and /gb/.