The influence of animacy, givenness, and focus on object order in Croatian ditransitives

In this talk we report how animacy, givenness, and focus influence word order in Croatian, which is considered a ‘free’ word order language and thus all word orders are attested but with varying frequencies (Siewierska, 1998). Previous research has shown that the preference for one word order over another is dictated by pragmatic factors and so factors like givenness, animacy, and focus influence word order (Branigan, Pickering, & Tanaka, 2008; Bresnan, 2003). We investigate the effect of these factors by using ditransitive structures in Croatian, more precisely the effect on indirect-direct (io-do) or direct-indirect object order (do-io).

Since corpus data provided a limited view on these categories, we have conducted a survey that provides the necessary contrasts that allow us to see how word orders are preferred based on the three properties. 82 native speakers (mean age=23.3) of Croatian completed the survey. The survey consisted in 18 context sentences based on which the objects were considered given/not given and focused/not focused in the target sentences; animacy had a human/inanimate binary distinction. The target sentence was then presented in four randomized word orders (vid, ivd, vdi, dvi) and each of them was judged on a 5-point scale.

We used Linear Mixed Effects to create models in which word order and each of the factors was a predictor, and a null model where only word order was a predictor. We conducted ANOVAs as comparisons of the factor models to the null model. The results for animacy and focus are significant (p-value=2.2e-16); the result for givenness is not significant (p-value=0.24). However, adding givenness to the animacy model improves it significantly (p-value=0.016). We can see the effect that a factor has on word order in the examples where all factors but one were neutralized. What we find is that, as expected, animacy influences an object to be placed forward while focus has the opposite effect. The effect of givenness was not as clear because in the condition of neutral animacy, the participants preferred do-io orders (vdi and dvi) regardless of which object was the given one. There was however a better judgment of io-do orders when the io was given (mean=3.38) than when the do was given (mean=2.09). In order to reveal the relative importance of the factors we used the conditions in which there was an interaction of two factors. Focus is a stronger factor than animacy because animacy influences object order only when neither object is focused. Surprisingly, in the condition where animacy and givenness were in competition (io animate and do given) there was a preference of do-io orders. Unfortunately, there was only one example of this condition. So, because of the unclear effect of givenness seen in the neutral animacy condition, and the insufficient power of its model, we conclude that animacy is a more relevant factor than givenness.  

Our conclusion is that, like in other languages, focus influences the object to be placed last, animacy and givenness influence it to be placed before objects without this property. However, the extent of the givenness effect is not clear and more research is needed in order to establish to which extent is givenness relevant in Croatian. We thus reveal that these three factors are ordered hierarchically in the following way: focus > animacy > givenness.

 

References

Branigan, H. P., Pickering, M. J., & Tanaka, M. (2008). Contributions of animacy to grammatical function assignment and word order during production. Lingua, 118(2), 172-189.

Bresnan, J. a. N., T. (2003). On the Gradience of Dative Alternation. 1-39.

Siewierska, A., and Ludmila Uhlirova. (1998). An overview of word order in Slavic languages. In A. Siewierska (Ed.), EMPIRICAL APPROACHES TO LANGUAGE TYPOLOGY (Vol. 20, pp. 46). Berlin: Mouton de Gruyter.