Dissociating the perception of speed and the perception of animacy: a functional approach

Abstract

Differences in acceleration, differences in constant speed and illusory speed differences are all associated with predictable differences in animacy perception. The current study describes a dissociation between perceived speed and perceived animacy, apparently resulting from the human visual system taking gravity into account. In Experiment 1, participants compared dots moving at the same speed up and down a vertically oriented computer screen. Dots moving up were judged as animate more often than dots moving down, while dots moving down were judged as faster most often. To test whether this pattern of results was sensitive to changes in the orientation of the stimuli relative to gravity, Experiment 2 presented the same stimuli on a screen oriented horizontally. The dissociation between the perception of speed and the perception of animacy was maintained: The difference between the perception of animacy of dots moving away vs. toward was much reduced, while the effect on speed perception was more pronounced, compared to the vertical orientation. These results are consistent with the idea that the human visual system is designed to perceive animacy in a functionally reasonable way given the terrestrial environment in which it evolved.

Introduction

Solutions to human adaptive problems require perceptual and cognitive systems that incorporate the statistical regularities found in the environment of evolutionary adaptedness (EEA). Because of this, we should expect adapted psychological processes to be organized functionally. There is evidence of this functional organization in, for example, the perception of colour (Shepard, 1992) and size, shape and brightness (see Palmer, 1999 for review).

Many of the adaptive challenges in our EEA were social in nature, and we appear to have evolved numerous perceptual and cognitive solutions to social adaptive problems (Dunbar, 1998, Humphrey, 1983). These include greater accuracy for solving problems presented in a social exchange context compared to a nonsocial context (Cosmides, 1989, Cosmides and Tooby, 1992), a preference for looking at faces that is present from birth (Mondloch, Lewis, Budreau, & Maurer, 1999) and that direction of eye gaze captures attention even when it is not predictive of a target's location (Friesen & Kingstone, 1998), for example. These perceptual and cognitive adaptations relating to social stimuli are primarily concerned with animate objects, including predators, prey and conspecifics, since there was likely great selection pressure for detecting these objects in the EEA. For example, New, Cosmides, and Tooby (2007) found that people are better at identifying a change in the visual scene when an animate object disappears (e.g., an elephant) than when an inanimate object disappears (e.g., a car) in a change-blindness paradigm.

Our social perceptions allow us to identify, anticipate and manipulate peoples' intentions, desires and goals (Baron-Cohen, 1995), beginning in the first year of life and for unfamiliar actions (e.g., Biro and Leslie, 2007, Johnson, 2000, Hamlin et al., 2007). These mental states (which cannot be accessed directly) often manifest themselves as actions, which even children can perceive (Baldwin, Baird, Saylor, & Clark, 2001). The attributes and qualities of these motions are highly informative and humans are very sensitive to variations in them. For example, Guajardo and Woodward (2004) observed that 7- and 12-month-olds interpret a bare hand as goal directed when moving towards an object, but do not interpret a gloved hand moving in an identical manner as goal directed unless they see the gloved hand at the end of an arm. Prior to attributing mental states to agents, one must first discriminate those objects that are alive from those that are not. A crucial question is: how do we identify objects that are alive?

This perception of animacy is a basic and fundamental social ability underlying other social cognitions and provides the starting point for more complex social abilities. There are a number of possible cues to animacy, including the presence of morphological features such as heads, eyes, limbs and asymmetry (Gelman, 1990). An alternative and ubiquitous cue available in our EEA would have been an objects' motion. Heider and Simmel (1944) were the first to demonstrate that animations of simple geometric shapes interacting with one another can elicit the perception of animacy, as well as more complex social behaviour such as chasing, cowering and protecting. Numerous studies since have shown that similar animations of geometric shapes can elicit a variety of social perceptions as complex as intentionality (Gergely, Nadasdy, Csibra, & Biro, 1995), chasing (Rochat, Morgan, & Carpenter, 1997) and helping or hindering (Kuhlmeier, Wynn, & Bloom, 2003). These perceptions have been shown to emerge early in life (Luo and Baillargeon, 2005, Hamlin et al., 2007) and to be stable cross culturally (Barrett, Todd, Miller, & Blythe, 2005). Manipulating motion cues allows researchers to selectively examine the underlying movements, both individual and relational, that influence our social perceptions and allows researchers to study our perceptual strategies. Most studies have examined our perceptions of animacy as defined by goal-oriented and interactive actions (e.g., chasing, fighting, avoiding); only a few studies to date have attempted to identify the most basic and rudimentary motion cues such as speed and direction that reliably trigger our perceptions of animacy.

Tremoulet and Feldman (2000) and Szego and Rutherford (2007) explored the association between simple motion cues, such as speed and acceleration, and the perception of animacy. In these studies, participants saw a simple geometric figure (a dot or line) travelling against an otherwise empty background. Features of the objects' motion (viz., speed, acceleration or deceleration, or a change in direction) were systematically manipulated to determine the features most predictive of a perception of animacy. The results of these experiments consistently found a strong and reliable association of speed and animacy, such that relatively greater speeds are associated with more frequent perceptions of animacy. This has been shown for differences in accelerations, constant speeds and even illusory differences in speeds. Regardless of whether the speed differences were actual or illusory, perceptions of animacy were similar: Participants in both experiments judged the (apparently) relatively faster object as appearing alive more often than the relatively slower object. Notably these studies presented stimuli on a monitor oriented horizontally — such that the objects were travelling across a horizontal plane while being viewed from above — to avoid any suggestion of a gravitational context.

There are sound evolutionary reasons why there should be a perceptual association between speed and animacy. In our EEA, a living organism might well have appeared that was small, fast-moving, far away, or partially obscured, and these organisms could have been predator or prey. Surfaces in the EEA were generally rough, rocky or bumpy rather than smooth. Under these circumstances, an inanimate object travelling across such a surface would be slowed by friction, so any object that could maintain speed across these natural surfaces was likely to be self-propelled. Self-propelled motion has been argued to be a cue to animacy (Biro and Leslie, 2007, Gelman et al., 1995, Leslie, 1994, Premack, 1990, Stewart, 1982).

Additionally, a bias to consider fast moving objects as animate may be advantageous. There would likely be a relatively small fitness cost associated with frequently mistaking an inanimate object for a possible predator or prey, while the cost of erroneously categorizing an animate object as inanimate could be great. This reliable and asymmetrical distribution of costs and benefits regarding speed as an indicator of animacy would be an adaptive form of error management (see Haselton & Buss, 2000).

The purpose of the current study was to investigate the relationship between the perception of speed and the perception of animacy with and without the context of an apparent gravitational field, using the methods similar to that of Tremoulet and Feldman (2000) and Szego and Rutherford (2007). We reasoned that since our visual strategies for perceiving animacy evolved in a terrestrial environment — specifically in the context of a constant gravitational field — introducing the context of gravity and the perception of a gravitation field (Experiment 1) should be sufficient to dissociate the previously reported perceptual association of speed and animacy. We predicted that objects that are able to move against gravity without an apparent external power source will be seen as having an internal power source and thus be seen as animate, regardless of perceived speed. If people judge a dot that appears to travel against gravity (by “rising” up a monitor) as animate more often than a dot that appears to move with gravity (by “falling” down a monitor), this dissociation would be evidence that peoples' perception of animacy is influenced by at least one ubiquitous and systematic influence on motion in our EEA (viz., gravity). Additionally, we reasoned that people would not have a bias to judge either the “rising” or “falling” dot as faster or slower, given equivalent speeds for both dots in a given trial. If different patterns of results for judgments of speed and judgments of animacy were observed, it would be evidence that the perceptions of animacy and speed are not rigidly linked and can be dissociated. Conversely, if participants' perceptions of speed and animacy are not dissociated, we should expect participants to be consistent in their judgments of speed and animacy for both directions, judging the same dots as both faster and animate regardless of the presence or absence of a gravitational context.

In both experiments, participants viewed objects travelling up and down a monitor that was oriented vertically (i.e., conventionally) or horizontally. Both experiments utilized a two-interval forced choice task, presenting a pair of dots whose motions were matched for all factors except direction; this allowed us to test whether the pattern of perceptions seen in Experiment 1 was sensitive to changes in the orientation of the stimuli when the context (and apparent influence) of gravity is absent in Experiment 2.