Academic journal article Style

Comparative Study of Verse: Language Probability Models

Academic journal article Style

Comparative Study of Verse: Language Probability Models

Article excerpt


The formal-statistical method in verse studies was developed in the Russian school of metrics and prosody. The establishment of this method has led to the formation of a particular field within cognitive poetics based on exact methods for studying how texts are generated. In Russia, the formal analysis of verse structure was initiated by Andrey Bely. Subsequently, this tradition was developed throughout all the 20th century in works by scholars representing different generations and countries: Boris Tomashevsky, Roman Jakobson, Victor Zhirmunsky, Andrey Kolmogorov, Mikhail Gasparov, James Bailey, Barry Scherr, Marina Tarlinskaja, el al. At the end of the past century, it served as the basis for a new direction, which can be said to fall within cognitive verse theory. (2) This direction arose thanks to the theory of reconstructive simulation of versification (RS) developed by Marina Krasnoperova. (3)

The apparatus of the RS theory represents a unique system of cognitive and probability models of text rhythm, allowing for the study and reconstruction of the processes for generating and perceiving rhythmical texts. The probability models comprise background models for verse study. They represent a link between the text and deeper cognitive models. The latter form the nucleus of the RS theory and among them the central model is that for the perception and generation of the rhythmic structure of poetic texts. Through it, a given set of conditions of versification are placed into correspondence with probability models. Cognitive models are than used to interpret the agreement or disagreement of the data with the structural characteristics of text provided by probability models.

The main group of probability models consists of so-called language models of meter. The first such model was constructed by the world-famous Russian mathematician Andrey Kolmogorov who had devoted many years to the study of verse. The language models are constructed using the basis of a rhythmic dictionary of prose, which is perceived as arhythmically neutral language background. The rhythmical dictionaries of prose give an idea of the frequencies of rhythmic (phonetic) words in texts. The phonetic word is a group of syllables united by one principle word stress (the table, she said, my name). Since the frequency of a rhythmic word, in a sufficiently large sampling, approaches its probability, the particular probability of the use of this word in speech (the language probability) is assigned to it in accordance with its frequency in prose.

Probability models have provided an essentially new perspective on the analysis of verse rhythm, making it possible to study the mechanism of versification. After Kolmogorov's model, others have been proposed. Of particular importance is the so-called language model of dependence (LMD) developed by Krasnoperova. This model is based on the principle of dependent combinations of rhythmic words in verse formation. The choice of a word depends on certain conditions such as meter, metric position, and the preceding context. The preceding rhythmical context is formed as a result of the action of certain syntactic conditions and can be also called the syntactic context. In this regard the model differs from the usual Kolmogorov language model (LM), which is constructed on the principle of the independent selection of rhythmic words. (4)

In the apparatus of the RS theory, different variants of LMD are developed: for example, symmetric and asymmetric models. These differ in the sequence by which the rhythmic line is formed. Moreover, an important role is played by the choice of words in the initial and final strong positions (SP or S-positions) of a line. The symmetric model is highly rigid. In it, the outer S-positions of a line are filled no later than the internal: if the line beginning is formed first (the word/words in the first SP are chosen), then the line end is formed next, and vice versa. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.