The acquisition of temporal event referencing, encoded by the temporal connectives: then, before, after, when, while, together, until, and since in English, Thai and Lisu was investigated using two acting-out comprehension tasks, a Marble task and a Toy task. Forty children aged 3.6-7.6 years from each language participated. The Marble and Toy tasks differed in their cognitive complexity: in the Marble task the child had to act out only one clause, whereas in the Toy task the child had to act out both clauses. This task manipulation affected performance in Lisu children only. Language-general effects were found, namely "then" and "together" were relatively early and "since" was relatively late in acquisition. Language-specific effects were found for Thai and Lisu. Results confirm that characteristics of task and test sentences affect children's comprehension of sentences expressing temporal relations and can partially account for the disparity in acquisition order found in previous studies.