This article challenges the view that the generation effects occurs only for items represented in semantic memory. We obtained generation effects regardless of the lexical status of the cue used to generate the target and regardless of the lexical status of the target. Generation effects were comparable for wordlike and unwordlike nonwords. The effect for nonwords depended on displaying the target at the end of each generate trial. We argue that such feedback gives subjects appropriate visual experience with nonwords; otherwise, the disadvantage of never seeing the generated nonwords could overwhelm any memorial advantage conferred by generating. Generated items also suffer when study and test formats differ more for generated than for read items; we demonstrated that changing format reduces recognition for nonwords. We conclude that previous failures to demonstrate generation effects with nonwords reflect confoundings with such familiarity factors and that the generation effect does exist for nonwords.

Studies disagree regarding the relationship between word frequency and apparent duration. The present experiments evaluate factors that might explain conflict in prior studies. In Experiment 1, word frequency was manipulated factorially with three stimulus durations. High-frequency words were judged longer in duration than low-frequency words at each exposure duration. When briefer durations were used in Experiment 2, frequency did not affect subjective duration. In Experiment 3, a wider range of frequency restored the longer apparent duration of high-frequency words. Use of a postexposure mask lengthened duration judgments but did not interact with the frequency effect. Use of a paired comparison procedure in Experiment 4 again showed the frequency effect. The results are consistent with an attentional model that suggests that subjective time estimation is directly related to the amount of attention remaining to evaluate the passage of time once the stimulus target has been cognitively processed.

No abstract