Info

Break location: a) 30%, b) 50%, c) 70% of target duration

FIGURE 9. 11 Trials in Tremblay et al.'s (2001) and in Tremblay and Fortin's (submitted) experiments. A short or long tone was presented, with a break in tone presentation. Break location and duration were varied.

beginning (or near the end) of tone presentation were compared. For example, in one experiment, the proportion of short responses was higher when the break occurred when 65% of tone presentation had elapsed than when it occurred at 50% of tone presentation (0.62 vs. 0.55, respectively), but was similar whether the break occurred at 65 or 80% of tone presentation (0.62 vs. 0.63, respectively).

These results from discrimination experiments may be interpreted in part as an effect of attentional time-sharing due to break expectation, as illustrated in Figures 9.3 and 9.4. In that respect, the interpretation is essentially the same as that used to explain the effect of varying location of empty breaks (Fortin and Massé, 2000) and of a concurrent task (Rousseau et al., 1984) in a production paradigm, as well as of a concurrent task in time discrimination (Casini and Macar, 1997). This interpretation could not account for the whole pattern of results in time discrimination with breaks, however, especially for the flattening of the function relating the proportion of short responses to break location at extreme values of break location. In fact, slower accumulation due to break expectation would predict strictly linear functions, with a regular increase in proportion of short responses with increasing prebreak duration at any value of break location.

A more accurate account of the results in time discrimination with empty breaks is provided by coupling the attentional time-sharing hypothesis illustrated in Figures 9.3 and 9.4 to a real-time criterion hypothesis (Kristofferson, 1977). According to this hypothesis, the response would be determined by the outcome of a race between a criterion value, possibly based on the short duration, and the presented stimulus. For example, on presentation of one of the two target durations, accumulation would start. As illustrated in Figure 9.12, it would stop if either one of the two following conditions is met: (a) the tone offset indicates the end of the duration to be estimated, or (b) the criterion is met. If (a) occurs before (b), the tone will be perceived as criterion t

(perceived duration)

Short

Subjective time

FIGURE 9.12 Perceived duration as a result of an accumulation process in time discrimination. In a two-choice time discrimination task, a tone will be perceived as short if (a) its presentation ends before (b) a criterion is met. (a) is more likely to occur before (b) when a short tone is presented or when there is subjective shortening caused by slower accumulation during break expectancy. When tone presentation ends well before (or well after) the criterion is crossed, as at extreme values of break location, a small increase in duration of break expectancy will not change the decision to classify the tone as short or long.

short. If (b) occurs before (a), the tone will be perceived as long. Of course, the end of accumulation because of tone offset (a) is more probable when the short duration is presented, whereas reaching the criterion (b) first is more likely in trials where a long duration is presented.

As in time production, expecting a break induces attentional shifts that cause brief interruptions in accumulation, hence slowing down pulse accumulation. For example, if a break is located at 75% of the target duration, accumulation would be slower during the prebreak period and would resume at its usual rate after the break. At the end of the target duration, the criterion might not be reached on these trials because of the relatively important loss in accumulation due to attentional timesharing in the prebreak period. In comparison with a trial where accumulation proceeds with no time-sharing, expecting a break would change the outcome: a decision for a short rather than a long response would be made.

Generally, increasing prebreak duration makes the relative loss more important, thus making the end of tone presentation (Figure 9.12a) occurring before the criterion is met (Figure 9.12b) more probable, hence increasing the probability of a short response. At extreme values of break location, however, increasing prebreak duration may not change the proportion of short responses, because increasing the duration of expectancy would not change the outcome of accumulation relative to the criterion. For example, if a break occurs when 75% of the target duration has elapsed, perceived duration will be quite short because a break was expected for the most part of tone presentation. In this case, as shown in Figure 9.12, (a) will occur well before (b), and a short response will be provided. Increasing further prebreak duration, for example, to 85% of tone duration, will shorten perceived duration even more, because of additional loss in pulse accumulation due to the break being expected for a longer period of time. This extra loss will not change the decision in classifying the tone as short or long, however, because (a) would still occur before (b) and, in fact, would even occur earlier, in subjective time, than when the break is located at 75% of tone presentation. At a break location of 85%, as at a break location of 75%, a short response would therefore be decided. Similarly, changing the break location from 15 to 25% of target duration could have no effect on the outcome of accumulation relative to the criterion. For example, when a long duration is presented, the criterion could be crossed in both cases before the end of target duration, leading to a long response at both break locations.

Finally, as with time production, the shortening of perceived duration may not be attributed to some weighting or comparison of prebreak and postbreak durations. In effect, there is a clear shortening of perceived duration in no-break trials, when a break is expected for the most part of the stimulus duration, but does not occur. Thus, in one experiment of Tremblay and Fortin's (submitted) study, the proportion of short responses was higher when a break was expected but did not happen than in most trials with breaks. Dividing tone presentation in prebreak and postbreak periods is not necessary for the subjective shortening of target duration to occur because the tone is also perceived shorter if a break is expected during its uninterrupted presentation.

0 0

Post a comment