Production of the double object construction: An experiment

Abstract

In the double object construction (e.g. the man gave the woman some flowers) a preference has been observed for placing definite arguments before indefinite arguments when both appear post-verbally. In Danish it has been reported that examples with the indefinite–definite order are read more slowly than those with the definite–indefinite order in speeded acceptability judgement tasks, and they are less frequent in corpus texts. This short communication presents a memory recall experiment showing that the preference observed in comprehension and written production is also observed in on-line oral production. Participants produce definite–indefinite orders when attempting to recall definite–indefinite orders in 95% of the cases, but when attempting to recall indefinite–definite orders they alter the definiteness of one or both of the arguments and produce indefinite–definite orders only in 6% of the cases.

The effect of givenness and referring expression on dative alternation in Norwegian: A reaction time study

“…The speeded acceptability tasks reveal significantly higher acceptability and faster RTs with DODs when the first object is definite and the second one is indefinite than the other way around; this difference is not found with the PD. These observations are corroborated in a memory recall experiment in Danish (Kizach 2017) where the participants had no trouble correctly repeating DOD sentences in which the first object was definite and the second indefinite, but they altered the definiteness of the sentences with indefinite-definite order at 94%. Clifton & Frazier (2004) tested the effect of givenness on both DODs and PDs in a task which included a context sentence introducing one of the object arguments, in addition to definiteness marking in the target sentences.…”

“…In order to investigate the effect of givenness on the DA, we developed a speeded acceptability judgment task in OpenSesame (Mathôt, Schreij & Theeuwes 2012) that measures the reaction times (RTs) of the participants' responses to test sentences which they were asked to judge as either 'good' or 'bad'. Based on previous studies conducted on languages with DA (Clifton & Frazier 2004, Brown, Savova & Gibson 2012, Kizach & Balling 2013, Kizach 2017, we expected there to be a significant difference in the RTs between the conditions conforming to and the conditions violating the given>new principle in the DOD, whereas no difference should be observed for the PD. All the former studies on dative alternation making use of RT data tested structures with noun phrase objects.…”