M. I. SHAPIR
”VERSUS” VS “PROSA”:
SPACE-TIME OF THE POETIC TEXT
In the word the aspects of form, content and material are bound together so closely that the problem of “verse and prose” (which in point of fact is the problem of the relationship between speech forms) is usually substituted or obscured by two other fundamental problems in versification theory: those of the relationship between form and material (i. e. “verse and language”) and between form and content (i. e. “verse and meaning” or “verse and sense”). This confusion of problems is also caused by the fact that, in certain cases, the difference between verse and prose may indeed be manifested in material or content: it is not infrequent that poetic language does not coincide, either in its structure or in its semantics, with the language of prose. However, those differences between verse and prose, which pertain to the material itself, are all, without exception, optional. No equivalent element of poetic language exists, which could not, at will, be transplanted into prose: as is generally known, prose can be metrical and rhymed, while verse, in its turn, can do without metre, rhyme and other instruments of obvious violence to language. The situation is exactly the same with all those differences between verse and prose, which pertain to semantics themselves: any abstract content (detachable from poetic form) can become the property of prose; and vice versa, there is no theme, motif, emotion etc. which would not be allowed to migrate from one speech type to the other. Aspects of semantics peculiar to poetry are only those which are directly linked to the “resolving capacity” of form form which enables the poet to express, in verse, that which is impossible to express and above all to do this in the way in which it is not possible to do the same using the means of prose. Therefore, if there is a border which separates verse from prose, one should look for it exclusively in the area of form.
The most authoritative of the various approaches to the formal definition of verse is today the idea of its “dual segmentation”: “Any text breaks up into subordinate syntactic segments; in the poetic text, however, this <...> is combined with the segmentation into lines of verse as well as verse entities which are larger or smaller than the line <...> the latter segmentation can either coincide with, or diverge from the former, thus creating innumerable possibilities of rhythmic-syntactic correlations” (Buxshtab). The most debatable point here is the description of verse segmentation as “dual”. Even in prose there are many other divisions besides syntactic ones; moreover, the boundaries of significant (bilateral) units do not necessarily coincide with those of non-significant (unilateral) units: syllable may be non-congruent with morpheme, the breath group with lexeme, and colon with syntagma. Verses do not form a network of necessarily subordinate segments either: the rhythmic unit (line) is not always identical with the metrical unit, the rhyme may be fractured by the clausula and transferred to the beginning of the next line, and so on.
Verse entities which are smaller than the line can equally be found in prose; those which are larger are derivative from the line. Therefore Buxshtab’s first proposition can, in principle, be reduced to the following: the difference between verse and prose is the division into verses itself. This definition of poetic discourse was formulated independently by Kenigsberg, Tomashevskij and Tynianov in the early 1920s. The shortcoming of their definition is not its seemingly tautological character, but its insufficiency: we do not know what is the peculiarity of the poetic line as compared with any other. Tomashevskij believed the specificum of verse was its state of being divided into comparable and commensurable segments: “<...> the breaking of poetic discourse into ‘lines of verse’ (‘stixi’) <...> the acoustic potential of which is comparable, or in the basic case <...> simply equal <...> is <...> the specific feature of poetic discourse”.
Nevertheless it is not only poetic lines but also prosaic colons that can be “comparable” and “commensurable”; on the other hand, adjacent lines sometimes contrast more than the colons of some specimens of prose. It is generally believed that words, syllables or feet can serve as a criterion of commensurability for poetry; these same yardsticks may, however, be applied when comparing the colons of metrical prose. At the same time, different lines of a single poem cannot always be measured using the same standards: thus some lines of eighteenth-century Russian “free verse” are measured in iambic feet, while others in trochaic feet or even syllables. Lines of verse are equated, rather than compared. Verse is distinguished by its ability to compare that which is not comparable in prose, and to make commensurate that which is incommensurable; in other words, verse has another (additional) dimension.
Tomashevskij was the first to conceive of verse as two-dimensional: “<...> if prose is linear, verse is two-dimensional speech”. In verse, “speech moves in two directions: from one word to another, and from one rhythmico-syntactic section of the line to the analogous section of the next line”. The scholar called this movement from one line to another “vertical” apparently he thought of the movement “from one word to another” as horizontal. Tomashevskij’s description requires an important amendment: any written text is bid out “on the plane” and, in this sense, it has two spatial co-ordinates. It should be realized that the dimensions of verse are not physical, and their number is, in fact, higher than two.
Semioticians are accustomed to considering space as text; for our purposes it is more fruitful to consider text as space. An observer perceives the “outer” space as “positive”: the distance between him/her and any point “outside” is expressed in terms of positive quantity. The “inner” space is thus “negative”; it reflects an outer world in a particular form, and is symmetric to it, relative to the axis of time. One of the means of exteriorizing inner space is a text in a “natural” language: like the physical world, the text normally has three orthogonal dimensions. The first dimension is formed by its overall length: from beginning to end; this co-ordinate can provisionally be called the speech co-ordinate. The second dimension is constituted by the hierarchy of grammatical levels (morphemic, lexemic, syntactic and others), parallel to which another, purely formal, hierarchy unfolds (distinctive features phonemes syllables breath groups colons etc. ); I would call this co-ordinate of the text: the language co-ordinate (in the narrow sense of this word). Significant and non-significant units, even if they occupy the same place in the text and are located at the same hierarchic level, are necessarily different in relation to the third, semiotic, co-ordinate: independently from the level any sign has its own position on the axis “object concept”.
The prosaic text is orientated along three axes of co-ordinates; in verse, the fourth, poetic, dimension (or the verse dimension) is added, which can in no way remove any of the other three. Each line of verse has its length, grammar and semantics, which are not totally pre-ordained by its verbal substance. The autonomization of the length of the poetic line is connected with its metre. The autonomization of poetic grammar manifests itself in various instances of parallelism, in enjambements, rhythmic-syntactic clichés, grammaticization (or degrammaticization) of the rhyme, and so on. Finally, the autonomization of the verse semantics is caused by the ability of verse to express metric meanings and communicate rhythmic sense. It is often the case that contiguous lines are strikingly different from each other from the viewpoint of their length, grammar or semantics: they can be discrepant relative to each of the first three co-ordinates, but as instances of verse they are equal relative to the fourth (it should be pointed out that they are equal, and not “comparable” or “commensurable”).
In the poetic text the fourth dimension supplements quasi-spatiality with quasi-temporality: the rhythm of verse itself (which forces “real” time out and replaces it) plays the role of poetic time. As Roman Jakobson emphasized, “the decisive role of time factor in the structuration of a poetic meter is the indispensable prerequisite of any meter”. “Only in poetry with its regular reiteration of equivalent units is the time of the speech flow experienced”. What is relevant here is the “inner time” of the text, rather than the real duration of reading or recitation: the poetic discourse is itself experienced as duration measured by the units of poetic rhythm. Each unit perceived at the present moment recalls the analogous units in the past and predicts their iteration in the future.
It is not its arrangement in columns that makes verse the fourth coordinate; much more important is the fact that verse does not comprise another level of language. In prose, the units of lower levels are fully integrated into those of higher levels: phonemes consist of distinctive features, syllables consist of phonemes, breath groups are composed of syllables and form themselves into colons. The situation is exactly the same with significant forms: words consist of morphemes, sentences are made up of words, utterance-length entities are composed of sentences. As to verse, it is able (being the main compositional unit, and having both usual and casual semantics) to cut the units of different levels into parts. The verse boundary can divide a sentence, a word-combination, a word, a morpheme (and even a breath group or a syllable), so it is not possible to say that the verse “consists” of them.
Breaking up of the lexical, morphemic or syllabic whole, and redistribution of its “fractions” between two rhythmic entities is the extreme realization of the possibility of non-coincidence between different articulations (segmentations), which was discussed above. Nevertheless, to say simply that such a possibility exists is to say nothing. According to Buxshtab, one “segmentation can either coincide with, or diverge from” another but one can only guess as to whether they tend to coincide more frequently than diverge, or exactly the opposite. It is generally accepted that syntactic and metric divisions coincide frequently enough; I make bold to contend that the potential non-coincidence between grammar and verse is realized in the overwhelming majority of literary poetic works. Its essence is in the discordance between rhythmic and grammatical hierarchies: more often than not, the verse positions belonging to the same level are filled with grammatical material belonging to different levels, and vice versa. The disjunction between verse and grammar does not limit itself to line-ends. It is known that it can penetrate inside the line, or embrace whole groups of lines: strophoids, stanzas etc. Zhirmunskij argued though that in any poem, enjambements, “however frequent they were, are ‘resolved’ (as a dissonance) in the subsequent instances, where the end of the sentence coincides with the line-end”. Nonetheless, notwithstanding Zhirmunskij, the contradiction between rhythm and grammar is sometimes located even on the border of the work: at the beginning or the end.
The syntactic incompleteness of the stanza or of the work may be analogized with the incompleteness of the line and is, as I have already pointed out, derivative from it. The rhythmic-syntactic disharmony within the line is also, to a large extent, analogous with enjambement. At the same time the similarity between the phenomena which can be observed at the clausula and those at the caesura should not disguise the cardinal difference between them: if a word is divided by the clausula, poetics take precedence, if by the caesura, then grammar is paramount, and the line is non-caesural. The caesura is deprived of the compulsion pertaining to the clausula, whereas it is only this compulsion that makes verse what it really is, creating the necessary prerequisites for the conflict between grammar and rhythm. But nonetheless the definition of verse is not thereby settled: the discordance between autonomous hierarchies of semiologically relevant units, while being a sufficient marker of verse, is not at the same time an indispensable marker. It is fairly easy to recognize verse when grammatical and rhythmic articulations diverge, but how can one distinguish it from prose if there is no contradiction between rhythm and grammar?
The absolute coincidence between poetic and grammatical segmentations is extremely rare, but it is essential to allow for the works in which the non-co-ordination of segmentations is close to zero: for it is evident, from the instances when all other properties of verse are absent, what is the distinction between verse and prose which will be manifested under any circumstances. As the contradictions between grammar and verse disappear, the paradigmatic nature of verse segmentation proper reveals itself: two hierarchically equal verse units emerge as the representatives of a single paradigm (rhythmic, rhythmic-syntactic, lexical-rhythmic-syntactic etc. ), and in this capacity they are equivalent to one another. Whatever interpretation of the concept of “paradigm” we choose, verse possesses almost all the characteristic features of paradigmatic structure, with the exception of only one feature: the very thing that determines the specificity of the poetic discourse. Theoretical linguistics conceive of syntagmatic relations as those in praesentia, and of paradigmatic relations as those in absentia: while the constituents of the syntagma coexist, the constituents of the paradigm are mutually exclusive. The relations are considered paradigmatic when they are established between the form which is present in the text and the forms which are absent from it. In verse, however, the paradigmatic relations pass from in absentia to in praesentia.
In my opinion, Jakobson came closer than anyone else to an understanding of this peculiarity of verse: for him, “the poetic function of language” “projects the principle of equivalence from the axis of selection into the axis of combination”. He did not use the terms syntagmatics and paradigmatics here, but (as is evident from his other works) he considered them synonyms of selection and combination. Therefore Jakobson’s proposition can be re-formulated as follows: in poetic language, syntagmatics are constructed following the principles of paradigmatics. For all the similarity of this assertion with what was discussed above, it cannot serve as the definition of verse: Jakobson’s objectives were the study of any conceivable symptoms of the aesthetic function of language, so it was not his intention to define verse as such. But in his reflections he departed from poetry, and this is why his formulae, which were insufficient to specify verse, were sometimes too stringent if applied to prose works.
The most important point missing in Jakobson, and which should form part of the definition of verse, is the appearance of another axis of co-ordinates. In poetry, Jakobson argues, “equivalence is promoted to the constitutive device of the sequence”. It would not be correct, however, to suppose that in verse the syntagmatic sequence simply materializes the paradigmatic structure of language. Rhythm and syntax, paradigmatics and syntagmatics, are two autonomous dimensions; they preserve their self-identity independently from whether they are antagonists or allies. The “equivalence” is proper exclusively to the units unfolding along the fourth co-ordinate; with reference to other dimensions, the speech segments are not equivalent to one another: at the best they are similar. The similarity of segments is possible in both poetry and prose; their identity is possible only in poetry. Verse is a sui generis projection from the axis of identity onto the axis of difference. Of course, any identity is manifest only against the background of difference and vice versa; thus from the aspect of paradigmatics speech forms are identical despite some differences between them, whereas from the aspect of syntagmatics they are different despite some instances of coincidence. While the constituents of the syntagma are parts of the whole, the constituents of the paradigm are modifications of the whole: in the syntagma, two words are different (e. g. as sentence parts), whereas in the paradigm, two word forms are identical (e. g. as case variants of one and the same word). Similarly, despite their factual difference (which can be quite considerable), syllables are correlated with other syllables (or feet with other feet, accentual groups with other accentual groups, lines with other lines, stanzas with other stanzas) as variants of a single invariable, and they are eo ipso identical on the grounds that they belong to a common paradigm.
As it is applied to verse, Jakobson’s theory has another shortcoming. The problem of the extent to which the poetic discourse is the sequence of equivalent segments (in full, from the beginning to the end, or in part, in a few places) remains unresolved. Paradigmatic relations between speech segments are also manifested in metrical prose, but there they are optional and superimposed onto syntagmatic relations. In verse, on the contrary, syntagmatic segmentations are superimposed onto paradigmatic ones: as Vinokur wrote, “the material of language” “is as if sifted” through “the metric form”. Unlike prose, verse contains the segmentations which are drawn through the whole length of the text, the pervasive compulsory paradigmatic segmentations. Their visual compulsoriness is ensured by the graphic form of verse, and their acoustic compulsoriness by the verse intonation. Where this pervasive and compulsory division into verses is absent, in the collision of rhythm and grammar the latter prevails over the former. On the boundary between the lines of verse, the autonomization of rhythm reaches its maximum; by contrast, within the line, as in prose, the rhythm is formed primarily by means of grammar itself. This is why verse re-written in continuo, as prose, indeed turns into prose: all clausulae, deprived of their compulsoriness, become a kind of caesurae.
It is known under which circumstances paradigmatic relations appear between the elements of prose; now it is necessary to find out whether syntagmatic relations can be established between the elements of verse. Syntagmatics dictate the rules according to which the parts form themselves into the whole: thus, syntagmatics impose restrictions on the combinations of language forms. Do analogous limitations remain in force as far as poetry is concerned? The general answer is no. The laws of verse primarily set up the inner construction of elements, and if two units are allowed in the poetic text, it is most likely that their unrestricted transposition is also possible. In iambics, for example, the minimal “verse-building” paradigmatic unit is a binary foot in several rhythmic variations. All of them are equal in rights and interchangeable, with one exception: in “classical” syllabic-accentual (syllabotonic) versification, stress on the ultimate ictus approximates to 100%, i. e. the last foot may be only an iambus or spondee. This circumstance imposes some restrictions on the combinability of paradigmatic units. But firstly, all syntagmatic relations between the units of verse are optional: verse can do without them, without ceasing to be itself (for instance, the above mentioned rule of the stressed constant in Russian iambics was occasionally violated from the late eighteenth century onwards). Secondly, all syntagmatic relations between verse units belonging to a lower level become paradigmatic at a higher level: the ultimate foot is not rhythmically identical with earlier feet in the same line, but is identical with the ultimate feet of other lines. These lines are correlated with each other as equal and interchangeable representatives of a single rhythmic paradigm.
This is the universal mechanism of the transformation of syntagmatics into paradigmatics: if feet are linked by syntagmatic relations, then lines will be linked by paradigmatic ones; if lines are linked by syntagmatic relations, then stanzas will be linked by paradigmatic relations. One of the most widespread types of syntagmatic relations in verse is rhyme (often coupled with alternance). Lines linked by rhyme function not only as variants of the whole (as forms of a single paradigm), but also as parts of the whole (of the stanza or strophoid). Restrictions imposed by the rhyme on the combination of lines are cognate with syntactic agreement; restrictions imposed by the alternance are somehow akin to syntactic government. But stanzas, constituents of which are linked by syntagmatic relations, manifest paradigmatic relations with contiguous stanzas: they are permutable variants of a single invariable.
The transformation of poetic syntagmatics into paradigmatics is distinctly marked in logaoedic verse. Here it is not rhythmically non-identical units but metrically non-identical units that are linked by syntagmatic relations. We must assume that in the sapphic stanza the combinability of dactyls and trochees is subordinate to strict limitations (akin to parataxis). However, the first line is correlated with the second (and the second with the third) as equal paradigmatic forms: they all have the same metric structure (trochee trochee dactyl trochee trochee). Again, the fourth line is linked with the other three by syntagmatic relations (it has a distinct configuration: dactyl trochee), but the four lines together form a stanza which establishes paradigmatic relations with all other sapphic stanzas, both within the same poem and outside it.
When stanzas are integrated in a syntagmatic whole, they form themselves into a suprastrophe or an even more complicated form of fixed stanzaic concatenation, as the sonnet of the French type. In it quatrain associates with another quatrain (or tercet with another tercet) by means of both paradigmatic and syntagmatic relations, for they are linked by a common rhyme. Above that, all four stanzas form a strict syntagmatic sequence: quatrain quatrain tercet tercet. However, the unchangeability of this sequence turns the syntagmatic “agreement”, “government” and “parataxis” of stanzas into an invariable, of which any sonnet is a modification: it is linked by paradigmatic relations with all other similarly rhymed sonnets. Any patterns in which the whole of the work is the paradigmatic constant can be called “strict (rigid) patterns”.
In connection with the problem of the border between verse and prose such a strict pattern as the monostich is of paramount interest. Normally, in monostichs the hierarchy of verse segmentations is not non-congruent with the syntax. What makes them an instance of verse is their inclusion into the system of paradigmatic relations: just as the sonnet does not cease to be itself even if it is written using super-short monosyllabic lines, instead of the traditional iambic pentameters or hexameters, so too, the monostich does not cease to be itself even if its line is not metrical. Vladimir Markov is right when reminding us that “there is no poetry without poetic tradition”: the iambic line “is perceived as iambic against the background of other lines, not only published, but also possible”. With due reservations made, one can admit that, among other “monostichs, the line which cannot be typified as belonging to any <...> metre <...> will be perceived as the monostichous vers libre”. The most important thing is that the monostich is recognized as such against the background of other single-line poems (irrespective of the versification).
It should be mentioned that the restrictions imposed on the combinability and interchangeability of units can be of both a syntagmatic and paradigmatic nature. There is nothing here which would contradict the essence of paradigmatic relations. So far linguists have not agreed as to whether the paradigm includes those forms which are found in the same position, or those which alternate, depending on the context, but at the same time are perceived as identical. Hjelmslev maintained that the paradigm is formed by interchangeable units, but his ideas on this have been justifiably opposed. Even if the paradigm is defined by the positional alternation of elements, there is nothing extraordinary in the fact that, in verse, paradigmatic forms coexist: this is its distinction from prose. Just as in prose where the constituents of the morphological paradigm (which are mutually exclusive in theory) can sometimes interchange, in verse the constituents of the rhythmical paradigm (mutually exclusive in theory) can from time to time lose their equivalence.
The elements of the same level of verse are normally interchangeable because they are forms belonging to the same paradigm. Nonetheless, they are different forms of this paradigm, and if this difference is played upon, they cease to be equivalent and therefore occupy their proper and peculiar place in the poetic framework. This can happen though only when the text is constructed as the enumeration of the rhythmical microparadigm, which is completely presented and exhausted by the individual lines of the poem. The purpose of these exposed microparadigms is to go step by step through all possible combinations of the variable element with constant ones. Poems of this kind, because they are not rigid forms, have fixed volume pre-ordained by the extent of the microparadigm: no one such poem can be longer or shorter than they are, even by one line or stanza.
Thus verse is the system of pervasive compulsory paradigmatic segmentations, which structure the fourth dimension of the text. Besides quasi-space, verse also has quasi-time measured by units of poetic rhythm. But the most striking thing is that, besides being the model of time, verse is the model of eternity (or what we understand by this word). Eternity is not given to us as a category of the physical world, and it is quite possible that it is only a mental construction. Nevertheless, poetic eternity is no less real than poetic time, and its image is created by the paradigmatic network of verse. Both potentially and actually, eternity includes all times: what has passed, what is, what is to come, as well as what might happen, but never will. Each line of verse is the constituent of a paradigm which is open and inexhaustible: as early as Edgar Allan Poe we find the view that “the possible varieties of metre <...> are absolutely infinite”. Their invariable includes at once all lines of a particular metre: those already composed, those which will be composed, those which might have been composed; and each of them exists as a rhythmic form only by virtue of the existence of the metre. If the line is a temporal unit, then the invariable embracing all times is nothing but poetic eternity: it resides “out of time and space” of texts which acquire verse dimension, entirely thanks to this invariable.