Demos: Causal Perception and Postdiction in 7-month-old Infants
This page contains demonstrations discussed in the following paper:
Newman, G. E., Choi, H., Wynn, K., & Scholl, B. J. (2008). The origins of causal perception: Evidence from postdictive processing in infancy. Cognitive Psychology, 57(3), 262-291.
These demonstrations are provided as Quicktime movies, which can be downloaded or viewed directly in most web-browsers. (To download a free Quicktime player, go here.) In some cases the spatiotemporal parameters may be slightly different than those used in the experiments, but they should be sufficient to illustrate the basic effects.
Note that some browsers may not be able to play these movies directly; instead, you may need to download them to your local hard drive, and play them from there.
In each case, you should judge the causal status of the specified event: does it look like a causal launch (wherein one item collides with another, causing its motion) or a non-causal pass (wherein one moving item simply passes across the entire display, passing over a stationary item)?
The overall goal of these studies is to demonstrate and explore several new aspects of causal perception in infancy. In particular, we demonstrate here for the first time that 7-month-old infants are sensitive to the same types of perceptual grouping effects that influence causal perception in adults. Moreover, this work provides a new way of demonstrating causal perception in infants that differs from previous strategies, and provides one of the first demonstrations of postdictive processing in infancy.
Basic Launching and Ambiguous Full-Overlap Displays
Experiment 1: Grouping in Causal Perception
This is a standard launching display, a la Michotte. Beyond the kinematics involved in the event, you see causality: one object collides with another, causing its motion.
Ambiguous Full-Overlap Event
This event -- first used in causal capture demonstrations, and later in causal grouping demonstrations -- serves as an example of the type of baseline condition used in many of our studies. It is identical to the basic launching event, except that the two discs fully overlap before the second motion. While this ambiguous event may sometimes still be seen as causal launching, many observers can perceive it as non-causal passing (as described above). (If you find it difficult to perceive passing, put the event in the periphery. Note that you can see passing here more easily, even when you can still discriminate the different colors -- though in the present study all objects in the display had the same color.)
Expt 1 Habituation: Synchronous Grouping (with green squares)
Experiment 2: Postdiction in Causal Perception
7-month-olds infants were habituated to a grouping display, in which the motion of an isolated object (C, appearing toward the bottom) leads adults to perceive a full-overlap event between A and B (on top) in terms of causal launching (wherein A is seen to cause B's motion), via C's perfectly synchronized common motion with B. We hypothesized that this manipulation would also be effective in infants, who would therefore be habituated to an instance of causal launching.
Experiment 1 Test #1: Synchronous Grouping (with red discs)
Infants were then tested with two test events. This first test event was identical to the habituation display, except that it involved red discs instead of green squares.
Experiment 1 Test #2: Grouping Destroyed by a 600 ms Offset (with red discs)
In this second test event (which also involved red discs), the context disc C began moving a full 600 ms after the onset of B. In adults, this temporal asynchrony (between B and C) destroys the grouping effect, which in turn readily allows them to see the interaction between A and B in terms noncausal passing (wherein A simply passes right over or through B, which in turn remains stationary). Infants looked significantly longer at this test event. We interpret this in terms of the change from causal perception (during habituation) to noncausal perception (during this test event), just as adults tend to perceive these displays.
Expt 2 Habituation: Synchronous Grouping (with green squares)
Experiment 3: The Salience of Small Temporal Offsets
This experiment involved the same habituation event as used in Experiment 1.
Experiment 2 Test #1: Synchronous Grouping (with red discs)
The first test event in this experiment was identical to that used in Experiment 1.
Experiment 2 Test #2: Grouping Preserved Despite a 120 ms Offset (with red discs)
The second test event in this experiment was identical to that used in Experiment 1, except that it used a shorter temporal asynchrony. This more modest asynchrony is not sufficient to disrupt the grouping in adults, who therefore still tend to perceive causal launching in this display. As noted in the initial report of this effect -- Choi & Scholl (2006), Perception -- the failure of this temporal offset to disrupt the grouping is itself interesting. Indeed, it seems paradoxical: If we tend to see noncausal 'passing' when the lower object is completely stationary, then that should occur here too -- since after a 120 ms temporal gap the second motion has already begun, is fully visible, and should need to be bound to one object or the other! Since we know that this second motion will tend to be bound to the initially moving object (yielding perceived passing) with no motion, that should also occur here. But then, when disc C does start moving, it should be too late to switch this percept to perceived launching. (By that point, it seems that we should already be perceiving passing!) Thus, the fact that adults can still readily see causal launching in this case illustrates a "postdictive" effect: our conscious perception of the world is not an instantaneous moment-by-moment construction, but rather is formed by integrating information over short temporal "chunks". Here, we can perceive a collision between two objects even after the moment of potential 'impact' has already passed. In this experiment infants did not look differentially longer at this test event. We interpret this in terms of maintained causal perception: the temporal gap here presumably did not change the causal status of the event, which in turn did not lead to longer looking.
Expt 3 Habituation: Synchronous Motion with No Launcher (with green squares)
Experiment 4: The Salience of Large Temporal Offsets
This experiment involved the same habituation event as used in Experiment 2, except that object A (the 'launcher') was not present.
Experiment 3 Test #1: Synchronous Motion with No Launcher (with red discs)
The first test event in this experiment was identical to that used in Experiment 2, except that object A (the 'launcher') was again not present.
Experiment 3 Test #2: A 120 ms Offset with No Launcher (with red discs)
The second test event in this experiment was also identical to that used in Experiment 2, except that object A (the 'launcher') was again not present. In this experiment infants did look longer at this second test event. This indicates that the 120 ms offset was visible, and perceptually salient in the absence of perceived causality. This in turn suggests that the failure to look longer at the second test event in Experiment 2 was not due to any failure to perceive the temporal offset in the first place.
Expt 4 Habituation: Synchronous Grouping (with green squares)
This experiment involved the same habituation event as used in Experiment 1.
Experiment 4 Test #1: Unambiguous Launching (with red discs)
The first test event in this experiment was identical to that used in Experiment 1, except that A and B did not fully overlap: as soon as A arrived adjacent to B, A stopped and B began moving.
Experiment 4 Test #2: Unambiguous Launching with a 600 ms Offset (with red discs)
The second test event in this experiment was identical to that used in Experiment 1, except that (as in the other test event), A and B did not fully overlap: as soon as A arrived adjacent to B, A stopped and B began moving. In adults, the lack of overlap between A and B renders the context motion irrelevant: this event will be seen in terms of causal launching no matter how or when the disc C moves. Infants did not look differentially longer to either test event. We interpret this in terms of infants enjoying the same causal percepts that adults do in these situations: infants saw all habituation and test events in terms of causal launching. This indicates that a 600 ms temporal offset is not sufficient by itself to cause dishabituation when it does not disrupt perceived causal launching. Accordingly, the dishabituation to the 600 ms gap in Experiment 1 did reflect the changed causal status of that display.
In sum, the infants in our experiments appear to have looked differentially longer at long temporal offsets when they did change causal percepts from launching to passing (in Experiment 1) but not when the display was unambiguously causal for other reasons (in Experiment 4). And infants did look differentially longer at short temporal offsets when none of the displays were perceived in causal terms (in Experiment 3), but not when the display was always perceived in causal terms (due to postdictive grouping, in Experiment 2). We conclude that infants (1) perceive causality, (2) do so on the basis of subtle and nuanced visual cues, (3) are influenced by the same cues, in the same ways, as are adults, and (4) exhibit postdictive perceptual processing.