Языки стран Дальнего Востока, Юго-Восточной Азии и Западной Африки

20 Языки Стран Дальнего Востока, Юго-Восточной Азии и Западной Африки | LESEWA-XIII exception of the variable “type of argument”, which is not aspectually relevant for predicates denoting states, repetitive, negative or irrealis situations. The factors that prove statistically significant in this model are: (1) The speech mode — dialogue, which was reported also significant for 了 2 with dynamic predicates. (2) In this model we observe for the first time a significance of the factor “End of discourse”, which is really defined as a specific text-forming func- tion of 了 2 , which is to mark the end of the speaker’s statement (Li, Thompson 1981: 283–290). This factor is not significant to the rest of the studies markers. We propose one more correlational model for the usage of 了 2 with non- dynamic predicates, in phrases to express the specific aspectual meaning “change of state”. Fig. 4. Variation of 了 2 with non-dynamic predicates —Correlational Model IV Model Summary Model 4 R R Square Adjusted R Square Std. Error of the Estimate 0,533 0,284 0,277 150,385 Coefficients Model 4 Unstandardized Coefficients Standardized Coefficients t Sig. B Std. Error Beta (Constant) 10,106 10,431 0,773 0,440 Type of predicate: Static 0,855 10,629 0,022 0,525 0,600 Type of event: Change of state 320,625 20,763 0,493 110,807 0,000 Discourse position: End 50,313 10,897 0,117 20,801 0,005 Speech mode: Dialogue 20,217 10,745 0,053 10,271 0,205 This model has the highest correlational coefficient among all tested models for both 了 1 and 了 2 , which is R = 0, 533, the correlation is also within the medium level. The coefficient of determination R 2 (R Square) = 0,284, which means that 28,4% of 了 2 marking under such conditions is determined by the factors in the model. We have to note the relatively high degree of correlation with the factor “type of event — change of state”, which means that in expressing this aspectual meaning with static predicates, the usage of the marker is relatively more regular and functionally motivated. Apart