Structural Variability of Indo-European Morphology

Structural Variability of Indo-European Morphology. By Cyril Babaev

I. The Sinusoid of the Indo-European Language's Development.

Typologically we refer the Proto-Indo-European language and the dialects which it generated to the languages of synthetic, nominative, flective structure (by morphological classification). Sometimes certain languages of the Indo-European family avoided the rule and acquired features which are not characteristic to the whole family. For instance, many of the modern Indo-Aryan languages lost the fusion system which existed in the ancient tongues of the group - Vedic and Sanskrit - and introduced elements of the agglutinative structure. The same has happened to some languages of the Iranian branch. Indo-European languages of Europe also demonstrate the collapse of their flectiveness, transforming their morphology to the analytical type, for example, English, French, Dutch. Practically all the history of development of Indo-European tongues is the history of their losing the inflections and the grammatical categories of the noun and the verb, the simplification of morphology. If we take, for example, a number of languages of the family which were spoken in the first millennium B.C. (Ancient Greek, Latin, Sanskrit, Luwian, Old Persian) then it is clear that in average the level of flectiveness was much higher at that time, than that of the contemporary tongues (English, French, Russian, Hindi, Persian, New Greek, Irish, Lithuanian). Here in this essay we will understand flectiveness (according to the Linguistic Encyclopedia 1990) as the set of certain categories of the noun and the verb, more exactly the number of flective noun case forms, types of nominal declension, gender, number, verb conjugation in person and number.

Let us turn to the chronological chart showing the dynamics of the number of noun cases in morphology of the principal Indo-European branches through history:

	Germanic	Celtic	Slavic	Baltic	Italic	Indic	Iranian	Greek	Anatolian
1000 BC					6-7	8	8	7	8
1 AD	6	7			6	6	8	5	4-5
1000	5	5	7	7	3-5		2	4
2000	1-4	1-3	1-7	5-7	1-2		2-9	4

On analyzing the chart, it becomes evident that the whole history of the Indo-European languages is the way towards simplification of the flective forms. Losing such important elements of the Indo-European system of morphology as the markers of gender, case, type of declension, has been going along with the elaboration of the analytical structure of modern Indo-European languages. Among the most typical means of replacing flective constructions with analytical ones we should call, first, merging forms into one, and second, inventing new constructions with auxiliary words.

As an example we can take the processes that took place in the nominal structure of Italic and Romance languages. The Old Latin language, known from inscriptions and texts of the 6th - the 3rd centuries B.C., uses forms of seven cases for thematic stems (see Appendix 1), has got 2 numbers and 3 genders, 5 basic types of declension. The Classical Latin period dative and locative cases merge into one, so that the dative case forms acquire another locative meaning. The Popular (or Vulgar) Latin suffered the final collapse of the case system of the Latin noun: the vocative form coincided with nominative, genitive with ablative, both replaced later by an analytical construction with the preposition de. The dative case was assimilated by accusative, preserved only in the Eastern dialects of the language. Thus, by the time the separate Romance languages were formed in Europe the noun case forms merge, and their number is decreased to two (common, or nominative, and genitive-possessive), which today are used only in Balkan Romance languages, while absent in other tongues of the group. The same can be said about the gender system: three Latin genders became two in Romance - masculine and feminine. The types of nominal declension are unified as well - there are three of them in Old French (the 12th century), none in Modern French.

In the sphere of the verb similar processes have been under way. The complicated system of Latin tenses, formed in a flective way on the basis of various verb stems, is replaced in Romance (already in Popular Latin) by analytical constructions with auxiliary verbs habere and esse.

Something like that happens in new Germanic tongues: forms either coincide with each other, or are transformed into complex constructions with auxiliary words. The Gothic language, the most ancient known Germanic dialect, presents a fully flective system, with a great variety of synthetic nominal and verbal forms, being, according to linguists, also the most archaic among the languages of the branch. The Old English (Anglo-Saxon) language already demonstrates a trend of simplifying the Indo-European inflections: in Late Old English personal verb endings merge, the remnants of the dual number disappear; in Middle English there is unification of nominal declension forms, followed by erasure of the gender category. In Modern English there are only two case forms, the common case and the possessive case, the latter formed agglutinatively, there is no gender at all, the categories of number and case are divided.

The analysis of the history of development of new Indo-European languages together with the above given chronological chart prove that absolutely all branches of the Indo-European language family have been moving towards simplification of the flective system of verbs and nouns which existed in ancient Indo-European languages.

Now let us turn to the period of the Indo-European language development that cannot be researched with the help of historical data - the period of the Proto-Indo-European language, the common ancestor of all existing languages and dialects of the family. Following world linguistics, we state that the Proto-language can be reconstructed, and therefore can be analyzed as well. How did the system of morphology develop in that common Proto-language period?

The majority of linguistic schools since the 19th century agree with the opinion that the morphology of the Late Proto-Indo-European language, right before the break-up, included the following punch of flective categories:

1. Eight noun cases (nominative, genitive, dative, accusative, instrumental, ablative, locative, vocative). This variety was seen best of all in the singular number of the thematic stems (or o-stems).
2. Three genders (masculine, feminine, neuter) with their strict opposition.
3. A certain number of types of declension, which were not always connected with the gender: á-stems, i-stems, u-stems, í-stems, ú-stems, different consonant stems and thematic o-stems.
4. Three numbers (singular, dual, plural), of which the dual number was used for only animate nouns and so-called 'pair' things (like 'eyes', 'parents', etc.).
5. Three principal forms of the verb (the present tense stem, the aorist stem, the perfect stem), the rest tenses and moods were formed using one of these three. For example, the imperative mood used the present stem, but the passive participle could only use the perfect form.
6. Three sets of verbal personal endings (the so-called "primary", "secondary" endings and the perfect endings). Moreover, primary and secondary endings varied in voice.
7. Two voices - active and medium forms. (Meillet 1938, Szemerenyi 1987, Tronsky 1960).

This list of categories - each of them expressed flectively - is characteristic for the synthetic system of the Proto-Indo-European language at the late stage of its existence. After the Proto-Indo-European community was broken and separate dialectal groups were formed, a constant process of simplification starts in the flective structure of morphology, so already in Proto-dialects, reconstructed on the basis of the most ancient historical languages (Common Celtic, Common Indo-Iranian, Common Germanic etc.) this process was going on.

But what was the Proto-Indo-European language like on the early stages of its development? Did it also possess a trend of simplifying the flective structure? If we suggest that such morphological changes had taken place for all the history of the Proto-language, then the Early Proto-Indo-European language had an even more complicated system of its grammar, with even more flective markers. This supposition cannot be true, because it is scientifically proved that the human language from the moment it appeared has become more complex, developing from the primitive stage.

In the 19th century linguists who studied different language families of the world concluded that the human language was gradually turning more and more complicated, from the isolative type to the flective one. August Schleicher, as well as his predecessors F.Schlegel and W. von Gumboldt, studying typology of the world languages came to a conclusion that the process of development for any language passes several stages from isolative (primitive) to the flective type of morphology. Later this was witnessed by the research in reconstructing the Early Proto-Indo-European language (Andreev 1986, Ioffe 1973, Tronsky 1960). Avoiding general conclusions about the laws of the human language's development, we would like to notice that this theory reflects exactly the process of prehistoric development of the Indo-European language - from affixless to flective.

Basing on the studies of the early stages of the Proto-Indo-European development the majority of scholars claim that the nominative structure with high level of flectiveness appeared in the Proto-language on its late stage. The Early Proto-Indo-European language moving from the ergative stage to the nominative structure, there was in fact the only opposition between nominative and accusative, cases of the subject and the direct object. The personal categories of the verb (person, number, voice, aspect, tense) were just beginning to form. The complex flective structure, which we reconstruct at the time of the split of the Proto-language, was forming gradually, while the morphological system was becoming more complicated. For example, it is doubtless that the dative case appeared much later than accusative; the locative case marker, in its turn, is in fact another ablaut grade of this very dative formant: *-ei / -i which used to mark both dative and locative before they broke into two. It is also clear that the feminine gender nouns appeared as an offspring of the "animate" gender, much later than all nouns were divided into two original genders: animate and neuter. The same can be said about the number of nouns: the dual number emerges later than the singular and the plural - for denoting "pairs".

Anatolian languages, as linguists state (Ivanov 1963, Bayun 1990), were the first to move apart from the Indo-European community, and therefore they preserved several important and quite archaic features in their phonetics (e.g., the laryngeal phoneme) and morphology. The most ancient of the Anatolian languages, Hittite and Luwian, do not have the dual number at all, nor have they any traces of feminine noun stems. This means that at the moment of the Anatolian exodus the Proto-Indo-European language did not use the dual number or the feminine gender of nouns. These two categories appeared later: the feminine gender as the one for vocal stems á-, í-, ú-, the dual number even later (to disappear soon in the majority of Indo-European languages: today only some Slavic tongues and Lithuanian dialects use dual).

The grounds specified above allow us to make the following conclusion:

The morphological changes of the flective structure of the Indo-European language passed two stages in its development: the stage where morphology was going more complicated since the language appeared and till it broke into dialects; and the stage where the flectiveness was going simpler since separate dialects appeared and till today.

This chain of development can be easily reflected on the following chart where point (1) marks morphology of Early Proto-Indo-European, point (2) denotes morphology of Late Proto-Indo-European, point (3) shows such modern analytical languages of the family as Afrikaans.

Chart 1

Thus, such contemporary progressive languages as English, Spanish, Persian, which completely lost their inflections, become typologically similar to the Early Proto-Indo-European language. But what is going to happen next? Will the development stop here?

One can answer this question only having in hand data of the most progressive languages of the Indo-European family, where the destruction of the inflections happened several centuries ago and where today the processes are going on further. Some of such languages are obviously New Indic and New Iranian.

Their ancestors - Avestan, Old Persian, Vedic and Sanskrit - were distinguished by their archaic morphology in comparison with other Indo-European languages of the period. But due to various conditions, and first of all the wide contacts with non-Indo-European languages, causing the penetration of alien grammar elements into them, Indo-Iranian dialects of the Middle Ages (from the 5th to the 15th century) demonstrate an almost absolute loss of inflections and the very nucleus of the fusion. Middle Indic and Middle Iranian dialects (Prakrits, Pahlavi, Sogdian etc.) acquire a lot of analytical features, therefore their morphological structure became fully analytical very early comparing to other Indo-European languages: complex verbal constructions, auxiliary words (prepositions or postpositions) instead of the nominal declension.

These are the languages (the most progressive among tem are Oriya, Bengali, Assamese, Ossetic) which today demonstrate a definite trend for acquiring new inflections - already based on postpositions. It becomes evident, therefore, that the language starts again its way towards the flective structure. The only thing is the conclusion: East Indic languages and the Ossetic language have gone through the whole cycle of their development, from the Early Proto-Indo-European primitive stage (point (1) on the chart below) via the flective system of Sanskrit (point (2)), via the analytical stage of the late Middle Ages (point (3)) and again to the formation of the new inflections.

Thus, our diagram representing the development of the Indo-European language includes now point (4) where a number of East Indic languages and Ossetic are situated at the moment:

Chart 2

and we can predict basing on the data of comparative linguistics that the diagram will further tend upwards approaching the highest point of this "sinusoid" curve - the point which will be typologically identical to point (2).

Accordingly, each of the Indo-European languages spoken on Earth today is situated at a definite point on the given diagram, in the process of moving along it. All the ancient and extinct languages of the family can also be placed in a certain position on the curve. It is interesting to note that practically all historically known Indo-European dialects are situated on the line between point (2) and point (3) while the period from (1) to (2) was passed by the Proto-language in prehistoric times. Those language which have existed for a long time in history and are still in use today clearly prove the form of the curve: e.g., Persian which has passed the distance from point (2) to point (3) for over two thousand years of its history.

The final variant of the diagram representing the theory described above is given on the following picture:

Chart 3

The dotted parts of the line here mark the reconstructed way of the sinusoid - the period of the Proto-language at the beginning and the predicted period of the future development of the language; the firm line marks the historical period of existence of Indo-European tongues since the 19th century BC (Old Hittite texts) to the present day. The points of the line denote the following examples:

point (1) - the Old Indic language of Rgveda (c. 16th century BC);
point (2) - the modern Lithuanian language;
point (3) - the modern Russian language;
point (4) - the Old English language of the Anglo-Saxon Chronicle (the 9th - the 11th centuries);
point (5) - the modern English language;
point (6) - the modern Afrikaans language;
point (7) - the modern Bengali language.

As we can see, numerous historical languages documentally state the period of simplification of the flectiveness, while the period of the formation of the new inflections is hard to confirm today by the linguistic data other than the material of the Indo-Iranian dialects. The great majority of the languages of the Indo-European family did not yet reach the lowest point of the curve (no inflection at all), some of them being still rather high on it - such archaic European languages as Sorbian, Slovene, Lithuanian. But the common thing is the direction all Indo-European languages are going on the curve of morphological changes.

The conclusion we make gives us every reason to look into the prospects of history of the language, to reconstruct the situation this or that Indo-European language will face in a certain period of time.

Let us take, for example, Russian - a comparatively archaic morphological structure. Morphology of its reconstructed ancestor - the Proto-Slavic language - is distinguished for its highly flective type, the noun declining in seven cases in practically all kinds of stems, with three genders and three numbers. Verbal tense and aspect forms are created using specific suffixes and/or endings to different stems (the present-future endings, the aorist endings, the imperfect endings). According to the majority of linguists, the Proto-Slavic period, i.e. the period of the Slavic language community, lasted approximately till the 8th or the 9th century. It was followed by the first historically known language of the Slavic group, the Old Church Slavonic language. Though it obviously belonged to the South Slavic branch, its structure is quite similar to those dialects spoken by East and West Slavs, including Old Russian dialects of the period. In Old Church Slavonic, the dual number is already reasonably weak, the system of noun inflections is being unified, analytic forms of the verb are created, like the Perfect and the Pluperfect tenses. The literature Old Russian language of the 10th - the 14th century the same processes take place, and by the beginning of the Modern Russian period the morphological structure contains only 2 numbers, all past tenses coincided with the participle -l form, the nominative case assimilated the vocative one, while the rest of cases often tend to coincided as well (e.g. put' 'way': genitive puti, dative puti, accusative put', locative puti).

This demonstrates a typical situation of simplification of inflections. It is clear that in another thousand years the Russian language will face further changes: the neuter gender will disappear, the nominal declension will be limited to two or three cases, the plural markers will also be unified, the system of the verb will become even simpler. Moreover, there is a number of languages in the Slavic group that already show some of the above changes - like Bulgarian and Macedonian, which have no nominal declension at all.

However, today we can hardly estimate chronological gaps of such transition of the Russian language into the analytical stage. The speed of language changes, as we clearly see, is different in all groups and in all languages of the Indo-European family. Why, for instance, Indic languages demonstrate such a dynamism in losing their flectiveness? On the contrary, why Baltic languages remain so archaic?

II. Speed of the Indo-European language's development.

The answer to this question is the definition of factors that influence the speed of morphological changes, the speed of the language's movement along the sinusoid of its development.

The experience of comparing various Indo-European languages demonstrates that the most progressive languages of the family are those which have had close and lengthy contacts with languages and dialects of other groups and families.

It is clear, for example, that Slavic and Baltic languages are still situated in the geographical center of the Indo-European world, having contacts just with their relative languages: like Lithuanian with Russian, German, Latvian, which have practically the same flective structure. As a result, Lithuanian preserved a highly flective morphological system. But the set of languages which had close contacts with Bengali includes Burmese, Tibetan, Dravidian languages; the neighbours of Afrikaans from the 17th century have been Zulu, Gottentote, Bushmen and Malayan languages. So, both Afrikaans and Bengali are at the moment the most analytic in the Indo-European family.

Those languages which historically appeared surrounded by alien peoples, changed really a lot, losing flective morphology: Tocharian, Ossetic, Armenian.

There is a number of examples in the history of the Indo-European language family when a sudden extension of the language contacts caused drastic changes in the structure of the language - contacts with a non-Indo-European language speeded up the process of the language's development. The example of Afrikaans is good here: together with the Dutch language Afrikaans was a descendant of Middle Dutch spoken in Holland, but despite their close relativeness the speed of development of the Dutch language in Europe surrounded by relative languages and the speed of development of the Afrikaans language in South Africa, in contact with aboriginal tongues, appeared quite different. The moment when Afrikaans became a separate language can be dated as the 17th century. 350 years passed, and the language has gone far in its development. The comparison between the Dutch language and its African counterpart demonstrates the following:

Dutch	Afrikaans
2 noun cases (common and genitive)	no cases
2 genders (common and neuter; feminine is preserved somehow in the written language)	no gender
verbs can be strong or weak; there is also the irregular conjugation	the strong verb type was lost, all verbs are conjugated as weak
verbs are conjugated in number and person	no conjugation; personal forms of the verb are equal to the infinitive

All the above stated changes happened in Afrikaans during 350 years of its independent development in close contact with non-Indo-European languages.

A sudden shift of analytization can happen, evidently, also within the language unity when several language of the region which belong to different groups or families acquire similar features due to close contacts. It does not matter in this case if some of them are flective or not: contacts speed up the language development. A typical example is the Balkan language unity which includes Greek, Albanian, Romanian, Bulgarian and Macedonian languages. These tongues belong to different Indo-European groups, but in the Early Medieval times when the unity was emerging they were all situated in different points on the curve of their development: aboriginal languages of the Balkans (Illyrian, Thracian, Dacian), the language of Romance colonists in Dacia, the Greek language of Byzantium and the dialects of Slavic immigrants. Still, in the 17th century we see the synchronization of the loss of inflections in the tongues of the Balkan unity. Bulgarian became the only Slavic language to lose the whole declension of nouns; Greek suffered drastic changes in its nominal and verbal system. The same can be said about Romanian, originally a Popular Latin dialect, and Albanian, a descendant of Illyrian.

The examples of the archaism of those languages which were isolated from communication, include Icelandic, a Germanic tongue preserved very well since the Old Norse period, while its relatives English, Dutch, Danish, situated closer to the European center and having much more contacts with the world, demonstrate dynamic processes of transition to the analytical stage.

Certainly, historical contacts of the language speakers are not the only factor which influences the changes of morphology. But the impact of the geographical position and the development of culture on the language is evident. Otherwise it would be hard to explain why different languages of the same group either suffer the destruction of the inflections or vice versa preserve it archaic.

We attempted to analyze chronologically the Indo-European dialects in order to show that even after thousands of years after the breakup of the Proto-Indo-European language community all the languages of the family still follow the law of structural development of morphology. The nominal and verbal systems examined in this essay are just a part of the complex of grammar changes which are amazingly similar in so many languages of the family, on various continents where people speak Indo-European. Otherwise how could one explain such a phenomenon as the definite article which clearly did not exist in the Proto-language but was developed independently by the majority of dialects after the community broke up: in Greek, Italic, Celtic, Germanic, Albanian, Iranian and South Slavic. Moreover, the source of its emerging in the language was the same in all of them - the Indo-European demonstrative pronoun. Such a parallel could not be explained by a coincidence or influence of one language on another.

The reasons for these common rules of the Indo-European language are still unclear. A deeper analysis is needed into the very nature of the language, along with comparison with other language families and the processes they undergo. But it appears possible that since the moment it was born the Indo-European language carries some nucleus of a general law, an algorithm of development which exists in every dialect of the family irrespective of their geographical or cultural transformations. This does not mean that the Indo-European language will preserve its nominative or flective structure forever - it can be analytical, agglutinative, ergative. But the history of its development tells us that the range of its changes will remain the same, stable as the curve above shows.

We would like to note that the processes of simplification in morphology which we observe in the majority of Indo-European tongues find their parallels both in phonetics and in syntax. But much work is still necessary to find out the laws of such changes in other spheres of the language which we did not touch in this essay. The language lives as a single organism, according to certain rules. Our task is to discover these rules, in order to be able not only to reconstruct the past of the language, but also to predict the future.

Moscow, June 2000

Reference.

1. Albretti A. Colloquial Slovene. London, New York, 1995.
2. O Sé D., Sheils J. Irish. Reading, 1993.
3. Schmalstieg W. An Old Prussian Grammar, Pennsylvania Univ., 1974
4. Aleksandravicius Y., The Lithuanian Language (Vilnius, 1984)
5. Andreev N., The Early Proto-Indo-European Language (Leningrad, 1986)
6. Arbuzova A. etc., Serbo-Croatian. Moscow, 1965.
7. Arsenyeva M., etc., Introduction to Germanic Philology (Moscow, 1980)
8. Bayun L., Experience of Comparative Studies in Anatolian Languages (Moscow, 1990)
9. Bayun L., The Language of Hittite Hieroglyphs.
10. Bernstein, S. The Builgarian Language Manual. Moscow, 1948.
11. Bourcier H. Basic Romance Linguistics. Moscow, 1952.
12. Veksler B., Jurik V., The Latvian Language. Riga, 1978.
13. Gamkrelidze T., Ivanov V., The Indo-European Language and Indo-Europeans. Tbilisi, 1984.
14. Gukhman M., The Gothic Language. Moscow, 1998.
15. Yelizarenkova T., The Vedic Language. Moscow, 1987.
16. Zhirmunsky V., History of the German Language. Moscow, 1965.
17. Zinder L., Stroeva T., Historical Morphology of the German Language. Leningrad, 1968.
18. Ivanov V., Hittite. Moscow, 1963.
19. Ivanova T., Old Church Slavonic. St. Petersbourg, 1998.
20. Ilyish B., History of the English Language. Moscow, 1968.
21. Ioffe V., Origin and Development of the Gender in Proto-Indo-European. Rostov, 1973.
22. History of the French Language. Moscow, 1963.
23. Kalygin V., The Language of Ancient Irish Poetry. Moscow, 1986.
24. Kalygin V., Korolev A., Introduction to Celtic Philology. Moscow, 1989.
25. Kochergina V., A Manual in Sanskrit. Moscow, 1994.
26. Krotovskaya Y., Kashkurevich L., A Polish Manual. Moscow, 1987.
27. The Latin Language. Moscow, 1969.
28. Linguistic Encyclopedia. Moscow, 1990.
29. Lindsey V. A Concise Historical Grammar of Latin. Moscow, 1948.
30. Lourier S. Language and Culture of Mycenaean Greece. Ooscow-Leningrad, 1957.
31. Lewis H., Pedersen H., A Concise Comparative Celtic Grammar. Moscow, 1954.
32. Meillet A. Introduction to Comparative Studies of Indo-European Languages. Moscow, 1938.
33. Meillet A. The Common Slavic Language. Moscow, 1938.
34. Miklosic F. Comparative Morphology of Slavic Languages. Moscow, 1887.
35. Mistrik J., Tugusheva R. A Slovak Manual. Moscow, 1981.
36. Molchanov A., etc. Ancient Greek Written Pieces. Moscow, 1988.
37. Novakovich A. Danish. Moscow, 1974.
38. Perelmuter I. The Proto-Indo-European and Greek Verb. Leningrad, 1977.
39. Peterson M. A Sketch of Lithuanian. Moscow, 1955.
40. Szemerenyi O. Introduction to Comparative Studies of Indo-European Languages. Moscow, 1987.
41. Slavyatinskaya M. An Ancient Greek Manual. Moscow, 1996.
42. Tronsky I. A Historical Grammar of the Latin Language. Moscow, 1960.
43. Tronsky I. The Common Indo-European Language Stage. Leningrad, 1967.
44. Tsvetaev I. Oscan Inscriptions. Kiev, 1877.
45. Shevoroshkin V. Lydian.
46. Shirokova A. etc. Czech. Moscow, 1973.
47. Eintrei G. Albanian. Leningrad, 1982.
48. Rytova M. New Greek. A Practical Course. St.-Petersbourg, 1994.