The participants task was first to read the passage and then perform a verification task. On the verification task, they were given sentences such as A discarded cies rette started a fire, and their reaction time to respond true or false was recorded The results are shown in Figure 7-t Note that although the verification times for explicit propositions are faster when given an immediate test, there is no differ ence between explicit and implicit propositions when the test is delayed by 15 min utes. Kintsch explains the results by appealing to the two levels of representabon we have discussed: a short-term surface representation that decays or is otherwise lost very quicklv and a long term propositional representation. Implicit propositi have only a propositional representation, and it is assumed that consulting a sur face representation is quicker than retrieving a propositional representation. The immediate test laps both representations, so there should be an advantage for the explicit propositions. However, because this surface representation is losu with a longer retention interval, lhere is no difference between explicit and implicit propo sitions in the delayed test.