In chapter 3, section 1, it was shown that two levels must be distinguished in a sentence: a functional level and a syntagmatic level. The functional level is called the genotype level, because it underlies the world’s languages as their universal semiotic basis; language universals are part of the genotype level. The syntagmatic level is called the phenotype level, because it is different in different languages; the same functional structure can be represented by a variety of different syntagmatic structures.
In accordance with the distinction of the genotype and the phenotype levels, applicative grammar consists of two parts: genotype grammar and phenotype grammar. An outline of genotype grammar was given in the foregoing chapter. In this chapter I will present an outline of phenotype grammar.
The basic sign of the phenotype level is the word. Every combination of words is called a word syntagm. The sentence as a functional unit is usually represented by a word syntagm, and sometimes it is represented by one word.
The task of phenotype grammar is the study of the structure of the word and of the syntagmatic structure of the sentence. What is called morphology constitutes only a part of phenotype grammar; phenotype grammar studies not only the structure of the word but the structure of word groups, as well. An important part of phenotype grammar is the theory of word order.
The notion of the word is common in current linguistic literature. But in defining this notion, one runs into difficulties.1
In Bloomfield’s classic formulation, forms such as -ess [es] in countess, lioness, duchess, etc.; -ish [1ʃ] in boyish, childish, and greenish; and -s [s] in hats, books, and cups, are bound forms, because they can appear only as part of a larger form or larger sequence of morphemes. However, big lioness, childish games, and so on are all free forms—capable, that is, of appearing on their own. A free form that consists of two or more free forms is called a phrase. Thus, big lioness and childish games are phrases, because they can be divided into smaller free forms. But lioness and childish cannot themselves be so divided: they are minimal free forms. Bloomfield defines the word as a free form that cannot be divided into smaller free forms. The word is a minimal free form (Bloomfield, 1933).
Bloomfield’s definition of the word is not satisfactory for several reasons.
First, Bloomfield conceives of his definition as an explication of what modern linguists normally call the word. But this definition conflicts with the normal use of the term word. For example, the pronoun my and the articles a and the are normally called words, but they cannot appear on their own.
Second, Bloomfield confounds the phonological representation of the word with the grammatical notion of the word. Thus, the phonological word lɪt and the corresponding orthographic word licked represent a particular grammatical word that can be characterized as the past tense of lick. But the phonological word ʃΛt and the corresponding orthographic word shut represent three different grammatical words: the present tense of shut, the past tense of shut, and the past participle of shut.
Third, it follows from Bloomfield’s definition that table and tables are different words rather than two forms of the same word. By the same token, he considers do, does, did, and done as four different words rather than four forms of the same word. Bloomfield, like many other modern linguists, condemns the abstract notion of the word.
In search of an adequate definition of the word, as well as of other linguistic notions, we must recognize the fact that concepts are defined within the context of a particular theory, and this context changes as different theories are advanced. The same term may signify different things within the context of different theories. Concepts change constantly as theories change. The terms space and time do not signify in the theory of relativity what they meant in classical physics. The evolution of a concept may involve a radical break with established ideas in a particular domain.
Our problem is to define the notion of the word within the theoretical matrix of applicative grammar.
I will advocate a functional approach to the morphological notion of the word. I mean that the main classes of words are morphological crystallizations of the basic syntaxemes: predicates crystallize into verbs, terms crystallize into nouns, modifiers of predicates crystallize into adverbs, modifiers of terms crystallize into adjectives. And subclasses of words are crystallizations of their different paradigmatic functions. A definition of the word must be independent of the notion of the morpheme. The word must be defined through the notion of syntactic function.
I propose the following definition of the word, which is independent of the notion of the morpheme and is based on the notion of syntactic function.
A word is a minimal linguistic sign that is capable of having various syntactic and paradigmatic functions either (1) by itself or (2) together with a word of type (1) meeting in the latter case the condition of separability. ‘Minimal’ means that a word contains no other word.
A word of type (1) I call an autonomous word; a word of type (2) I call a nonautonomous word.
Let us take some examples. Run, runs, ran, run, and running are different forms of the same autonomous word RUN, because they signify different syntactic or paradigmatic functions, and the symbolic function of RUN is invariant of changes in its syntactic and paradigmatic functions and forms. (The capital letters of RUN mean that it signifies an abstract word, that is, a class of different forms of a word.) On the other hand, runner and runners are forms of the autonomous word RUNNER rather than forms of the word RUN, because RUN and RUNNER have different symbolic functions. Since RUNNER is derived from RUN, these different words are related derivationally.
Book and books are different forms of the autonomous word BOOK, whose symbolic function is invariant of these forms signifying its different paradigmatic functions.
The Russian word KNIGA ‘book’ has the following forms: kniga (‘book’, nominative singular), knigi (‘of the book’, genitive singular), knige (‘to the book’, dative singular), knigu (‘the book’, accusative singular), knigoj (‘with the book’, instrumental singular), [o] knige (‘[about] the book’, prepositional singular), knigi (‘books’, nominative plural), knig (‘of the books’, genitive plural), knigam (‘to the books’, dative plural), knigi (‘the books’, accusative plural), knigami (‘with the books’, instrumental plural), and [o] knigax (‘[about] the books’, prepositional plural).
Let us compare these forms with their English equivalents, for example, genitive singular knigi and the English prepositional phrase of the book. Although the English preposition of corresponds to the genitive singular suffix -i of knigi, of cannot be considered a suffix because it is separable from the book. Owing to separability, the preposition of, like other prepositions, assigns syntactic function not only to nouns but to noun phrases, as well: compare of the book, of the large book, of a very interesting large book. By contrast, inseparable case suffixes assign syntactic functions only to the words they are an integral part of. The preposition of is a word of type (2), that is, a nonautonomous word.
Under the above definition of the word, autonomous words are verbs, nouns, adjectives, adverbs, pronouns, and interjections; nonautonomous words are prepositions, articles, and conjunctions.
According to the sign functions described in chapter 1, section 1, autonomous words are classified into words having a representational function—verbs, nouns, adjectives, adverbs, pronouns—and words having an expressive or vocative function—interjections (such as oh, wow, or hey).
Autonomous words having a representational function are classified into words having a symbolic function (verbs, nouns, adjectives, adverbs) and words having a deictic function (pronouns).
The classification of autonomous words according to sign functions can be represented by the following tree diagram:
The above definition of the word has the following advantages:
1) It conceives of the autonomous word as an abstract unit with all its syntactic and paradigmatic functions expressed by its different forms. For example, walk, walks, walking, and walked are not four different words but four different forms of one word, WALK.
2) It establishes correlations between autonomous words and syntactic functional units, or syntaxemes. This correlation has significant typological implications; one of the basic questions of linguistic typology must be: How are primary syntactic functions of words transposed into secondary syntactic functions? To answer this question, we must a) describe forms of words having primary syntactic functions, and b) describe formal processes that serve to transpose primary syntactic functions of words into secondary syntactic functions.
3) Since the word is defined independently of the notion of the morpheme, it is conceived as the central unit of morphology. That makes it possible to regard morphemes not as sequentially organized units but as properties of each word as a whole. This insight increases our understanding of the word.
In post-Bloomfieldian linguistics, only the utterance and the morpheme were taken for granted. The word was no longer considered a theoretical concept. Since the word was denied a theoretical status, the division between morphology and syntax could not have a theoretical status anymore, either. Morphology, which was traditionally defined as the branch of grammar that is concerned with the structure of the word, was to go into limbo.
This neglect of morphology—characteristic not only of post-Bloomfieldian structural linguistics but of generative-transformational grammar, as well—cannot be justified. If we accept the notion of syntactic functional units, we must accept the word as the basic morphological unit that can serve as a syntactic functional unit. Morphemes in themselves cannot serve as syntactic functional units. A morpheme can have syntactic functions solely as a part of a word. Only a word or a combination of words can serve as a syntactic functional unit.
Every word can be analyzed into smaller units, customarily referred to as morphs. For example, the English word renewable is made up of three morphs: re, new, and able. The morph is a minimal unit that comprises a minimal sign represented by a phoneme sequence or one phoneme and its meaning. A class of morphs with the same meaning but different phonemic shapes is called a morpheme, and the morphs representing the same morpheme are called the allomorphs, or variants, of that morpheme.
The difference between the variants of a morpheme can be conditioned phonetically or phonologically.
An example of phonetically conditioned variants of a morpheme is the alternation resulting from the neutralization of the opposition voiced consonant: voiceless consonant. In Russian at the end of a word or before a voiceless consonant, b, d, g, v, z, ʒ, etc. are replaced by p, t, k, f, s, f, etc.; for instance,
|(1)||sad [sat] ‘garden’ : sadu [sád-u] ‘garden (Dat.)’|
|gudok [gud-ók] ‘whistle’: gudka [gut-ká] ‘whistle (Gen.)’|
Before voiced consonants, p, t, k, f, s, ʃ, etc. are replaced by b, d, g, v, z, ʒ, etc.; for instance,
|(2)||molotit’ [mǝlat’-ít’] ‘to thresh’: molot’ba [mǝlad’-bá] ‘threshing’|
|kosit’ [kas’-ít’] ‘to mow’ :kos’ba [kaz’-bá] ‘mowing’.|
An example of morphologically conditioned variants of a morpheme is the alternations of the back vowels a, o, u, and au and the respective front vowels ä, ö, ü, and äu in German; for instance,
|(3)||Bad: Bäder ‘bath: baths’|
|Wort: Wörter ‘word: words’|
|Haus: Häuser ‘house: houses’|
It is common to distinguish between lexical and grammatical morphemes and between free and bound morphemes.
Lexical morphemes are units such as big, walk, and table; grammatical morphemes are units such as of, for, and with, or Russian ‘genitive’, ‘dative’, and ‘instrumental’. A lexical morpheme indicates a fundamental concept, a concept of subject matter (lexical meaning). A grammatical morpheme indicates a subsidiary, usually a more abstract concept (grammatical meaning). Grammatical morphemes comprise closed-set items. The sets of items are closed in the sense that, as a rule, no new items can be created to extend these sets. Lexical morphemes comprise open-set items. The sets of items are open in the sense that new items can be created to extend these sets. A list of grammatical morphemes is limited; it can take a few pages of a book on the grammar of a particular language. By contrast, lexical morphemes of a language belong to a lexicon that may comprise thousands of items and is open to further extension.
There are two types of grammatical morphemes: functional morphemes and derivational morphemes. Functional morphemes (also called inflectional morphemes) indicate different functions of a word. Derivational morphemes are used to derive new words from a given word. For instance, s, ed, ing in works, worked, working are functional morphemes, because they indicate different functions of the word (to) work. But er in worker is a derivational morpheme, because it is used to derive the word worker from the word work. The morpheme s in boys is a functional morpheme, because it indicates a paradigmatic function of the word boy, while the morpheme hood in boyhood is a derivational morpheme, because it is used to derive the word boyhood from the word boy.
Derivational morphemes modify the meaning of lexical morphemes. A lexical morpheme unmodified by a derivational morpheme is called a root. Roots modified by one or more derivational morphemes are called stems. For instance, work is a root and worker is a stem in workers.
Grammatical morphemes are also called affixes. Derivational affixes serve to derive new words, and inflectional affixes serve to differentiate different forms of the same word.
Affixes are only one type of morphological formative, even though they are the most frequently used.
A morphological formative is an elementary linguistic unit that, being a part of a word, is not a root. Given a root X and a meaning Y that is not expressed by a root, the question is: In what ways can X be combined with Y so that the meaning of X and the meaning of Y constitute the meaning of a single word W, and how should the forms of X and W differ?
For example, let the meaning of X be dog and the meaning of Y be plural’. To obtain a word that means dogs, we can use two types of morphological processes: 1) adding to X a separate unit Z whose meaning is Y; this unit is an affix; or 2) changing X in a certain way; as a result, we obtain a unit that I call a suprafix.
We can distinguish five types of suprafixes:
1) A suprafix whose meaning is correlated with a shift of a prosodic feature. For example, when in English nouns are derived from verbs of two syllables, the stress is sometimes shifted from the second to the first syllable. Thus, the nouns mísprint, récord, and ábstract are derived from the verbs misprínt, recórd, and abstráct (Quirk, Greenbaum, and Leech, 1972: 1018)
2) A suprafix whose meaning is correlated with subtraction. For example, in French, the masculine adjectives are said to be derived from the feminine adjectives by a process of subtraction. Thus, the masculine blanc [blā] is derived from the feminine blanche [blā: ʃ] by the subtraction of final ʃ; long [1ɔ̄] is derived from longue [lɔ̄: g] by the subtraction of g; bon [bɔ̄] is derived from bonne [bɔn] (with accompanying nasalization of ɔ). One may wonder whether we could treat the relation between French masculine and feminine adjectives as a reverse process: an expansion of the stems of the masculine adjectives by the respective consonants. We could, but in that case we should posit as many rules as there are final consonants in feminine adjectives, while the adopted description requires only one rule: subtract the final consonant. This process can be generalized in one direction, but not in the other (Matthews, 1974:134).
3) A suprafix whose meaning is correlated with an alternation of phonemes. Examples: take : took, man:men, foot:feet. In Rumanian: [lup] lup ‘wolf’: [lup’] lupi ‘wolves’, [kopák] copac ‘tree’: [kopač’] copaci ‘trees’. In English: the verb hauz from the noun haus (house).
4) A suprafix whose meaning is correlated with reduplication. Reduplication may be partial, as in Latin tango ‘I touch’: tetigi ‘I have touched’, cado I fall’: cecidi ‘I have fallen’, or complete, as in Malaysian orang man’: orang-orang ‘men, people’.
5) A suprafix whose meaning is correlated with transference (meaning converting change of the context of the word). For example, in English, the noun play is derived from the verb play by combining play with the articles: a play, the play. By changing the context, we obtain the transitive verb stop from the intransitive stop:the train stopped:John stopped the train.
A word is a root plus a bundle of formatives—affixes and suprafixes. Affixes are linearly ordered, but suprafixes are not. Therefore, derivational analysis of a word cannot be reduced to an analysis of immediate constituents of its stem. The partial nonlinearity of the formation of a word is a problem for linguists who attempt to study the structure of the word in the framework of generative grammar, which regards the word structure as a system of internally bracketed immediate constituents. Thus, Aronoff writes:
According to our theory of morphology, every new word, if it is derived by a regular rule, must have a cyclic structure: that is, it must be bracketed internally. However, [proәbišon] has been shown not to have cyclic structure. This seems to be a problem for our theory. According to it, shouldn’t all complex words be derived cyclically? (Aronoff, 1976:26)
It is difficult to see how the conflict between the partial nonlinearity of the word structure and the linearity of the linguistic model of generative grammar can be resolved satisfactorily.2
Morphological formatives may differ with respect to two types of affixation, called agglutination and fusion. To see the difference between the two types of affixation, let us compare parallel case forms of the Russian word konec ‘end’ and the Turkish word son ‘end’.
We can see the following difference between the Russian and Turkish word forms:
1) In Russian the phonemic shape of the root may change: in the nominative singular, the root of the word konec has the vowel e, which is lost in other forms of this word.
In Turkish the phonemic shape of the root does not change.
2) In Russian, affixes may have more than one meaning; for instance, the inflectional suffix -am signifies both the dative and the plural.
In Turkish, there is a one-one correspondence between affixes and their meanings; that is, each affix has only one meaning: -a signifies the dative, and -lar signifies the plural; therefore, to represent the dative and the plural, these affixes constitute a linear sequence -lar-a.
3) In Russian the affixes are nonstandard; that is, to express a grammatical meaning, we cannot use one and the same affix that fits all contexts: in our example the dative singular is expressed by the suffix -u, and the dative plural by the suffix -am.
In Turkish, affixes are standard; that is, one and the same affix is used for expressing a given grammatical meaning. In our example, the affix -a indicates the dative both in the singular and in the plural; the affix -lar indicates the plural in all cases.
4) In Russian, affixes are combined with a stem that normally is not used without these affixes: in our example, konc- is not used by itself without suffixes.
In Turkish, affixes are combined with a stem that can be used as an autonomous word.
Affixation in Russian involves a blend, a fusion, of affixes and roots, while affixation in Turkish is a mechanical process of combining monosemantic standard affixes with regular stems that can be used as autonomous words.
In view of these differences, the type of affixation represented by Russian examples is called fusion, and the type of affixation represented by Turkish examples is called agglutination.
Both types of affixation can occur in one language. For example, the formalization of plural nouns in English normally involves agglutination, as in book-s, head-s, etc. But the plural forms, such as feet or mice, involve fusion, because these forms simultaneously signify both some lexical concepts and the grammatical meaning ‘plural’. Similarly, the past tense forms of English verbs such as (he) walk-ed, (he) cover-ed, etc. involve agglutination, while the past tense forms such as (he) took, (he) came, etc. involve fusion.
Depending on which type of affixation prevails in a given language, the world’s languages are classified into fusional and agglutinating languages. Indo-European languages belong to fusional languages (although they may use agglutination, as in the above English examples). Asiatic, African, Malayo-Polynesian, and Australian languages belong to agglutinating languages.
The term grammatical formatives covers the morphological formatives, which were considered in the foregoing sections, and formatives that can be called syntagmatic formatives.
Morphological formatives are units that indicate grammatical meanings inside of a word. By contrast, syntagmatic formatives are units that express grammatical meanings outside of words. Morphological formatives are internal and syntagmatic formatives are external in regard to words.
Morphological formatives synthesize grammatical meanings and lexical meanings inside of a word, while syntagmatic formatives separate grammatical meanings from lexical meanings—grammatical meanings are indicated by three types of formatives: 1) prepositions and other nonautonomous words, 2) word order, and 3) sentence intonation.
The nonautonomous words and word order can be called segmental syntagmatic formatives, and sentence intonation can be called a suprasegmental syntagmatic formative.
Depending on whether morphological formatives play the chief part in a language or a language employs nonautonomous words and word order for expressing grammatical meanings, the language is called synthetic or analytic.
In synthetic languages, isolated words taken out of a sentence retain their grammatical characteristics. For instance, the Latin word lupum has the lexical meaning ‘wolf’, and besides, it has signs that indicate that it is 1) a noun, 2) singular, 3) in the accusative case, 4) a secondary term depending on a transitive verb, etc. The word in synthetic languages is independent of environment; it has a full value both lexically and grammatically and requires a morphological analysis in the first place, and its syntactic properties are direct consequences of its morphological structure.
The word in analytic languages expresses solely a lexical meaning; it acquires grammatical characteristics only as a part of a sentence. In English, play signifies solely a certain general notion. Only in syntactic contexts can we distinguish different words: the play was good, children play, etc. But the Russian words igra ‘the play’, igrat’ ‘to play’, etc. are clear outside of their contexts, and therefore they are incommensurate with the English play, although they have the same lexical meaning.
Analytic languages use word order as a means for expressing grammatical meanings, while in synthetic languages word order normally is used not for expressing grammatical meanings but mainly for stylistic purposes, as a means for expressing different shades of the same grammatical meaning. Compare the English sentence
(1) John frightened Mary.
with the Russian synonymous sentence
(2) Ivan ispugal Mariju.
In (1) the agent is expressed by the primary term John, which precedes the verb frightened, and the patient is expressed by the secondary term Mary. If we exchange the positions of both terms, we will get
(3) Mary frightened John.
The meaning of (3) is opposite to the meaning of (1), because of the change of the word order.
By contrast, in Russian the exchange of the positions of the primary and secondary terms in (2) does not change its meaning:
(4) Mariju ispugal Ivan.
The difference between (2) and (4) is stylistic; in (4) the emphasis is on Mariju.
Typical synthetic languages are Greek, Latin, Old Slavonic, Russian, German, and some others. Typical analytic languages are English, French, Danish, Modern Greek, Bulgarian, and others.
Some languages synthesize words that are equivalent to parts of sentences, or even to whole sentences, in regular analytic languages. For example, in Chukchee the following words occur (Uspenskij, 1965:99):
quickly - walk
lake - quickly - walk
In these words, ny- . . . -k’inet is a confix that signifies third person plural, present tense.
(5a), (5b), and (5c) are equivalent to the sentences They walk, They walk quickly, and They walk quickly to the lake.
Languages of the Chukchee type are called polysynthetic languages. Poly-synthetic languages are found among the Indian languages of North America and the Paleosiberian languages of Russia (Chukchee is one of them).
In some languages the connection between words constituting a syntagm is characterized by an obligatory repetition of some grammatical meanings. Compare the Russian sentence
(1) Malen’kie deti pridut.
with its English translation
(2) The small children will come.
In the Russian sentence, plural is expressed three times: in the noun det-i, in the adjective malen’k-ie, and in the verb prid-ut, while in the English translation, plural is expressed only once, in childr-en.
We say that in the Russian example the adjective malen’k-ie and the verb prid-ut are in concord, or agree, with respect to number with the noun det-i.
The connection between words characterized by an obligatory repetition of some grammatical meanings is called concord, or agreement.
Concord is characteristic of synthetic languages; analytic languages have relatively little concord. This difference can be explained as follows: since synthetic languages have a free word order, they need a formal means to indicate connections between words separated by other words, and concord serves this purpose; it is a formal means indicating connections between words. Concord is much less important for analytic languages, where words that are grammatically connected normally are not separated from one another, and the connections between words are indicated by the fact that they are adjacent.
Here are some examples of concord in English:
|(3)||a.||this (that) man: these (those) men|
|b.||he walks:I (you, we, they) walk|
|c.||he has walked:I (you, we, they) have walked|
|d.||he is walking:you (we, they) are walking:I am walking|
|e.||he (I) was walking:you (we, they) were walking|
In (3a) the demonstrative pronouns agree with the nouns with respect to number.
In (3b) and (3c), the verb is third person singular; and if the primary term is not third person singular, the verb has an unmarked form.
In (3d), besides third person singular concord, there is also first person singular concord (I am walking).
In (3e) the verb is singular when the primary term is singular and plural when the primary term is plural. There is no concord with respect to person here, because first person singular and third person singular fall together.
Another type of formal connection between words is a phenomenon called government.
Government is a connection between two words that belong to two different grammatical classes, with the grammatical class of one word determining the grammatical class of another word.
Examples of government: In many languages the verb governs a noun in a particular case: in the Russian sentences
|(4)||a.||Ja beru karandaš.|
‘I take a pencil’.
|b.||Ja pišu karandašom.|
‘I write with a pencil’.
the opposition karandaš versus karandašom—accusative versus instrumental—is governed by the opposition of the verbs beru versus pišu.
In Russian, Latin, German, etc., not only verbs but prepositions can govern a noun. Compare Latin ad urbem ‘to the city’ (ad governs the accusative: urbem) versus ab urbe ‘from the city’ (ab takes the ablative: urbe).
In current linguistic literature, concord and government are treated as phenomena that correspond to syntactic dependencies between functional units. But as a matter of fact, concord and government are morphologically conditioned connections between words that do not necessarily reflect dependencies between syntactic functional units.
Morphologically conditioned dependencies between words may conflict with syntactic dependencies between functional units. For example, in Latin and Russian, the predicate is characterized by morphological markers of agreement with the primary term. So, from a morphological point of view, the predicate depends on the primary term, but from a syntactic point of view, the reverse is true: the predicate is the head, and the primary term is its dependent. In Abkhaz, a Caucasian language, the verb may agree not only with the primary term but also with the secondary and tertiary terms. So, from a morphological point of view, the Abkhazian verb may depend simultaneously on three nouns. But from a syntactic point of view, it is not the verb that depends on the nouns but the nouns that depend on the verb.
In the Turkish syntagm
|‘the rooms of the students’|
odalarī ‘rooms’ is the head, and talebelerin ‘of the students’ is its dependent. But from the standpoint of Turkish morphology, both the head and the modifier depend on each other, since both the head and the dependent have morphological markers of dependency: the dependent is marked by the genitive suffix -in, and the head is marked by the third person possessive suffix -i.
Given a syntagm A + B where A is the head and B is its dependent, there are the following possibilities of morphological marking: 1) only the head is morphologically marked; 2) only the dependent is morphologically marked; 3) both the head and its dependent are morphologically marked; 4) neither the head nor its dependent is morphologically marked. All of these possibilities are realized in the world’s languages.
Syntactic connections between functional units, on the one hand, and concord and government, on the other hand, are completely independent of each other.
We cannot explain either syntactic connections between functional units by facts of concord and government, or concord and government by syntactic connections between functional units.
Confounding the level of syntactic connections between functional units and the level of morphologically conditioned connections between words can have grave consequences. Here are two examples:
Some linguists question the validity of the word as a theoretical notion, because they explain syntagmatic connections between words in terms of connections between functional units. Thus, J. Kuryłowicz considers prepositional phrases such as French sur la table or English on the table to be one word rather than a group of words. He writes:
A prepositional phrase such as, for example, sur la table, is a word rather than a group of words. If it were a group, the member which governs the group could be put to the same syntactic use as the group itself. But sur cannot be put to this use. Therefore the preposition is not a word, but rather a morpheme (and sometimes a sub-morpheme which constitutes a unit with a case marker). (Kuryłowicz, 1973: 47)
Kurylowicz confounds the functional unit with the word—a morphological unit. True, the English prepositional phrase of the wolf has the same syntactic function as its Latin synonym lupi, but from a morphological point of view, lupi is one word and of the wolf consists of three words. By lumping syntactic functional units and words together, Kuryłowicz disregards the fundamental contrast between analytic languages, such as English, and synthetic languages, such as Latin or Russian.
A. Martinet draws another conclusion from the functional equivalence of prepositional phrases, such as English of the wolf, and words such as Latin lupi. He rejects the word as a theoretical notion and claims that the real meaningful units of language are so-called monemes. (The moneme is nothing but a refined notion of the morpheme. Martinet seems to have chosen the term moneme because in the French linguistic tradition the term morpheme means what is called grammatical morpheme in linguistic science of other countries, while French linguists use the term semanteme for what is called lexical morpheme in other countries.) Martinet writes:
Since we are not inclined to salvage the concept of word, we must operate with ‘minimal significant units’ and therefore my definition of morphology will have to be couched in terms involving this concept. Now what sort of minimal significant units should we retain? Both in Europe and America, the minimal significant unit has been called a morpheme. For a number of reasons, I was led to choose for it another designation, namely ‘moneme.’ (Martinet, 1975:154)
By eliminating the notion of word and introducing the notion of moneme as the basic significant unit of language, Martinet blends the main part of morphology with syntax and restricts morphology to the study of formal variants of morphemes.
The term category generally denotes a class of objects that share a common feature. The term linguistic category can be defined as a class of expressions that share a common meaning or function denoted by common morphological or syntagmatic formatives.3 Thus, the words book-s, citi-es, hous-es, glass-es, etc. belong in the category of the plural number; this category is denoted by a common morphological formative, a suffix having variants: s, z, ɪz. The words (he) walk-ed, danc-ed, explain-ed, extend-ed, etc. belong in the category of the past tense; this category is denoted by a common morphological formative, a suffix having variants: t, d, ɪd.
Forms such as feet, lice, etc., on the one hand, and (he) wrote, bought, went, etc., on the other hand, are also members of the category of the plural or the past respectively, since the relation between foot and feet, etc. is the same as that between book and books, and the relation between write and wrote, etc. is the same as that between walk and walked. What is crucial is proportions, such as foot:feet=book:books and write:wrote=walk:walked. It follows that the plural category is denoted in English by two morphological formatives: a plural suffix having variants s, z, ɪz, and apophonies.
The categories of primary and secondary terms are denoted in English by a syntagmatic formative of word order—primary terms precede the verbs, and secondary terms follow the verbs. In Russian and Latin, categories of primary and secondary terms are denoted by morphological formatives—primary terms are denoted by a set of nominative case markers, and secondary terms are denoted by a set of accusative case markers.
The category of modifiers of a term is denoted in English by two syntagmatic formatives: the preposition of and word order. Compare
|(1)||a.||the wall of stone|
|b.||the stone wall|
Both of stone in (1a) and stone in (1b) belong in the category of modifiers of terms.
A word can be a member of different categories. Thus, in English, the word writes is a member of the category of the present tense, the category of the third person, and the category of the singular number.
Phonological units, such as phonemes or syllables, also constitute categories. There are categories of vowels, categories of consonants, categories of voiced and voiceless consonants, etc.
In accordance with the distinction of derivational and inflectional morphemes, we must distinguish derivational and inflectional grammatical categories. Since the range of occurrence of inflectional categories is wider than that of derivational categories, the latter are subordinated to the former.
The essential changes of the grammatical system of a language result from the shifts between related inflectional and derivational categories. The change of an inflectional category into a derivative category is called lexicalization. The opposite change of a derivational category into an inflectional category is called grammaticalization.
By case I understand systematic differences in the forms of nouns, adjectives, and participles corresponding to the differences between their roles as functional units in a sentence.
There are three basic types of case systems: a) the active system, b) the ergative system, and c) the accusative system.
The active system is based on case markings for the contrasting syntactic functions agent: nonagent.
Agent terms are expressed in the active case. Nonagent terms are expressed in the inactive case.
One finds the following distinction of the active and inactive case:
|(1)||Active case||Inactive case|
|Maryagent walked.||Marynonagent is beautiful.|
|Maryagent deceived Billnonagent.||Maryagent deceived Bill nonagent.|
In the active system, verbs are divided into active verbs and stative verbs rather than into transitive and intransitive verbs, as in the ergative and the accusative systems.
In Dakota, the prefix wa marks the first person singular agent, while mā marks the first person singular nonagent, as in
|(2)||wa-kaśka||‘I bind him’|
|mā-waste||‘I am good’|
|mā-kaśka||‘he binds me’|
In the ergative system, the choice of cases is determined by the contrast between transitive and intransitive verbs.
The primary terms of transitive and intransitive verbs are denoted by the accusative case, and the secondary term of the transitive verb is denoted by the ergative case.
As in the ergative system, the choice of cases in the accusative system is determined by the contrast between transitive and intransitive verbs.
The primary terms of transitive and intransitive verbs are denoted by the nominative case, and the secondary term of the transitive verb is denoted by the accusative case.
The ergative and accusative systems relate to each other as mirror images. If we compare transitive sentences in both systems, we find that the accusative corresponds to the absolutive and the nominative corresponds to the ergative.
I will focus on the analysis of case in the accusative system.
The cases of the accusative system are divided into two classes: 1) grammatical cases and 2) concrete cases. Grammatical cases are the nominative, the accusative, and the genitive. There are four basic concrete cases: the instrumental, the allative, the ablative, and the locative. Both grammatical and concrete cases have primary and secondary functions.
The primary function of the nominative is to denote the primary term in a sentence; the primary function of the accusative is to denote the secondary term in a transitive sentence. The primary function of the genitive is to serve as an instrument of the transposition of a sentence into a nominal group, for example, in Russian
|(3)||Ivan priexal ‘John arrived’ → priezd Ivana (Gen.) ‘John’s arrival’|
|Ivana ubili ‘They killed John’ → ubijstvo Ivana (Gen.) ‘the murder of John’|
The primary functions of concrete cases are as follows: the locative means ‘lack of movement’, the ablative means ‘movement from’, the allative means ‘movement to’, and the instrumental means movement across’.
The important thing is a distinction of central and marginal positions in the case system. The grammatical cases occupy central positions, while the concrete cases occupy marginal positions. The hierarchy of positions can be represented by the following scheme:
(4) Verb > Grammatical Case > Concrete Case > Adverb
This hierarchy is purely relational and is independent of the linear order of words. Concrete cases are intermediary categories between grammatical cases and adverbs.
To understand the meaning of the positional hierarchy of cases, consider the following sentence in Russian:
|(5)||On čitaet knigu v sadu.|
|‘He is reading the book in the garden’.|
The accusative in knigu is a mere exponent of a syntactic dependency of the noun, while the prepositional group v sadu in the locative has a concrete meaning. The expressions knigu and v sadu cannot be substituted one for another. If we omit knigu in the above sentence, we will get a structure with one free position:
(6) On čitaet _____ v sadu.
The secondary function of grammatical cases is to serve as concrete cases. By contrast, the secondary function of concrete cases is to serve as grammatical cases. Here are some examples from Russian:
|(7)||On pisal pis’mo čas.|
|‘He was writing the letter for an hour’.|
In this sentence there are two accusatives, pis’mo and čas. The second accusative has the secondary function of a concrete case and therefore occupies a marginal position.
|(8)||On upravljaet zavodom.|
|‘He manages the factory’.|
The instrumental zavodom has a secondary function here. It functions as a grammatical case equivalent to an accusative.
An understanding of the interplay of the primary and secondary functions of the grammatical and concrete cases is of paramount importance for explaining many synchronic and diachronic syntactic phenomena, in particular the phenomena relating to passivization.
The agentive term in the instrumental or other concrete cases in passive constructions can be explained as a secondary term resulting from the use of the instrumental or other concrete cases in their secondary function of a grammatical case equivalent to the accusative.