Short term morphosyntactic change

Mustajoki & Hannes (1991) - STMC bibliography

Reference

Mustajoki, Arto and Hannes Heino. 1991. Case selection for the direct object in Russian negative clauses (Slavica Helsingiensia 9). Helsinki: Helsinki University Press.

Summary

This book is an extensive statistical analysis of the variation between the genitive and accusative with direct objects of negated transitive verbs, as in (1):

The authors investigate the factors which influence the choice of case, and their interaction. The study is based on a sample of 4037 examples extracted from a corpus of stylistically diverse texts (fictional prose, drama, journalism and scholarly works). Given that morphosyntactic choices are typically conditioned by multiple factors, each example was coded with respect to a large number of variables (51 in total) and the effect of each individual variable was evaluated separately. Some of the variables, together with examples in which they were found, were later excluded from the analysis, as they generally required one of the case forms and allowed no variation or only insignificant variation. The analysis presented in the monograph deals with the remaining 33 variables. These variables fall into nine groups (which, as the authors point out, may include heterogeneous phenomena):

• the force and scope of negation (double negation, various particles and modifiers which intensify the negation);
• clause types (e.g., declarative, exclamatory or interrogative; types of main and subordinate clauses);
• variables connected with the predicate (mood, tense, aspect, person and number, semantics of the predicate, etc.);
• infinitival constructions;
• degree of idiomaticy;
• variables connected with the object (such as part of speech, gender, inflectional class, semantic properties);
• variables connected with the subject;
• syntactic factors (word order, presence of particular clause constituents);
• stylistic factors.

These variables are further analysed, for example, the variable “Postmodifier connected to the object” includes: “Postmodifier in the genitive case”, “Relative clause”, “Participial construction”, “Prepositional phrase or adverb” and “Other cases”.

Mustajoki and Heino include variables that have been investigated by previous authors. For those that are well known, they provide a very detailed elaboration that has not been suggested before. Thus, for the variable “Semantic type of the object noun” the study suggests a set of twenty five factors, which form a scale with respect to concreteness/abstractness. The preference for the accusative gradually decreases, as the object becomes less concrete. Thus concrete countable nouns take the accusative in 59% of instances, “relatively abstract countable products of  human activity” (e.g., zadača 'task', ošibka ‘mistake’, vopros ‘question’) give 30% for the accusative, while with nouns denoting emotional or physical state the frequency of the accusative is about 5%. Although concreteness/abstractness as a condition on direct object case marking has been studied previously, Mustajoki and Heino were the first to show this gradual transition in detail.

The authors tested a number of factors that had not been investigated in previous studies of direct object case in Russian. These factors include the number and person of the predicate verb, the prefix of the verb, negative meaning of the object and semantic properties of the subject. With some of them, such as the number and person of the predicate verb, statistical differences emerged due to the effect of other co-occurring factors. The same is true for nouns with positive and negative meaning (e.g., udovol´stvie ‘pleasure’ vs. toska ‘melancholy’, ‘depression’). Positive nouns have a much higher preference for the genitive; the authors suggest however that this may be due to different collocation patterns characteristic of these two groups of nouns. Other factors, such as the semantics of the subject, might have an effect on the object noun case, but this requires further investigation.

Texts investigated

The corpus investigated represents four types of text created within the overall period 1948-1982:

• fictional prose (from the period between 1973 and 1978);
• drama (1953-1975);
• scholarly works (linguistic texts written between 1948 and 1982);
• journalism (1981-1982).

The corpus is purely synchronic.

Statistics

The original sample of 4037 examples was reduced to 2722 after eliminating Marginal Cases (for example idiomatic expression, or those in which the case of the object could not be unambiguously identified), and Strong Factors (such factors which determine the case of the object irrespective of other factors, as with human object nouns). With each factor, calculations are presented separately for constructions with a governing finite verb, those with a governing infinitive and for both groups taken together. Raw number and percentages are provided.

Which data from the source were used

Selected data on object properties, the type of the governing verb and the properties of the clause. The criterion for selection was the possibility of correlating the data with the diachronic statistics presented in the database.