Experiencing Time—A Commentary on Recent Perspectives on the Perception of Time

Efforts to Design Time Have Ancient Origins

Sheridan (2020) points out that the idea that we experience time dates back to the ancient Greeks. However, his article focuses on methods devised by humans to measure the passage of time, that he terms physical time. Human efforts to measure time date back many thousands of years (Mondschein, 2020)—a notched tablet putatively identified as a pregnancy calendar, found in a German cave, has a suggested date of 32,500 to 38,000 years ago (Vince, 2020).

On the other hand, designing experiential time poses a more problematic challenge, given the subjective nature of the experience of time. This is suggested by the various metaphors applied to experiential time (Rigotti, 1986): time’s arrow, time flies, time lines, etc. Flach (2018) cites an observation of Gibson that, “Events are perceived, but time is not.” Sheridan (2020) concludes his article by noting that, “Models for designing with respect to experiential time are lacking and pose a current research challenge.”

Experiential Time Under Delayed Visual Feedback Conditions

The dramatic effects of visual feedback delay on performance first were documented during World War II, with the introduction of motor-assisted velocity and aided tracking systems (Smith, 1962) (disclosure: K.U. Smith was my father). Findings from many experimental studies conducted over the past 75 years indicate that visual feedback delay is almost unsurpassed in its ability to create profound and immediate disturbances in behavior (Smith et al., 2015, Chap. 5). From a behavioral cybernetic perspective, visual feedback delay effects occur because effective guidance of behavior is critically dependent on close temporal alignment between movements and sensory feedback generated by those movements (Smith, 1962).

One of these aforementioned studies was aimed at assessing how performance of a standard visual-manual tracking task was affected by imposition of delayed visual feedback with different delay durations (T. J. Smith et al., 1998). Subjects (nine) tracked movements of a moving target presented on a computer display by using a joystick to control movements of a displayed tracking cursor. A computer memory buffering technique was used to introduce visual feedback delay between movements of the target, and joystick-controlled movements of the cursor used to track the target.

Results of this study are shown in Figure 1, with delay interval in seconds plotted on the abscissa, and root-mean-square (RMS) error (in cursor movement relative to target movement) plotted on the ordinate. Each point (with standard deviation error bars) represents averaged results for nine subjects, each of whom completed ten tracking trials. Average error increased over eightfold with an increase in delay from 0 and 3.0 seconds. The standard deviation in RMS error increases markedly with longer delay intervals. There appears to be no threshold below which delay intervals greater than zero do not evoke increased error, and error does not level off at 3.0 second delay.

OPEN IN VIEWER

Figure 1.

Tracking error in a pursuit tracking task as a function of visual feedback delay interval.

Analysis of the influence of delayed visual feedback on visual-manual tracking, illustrated in Figure 1, recapitulates earlier work (Levine, 1953) and indicates that sensory feedback from different discrete intervals of temporal delay differentially affects performance. Results in Figure 1 show that tracking error increases monotonically as the magnitude of feedback delay increases.

These data support the conclusion that feedback delay is sensed—a smooth, monotonic function supports no other interpretation. The explanation offered of these data references the behavioral cybernetic theory of my father Smith (1962)—visual feedback delay effects occur because effective guidance of behavior is critically dependent on close temporal alignment between movements and sensory feedback generated by those movements.

Another intriguing example of delayed sensory feedback effects is the control of speech under conditions of delayed auditory feedback. Studies of this type of delay date back to 1950 (Smith, 1962, Chap. 3). Delaying the auditory feedback a person receives of her/his own speech has disruptive speech control consequences, including slowing, increased intensity and higher pitch, and a serious disturbance of the pattern, of that person’s speech.

Broadly speaking, the findings cited above support the thesis advanced by Smith (1962) that delay effects are insidious, because subjects perceive delay as an effect of spatial displacement rather than of timing.

Is the Passage of Time Directly Perceived?

Sheridan (2020) observes that experiential time is measured by the cognitive sense of occurrence of events—what is most vividly remembered or anticipated. This perspective assumes that experiential time relies upon sensory feedback from environmental events in order for memory and projective planning to occur. Two historical examples, as well as more recent findings from a second delayed visual feedback study, collectively support the conclusion that we do not directly perceive the passage of time.

One example is the challenge of determining longitude for navigational purposes. The key breakthrough was the invention of the pendulum clock that could be carried on sailing ships of the time. Huygens in 1657 invented the clock (a physical timekeeping device) explicitly to aid longitudinal measurement at sea (Gould, 1935), because sailors experientially were incapable of gauging the passage of time on their own.

Another dramatic exploration of experiential time is the adventure of Michel Siffre (Vince, 2020). In 1962, this French geologist holed up alone in a deep cave in the Alps for 2 months without light, with the goal of ascertaining whether we need outside stimuli (such as sunlight) to maintain natural rhythms, or whether we have internal clocks (scientists now believe that nearly every cell in nearly every tissue in the body keeps time, by means of which the body’s circadian rhythms are controlled (https://www.quantamagazine.org/how-the-bodys-trillions-of-clocks-keep-time-20150915/; Accessed April, 2020). Experientially, time “slowed” for him. After 63 days, the planned duration of his immersion, he believed only 39 days had elapsed, underscoring the subjective nature of experiential time. On the other hand, his cellular time-keepers remained in sync with surface time, suggesting that circadian rhythms gauge physical time.

Results in Figure 1 indicate that when durations of visual feedback delay are fixed, these durations can be sensed and thus controlled. From a behavioral cybernetic perspective, behavioral control under delay is compromised, but still persists. In contrast, the second delayed visual feedback study referred to above suggests that the passage of time cannot be directly perceived.

This study (Keran et al., 1994) imposed delay in a continuously varying pattern called a forcing function. (Jagacinski & Flach, 2003, pp. 137–157). Feedback delay with a visual-manual tracking task was continuously varied between 0 and 0.4 seconds in a sinusoidal manner.

Subjects (11) completed a total of eight trials at each of eight delay conditions: no delay plus seven variable delay trials at sinusoidal frequencies varying between .05 and 2 Hz.

Figure 2 illustrates the key result from this study, termed a Bode gain plot (https://en.wikipedia.org/wiki/Bode_plot; retrieved 12/6/21). Each point represents the frequency of the sinusoidally varying delay condition (averaged over subjects and trials, with standard deviation error bars), plotted on the abscissa against gain on the ordinate. Gain is calculated (using Fast Fourier transform) as the ratio between the magnitude of the prominent frequency response of tracking movements relative to that of the sinusoidally varying delay frequency.

OPEN IN VIEWER

Figure 2.

Mean gain, and associated standard deviation, as a function of variable delay frequency.

The concept of gain can be understood in the context of visual-manual tracking tasks based on other sensory feedback modalities. We are able to track visual feedback from visual targets (varying in accord with movement capabilities) displayed at electromagnetic frequencies within the range of human vision—gain would be close to one. At ultraviolet or infrared visual target frequencies, however, gain would drop to zero. Similarly, gain for tracking visual analogs of auditory feedback would be close to one only for auditory frequencies within the range of human hearing.

The plot in Figure 2, showing gain observed in tracking of sensory feedback from variable delay, indicates little change in gain across the variable delay frequencies imposed. In particular: 1) the standard deviation of gain at the highest variable delay frequency (2 Hz) overlaps with that of all other variable delay frequencies from 0.05 to 1 Hz, but is higher than that of the gain at the fixed delay level of 0 Hz; and 2) the standard deviations of gain at all variable delay frequencies below 2 Hz overlap with one another, and also with that of gain at the fixed delay level of 0 Hz.

Interpretation

A priori, what would be the predicted result from this tracking experiment? Past research provides no guidance—the experimental paradigm had not been replicated previously. To yield each of the points in Figure 2, feedback delay was sinusoidally varied between 0 and 0.4 seconds at the frequencies indicated. Results in Figure 1 indicate that, when these two delay intervals are fixed, tracking error approximately doubled. Arguably, relative to the higher frequency trials, varying delay between these two temporal levels at a low frequency (say .05 Hz, Figure 2) more closely approximates imposing fixed delays at delay levels across the range. Conversely, varying delay at a higher frequency of 2 Hz should provide 40-fold less time for subjects to experience the effects of a given delay level across the same range.

Based on this a priori analysis, given results in Figure 1, tracking error should be higher, and gain lower, at the lower variable feedback delay frequency, relative to the higher variable feedback delay frequency.

Of course, this analysis is purely speculative, given that the experiment had no precedent. As noted above, the actual results in Figure 2 indicate little change in gain across the variable delay frequencies imposed.

There are two alternative interpretations of this finding: (1) the control system is stable over the range of variable delay frequencies employed, but gain will start to drop (or increase) at some point as variable delay frequency is increased above 2 Hz; or (2) performance gain is independent of variable delay frequency.

Neither alternative can be ruled out at present. However, the second alternative is provocative, in that it suggests that even though behavioral control is sensitive to delayed visual feedback delay imposed at discrete temporal durations (Figure 1), the system does not control delayed visual feedback imposed in a continuously varying manner. If this were the case, the gain levels observed for different variable feedback delay frequencies would not be relatively constant. Invoking Occam’s razor, a conservative conclusion from these results (and recalling that for millennia humans have relied on ergonomic designs to track time), is that we are unable to perceptually track the continuous passage of time. In contrast, as Sheridan (2020) documents, sensory feedback from physical time can be effectively controlled behaviorally.

Implications

The paper by Hancock (2018) calls attention to the idea that time can be designed. As noted above, for a number of millennia humans have been devising various designs for keeping track of time, ranging from early pyramids and temples (i.e., Stonehenge) to more recent timekeeping technologies, such as the hourglass, clocks and watches, or the atomic clock. These designs serve as proxies for the passage of time itself, for purposes of allowing us to control temporal influences on our behavior. In temperate and cold zones of the earth, thousands of animals and plants use another proxy for time—the amount of light energy from the sun falling upon a given area of the earth, and consequent effects on atmospheric temperatures—to control behaviors such as migration, hibernation, caloric ingestion, metabolic slowing, and so forth. The cave adventure cited above suggests that humans themselves also use light exposure as a proxy for judging time. Thus, the idea of experiential time may be considered to be well established.

Indeed, the inherent importance of keeping track of time appears to be so fundamental to biology that time-keeping is built into the cellular functioning and genetic expression of myriad animals and plants, including us.

It can be argued, as noted above, that with one exception every sensory modality of the body shows a frequency response in visual-manual tracking wherein the gain shows a non-zero slope (i.e., is not stable). For these modalities, perception-action coupling is obeyed.

The one sensory feedback exception, as this paper argues, is time. Figure 2 indicates that when sensory feedback delay is imposed as a continuously varying forcing function, gain is relatively stable. As the standard deviations in Figure 2 show, over a forty-fold frequency range (.05 Hz–2 Hz), the gains are statistically indistinguishable from one another, as well as statistically indistinguishable from a gain of 1. In other words, visual-manual tracking was equally successful over this range and was demonstrably not affected by delayed sensory feedback imposed as a forcing function.

From these perspectives, the significance of this paper for understanding experiential time may be considered fundamental. The research reported herein—which has not yet been replicated—clearly shows that an experimental protocol can be devised demonstrating that the passage of time is not directly perceived. The findings thus support the observation of Gibson cited earlier (Flach, 2018) that time is not perceived. The implication is that, down through the millennia, behavioral efforts to design proxies for detecting the passage of time and thus experiencing time may reflect a consequence of this reality. One implication of the research is that time represents the only environmental stimulus that cannot be directly experienced as perceived sensory feedback.

Limitations

There are evident limitations of this research, inherent to any empirical study, related to the number and the demographics of subjects evaluated. The analysis was limited to evaluation of effects at a peak delayed feedback forcing function frequency of 2 Hz imposed across a feedback delay range of 0–0.4 sec. Analyses at a broader feedback delay range and at higher forcing function frequencies are warranted. Finally, the findings reported here have not yet been replicated.

Conclusions

The foregoing account supports the following conclusions, some of which must be considered tentative at this point.

1. Our perception of experiential time is subjective, suggested by the metaphors and by the perspectives of various observers applied to the idea, as well as the cave experience story outlined above.