Change Blindness Paper

Posted by on July 9, 2011

Our experience of the world is vivid and filled with rich details.  We view the world as one continuous, smooth motion, not as a rapid interaction of stimuli colliding with our visual sensory system.  It is not like watching a homemade video where the camera continually shakes and the film is out of focus, rather it is like watching a movie in a theater with high definition video equipment that shows every scene in clear detail.  If the world was not so, our minds would be mesmerized by the continual fluctuation in detail every time we blink.

Instead of choppy and divided visual photos, the mind has a system through which it smoothes out the continuous number of stimuli colliding with the eye to produce a stable, colorful, detailed view of the world.  This system though has some anomalies which make the study of the mind and the visual system curiously interesting.  Anomalies such as the ability to throw away information, take mental shortcuts, and create seemingly logical stories about two things exist in our mind.  All of these come together to create the great mystery today that is called change blindness (Aamodt and Wang 5).  Through this paper, the study of change blindness, its two paradigms, and two of the five hypotheses explaining change blindness will be discussed.

Change blindness has been tested “across saccades, blinks, blank screens, movie cuts, and other interruptions,” and is the occurrence where the observers continually fail to notice from the smallest to the most substantial changes of the visually depicted objects.  One such example is that of a demonstration where the observer is told to view a photograph for a later memory test and while they study the scene, the details are changed during a saccade.  The observers continually fail to notice a change unless the images are viewed during fixation.  Though, even during fixation, it has been shown that if there is a simulation of the effect of saccade, change blindness is still experienced.  For example if there is a blank screen, “mudsplash,” a cut, or a pan in a motion picture presented between the original and changed image, change blindness occurs.  All of these demonstrations seem to illustrate inattentional and repetition blindness, but it is unclear whether they could have predicted our inability “to detect large, meaningful changes to natural scenes” (Simons 3).  Therefore, the goal of recent study on change blindness has been to discover the limits and conditions of it; also, the goal is to discover the limits to and precision in remembering scenes.

In studying change blindness, the two main paradigms present are the flicker paradigm and forced choice detection paradigm.  Participants in the first paradigm are presented with an original and modified image that rapidly alternate with a black screen placed between them; the participant’s goal is to respond quickly when they detect the changing object.

Research using this paradigm has produced two primary findings: (1) observers rarely detect changes during the first cycle of alternation, and some changes are not detected even after nearly 1 minute of alternation (Rensink et al., 1997); and (2) changes to objects in the “centre of interest” of a scene are detected more readily than peripheral or “marginal interest” changes (Rensink et al., 1997), suggesting that attention is focused on central objects either more rapidly or more often, thereby allowing faster change detection, (Simons 3).

Meanwhile, in the force choice detection paradigm the participant is only given one view of each image before they respond.  Through this method, the amount of exposure time to the initial image is controlled, and also allows for signal detection analyses and dependent measures such as accuracy and latency.  Simons states that through the study of these two paradigms it is understood that attention is necessary for visual memory but might not be sufficient.  In fact, the studies mentioned by Simons demonstrate an incredible infallibility of our visual memory, such that even when central objects to our perception of the real world change, we do not necessarily notice the change.  One infamous example of this is the experimenter that asks the pedestrian for directions.  While the experimenter is asking and the pedestrian is pointing, construction workers come between the two and the initial experimenter gets replaced in position by a second distinguishingly different experimenter; 50% of the time in this experiment the pedestrian does not notice the change

In order to explain this incredible phenomenon of change blindness, three of the theories posited are “overwriting,” “nothing is stored,” and “nothing is compared.”  Overwriting is the most common one that is also dominant in current literature on change blindness.  This theory suggests that whenever a new image is viewed, the previous image is lost, and only the details of the new image are known.  Though this theory explains the studies of object masking, it does not thoroughly explain change blindness.  For example, in the case of the flicker paradigm, if this theory is to be held true, the image that the observer represents may be neither the first nor the last, rather it may be the blank or mask in between the two.

In “nothing is stored,” the theory that is posited is that the best representation of the world is the world itself; therefore it is not necessary for the subject to have a representation.  This view suggests that only the details of a scene that are abstracted by the subject are the ones that are retained, all of the rest are lost.  This predicts an identical performance to the overwriting theory in examples in which the second image remains visible till a response.  Contrary to this, the theory would also suggest that details of the second image would be lost when it is removed.  A weaker form of this theory, that may be more plausible, suggests that the only information that is abstracted from a view of the scene is information that is necessary for the next.  An example of this is that of experimenters dressed up as construction workers asking the pedestrian for directions.  When the experimenters switch on the pedestrian that is giving them directions, the pedestrian may have never noticed that the construction workers changed because the only necessary information was that the person being given directions to was a construction worker.  Something to note here is that this is also an example of inattentional blindness in the aspect of a seemingly large, meaningful change to a natural scene.

The “nothing is compared” perspective states that everything we view is in a way stored, but not compared.  This is explained in the same way that research shows how it is possible for a person to maintain two contradictory views without ever noticing until someone makes it apparent for them.  It could very well be possible in visual representation for a person to maintain two distinct visual representations from two scenes or images, till either something about the meaning of the scene or a person’s questioning prompts the comparison.  A real world example of this might be going to a friend’s house and not noticing how they have shifted the living room or added something to it till they ask, and in the event of which, one responds with “oh yeah I remember that ‘x’ used to be there at that ‘y’,” or “oh yeah that ‘x’ didn’t use to be here before, where did you get that ‘x’?”

Simons states that “evidence from a number of literatures suggests the possibility that an implicit trace from a feature or object can be preserved, even when observers do not consciously perceive it” (Simons 11).   Essentially, this verifies that it is possible for the dual existence of trace representations of both scenes in the person’s memory, which are brought out when properly induced.  One experimental example would be yet another setup of the experimenter asking the pedestrian for directions.  In this situation while the pedestrian gives directions, a group of students pass through them and one of them stealthily takes the basketball from the experimenter’s hands.  When asked in this situation if the subjects had noticed anything unusual, only three of them spontaneously reported the basketball missing.  Most of them did not respond so till they were asked specifically if the experimenter had a basketball, in which case more than half would respond by stating something like “oh yeah you did have a ball that was red and white.”  Experiments such as these successfully demonstrate how people do maintain two representations, and bring the appropriate one out when cued.

Of the three theories, “nothing is compared” seems to be the best possibility, because this theory seems to be able to explain both the theories “nothing is stored” and “overwriting,” while also explaining how people may be able to maintain two distinct representations which are evident in real world experience.  This dual maintenance of distinct representations is contradictory to both “nothing is stored” and “overwriting” because both theories would suggest that the representation of the first image must be lost, and the evidence for “nothing is compared” suggests otherwise.  The similarity between the “nothing is compared” and “nothing is stored” theories lies in that they share a common experimental example that explains both theories; the pedestrian and experiment example is common to both.  The difference lies in the fact that “nothing is compared” explains most of which the other theories explain, while also providing the important information that we might not be consciously aware of the two representations of the scenes we hold in our memory.  Such a preference for the “nothing is compared” theory would also be expected because reductive explanations are preferred in place of ad hoc.  In this situation, “nothing is compared” seems to be the reductive theory that explains the phenomena of the other theories while adding its own novel aspect which has been discussed above.  Therefore “nothing is compared” appears to be the better theory.

None of these results for change blindness would be considered common sense; in fact all of this is surprising.  It is a common practice in law to consider eye-witness testimony on cases, but it is becoming increasingly evident that might not be so infallible; there is increasing scientific evidence accumulating that suggests we might not remember that which we think we remember, and we might not see that which we think we saw.  This all seems surprising because our view of the world is so rich with detail; we are able to see with intense smoothness, clarity, and colorfulness.  It may be possible that we evolved only be able to detect changes when they change the meaning of the scene, like a primitive human being picking berries from a bush till he notices a large bear approaching him, which elicits a flight response from his sympathetic nervous system.  This evolutionary perspective would also support the aforementioned “nothing is compared” theory.

Change blindness continues to be a topic of fascination and scientific discovery.  Through this paper I have successfully discussed the study of the subject, the paradigms associated with it, and some of the possible theories posited to explain it.  Nothing is compared may very well be the best possible theory right now.


Work Cited

Aamodt, Sandra, and Sam Wang. Welcome to your Brain. New York:
Bloomsbury, 2008.

Simons, Daniel J.  “Current approaches to Change Blindness.”  VISUAL COGNITION, 2000, 7
(1/2/3), 1–15

Leave a Reply

Your email address will not be published. Required fields are marked *