Imitation and Affordance Learning by Pigeons (Columba livia).

Abstract

The bidirectional control procedure was used to determine whether pigeons (Columba livia) would imitate a demonstrator that pushed a sliding screen for food. One group of observers saw a trained demonstrator push a sliding screen door with its beak (imitation group), whereas 2 other groups watched the screen move independently (possibly learning how the environment works) with a conspecific either present (affordance learning with social facilitation) or absent (affordance learning alone). A 4th group could not see the screen being pushed (sound and odor control). Imitation was evidenced by the finding that pigeons that saw a demonstrator push the screen made a higher proportion of matching screen pushes than observers in 2 appropriate control conditions. Further, observers that watched a screen move without a demonstrator present made a significantly higher proportion of matching screen pushes than would be expected by chance. Thus, these pigeons were capable of affordance learning.

Imitation of a two-action sequence by pigeons

“…It is possible that the observation of a demonstrator moving the treadle with a distinctive part of its body is more detectable than the observation of it moving the screen in a distinctive direction. However, earlier research with this same apparatus found that pigeons demonstrated significant matching of left-right screen pushing (Klein & Zentall, 2003). In fact, in that study, pigeons tended to match the direction of their demonstrator's screen push (85.0%) to the same extent as pigeons in a previous study tended to match their demonstrator's treadle response (84.2%; Dorrance & Zentall, 2001).…”

“…When tested, the observer quail showed a significant tendency to push the screen in the same direction as their demonstrator. Furthermore, when control observers saw the screen move without being pushed by the demonstrator (a control for object-movement reenactment or learned affordances), they did not match the direction of screen movement (see also Klein & Zentall, 2003, for inclusion of a control for possible odor cues). Thus, the bidirectional control procedure appears to provide a valid alternative procedure for demonstrating imitation in birds.…”

Perspectives on observational learning in animals.

“…When observation of a demonstrator allows an animal to learn how the environment functions, a sophisticated form of learning may be involved. For example, if a pigeon observes a screen (that is capable of moving to the left or to the right) move to the left to allow access to food, when the pigeon is given access to the screen, it is more likely to move the screen in the same direction (Campbell, Heyes, & Goldsmith, 1999;Klein & Zentall, 2003). However, because learning how the environment works may occur in the absence of the behavior of another animal, one would not want to view such learning as social learning.…”

“…Learning about things and how they work by observation has important implications for cognitive learning. There is evidence not only that chimpanzees can emulate the movement of a tool (Nagell, Olguin, & Tomasello, 1993), as well as increasing the level of water in a cylinder to gain access to a reward (Tennie et al, 2010), but also that pigeons (Klein & Zentall, 2003) and dogs (Miller, Rayburn-Reeves, & Zentall, 2009) can emulate the direction of movement of a screen permitting access to food.…”

“…To distinguish this learning from emulation, a control condition (sometimes referred to as a "ghost" control), in which the manipulandum appears to move by itself without a demonstrator pushing it, is needed (see Hopper, 2010;Hopper, Lambeth, Schapiro, & Whiten, 2008). Although the ghost control provides a necessary comparison condition, it can be argued that it does not control for social facilitation (the mere presence of a conspecific, see Klein & Zentall, 2003) nor does it control adequately for observation of the outcome or goal (the observed "demonstrator" should be provided with reinforcement to control for the motivation induced by seeing an conspecific eating as well as potential associations between the moving manipulandum and the sight a conspecific eating-what Hopper et al, 2008, call an enhanced ghost control). But if one controls for social facilitation by including the presence of an inactive conspecific (Hopper, 2010), the conspecific may be distracting or imitation of the conspecific could result in the absence of emulation but for the wrong reason.…”