Infants’ perceptions of constraints on object motion as a function of object shape

Cognition 165 (2017) 126–136

Contents lists available at ScienceDirect

Cognition journal homepage: www.elsevier.com/locate/COGNIT

Original Articles

Infants’ perceptions of constraints on object motion as a function of object shape Gelareh Jowkar-Baniani, Angelina Paolozza, Anishka Greene, Cho Kin Cheng, Mark A. Schmuckler ⇑ University of Toronto Scarborough, Canada

a r t i c l e

i n f o

Article history: Received 4 August 2016 Revised 27 March 2017 Accepted 29 April 2017

Keywords: Object perception Motion perception Object shape Perceptual development Visual development

a b s t r a c t Three studies examined young infants’ ability to distinguish between expected and unexpected motion of objects based on their shape. Using a preferential-looking paradigm, 8- and 12-month-old infants’ looking time towards expected and unexpected motion displays of familiar, everyday objects (e.g., balls and cubes) was examined. Experiment 1 demonstrated that two factors drive infants’ preferential fixations of object motion displays. Both 8- and 12-month-olds displayed a tendency to look at rotating information over non-rotating, stationary visual information. In contrast, only 12-month-olds showed a tendency to look at object motions that were inconsistent or ‘‘unexpected” based on shape. After controlling for the preference for more complex (rolling) by adding rolling motion to both displays (Experiment 2), 12month-olds’ ability to distinguish between expected and unexpected motion displays was facilitated. Experiment 3 provided a control by demonstrating that the preference for the unexpected object motion was not due to any other motion properties of the objects. Overall, these results indicate that 12-monthold infants have the ability to recognize the role that object shape plays in constraining object motion, which has important theoretical implications for the development of object perception. Ó 2017 Elsevier B.V. All rights reserved.

1. Introduction Infants’ perceptions of objects is arguably one of the most thoroughly studied questions in the history of research in developmental psychology (see Johnson, 2010, 2011, 2013; Slater et al., 2010, for reviews). As aptly described by Johnson (2013, p. 371), ‘‘. . .(o) bject perception is the raison d’être of the visual system.” The role of object motion in driving object perception is a particularly important, and compelling property that has been examined by a wide number of researchers in a wide array of contexts. Object motion provides critical information for the perception of threedimensional object shape and structure (Arterberry, 1992; Arterberry, Craton, & Yonas, 1993; Arterberry & Yonas, 1988, 2000; Hirshkowitz & Wilcox, 2013; Kellman, 1984; Kellman & Short, 1987; Owsley, 1983; Schmuckler & Proffitt, 1994; Soska & Johnson, 2008, 2013; Wallach & O’Connell, 1953; Yonas, Arterberry, & Granrud, 1987), the three-dimensional layout of objects and surfaces (Johnson, 2000; Johnson, Davidow, HallHaro, & Frank, 2008; Johnson & Mason, 2002), and the threedimensional completion of partly-occluded objects (Johnson, 2004; Johnson & Aslin, 1995, 1996; Kellman & Spelke, 1983; ⇑ Corresponding author at: Department of Psychology, University of Toronto Scarborough, 1265 Military Trail, Scarborough, ON M1C 1A4, Canada. E-mail address: [email protected] (M.A. Schmuckler). http://dx.doi.org/10.1016/j.cognition.2017.04.011 0010-0277/Ó 2017 Elsevier B.V. All rights reserved.

Mareschal & Johnson, 2002; Soska & Johnson, 2013). Object motion is similarly fundamental in the understanding of complex object properties such as animacy (Di Giorgio, Lunghi, Simion, & Vallortigara, 2016; Gelman, 1990; Mandler, 1992, 2003; Markson & Spelke, 2006; Poulin-Dubois, Crivello, & Wright, 2015; Rakison & Poulin-Dubois, 2001; Träuble & Pauen, 2011; Träuble, Pauen, & Poulin-Dubois, 2014). Clearly, object motion is one of the richest sources of information regarding innumerable object properties. Motion is also critical for providing information regarding fundamental constraints on object properties and behavior in the environment, which in turn provides insight into foundational aspects of infants’ knowledge and understanding of the world. Spelke and colleagues, in their theorizing on and investigations of infants’ ‘‘core knowledge” (Dillon & Spelke, 2015; Spelke, Breinlinger, Macomber, & Jacobson, 1992; Spelke & Kinzler, 2007; Spelke, Lee, & Izard, 2010; Spelke, Phillips, & Woodward, 1995; Spelke & Van de Walle, 1993) provide one of the most intriguing, and elegantly articulated, examples of the consequence of the perception of object motion. Spelke et al. (1992), for instance, found that 4-month-old infants understood basic object constraints such as solidity (solid objects do not pass through other solid objects) and continuity (objects continue to move on a given path unless obstructed in some fashion), but failed to understand effects of gravity (objects will fall unless supported by other objects or surfaces) and inertia (a free falling object will continue to fall towards

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

a support surface). According to Kim and Spelke (1992), it is not until 7 months that infants began to understand such properties, with full developmental understanding of these properties occurring between 7 months and two years of age (Kim & Spelke, 1999). Such findings demonstrate that although young infants understand basic constraints about the physical properties of objects (i.e., solidity, continuity), their understanding of the reasons underlying the behavior of objects and object movements (i.e., inertia, gravity) is a more advanced, later developmental achievement. One characteristic common to much of this work is that the relation between object motion and understanding has been primarily unidirectional, with object motion providing the vehicle for insight into a host of foundational object and world knowledge (e.g., Mandler, 1992, 2003, 2008a, 2008b, 2010, 2012). Interestingly, little work (with a handful of notable exceptions, discussed momentarily) has considered the bidirectional implications of this relation. In this case, not only does object motion illuminate critical object properties, but basic object properties can also highlight, and potentially constrain, fundamental aspects of object motion. Probably the best known example of this type of relation is seen in the perception of biological motion based on point-light visual information (Cutting, 1978, 1981, 1986; Cutting, Moore, & Morrison, 1988; Johansson, 1973; Kozlowski & Cutting, 1977; Pavlova, Krägeloh-Mann, Birbaumer, & Sokolov, 2002; Pavlova & Sokolov, 2000; Troje, 2002, 2013; Westhoff & Troje, 2007). Stemming from the classic work by Johansson (1973), point-light biological motion displays are created by attaching spots of light and/or reflective markers to the major joints of the body. Despite their relative paucity of visual information, compared to viewing these same movements under normal illumination, such displays are easily recognized by adult observers as to the behaviors undertaken by actors. Beginning with work by Fox and McDaniel (1982) researchers have found that infants are similarly responsive to biological motion information. Within the first few months of life infants can discriminate human point-light displays from nonhuman (Bertenthal, Proffitt, & Kramer, 1987), are sensitive to the orientation of biological movement (Booth, Pinto, & Bertenthal, 2002), and so on. This research is significant in that it provides indirect evidence that such structure can indeed constrain the perception of object motion. Of most direct relevance, Baker, Pettrigrew, and Poulin-Dubois (2014) investigated 10- to 20-month-old infants’ expectations for motion paths as a function of object animacy. Building from previous research examining infants’ expectations for animate actions to be extended more to animals than vehicles (Poulin-Dubois, Frenkiel-Fishman, Nayer, & Johnson, 2006), this work found that infants in this age range associated non-linear paths of motion with animate objects (animals) and linear paths of motion with inanimate moving (vehicles) and stationary (furniture) objects. Thus, infants do generate expectations for the types of motion that objects will display, with such expectations constrained by infants’ conceptual understanding of the motion cues involved with animacy (Mandler, 1992, 2000, 2010, 2012). Accordingly, there is evidence to suggest that basic object properties might influence ones’ expectations for object movement. The goal of the current series of experiments was to examine this question more directly, within the context of infants’ expectations for object movement based on physical shape.

2. Experiment 1: The role of object shape in expectations of object motion The principal goal of this first experiment was to examine whether an object’s shape would influence infants’ expectations

127

for the type of motion typically produced by this object. As an example, objects that are round, such as balls, move with a characteristic motion – they roll. Given its lack of flat surfaces, a significantly more non-characteristic motion for such an object would be for it to slide. In this case, the physical shape of the object drives expectations for the type of motion that should be produced by the object. By way of contrast, a differently structured object, such as a cube, because of its edges, corners, and flat surfaces, is more likely to move in a sliding fashion than a rolling fashion. Accordingly, the physical structure of a cube drives different, and opposite, expectations for object motion. Expectations for the motion of these two objects, a ball versus a cube, were examined using the violation-of-expectation preferential looking paradigm (Baillargeon, Spelke, & Wasserman, 1985; Spelke et al., 1992). Based on previous findings, two target ages were identified – 8- and 12-month-old infants. Eight months is of interest given the literature suggesting that infants between 5 and 7 months become sensitive to the impact of physical structure on perceived motion in biological displays (e.g., Bertenthal, Proffitt, & Cutting, 1984; Bertenthal et al., 1987; Pinto, Bertenthal, & Booth, 1996). In contrast, research on infants’ perceptions of complex object constraints such as gravity and inertia (Kim & Spelke, 1992, 1999; Spelke et al., 1992) indicate that such knowledge is not available until later in development, between 10 and 16 months (Kanass, Oakes, & Wiese, 1999). Based on these findings, if infants are sensitive to the impact of object shape on object movement we would predict that by 12 months infants will preferentially fixate unexpected object movements (a sliding ball and a rolling cube) over expected movements (a rolling ball and a sliding cube). Predictions for 8-month-olds are more variable. If 8-montholds are generally sensitive to constraints on object movement imposed by shape, they should preferentially fixate unexpected displays. Alternatively, if an understanding of how object shape constrains object movement aligns with knowledge of complex object properties such as gravity and inertia, 8-month-olds will not have strong expectations for object movement based on shape. Instead, looking may be driven by other motion characteristics of the objects. 2.1. Methods 2.1.1. Participants Sixteen 8-month-olds (M = 7.74 months, SD = 0.35 months) and sixteen 12-month-olds (M = 11.92 months, SD = 0.39 months) participated in this study. Ten additional 8-month-olds and one 12month-old also participated but their data were excluded due to fussiness (N = 3) and technical errors (N = 8). The names of the participants were obtained from a database maintained at the Laboratory for Infant Studies at the University of Toronto Scarborough and parents were contacted by telephone. All participants were recruited from the demographically diverse Greater Toronto area and received a certificate and a toy for their participation. 2.1.2. Stimuli Using the animation program 3DS Max, four video displays, shown schematically in Fig. 1, were created for this experiment. Two of the displays involved a colorful checkered ball with a diameter of 6.5 cm rolling (Fig. 1a) or sliding (Fig. 1b) across a flat surface and the other two showed a similarly patterned cube with a width of 6.5 cm sliding (Fig. 1c) or rolling (Fig. 1d) across a flat surface. Because both stimuli had comparable diameters, the visual angle for both the ball and cube was 5.7°
5.7°. Rolling is considered the typical movement of a ball (expected event), whereas sliding is an atypical motion (unexpected event). For the cube, the motion patterns were reversed; sliding is considered the typical movement (expected event) whereas rolling is atypical

128

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

Fig. 1. Motion stimuli used in Experiment 1. The Ball stimulus: A: Expected (rolling) motion, and B: Unexpected (sliding) motion. The Cube Stimulus: C: Expected (sliding) motion, and D: Unexpected (rolling) motion.

(unexpected event). Each motion display was 7 s in duration. In all displays, the object entered the screen in motion, meaning that there was no information regarding motion onset. Based on these displays, four experimental trials were created. Two trials consisted of simultaneous presentations of the expected and unexpected ball displays, and two trials consisted of presentations of the expected and unexpected cube displays. Within each of these two presentations, the side of appearance of expected and unexpected displays was counterbalanced (i.e., expected display on the left side, unexpected display on the right, versus unexpected on the left, expected on the right). Each motion was repeated four times continuously, resulting in trials that were 28 s in total length

mouse when infants looked away. The experimenter controlling the computer was unaware of the condition being presented on any given trial.

2.1.3. Apparatus During the experiment, infants faced a pair of Samsung (1700 ) CT-3312 VC colour video monitors, placed directly in front of them. These monitors were positioned on a 75 cm high table with a distance of about 20 cm between them. The monitors were connected to a G4 Macintosh computer which controlled the stimulus presentation, using Habit 2000 infant presentation software (Cohen, Atkinson, & Chaput, 2000). A Sony Digital Handycam DCR-TRV 340 video camera (Sony Corporation, Japan) was positioned between the two monitors to record the infant’s face. This camera allowed an observer outside of the testing area to maintain an image of the infant’s face via a 13-inch monitor. The infant’s looking behavior was recorded on an PC compatible computer. Fixations to the left or right monitor were recorded by pressing the left or right button on the mouse, respectively, and releasing the

2.1.5. Reliability Reliability was assessed by a second observer, who was similarly unaware of the conditions on each trial, for all but four participants (due to poor quality session recording) using three indices. First, each observer recorded which stimulus was looked at longer to determine an overall pattern of preferential looking, with looking potentially unreliable if there was a difference in the overall pattern of preferential fixation; no participants were flagged by this criterion. Second, the mean absolute difference between coders for the total fixation time was calculated for each infant. Neither age group exceeded a cutoff value of 2.0 s (Ms = 1.10 and 1.35, SDs = 0.86 and 1.08, for 8- and 16-month-olds, respectively). Finally, for each infant, the correlation between the two sets of looking times was calculated; any correlation <0.90 was flagged as potentially unreliable. Codings for the twelve-month-olds were

2.1.4. Procedure Infants were seated on their parent’s lap about 65 cm away from the monitors, with parents sitting on a chair positioned at a right angle to the monitors. Thus, infants faced the monitors while parents faced the wall, and did not look at the displays during the experiment. The observer outside of the testing area initiated the experiment, which began with an attention-getting display. Once the infant was looking at the monitors, the first trial was presented.

129

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

highly correlated, r’s(8) = 0.90–0.98, p’s < 0.01–0.001. For the 8month-olds, all codings but one had strong correlations, r’s(8) = 0.90–0.90, p’s < 0.01–0.001, with the coding for this remaining infant having a weaker the reliability, r(8) = 0.80, p < 0.025. Given that the previous reliability measures did not flag this participant, the data for this infant was included in subsequent analyses.

Motion Type Expected Unexpected

2.2. Results and discussion A mixed model Analysis of Variance (ANOVA) with three within-subject variables of Stimulus Type (ball versus cube), Motion Type (expected versus unexpected) and Trial (first versus second), and a single between-subject variable of Age (8- versus 12month-olds) was used to analyze looking behaviors. The ANOVA revealed a main effect of Motion Type, F(1,30) = 9.52, MSE = 148.25, p < 0.005, g2p = 0.24, with longer looking towards the unexpected displays (M = 8.82, SE = 0.52) compared to the expected displays (M = 7.30, SE = 0.28). There was also a main effect of Trial, F(1,30) = 7.88, MSE = 57.40, p < 0.01, g2p = 0.21, with longer looking on the first presentation of the stimuli (M = 8.53, SE = 0.40) compared to the second (M = 7.58, SE = 0.37). There was also a significant Stimulus Type by Motion Type interaction, F (1,30) = 76.04, MSE = 994.44, p < 0.001, g2p = 0.72, with longer looking towards the expected motion for the ball, t(31) = 4.13, p < 0.01, and longer looking towards the unexpected motion for the cube, t (31) = 7.18, p < 0.01. All of these findings were qualified by the significant Stimulus Type by Motion Type by Age interaction, F (1,30) = 4.19, MSE = 54.84, p < 0.05, g2p = 0.12. This interaction, shown in Fig. 2, indicated that 8-month-olds look longer towards the expected (rolling) motion for the ball, t(15) = 5.08, p < 0.001, and the unexpected (rolling) motion for the cube, t(15) = 4.80, p < 0.001. In contrast 12-month-olds looked longer at the unexpected (rolling) cube display, t(15) = 5.86, p < 0.001, but looked equally towards the expected and unexpected ball display, t(15) = 1.49, ns. No other main effects or interactions were significant. The findings of Experiment 1 provide initial insight into infants’ expectations for object motion based on shape. Specifically, 8month-olds do not appear to expect object motion to be constrained by shape. Instead, these infants showed a preference for rolling versus sliding movements. One obvious basis for such a preference is that rolling motion provides a combination of rotational and translational movement, as opposed to the strict translational motion present in sliding displays. One possibility is that the simultaneous availability of the rotational and translational information is perceived as being more complex than the simple translational motion; such preferential fixation would thus indicate the well-known preference for more versus less complex displays (Cohen, 1988, 1991; Cohen, DeLoache, & Rissman, 1975; Karmel & Maisel, 1974; Kidd, Piantadosi, & Aslin, 2012; see Richards, 2010, for a review). The findings for the 12-month-olds are trickier to interpret, and appear to show the influence of two dissociable processes driving looking behavior. On the one hand, as with the 8-month-olds, there is a preference for more complex displays over simpler ones. This preference drives looking towards the rolling display for both the ball and cube. On the other hand, there is a tendency to look at the unexpected motion of the object, based on the physical shape. This preference drives looking towards the sliding ball, and the rolling cube. When these two processes combine they produce exactly the pattern of findings observed for the 12-month-olds. For the ball, the preference for complexity drives looking to the expected display of the rolling ball, whereas the preference for the novelty of the unexpected display drives an interest in the sliding ball. Assuming these processes are relatively equal in strength these two tendencies thus produce equivalent interest in the expected and unexpected displays for the ball, albeit for different

Fig. 2. Mean (and standard error) looking times towards the expected and unexpected motion displays as a function of the object’s shape for 8-month-olds and 12-month-olds in Experiment 1.

reasons. In contrast, for the cube the preference for complexity drives looking towards the unexpected rolling cube, and the preference for novelty similarly drives looking to the unexpected rolling cube. Accordingly, 12-month-old infants showed a strong looking preference for the unexpected cube display. Although speculative, the operation of distinctive, simultaneous processes driving looking behavior has been noted in previous research (Jowkar-Baniani & Schmuckler, 2011; Kavšek & Yonas, 2006). For instance, Jowkar-Baniani and Schmuckler (2011), in an investigation of infants’ ability to generalize from 2D line drawings to 3D objects observed two simultaneous processes at work in driving infants’ looking behavior – a general preference for human type objects (i.e., a doll) over animal displays (i.e., a sheep), and the classic novelty effect typically associated with infant habituation paradigms. When these two processes converged in driving looking towards a single display infants showed strong visual discrimination; however, when they diverged in driving interest towards different displays infants exhibited what appeared to be a failure of discrimination. Interestingly, when Jowkar-Baniani and Schmuckler reduced the preference for the human object infants then showed a pattern of novelty preferences only. 3. Experiment 2: The influence of rotational motion in object movement displays The goal of Experiment 2 was to examine whether 12-monthold infants’ failure to fully discriminate the implications of object shape on object motion was driven by simultaneous interest in the novelty of unexpected events and a general preference for rolling displays. Reflecting on the rolling displays of Experiment 1, the most obvious aspect that could drive interest in such movement involves the rotational component of the visual structure contained on the surface of the ball and cube. The most straightforward means of examining this issue involves removing this preference by eliminating the rotational component of this display;

130

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

this would then enable an interest in the unexpected motion of the display to emerge unmasked by competing interests. Unfortunately, one cannot simply eliminate rotation from the rolling displays, as this would, by definition, remove the information that is critical for determining rolling versus sliding movements. Given this limitation, the current study created two complementary displays to be used in assessing the impact of a general preference for rotational information. The first involved decreasing the amount of rotational information within a rolling ball display. This goal was achieved by modifying the ball stimulus to include a stationary, inner circle that did not rotate with the ball’s movement, while maintaining a rotating outer ring that was in contact with the support surface.1 The second display inserted rotational motion into a sliding ball display by including an inner circle that rotated with the movement of the ball, with an outer ring that remained stationary during movement. Using these two displays it is possible to test infants’ preferences for rotational information per se, as well as to examine interest in rotating displays that move in an expected versus unexpected fashion, based on object shape. Specifically, if 12month-old infants have a general preference for rotational information then we would predict that infants will prefer a display containing rotational information over a display without rotational information, even when the amount of rotation has been dramatically reduced. Moreover, if 12-month-olds are sensitive to the expectedness of motion based on object shape then we would predict that infants will prefer a display containing unexpected object motion over a display containing expected object motion, even when both displays contain some degree of rotating information. 3.1. Method 3.1.1. Participants Because only 12-month-olds in Experiment 1 appeared to be sensitive to the constraints on object movement as a function of object shape, this study restricted its focus to this age group. Accordingly, twenty 12-month-olds (M = 11.91, SD = 0.32) participated in this study. The names of the participants were obtained via the same means as Experiment 1, and participants received comparable compensation for their participation. 3.1.2. Stimuli, conditions, apparatus, procedure, and reliability Three new animated video displays, shown schematically in Fig. 3, were created for this experiment using Adobe Illustrator. All three stimuli had comparable 6.5 cm diameters to the ball stimulus of Experiment 1, thus subtending a visual angle of 5.7°
5.7°. The first, or the rolling ball with the stationary inner ring (R + St) display consisted of a colorful ball rolling across the screen. In contrast to the rolling ball display of Experiment 1, this display contained a stationary inner circle that did not rotate as the ball translated across the screen. Based on its two-dimensional projection, the ball stimulus had an inner, stationary circle with a radius of 2.2 cm, producing a stationary area of 15.2 cm2 and an outer radius of 3.2 cm on the screen, thereby producing a surface area of 33.2 cm2 with a surface area of 18.0 cm2 of rotating information (33.2–15.2 cm2). Accordingly, the ratio of rotating to stationary information was approximately 1:1. As in Experiment 1 all ball displays entered the screen in motion. The second display consisted of a sliding ball with a rotating inner ring (S + Ro). Accordingly, the outer area of the ball in contact with the support surface was stationary, whereas the inner circle rotated. The respective radii of the 2D projection of the ball and 1 For readers who reside, or frequent, urban centers this stimulus is comparable to what is often seen on the hubcaps of taxis and buses, which contain stationary advertising information affixed to the tires of these vehicles.

the inner circle were the same as in the R + St display, producing a ratio of rotating to stationary information of approximately 1:1. Finally, the third display consisted of a sliding ball with a stationary inner ring (S + St). Because both inner and outer components of this display did not rotate this display was equivalent to the sliding ball of Experiment 1, although there were noticeable texture differences between the inner and outer rings for this display (see Fig. 3). Given the more unusual nature of the stimuli in this study it was felt desirable to provide infants with increased experience with the translation of the ball across the screen. However, to maintain comparability with the previous study in terms of potential exposure to the moving stimuli, it was also important to keep the overall trial length roughly equivalent. Accordingly, the speed of translation across the screen was increased to 3 s. By looping the 3 s displays, this increased speed provided 10 repetitions of the movement in a 30 s trial. These three motion displays enabled two critical looking comparisons, each of which tests the two processes hypothesized to be operative in the previous experiment. In the first comparison, the rolling ball with the stationary inner ring (R + St) display was presented simultaneously with the sliding ball with the stationary inner ring (S + St). If infants truly do prefer displays containing rotating motion, then they should show a clear preference for the R + St display compared with the S + St display, given that the former contains rotational information whereas the latter does not. Because this display has less rotation than the corresponding rolling ball display in Experiment 1, this comparison provides a strong test of 12-month-olds’ preference for rotational information over non-rotational information. In the second comparison, the rolling ball with the stationary ring (R + St) display was paired with the sliding ball with the rotating inner ring (S + Ro) display. This comparison equates the two displays in terms of the general presence of rotational information in the display, but places this rotating information in different parts of the display. If 12-month-olds are indeed sensitive to the consequences of object shape on object motion, then infants should show a preference for the S + Ro relative to the R + St display, given that because the rotating information in the S + Ro display is not in contact with the support surface the motion of this object would be considered unexpected. In contrast, because the rotating information in the R + St display is in contact with the support surface, the motion of this object would be considered expected. Parenthetically, this comparison tests whether infants are indeed aware that for object movement through the world, one critical source of information for expected motion is whether the object surface in contact with the support surface moves in the ‘‘expected” manner. The same apparatus was used as in Experiment 1, with the following exceptions: the display monitors were connected to an AMD 1.8 GHz IBM PC computer which controlled the stimulus presentation, and that a JVC Digital Handycam GZ-MG330 video camera (JVC Corporation, Japan) recorded the infant’s face and looking behavior. The procedure was comparable to Experiment 1, with two critical comparisons (R + St versus S + St, and R + St versus S + Ro), each counterbalanced for left – right position. The order of presentation of the two comparisons was also counterbalanced, with half the infants receiving the R + St/S + St comparison first. Because trials were 30 s long, the total experiment lasted less than five minutes. Reliability was calculated using the same three indices as in Experiment 1. For the first criterion the looking patterns were identical between the two coders, and for the second criterion there were no codings with a mean difference value of greater than 2.0 s (M = 1.20, SD = 1.08). Finally, for all but two sets of codings the correlations ranged from r(8) = 0.90–0.97 (all p’s < 0.01). For the remaining two infants because the previous two criteria indicated

131

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

Fig. 3. Motion stimuli employed in Experiment 2. A: The rolling ball with the stationary inner ring (R + St) stimulus; B: The sliding ball with the rotating inner ring (S + Ro) stimulus; C: The sliding ball with the stationary inner ring (S + St) stimulus.

their codings were reliable, their data were included in all subsequent analyses.

Motion Type Expected Unexpected

3.2. Results and discussion Looking behavior was analyzed using a mixed model ANOVA with three within-subject variables of Stimulus Pair (R + St/S + St versus R + St/S + Ro), Motion Type (expected vs. unexpected), and Trial (first stimulus comparison versus second stimulus comparison), and a between-subject variable of Order (R + St/S + St versus R + St/S + Ro first). Of the four main effects, the only significant result was a main effect of Trial, F(1,18) = 19.30, MSE = 42.88, p < 0.001, g2p = 0.52, indicating that the infants looked longer to the initial stimulus pair (M = 5.25 s, SE = 0.37 s) compared to the subsequent stimulus pair (M = 4.21 s, SE = 0.35 s). Three of the two-way interactions were significant, including the Stimulus Pair
Order interaction, F(1,18) = 21.30, MSE = 273.47, p < 0.001, g2p = 0.54, and the Trial
Order interaction, F(1,18) = 9.00, MSE = 20.01, p < 0.01, g2p = 0.33. The first of these interactions indicates average looking to the R + St/S + St stimulus pair, relative to the R + St/S + Ro stimulus pair, regardless of which pair appeared first. The second interaction reveals that infants who received the R + St/S + St stimulus pair first looked longer on the first presentation of these stimuli than the second, whereas infants who saw the R + St/S + Ro stimulus pair first show no difference between their two presentations. Most critical was the Stimulus Pair
Motion Type interaction, F (1,18) = 9.43, MSE = 51.89, p < 0.01, g2p = 0.34. This interaction appears in Fig. 4, and indicates that when the displays varied in rotational information (i.e., R + St versus S + St) infants looked longer at the display with the rotational information (the R + St display), M = 4.95 s, SE = 0.62, than the display with no rotational information (the S + St display), M = 3.82 s, SE = 0.49, t(19) = 3.60, p < 0.01. However, when rotating information was injected into both rolling and sliding displays (R + St versus S + Ro), infants looked longer at the sliding ball (S + Ro) display, M = 5.65 s, SE = 0.68, compared to the rolling ball (R + St) display, M = 4.50 s,

Fig. 4. Mean (and standard error) looking times towards the expected and unexpected motion displays in Experiment 2. R + St = rolling ball with stationary inner ring; S + St = sliding ball with stationary inner ring; S + Ro = sliding ball with rolling inner ring.

SE = 0.46, t(19) = 2.10, p < 0.05. No other significant interactions were found. Overall, Experiment 2 clarifies and extends the findings of Experiment 1. The fact that infants continued to prefer the rolling ball display compared to the sliding display is consistent with the previously hypothesized preference for rotating versus nonrotating information, even when the amount of rotation in the former display was reduced by adding the stationary inner circle. This finding converges well with the implications of the previous experiment regarding the availability of motion information and the perceived complexity of a visual display, thus representing a preference for displays containing more versus less complex visual information. Of even greater interest is that 12-month-olds showed a preference for the sliding ball display, compared to the rolling ball display, when both displays contained rotating information. This result strongly demonstrates that infants of this age are sensitive to the expected versus unexpected motions of objects based on

132

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

shape. Because these displays were roughly equated in the presence of rotating information, infants’ attention was free to focus on the expected versus unexpected nature of the stimulus motion. Looking across these two comparisons, then, this experiment confirms the previous hypothesis of two processes driving infant looking towards these displays – a general interest in rotating versus non-rotating information, and a recognition of the expected nature of object motion based on object shape. Comparing across Experiments 1 and 2, one can conclude that 12-month-old infants do indeed use object shape to generate expectations for multiple objects with different shapes (cubes – Experiment 1; balls – Experiment 2). Also, generalizing from the current findings these studies suggest that infants understand that it is the outer surface of the object – the surface that is in contact with the support surface – that needs to move consistently with the global translational movement of the object. Before drawing any final conclusions, however, it is important to assess a possible alternative explanation for infants’ preferential looking when comparing the outer versus inner rotation information. Specifically, one relatively uninteresting alternative account of the previous finding is that infants’ preference for the sliding ball with the rotating inner center (the S + Ro display) arose due to a general preference for interior circular motion, relatively to the exterior circular motion of the rolling ball with the stationary inner center (the R + St display). If true, this account would undermine any explanation based on the perceived expected versus unexpected nature of the object motion. Experiment 3 assessed this alternative hypothesis. 4. Experiment 3: Preference for inner vs. outer rotations One straightforward means of testing the just described alternative explanation is to eliminate the translatory component of the displays, presenting balls with rotary inner versus outer motion that do not engage in movement across the screen. If infants do have a general preference for rotating inner motion (in the absence of global object translation), relative to rotating exterior motion, then we would predict that infants will preferentially fixate a display with rotating inner motion, relative to a display with rotating exterior motion. In contrast, if the pattern of preferential looking observed in the previous experiment was driven by aspects related to object motion per se, then we would predict that removing translational movement will eliminate preferences for one versus the other display. The current study tested these predictions. 4.1. Method 4.1.1. Participants Twenty 12-month-olds (M = 12.21, SD = 0.58) participated in this study. The names of the participants were obtained via the same means as Experiment 1 and 2, and participants received comparable compensation for their participation. 4.1.2. Stimuli, conditions, apparatus, procedure and reliability Two video displays were created for this experiment using Adobe Illustrator. The displays were similar to those used in Experiment 2 with a colorful ball containing an inner ring. In one display, the outer ring of the ball with a diameter of 12.5 cm rotated while the inner ring with a diameter of 6.2 cm was stationary, whereas in the second display the same size outer ring was stationary while the same size inner ring rotated. As in the previous experiments, these displays were seen rotating on a visible ground surface. To better attract and hold infants’ attention the stimuli were larger that in the previous two experiments, thereby subtending larger

visual angles of 11.8°
11.8°. However, the approximate 1:1 ratio of rotating to stationary information from the previous two studies was maintained. In contrast to the previous two experiments, both balls remained in the center of the display. As in Experiment 2, individual displays rotated for 3 s, which was then looped 10 times to produce 30 s trials. The two displays were presented two times, randomly counterbalancing for side. All remaining aspects of the apparatus and procedure were identical to Experiment 2. Reliability was calculated for all participants using two of the indices of the previous studies; because there were only a small number of critical trials here (based on counterbalancing side of presentation of the two displays) correlations were not calculated. Again, the looking patterns of participants did not differ between the two coders, nor did reliability differences exceed a mean cutoff value of 2.0 s (M = 0.89, SD = 0.79). 4.2. Results and discussion Looking behavior was analyzed in a repeated-measures ANOVA, with two within-subject variables of Motion Type (outer rotation versus inner rotation), and Trial (first versus second trial). The only significant effect was a main effect of Trial, F(1,19) = 11.07, MSE = 31.84, p < 0.005, g2p = 0.37, with infants displaying longer looking in the first presentation of the stimuli (M = 6.70, SE = 0.63) compared to the second (M = 5.44, SE = 0.52). Most importantly, there was no effect of Motion Type, F(1,19) = 0.44, MSE = 4.14, p = 0.52, with infants looking equally towards the outer (M = 6.30, SE = 0.63) versus inner rotations (M = 5.84, SE = 0.66). Briefly put, Experiment 3 demonstrates that infants do not have any a prior preference for exterior versus interior rotations of an object. As such, the most parsimonious explanation for the observed preferences in Experiment 2 focuses on infants’ recognition of sliding as an unexpected motion for the translation of the ball. Overall, the results of all three studies converge in their implications that, by 12 months, infants can use the shape of an object to constrain their expectations for the motion of that object in the world. As a caveat to this interpretation, it is important point to note that in Experiment 2 the objects combined translational and rotational motion, whereas in Experiment 3 the objects contained only rotational information. Accordingly, one underlying assumption to this study is that translational and rotational movements engage motion processes similarly in infants of this age, a potentially worrisome assumption given the evidence that infants process objects and object characteristics (e.g., shape, speed of movement) differently depending on the nature of the motion information presented (Bower, Broughton, & Moore, 1971; Burnham, 1987; Burnham, Vignes, & Ihsen, 1988; Day & Burnham, 1981; Kaufman, Mareschal, & Johnson, 2003; Ruff, 1982, 1985, 1987). Ruff (1982, 1985, 1987), for instance, has demonstrated that 6month-olds were best at extracting object form when observing translating object movement, as opposed to rotations, or combined translations and rotations. Such findings fit into theoretical frameworks distinguishing dorsal versus ventral processing (e.g., Milner & Goodale, 2006; Ungerlieder & Mishkin, 1982) with dorsal streams processing moving stimuli and ventral streams processing stationary objects (see Kaufman et al., 2003, for an application to infant object processing). Fortunately, these differences in object processing as a function of motion information, although interesting, do not undermine the usefulness of Experiment 3 as a control condition for a few reasons. First, the previous work has typically examined younger infants ranging in age from 2 to 6 months, whereas the current studies have focused principally on 12-month-old infants; it is not clear whether the previously found differences are, in fact, operative with older infants. Second, despite this earlier work demonstrating

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

variation in object processing as a function of motion information, there is no suggestion that infants at these younger ages are in any critical sense unable to make use of rotational information. In fact, multiple authors (Gilmore, Hou, Pettet, & Norcia, 2007; Ruff, 1985, 1987; Shirai, Kanazawa, & Yamaguchi, 2008) have demonstrated that very young infants are exquisitely sensitive to rotational motion, and can and do discriminate such movement from translatory motion (e.g., Ruff, 1985). Third, and finally, it should be remembered that the goal of Experiment 3 was limited to providing a control for the alternative explanation that the observed preference in Experiment 2 for a stationary ball with a rotating interior arose due to an a prior preference for interior versus exterior rotations that had nothing to do with the implications of object translation across a display. Given that Experiment 3 was successful in demonstrating that infants had no preference for interior versus exterior rotations per se, any explanation for this earlier preference must incorporate the implications of the translational motion across the display. To our minds, the most straightforward explanation involves what such translation implies vis a vis object shape.

5. General discussion In summary, three experiments provided evidence that 12month-old, but not 8-month-old infants, generated expectations for object motion based on shape. Specifically, Experiment 1 demonstrated the operation of two factors driving infant preferential fixation of object motion displays – a tendency to look at rotating information over non-rotating information (presumably due to the increased complexity of the former relative to the latter), and a tendency to look at object motions that were inconsistent or ‘‘unexpected” based on the shape of an object (presumably based on the novelty of the motion for this shape). Eight-month-olds’ preferential fixation was found to be driven by only the first of these two factors, with infants looking more at displays containing rotating information irrespective of the actual object and its shape. Twelve-month-olds, however, showed evidence of both factors driving looking, leading them to discriminate the expected from unexpected motion of a sliding versus rolling cube (in which both factors drive looking towards the unexpected display), but failing to discriminate the expected versus unexpected motion of a rolling versus sliding ball (in which the preference for rotation drives looking towards the rolling ball whereas the preference for novelty drives looking towards the sliding ball). Experiment 2 confirmed this hypothesis of two factors driving looking, demonstrating a preference in 12-month-olds for rotating information when rolling versus sliding balls were clearly dissociated by the presence versus absence of rotation, and a preference for the novelty of the unexpected display when rolling versus sliding balls both contained some degree of rotating information. Finally, Experiment 3 demonstrated that the previously observed preference for the unexpected display was not due to an a priori preference for rotation occurring in some specific area of the visual display. Together, these results are persuasive as to 12-month-olds’ sensitivity to perceived constraints on object movement as a function of object shape. Of course, the parameter of object shape is likely not the only component that contributes to infants’ (or adults’ for that matter) expectations for object movement. Indeed, any number of factors might play a role in this regard. One likely candidate involves the propulsive agent for object movement, with different forms and strengths of movement initiation leading to different expectations for object movement. For instance, lightly to moderately pushing an object across a flat surface is likely to give rise to just the expectations for object movements as examined here. For a ball the expectation would be for rolling. For a cube the expectation would

133

be for smooth sliding, although even this expectation is further constrained by assumptions regarding the texture/smoothness and friction of the support surface. However, a more forceful push, or the actual throwing of the objects, could reasonably produce different expectations for object movements, including a rolling motion for the cube (think a thrown set of dice), or a possible bouncing movement for either object. If objects were placed on top of a ramp, propulsion due to gravity could also produce different expectations for object movement, again with these expectancies varying in terms of the steepness of the ramp, the friction of the ramp, the smoothness of the ramp, along with, of course, object shape. Clearly, the factors influencing expectations for object movement are complex, relying on a host of divergent information. Given the potential complexity of these interactions, it is important to remember that the object movement stimuli examined in this study were purposefully non-informative as to their propulsive event. Given that these objects entered the visible display in motion, and moved across the display at a moderate pace, we would argue that such motion information is strongly consistent with a light to moderate propulsion, which then drove expectations for sliding motions of a cube versus rolling motions of a ball. It would be of significant interest in future work to more explicitly manipulate information regarding object propulsion to explore the nature of its inter-relation with properties such as object shape, surface friction, surface texture, and so on. Of course, this assumption leads naturally to the question of why infants would assume a light to moderate propulsion event, and thus that the ball should roll while the square should slide? In this regard there are at least two possibilities for this assumption, both, interestingly, based on statistical learning processes, which have been found to be ubiquitously powerful across a wide array of domains, including both speech and visual events (Aslin, Saffran, & Newport, 1998; Kirkham, Slemmer, & Johnson, 2002; Kirkham, Slemmer, Richardson, & Johnson, 2007; Marcovitch & Lewkowicz, 2009; Saffran, Aslin, & Newport, 1996; Saffran, Johnson, Aslin, & Newport, 1999; see Lany & Saffran, 2013, for a review). First, it is possible that infants have learned over the course of development that most objects when moving through the world do so based on relatively moderate propulsion events. As such, they simply assume that the most likely propulsion event to have occurred is one that was moderate. Alternatively, and relatedly, it may be that infants learning is not focused on the propulsion event per se, but rather on having learned that rolling objects are typically circular, whereas sliding objects typically have a flat face. In this sense a sliding ball or a rolling square would be anomalous based on previous experiences with the world, and would be detected and responded to as such. Unfortunately, the current study cannot distinguish between these two interesting alternatives. Ultimately, expectations for object movement based on either the statistically typical type of propulsive event or movement of objects are constrained by the perceived animacy of the displays. If infants were perceiving these objects to be animate, then any a priori expectations for types of object movement are complicated significantly. Given that infants within the first year of life can distinguish between animate and inanimate objects (e.g., Baker et al., 2014; Behl-Chadha, 1996; Di Giorgio et al., 2016; Mandler & McDonough, 1993, 1998; Poulin-Dubois, Lepage, & Ferland, 1996; Rakison & Poulin-Dubois, 2001; Träuble et al., 2014; see Simion, Bardi, Mascalzoni, & Regolin, 2013, for a review), the question of whether or not infants perceived these displays is animate is important. Based on previous research, the principal cues for animacy are self-propelled motion (Di Giorgio et al., 2016; Markson & Spelke, 2006; Poulin-Dubois et al., 1996; Saxe, Tennenbaum, & Carey, 2005; Träuble & Pauen, 2011; Träuble et al., 2014) and goal-

134

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

directedness (Gergely, Nadasdy, Csibra, & Bíró, 1996; Luo & Baillargeon, 2005; Rochat, Morgan, & Carpenter, 1997; Schlottmann & Ray, 2010; Shimizu & Johnson, 2004), with goaldirectedness the more critical of the two criteria. With regard to the current study although the present stimuli could be seen as self-propelled, they do not provide any information for goaldirected motion (Csibra, Biró, Koós, & Gergely, 2003; Gergely et al., 1996; Guajardo & Woodward, 2004; Woodward, 1998, 1999, 2009; Woodward, Sommerville, & Guajardo, 2001). Accordingly, it seems unlikely that infants in this study attributed any form of animacy to the current displays. One of the more interesting findings in this project was the developmental difference between 8- and 12-month-olds in their sensitivity to the expected versus unexpected motions. Whereas both age groups showed an interest in the rotational information, only the 12-month-olds discriminated displays based on the expected versus unexpected nature of the object motion. As already suggested this first finding is not surprising given the wealth of data demonstrating that infants (Cohen, 1988, 1991; Cohen et al., 1975; Kidd et al., 2012; see Richards, 2010, for a review) and adults (e.g., Berlyne, 1958a, 1958b, 1960, 1963; Karmel & Maisel, 1974) vary their attention as a function of the perceived complexity of stimuli. Of course, the relation between complexity and attention/preference is not necessarily linear, with some arguing that complexity and preference are better described with an inverted U-shape (e.g., Dember & Earl, 1957; Vitz, 1966); this issue has yet to be fully investigated in developmental work (but see Lewkowicz & Turkewitz, 1981 and Mendelsohn, 1986, for evidence in this regard). What is relatively new with regards to this finding vis a vis complexity is the explicit relation between perceived complexity (as indicated by preferential fixation) and the amount of observable motion information. Interestingly, research on motion perception suggests that global motion information is generally decomposable into dimensions of translational, circular, and radial motion (Burr, Badcock, & Ross, 2001; Lee & Lu, 2010), with researchers debating whether the processing of such information is comparable across motions (Aaen-Stockdale, Ledgeway, & Hess, 2007; Blake & Aiba, 1998) or varies with the form of the motion (Edwards & Badcock, 1996; Freeman & Harris, 1992; Lee & Lu, 2010). Accordingly, further investigation of the relation between perceived complexity/preference and the amount and type of motion information in infants would seem to be a fruitful avenue for future research. The second of these findings – the limitation of sensitivity to the expected versus unexpected nature of the displays by 12-montholds – is both intriguing and novel, This result fits well with previous research investigating core knowledge (Dillon & Spelke, 2015; Spelke & Kinzler, 2007; Spelke & Van de Walle, 1993; Spelke et al., 1992, 1995, 2010), and their sensitivity to the effect of differing physical principles and parameters on object movement. Spelke et al. (1992), for instance, found that although young infants were sensitive to constraints that lay at the center of physical conceptions, such as continuity and solidity, other constraints, such as gravity and inertia were later development conceptions. Indeed, according to Kim and Spelke (1992, 1999), understanding of object motion constraints may not be in place until anywhere from 2 to 5–6 years. The critical point here is not as much the exact timing and acquisition of such understanding, but instead, this finding is important in demonstrating another example of decalage in children’s understandings of different forms of conceptual knowledge that share certain critical underlying similarities (e.g., Keil, 1989; Keil & Banterman, 1984; Morton & Munakata, 2002; Satlow & Newcombe, 1998). Given such findings, it does become of interest to look more closely at the parameters of the different types of knowledge constraints examined in infancy work, in an attempt

to determine a theoretical framework for the observed staggered acquisition of such knowledge. In conclusion, the current studies have investigated an intriguing constraint on infants’ perceptions of object motion – the role that object shape may or may not play on infants’ expectations for the typical motion of objects. These results have not only outlined a particular developmental progression in the use of such information, but have also provided potential insight into constraints in an array of peripheral domains, such as basic mechanisms of conceptual development and the understanding of animacy. In general, such findings, and their broader implications, highlight the central contribution that the perception of objects and their movements makes to infants’ growing understanding of the world around them. Although clearly not a surprising result when conceived in this fashion, this recognition further underscores the integrated nature of perceptual and cognitive processing early, and throughout, development. Acknowledgements The research was supported by a grant from the Natural Sciences and Engineering Research Council to Mark A. Schmuckler. The authors wish to thank Scott Johnson and two anonymous reviews for helpful comments on an earlier draft of this manuscript. Correspondence concerning this article should be addressed to Mark Schmuckler, at [email protected]. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.cognition.2017. 04.011. References Aaen-Stockdale, C., Ledgeway, T., & Hess, R. F. (2007). Second-order optic flow processing. Vision Research, 47, 1798–1808. Arterberry, M. E. (1992). Infants’ perception of three-dimensional shape specified by motion-carried information. Bulletin of the Psychonomic Society, 30, 337–339. Arterberry, M. E., Craton, L. G., & Yonas, A. (1993). Infants’ sensitivity ot motioncarried information for depth and object properties. In C. E. Granrud (Ed.), Visual perception and cognition in infancy. Carnegie Mellon symposia on cognition (pp. 215–234). Hillsdale, NJ: Erlbaum. Arterberry, M. E., & Yonas, A. (1988). Infants’ sensitivity to kinetic information for three-dimensional object shape. Perception & Psychophysics, 44, 1–6. Arterberry, M. E., & Yonas, A. (2000). Perception of three-dimensional shape specified by optic flow by 8-week-old infants. Perception & Psychophysics, 62, 550–556. Aslin, R. N., Saffran, J. R., & Newport, E. L. (1998). Computation of conditional probability statistics by 8-month-old infants. Psychological Science, 9, 321–324. Baillargeon, R., Spelke, E. S., & Wasserman, S. (1985). Object permanence in fivemonth-old infants. Cognition, 20, 191–208. Baker, R. K., Pettrigrew, T. L., & Poulin-Dubois, D. (2014). Infants’ ability to associate motion paths with object kinds. Infant Behavior and Development, 37, 119–129. http://dx.doi.org/10.1016/j.infbeh.2013.12.005. Behl-Chadha, G. (1996). Basic-level and superordinate-like categorical representations in early infancy. Cognition, 60, 105–141. http://dx.doi.org/ 10.1016/0010-0277(96)00706-8. Berlyne, D. E. (1958a). The influence of complexity and novelty in visual figures on orienting responses. Journal of Experimental Psychology, 55, 289–296. Berlyne, D. E. (1958b). The influence of the albedo and complexity of stimuli on visual fixation in the human infant. British Journal of Psychology, 39, 315–318. Berlyne, D. E. (1960). Conflict, arousal, and curiosity. New York: McGraw-Hill. Berlyne, D. E. (1963). Complexity and incongruity variables as determinants of exploratory choice and evaluative ratings. Canadian Journal of Psychology, 17, 274–290. Bertenthal, B. I., Proffitt, D. R., & Cutting, J. E. (1984). Infant sensitivity to figural coherence in biomechanical motions. Journal of Experimental Child Psychology, 37, 213–230. Bertenthal, B. I., Proffitt, D. R., & Kramer, S. J. (1987). The perception of biomechanical motions: Implementation of various processing constraints. Journal of Experimental Psychology: Human Perception and Performance, 13, 577–585. Blake, R., & Aiba, T. S. (1998). Detection and discriminatin of optical flow components. Japanese Psychological Research, 40, 19–30.

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136 Booth, A. E., Pinto, J., & Bertenthal, B. I. (2002). Perception of the symmetrical patterning of human gait by infants. Developmental Psychology, 38, 554–563. http://dx.doi.org/10.1037//0012-1649.38.4.554. Bower, T. G. R., Broughton, J. M., & Moore, M. K. (1971). Development of the object concept as manifested by changes in tracking behavior of infants between 7 and 20 weeks of age. Journal of Experimental Child Psychology, 11, 182–193. Burnham, D. K. (1987). The role of movement in object perception in infants. In B. E. McKenzie & R. H. Day (Eds.), Perceptual development in early infancy: Problems and issues (pp. 143–173). Hillsdale, NJ: Erlbaum. Burnham, D. K., Vignes, G., & Ihsen, E. (1988). The effect of movement on infants’ memory for visual compounds. British Journal of Developmental Psychology, 6. 351-350. Burr, D. C., Badcock, D. R., & Ross, J. (2001). Cardinal axes for radial and circular motion, revealed by summation and by masking. Vision Research, 41, 473–481. Cohen, L. B., Atkinson, D. J., & Chaput, H. H. (2000). Habit 2000: A new program for testing infant perception and cognition (Version 2.2.5c). Austin, TX: University of Texas. Cohen, L. B., DeLoache, J. S., & Rissman, M. W. (1975). The effect of stimulus complexity on infant visual attention and habituation. Child Development, 46, 611–617. Cohen, L. B. (1988). An information processing approach to infant cognitive development. In L. Weiskrantz (Ed.), Thought without language (pp. 211–228). Oxford, UK: Oxford University Press. Cohen, L. B. (1991). Infant attention: An information processing approach. In M. J. Weiss & P. R. Zalazo (Eds.), Newborn attention: Biological constraints and the influence of experience (pp. 1–21). Norword, NJ: Ablex Publishing. Csibra, G., Biró, S., Koós, O., & Gergely, G. (2003). One-year-old infants use teleological representations of actions productively. Cognitive Science, 27, 111–133. Cutting, J. E. (1978). Generation of synthetic male and female walkers through the manipulations of a biomechanical invariant. Perception, 75, 393–405. Cutting, J. E. (1981). Coding theory adapted to gait perception. Journal of Experimental Psychology: Human Perception and Performance, 7, 71–87. Cutting, J. E. (1986). Perception with an eye for motion. Cambridge, MA: MIT Press. Cutting, J. E., Moore, C., & Morrison, R. (1988). Masking the motions of human gait. Perception & Psychophysics, 44, 339–347. Day, R. H., & Burnham, D. K. (1981). Infants’ perception of shape and color in laterally moving objects. Infant Behavior and Development, 4, 341–357. Dember, W. N., & Earl, R. W. (1957). Analysis of exploratory, manipulative, and curiosity behavior. Psychological Review, 64, 91–96. Di Giorgio, E., Lunghi, M., Simion, F., & Vallortigara, G. (2016). Visual cues of motion that trigger animacy perception at birth: The case of self-propulsion. Developmental Science, Advanced online publication:. http://dx.doi.org/10.1111/ desc.12394. Dillon, M. R., & Spelke, E. S. (2015). Core geometry in perspective. Developmental Science, 18, 894–908. http://dx.doi.org/10.1111/desc.12266. Edwards, M., & Badcock, D. R. (1996). Global-motion perception: Interaction of chromatic and luminance signals. Vision Research, 36, 2423–2431. Fox, R., & McDaniel, C. (1982). The perception of biologiocal motion by human infants. Science, 218, 486–487. Freeman, T. C. A., & Harris, M. G. (1992). Human sensitivity to expanding and rotating motion – Effects of complementary masking and directional structure. Vision Research, 32, 81–87. Gelman, R. (1990). First principles to organize attention to and learning about relevant data: Number and the animate-inanimate distinction as examples. Cognitive Science, 14, 79–106. Gergely, G., Nadasdy, Z., Csibra, G., & Bíró, S. (1996). Taking the intentional stance at 12 months of age. Cognition, 56, 165–193. http://dx.doi.org/10.1016/0010-0277 (95)00661-H. Gilmore, R. O., Hou, C., Pettet, M. W., & Norcia, A. M. (2007). Development of cortical responses to optic flow. Visual Neuroscience, 24, 845–856. doi: 10.10170S0952523807070769. Guajardo, J. J., & Woodward, A. L. (2004). Is agency skin deep? Surface attributes influence infants’ sensitivity to goal-directed action. Infancy, 6, 361–384. Hirshkowitz, A., & Wilcox, T. (2013). Infants’ ability to extract three-dimensional shape from coherent motion. Infant Behavior and Development, 36, 863–872. Johansson, G. (1973). Visual perception of biologiocal motion and a model for its analysis. Perception & Psychophysics, 14, 201–211. Johnson, S. P. (2000). The development of visual surface perception: Insights into the ontogeny of knowledge. In C. K. Rovee-Collier, L. Lipsitt, & H. Hayne (Eds.). Progress in infancy research (Vol. 1, pp. 113–154). Mahwah, NJ: Erlbaum. Johnson, S. P. (2004). Development of perceptual completion in infancy. Psychological Science, 15, 769–775. Johnson, S. P. (2010). How infants learn about the visual world. Cognitive Science, 34, 1158–1184. Johnson, S. P., & Aslin, R. N. (1995). Perception of object unity in 2-month-old infants. Developmental Psychology, 31, 739–745. Johnson, S. P., & Aslin, R. N. (1996). Perception of object unity in young infants: The roles of motion, depth, and orientation. Cognitive Development, 11, 161–180. Johnson, S. P., Davidow, J., Hall-Haro, C., & Frank, M. C. (2008). Perceptual completion originates in information acquisition. Developmental Psychology, 44, 1214–1224. Johnson, S. P., & Mason, U. (2002). Perception of kinetic illusory contours by 2month-old infants. Child Development, 73, 22–34. Johnson, S. P. (2011). A constructivist view of object perception in infancy. In L. M. Oakes, C. H. Cashon, M. Casasola, & D. H. Rakison (Eds.), Infant perception and

135

cognition: Recent advances, emerging themes, and future directions (pp. 51–68). New York: Oxford University Press. Johnson, S. P. (2013). Object perception. In P. D. Zelazo (Ed.), Handbook of developmental psychology (pp. 371–379). New York: Oxford University Press. Vol. 1: Body and Mind. Jowkar-Baniani, G., & Schmuckler, M. A. (2011). Picture-perception in infancy: Generalizing from three-dimensional to two-dimensional displays. Infancy, 16, 211–226. Kanass, K. N., Oakes, L. M., & Wiese, D. (1999). The development of infants’ perception of object movement along inclines. Cognitive Development, 14, 215–240. Karmel, B. Z., & Maisel, E. G. (1974). A neuronal activity model for infant visual attention. In L. B. Cohen & P. Salapatek (Eds.). Basic visual processes (Vol. 1, pp. 77–131). New York: Academic Press. Kaufman, J., Mareschal, D., & Johnson, M. H. (2003). Graspability and object processing in infants. Infant Behavior and Development, 26, 516–528. Kavšek, M., & Yonas, A. (2006). The perception of moving subjective contours by 4month-old infants. Perception, 35, 215–227. Keil, F. C. (1989). Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press. Keil, F. C., & Banterman, N. (1984). A characteristic-to-defining shift in the development of word meaning. Journal of Verbal Learning and Verbal Behavior, 23, 221–236. Kellman, P. J. (1984). Perception of three-dimensional form by human infants. Perception & Psychophysics, 36, 353–358. Kellman, P. J., & Short, K. R. (1987). Development of three-dimensional form perception. Journal of Experimental Psychology: Human Perception and Performance, 13, 545–557. Kellman, P. J., & Spelke, E. S. (1983). Perception of partly occluded objects in infancy. Cognitive Psychology, 15, 483–524. Kidd, C., Piantadosi, S. T., & Aslin, R. N. (2012). The Goldilocks Effect: Human infants allocate attention to visual sequences that are neither too simple nor too complex. PLoS ONE, 7, e36399. http://dx.doi.org/10.1371/journal.pone.0036399. Kim, I.-K., & Spelke, E. S. (1992). Infants’ sensitivity to the effects of gravity on visible object motion. Journal of Experimental Psychology: Human Perception and Performance, 18, 385–393. Kim, I.-K., & Spelke, E. S. (1999). Perception and understanding of effects of gravity and inertia on object motion. Developmental Science, 2, 339–362. Kirkham, N. Z., Slemmer, J. A., & Johnson, S. P. (2002). Visual statistical learning in infancy: Evidence for a domain general learning mechanism. Cognition, 83, B35–B42. Kirkham, N. Z., Slemmer, J. A., Richardson, D. K., & Johnson, S. P. (2007). Location, location, location: Development of spatiotemporal sequence learning in infants. Child Development, 78, 1559–1571. Kozlowski, L. T., & Cutting, J. E. (1977). Recognizing the gender of walkers from dynamic point-light displays. Perception & Psychophysics, 21, 275–280. Lany, J., & Saffran, J. R. (2013). Statistical learning mechanisms in infancy. In J. L. R. Rubenstein & P. Rakic (Eds.). Comprehensive developmental neuroscience: Neural circuit development and function in the brain (Vol. 3, pp. 231–248). Amsterdam: Elsevier. Lee, A. L. F., & Lu, H. (2010). A comprison of global motion perception using a multiple-aperture stimulus. Journal of Vision, 10, 1–16. Lewkowicz, D. J., & Turkewitz, G. (1981). Intersensory interactions in newborns: Modification of visual preferences following exposure to sound. Child Development, 52, 827–835. Luo, Y., & Baillargeon, R. (2005). Can a self-propelled box have a goal? Psychological reasoning in 5-month-old infants. Psychological Science, 16, 601–608. http://dx. doi.org/10.1111/j.1467-9280.2005.01582.x. Mandler, J. M. (1992). How to build a baby II: Conceptual primitives. Psychological Review, 99, 587–604. Mandler, J. M. (2000). Perceptual and conceptual processes in infancy. Journal of Cognition and Development, 1, 3–36. Mandler, J. M. (2003). Conceptual categorization. In D. H. Rakison & L. M. Oakes (Eds.), Early category and concept development (pp. 103–131). Oxford: Oxford University Press. Mandler, J. M. (2008a). Infant concepts revisited. Philosophical Psychology, 21, 269–280. http://dx.doi.org/10.1080/09515080801980229. Mandler, J. M. (2008b). On the birth and growth of concepts. Philosophical Psychology, 21, 207–230. http://dx.doi.org/10.1080/09515080801980179. Mandler, J. M. (2010). The spatial foundations of the conceptual system. Language and Cognition, 2, 21–44. http://dx.doi.org/10.1515/LANGCOG.2010.002. Mandler, J. M. (2012). On the spatial foundations of the conceptual system and its enrichment. Cognitive Science, 36, 421–451. http://dx.doi.org/10.1111/j.15516709.2012.01241. Mandler, J. M., & McDonough, L. (1993). Concept formation in infancy. Cognitive Development, 8, 291–317. http://dx.doi.org/10.1016/S0885-2014(93)80003-C. Mandler, J. M., & McDonough, L. (1998). On developing a knowledge base in infancy. Developmental Psychology, 34, 1274–1288. http://dx.doi.org/10.1037/00121649.34.6.127. Marcovitch, S., & Lewkowicz, D. J. (2009). Sequence learning in infancy: The independent contributions of conditional probability and pair frequency information. Developmental Science, 12, 1020–1025. Mareschal, D., & Johnson, S. P. (2002). Learning to perceive object unity: A connectionist account. Developmental Science, 5, 151–185. Markson, L., & Spelke, E. S. (2006). Infants’ rapid learning about self-propelled objects. Infancy, 9, 45–71.

136

G. Jowkar-Baniani et al. / Cognition 165 (2017) 126–136

Mendelsohn, M. J. (1986). Perception of the temporal pattern of motion in infancy. Infant Behavior and Development, 9, 231–243. Milner, A. D., & Goodale, M. A. (2006). The visual brain in action (2nd ed.). New York: Oxford University Press. Morton, J. B., & Munakata, Y. (2002). Active versus latent representations: A neural network model of perseveration, dissociation, and decalage. Developmental Psychobiology, 40, 255–265. http://dx.doi.org/10.1002/dev.10033. Owsley, C. (1983). The role of motion in infants’ perception of solid shape. Perception, 12, 707–717. Pavlova, M., Krägeloh-Mann, I., Birbaumer, N., & Sokolov, A. (2002). Biological motion shown backwards: The apparent-facing effect. Perception, 31, 435–443. http://dx.doi.org/10.1068/p3262. Pavlova, M., & Sokolov, A. (2000). Orientation specificity in biological motion perception. Perception & Psychophysics, 62, 889–899. Pinto, J., Bertenthal, B. I., & Booth, A. E. (1996). Developmental changes in infants’ responses to biological motion displays. Investigative Ophthalmology & Visual Science, 37, S 63. Poulin-Dubois, D., Crivello, C., & Wright, K. (2015). Biological motion primes the animate/inanimate distinction in infancy. PLoS ONE, 10, e0116910. Poulin-Dubois, D., Frenkiel-Fishman, S., Nayer, S., & Johnson, S. (2006). Infants’ inductive generalization of bodily, motion, and sensory properties to animals and people. Journal of Cognition and Development, 7, 431–453. http://dx.doi.org/ 10.1207/s15327647jcd0704 1. Poulin-Dubois, D., Lepage, A., & Ferland, D. (1996). Infants’ concept of animacy. Cognitive Development, 11, 19–36. http://dx.doi.org/10.1016/S0885-2014(96) 90026-X. Rakison, D. H., & Poulin-Dubois, D. (2001). Developmental origin of the animateinanimate distinction. Psychological Bulletin, 127, 209–228. Richards, J. E. (2010). The development of attention to simple and complex visual stimuli in infants: Behavioral and psychophysiological measures. Developmental Review, 30, 203–219. http://dx.doi.org/10.1016/j.dr.2010.03.00. Rochat, P., Morgan, R., & Carpenter, M. (1997). Young infants’ sensitivity to movement information specifying social causality. Cognitive Development, 12, 536–561. Ruff, H. A. (1982). Effect of object movement on infants’ detection of object structure. Developmental Psychology, 18, 462–472. Ruff, H. A. (1985). Detection of information specifying the motion of objects by 3and 5-month-old infants. Developmental Psychology, 21, 295–305. Ruff, H. A. (1987). Preference for rotating objects in 5-month-old infants. Infant Behavior and Development, 10, 365–369. Saffran, J. R., Aslin, R. N., & Newport, E. L. (1996). Statistical learning by 8-month-old infants. Science, 274, 1926–1928. Saffran, J. R., Johnson, E. K., Aslin, R. N., & Newport, E. L. (1999). Statistical learning of tone sequences by human infants and adults. Cognition, 70, 27–52. Satlow, E., & Newcombe, N. S. (1998). When is a triangle not a triangle? Young children’s developing conceptions of geometric shape. Cognitive Development, 13, 547–559. Saxe, R., Tennenbaum, J., & Carey, S. (2005). Secret agents: Inferences about hidden causes by 10- and 12-month-old infants. Psychological Science, 16, 995–1001. http://dx.doi.org/10.1111/j.1467-9280.2005.01649.x. Schlottmann, A., & Ray, E. (2010). Goal attribution to schematic animals: Do 6month-olds perceive biological motion as animate? Developmental Science, 13, 1–10. http://dx.doi.org/10.1111/j.1467-7687.2009.00854.x. Schmuckler, M. A., & Proffitt, D. R. (1994). Infants’ perception of kinetic depth and stereokinetic displays. Journal of Experimental Psychology: Human Perception and Performance, 20, 122–130. Shimizu, Y. A., & Johnson, S. C. (2004). Infants’ attribution of a goal to a morphologically unfamiliar agent. Developmental Science, 7, 425–430. http:// dx.doi.org/10.1111/j.1467-7687.2004.00362.x. Shirai, N., Kanazawa, S., & Yamaguchi, M. K. (2008). Sensitivity to rotational motion in early infancy. Experimental Brain Research, 190, 201–206. http://dx.doi.org/ 10.1007/s00221-008-1461-2.

Simion, F., Bardi, L., Mascalzoni, E., & Regolin, L. (2013). From motion cues to social perception: Innate predispositions. In M. D. Rutherford & V. Kuhlmeier (Eds.), Social perception: Detection and interpretation of animacy, agency, and intention (pp. 37–60). Cambridge, MA: The MIT Press. Slater, A., Riddell, P., Quinn, P. C., Pascalis, O., Lee, K., & Kelly, D. J. (2010). Visual perception. In J. G. Bremner & T. D. Wachs (Eds.). The Wiley-Blackwell Handbook of Infant Development, Second Edition (Vol. 2, pp. 40–80). West Sussex, UK: John Wiley & Sons. Soska, K. C., & Johnson, S. P. (2008). Development of three-dimensional object completion in infancy. Child Development, 79. 1230-1126. Soska, K. C., & Johnson, S. P. (2013). Development of three-dimensional completion of complex objects. Infancy, 18, 325–344. Spelke, E. S., Breinlinger, K., Macomber, J., & Jacobson, K. (1992). Origins of knowledge. Psychological Review, 99, 605–632. Spelke, E. S., & Kinzler, K. D. (2007). Core knowledge. Developmental Science, 10, 89–96. Spelke, E. S., Lee, S. A., & Izard, V. (2010). Beyond core knowledge: Natural geometry. Cognitive Science, 34, 863–994. Spelke, E. S., Phillips, A., & Woodward, A. L. (1995). Infants’ knowledge of object motion and human action. In D. Sperber, D. Premack, & A. J. Premack (Eds.), Causal cognition: A multidisciplinary debate (pp. 45–78). Oxford: Clarendon Press. Spelke, E. S., & Van de Walle, G. (1993). Perceiving and reasoning about objects: Insights from infants. In N. Eilan, W. Brewer, & R. McCarthy (Eds.), Spatial representation. New York: Blackwell. Träuble, B., Pauen, S., & Poulin-Dubois, D. (2014). Speed and direction changes induce the perception of animacy in 7-month-old infants. Frontiers in Psychology, Published: 10 October 2014. http://dx.doi.org/10.3389/fpsyg.2014. 01141. Träuble, B., & Pauen, S. (2011). Infants’ reasoning about ambiguous motion events: The role of spatiotemporal and dispositional status information. Cognitive Development, 26, 1–15. Troje, N. F. (2002). Decomposing biological motion: A framework for analysis and synthesis of human gait patterns. Journal of Vision, 2, 371–387. http://dx.doi.org/ 10.1167/2.5.2. Troje, N. F. (2013). What is biological motion? Definition, stimuli, and paradigms. In M. D. Rutherford & V. Kuhlmeier (Eds.), Social perception: Detection and interpretation of animacy, agency, and intention (pp. 13–36). Cambridge, MA: The MIT Press. Ungerlieder, L. G., & Mishkin, M. (1982). Two cortical visual systems. In D. J. Ingle, M. A. Goodale, & R. J. W. Mansfield (Eds.), Analysis of visual behavior (pp. 549–586). Cambridge, MA: MIT Press. Vitz, P. C. (1966). Preference for different amounts of visual complexity. Behavioral Science, 11, 105–114. http://dx.doi.org/10.1002/bs.3830110204. Wallach, H., & O’Connell, D. N. (1953). The kinetic depth effect. Journal of Experimental Psychology, 45, 205–217. Westhoff, C., & Troje, N. F. (2007). Kinematic cues for person identification from biological motion. Perception & Psychophysics, 69, 241–253. http://dx.doi.org/ 10.3758/BF03193746. Woodward, A. L. (1998). Infants selectively encode the goal object of an actor’s reach. Cognition, 69, 1–34. Woodward, A. L. (1999). Infants’ ability to distinguish between purposeful and nonpurposeful behaviors. Infant Behavior and Development, 22, 145–160. Woodward, A. L. (2009). Infants’ grasp of others’ intentions. Current Directions in Psychological Science, 18, 53–57. http://dx.doi.org/10.1111/j.14678721.2009.01605.x. Woodward, A. L., Sommerville, J. A., & Guajardo, J. J. (2001). How infants make sense of intentional action. In B. Malle, L. Moses, & D. Baldwin (Eds.), Intentions and intentionality: Foundations of social cognition (pp. 149–169). Cambridge, MA: The MIT Press. Yonas, A., Arterberry, M. E., & Granrud, C. E. (1987). Four-month-old infants’ sensitivity to binocular and kinetic information for three-dimensional object shape. Child Development, 58, 910–917.