Some Thoughts on Structural Morphology*

B. Trnka

In treating morphology as a special part of grammar, we are well aware of the fact that the forms cannot be isolated from their meanings. Both aspects of the language, morphological and semasiological, maybe, however, regarded as autonomous in the sense that alterations in the meaning of a form are not directly dependent on its outward aspect.¹ The same form which denotes, e.g., the present tense may be used to express the past without any change, by which the shifting of the temporal meaning would be formally expressed. The mutual relation of the various meanings of words and groups of words is the object of the semasiological analysis of the language. Morphology is concerned only with fixing the principal meanings of the forms and constructions which may occur in the language, and with stating their formal interdependence in the given linguistic system, irrespective of their additional meanings.

All formal expedients that are able to express a morphological function in the language may be called morphological exponents. Languages differ in the choice of the exponents: the same formal expedients which have the functions of exponents in one language may be devoid of this role in another. Thus the direct object is expressed in English by putting the noun after the verb (in opposition to the subject, which is marked by the pre-verbal position), whereas the word-order in Czech and other “synthetical” languages does not represent a morphological exponent, its function being to denote stylistic variations of meaning only. Even if the accusative form of a noun happens to be the same as the nominative, e.g., všechny bolesti hojí’ čas = time cures all troubles, the position of words in the sentence is not felt to be instrumental in transmitting the correct meaning intended by the speaker. The true sense cannot, however, be misunderstood by the hearer in this case, because any alternative is naturally out of the question. Differences between čas hojí všechny bolesti, čas všechny bolesti hojí, and všechny bolesti hojí čas, are merely of a stylistic nature.

The morphological exponents represented in the language are of very different kinds. They may be 1. phonological (cf. the alternation of vowel phonemes in E. sing: sang: sung; foot: feet; imprudent: imprudence; Czech ved’ : ved’), 2. synthetic (i.e., the modifying morphemes in E. be-come, come-ly, come-s, com-ing), 3. analytic (e.g., word-order, groups of words, auxiliary words, grammatical concord), 4. compounds (i.e., groups of words expressed synthetically). In addition, some morphological exponents may consist in the combination of two, or more, formal means which have not the function of the expedients by themselves. Thus in English the possessive genitive is denoted by the ending -s (z and iz, respectively) + position of the substantive before its head-noun, or the combination of the same ending + the preposition of. Numerous morphological exponents of this kind in Indo-European languages are composed of the alternation of one or two phonemes in the basic morpheme + suffix (or prefix), e.g., E. life: live-s, keep: kept, leave: left, clement: clemency; Czech ruka (nominative singular) “hand”: ruce (dative sing.) “to the hand”: ruční (adjective); Jan John: Jene ! (vocative); hora “mountain”: po- hoř-í “(chain of) mountains”, etc.; cf. also women formed from woman by means of two alternations. The single elements of the combined morphological exponent do not possess any morphological value outside the combination, which must be, therefore, regarded as a unit from the point of view of structural morphology.

According to which morphological exponents are typical of various languages, four linguistic groups may be distinguished: 1. languages in which the typical morphological exponents are analytic or compounds, e.g., Chinese, Siamese, and other “isolating” languages. 2. languages in which alternations are typical morphological exponents, e.g., Semitic languages. 3. languages whose typical morphological exponents are synthetical, or even compounds, without alternations in the basic morphemes of words,² e.g., Bantu, American, and Ural-Altaic languages. 4. languages in which the combined morphological exponents are used, in addition to compounds, synthetical, phonological, and analytical expedients, e.g., Indo-European languages. Typical subgroups may be further distinguished according to the relative use of the latter morphological exponents. Thus Latin, Czech, Greek, etc., are synthetical languages, whereas Modern Germanic languages, especially English, are characterized by the greater use of analytical and phonological exponents employed side by side with the synthetical.

Morphological exponents group themselves into two classes according to whether they are used to express the relational and semasiological modifications of the same basis, e.g., wait, waits, waited, waiting, waiter, waiters, waiter’s; shall, will wait, etc., or whether they denote the same relations or semasiological modifications of differ ent words, e.g. kings, wives, men; the king’s garden, the garden of the king. The unifying element in the former group, which may be called the morphological group, is the common basis which is either a word or a basic morpheme. The forms of words are derived either directly from the basis or from another member of the group that serves as a secondary basis for a smaller morphological group, so that they may belong to the various grades of relationship which must be stated for a given linguistic system synchronously, without any regard to the preceding or subsequent stages of the development of the same language.³

If a word is felt to contain the modifying morpheme, the rest of it is always inferred to be the morphological basis, even in spite of absence of any meaning. Thus E. flabby is analyzed into the suffix — clearly denoting the adjective, and the basis *flab, which is not found in any other word. Cf. also Czech roz-man-itý. Such morphological bases that cannot be associated with any other morphemes, owing to their isolation, may be called virtual. It must be added that in the development of a language the virtual basis may become real, e.g. scavenge in scavenger, and vice versa. The same designation may be used to denote the phoneme or the group of phonemes which must be regarded as modifying the real basis although they do not occur elsewhere, e.g. fort-ress, kin-dred, front-ier.

The common element of the other group is the same meaning of the morphological exponents. Thus the modifying morphemes -i (-y), -e, í, alternation í, -i (cf. chlapi, buky, meče, bratfí, páni) serve to denote, as synonymous morphological exponents, the plural nominative of masculine substantives in Modern Czech. The preterite in English is formed not only with the ending -t (-d and -id, respectively), but also with the alternation of the vowel or diphthong (e.g. sing: sang, read: read), or with the ending t + alternation, e.g. keep: kep-t. The synonymous morphological exponents combine, in a complementary way, to express the same functions, none of them being employed in place of another. If a synonymous formation is limited to one or two words (e.g., him, them, whom), it may better be called “suppletiveness”. The morphological exponent -m (plus the alternation of the vowel or diphthong, respectively) is suppletive on account of the fact that the regular exponent of the direct object is word-order. Suppletiveness is a well-known phenomenon, if it occurs in the morphological group,⁴ two basic morphemes being used to supplement each other, (e.g. good: better, go: wen-t), instead of one. Such cases as worse and less, which are in the same semasiological relation to bad and little as colder and better are to cold and good, must be regarded as a sort of double suppletiveness, the idea of comparison being expressed by neither of the regular complementary exponents -er and more.

The existence of the complementary types may be regarded as an unnecessary complication of the linguistic system. In fact, the artificial auxiliary languages, such as Esperanto or Novial, may be considered more perfect than the living ones because they use one morphological exponent for each function. The reason for the linguistic phenomenon characteristic of natural languages in a greater or lesser degree is, perhaps, that the synonymous forms used, e.g., for the dative singular in Modern Czech, are closely associated as contrasting not only with those of the nominative, which is used in the most important syntactical functions, but also with the other case forms. If the dative singular were associated in morphological contrast with the nominative only, it would be hardly necessary to employ so many morphological exponents (which are, by the way, mostly homonymous of themselves) to denote the same function, and the rearrangement of the dative endings would take place with regard to the forms of the nominative and the grammatical genders. In other words: The tendency to use one morphological exponent for the same function is crossed by the associations existing between the members of the same morphological (sub)groups.

Some of the synonymous morphological exponents may substitute each other in some of their functions. The relation of such exponents is that of parallelism, one of them representing, so to say, another track performing, at some length, the same role as the main one. Thus the possessive case in English is a morphological expedient parallel to that of the preposition of + noun. The morphological exponent denoting the dative by placing the noun between the verb and the direct object may substitute the main exponent to + noun in many instances. In Czech the suffix -ove, which is used to denote the nominative plural of masculine substantives, may be employed instead of the other endings, e.g. rtové, mečové, Čechové, instead of rty, meče, Češi. The substitution of the morphological exponents by the parallel ones often implies different shades of meaning, but this is of no concern to morphology.

The same morphological exponent may be both complementary and parallel. Thus the adverbs more and most are used as complementary exponents expressing the comparative and the superlative respectively, the suffixes being limited to certain types of adjectives; at the same time they may stand for the suffixes as parallel means of expression, e.g. more, most poor instead of poorer, poorest. The same may be stated of the Czech suffix -ové, which is also a complementary suffix with a number of substantives, such as ve vo do vévodové, synové, etc.

In conclusion, I should like to point out that morphology must be treated from the point of view of the sentence, as well as of the word. All that has been stated refers to the morphological analysis of the language as the static system of means of expression and necessarily represents only the rudimentary stage of structural morphology, which promises to yield a rich harvest for future investigation.

Notes

1. The morphological structure of a language seems to be determined, to a great extent, by what the speakers feel to be basic or modified in the limits of the grammatical categories, e.g. those of gender and number, irrespective of the actual means of expression. This idea, which Dr. Jakobson develops in his interesting article written for this volume, deserves the closest attention. In this paper (read before the Prague Linguistic Circle in the Czech version in May) I confine my observations to what I call morphological exponents. Owing to lack of space, some of the interesting points, such as homonymy in morphological structure, the cumulation of functions in the exponents, and the relation of morphology to semasiology and syntax, must be left unmentioned.

2. The alterations of suffix vowel phonemes in Ural-Altaic languages (e.g., kések: házak in Magyar; cf. the similar phenomena in Old Norse dialects) and those of suffix consonants in English (bits: beads, laughed: loved) depend on the vowel or consonant contained in the basic morpheme of the word, and must be kept strictly apart from the alternations, which take place independently, e.g. sink: sank: sunk, foot: feet. It may be mentioned here that the alternations may also consist in the addition, or absence, of a phoneme.

3. Owing to alternation, some of the members of the same morphological group may have only one sound in common. Cf. dear: dearth, darling; Czech roznesu: přináším. —

4. The analysis of the morphological bases from the semasiological point of view will be reserved for a special article.

*From Charisteria Guilelmo Mathesio quinquagenario. . . oblata (Prague, 1932). pp. 57-61.