Do cats (Felis catus) predict the presence of an invisible object from sound?

Abstract

Recognizing invisible entities from auditory information is advantageous to animals in various situations including predator avoidance and foraging. In 2 experiments we asked whether cats could predict the presence of an unseen object on hearing noise it made, based on a causal-logical rule. After observing an experimenter shaking an opaque container for 15 seconds (observation phase), the cats freely explored the environment for 15 seconds (response phase). Experiment 1 tested 3 conditions. In the first, “contingent noise” condition, the object inside the container made a rattling noise when shaken. In the second, “irrelevant noise” condition, white noise accompanied the shaking action. In the third, “no noise” condition, the shaking action was silent. Experiment 2 tested a “noncontingent noise” condition, in which the rattling noise and movement of the container were out of synchrony. In both experiments cats looked at the container for longer in the “contingent noise” condition than the other conditions. These results suggest that cats used a causal-logical understanding of auditory stimuli to predict the presence of invisible objects. This ability may be related to the ecology of cats' natural hunting style.

Introduction

Information obtained via the sensory organs is often ambiguous or fragmentary. For example, an animal hunting in the bush by sight might hear only the noise the prey makes. In this case inferring the presence of a prey from the noise would be advantageous to the hunter's survival. Similarly, potential prey may be more likely to survive if they can predict the presence of predators from indirect clues such as odor and noise.

Inferential reasoning refers to the ability to use available information to draw conclusions about circumstances that are not directly observable (Heimbauer et al., 2012). Call (2004) explored this ability in great apes. He presented apes with 2 opaque cups, one of which they knew to be baited. The apes were given visual or auditory cues about the contents of both cups (full information) or only one of the cups (partial information) before making a choice. The subjects were able to see the contents of the cups in a visual domain test and to hear a rattling noise when the cup was shaken in an auditory domain test. The latter test required an understanding of the causal-logical rule between the noise and movement of the containers. In contrast to the full visual information task on which all apes succeeded, fewer subjects passed the auditory tests even with full information (16.6% (2/12), 50% (2/4), 0% (0/6), and 62.5% (5/8) of chimpanzees, bonobos, orangutans, and gorillas, respectively). Similar results have been obtained in other nonhuman primate species: 0% (0/8) in rhesus macaques (Petit et al., 2015), 33% (7/21) in olive baboons (Schmitt and Fischer, 2009, Petit et al., 2015), 50% (2/4) in lemurs (Maille and Roeder, 2012), and 30% (8/26) in capuchin monkeys (Sabbatini and Visalberghi, 2008, Paukner et al., 2009, Heimbauer et al., 2012), although 100% (8/8) of tonkean macaques succeeded in an auditory test (Petit et al., 2015). It appears surprising that this causal relationship is so poorly understood by primates.

Several researchers have related this poverty of causal understanding to the ecological importance of auditory information of each species (Maille and Roeder, 2012, Plotnik et al., 2014). Nonhuman primates are generally poor at auditory as opposed to visual tasks (Schmitt and Fischer, 2009). D'Amato and Salmon (1982) suggested that cats use auditory stimuli to locate prey, whereas primates often use sounds as cues to avoid rather than approach the source. Given that cats often use auditory cues when hunting (Turner and Meister, 1988), investigating cats' predictions about invisible objects from noise can contribute to understanding how ecological factors influence functional differences among sensory modalities.

It has been suggested that cats' causal-logical understanding in the physical domain is not sophisticated (Bradshaw, 2013). Whitt et al (2009) tested domestic cats on string-pulling tasks to explore their understanding of physical causality. After the cats were initially trained to pull a string to obtain a food reward, 3 tests were conducted. In “longer string” tests, cats were rewarded for choosing a baited string that was longer than the one used in training. In “parallel” and “crossed strings” tests, cats were required to choose between 2 strings, only one of which was baited. The cats failed to choose the baited string in both tests; no causal understanding was demonstrated. Bradshaw (2013) pointed out that string-pulling tasks lack ecological validity and that they may not be an appropriate test of cats' physical understanding. It may be advantageous to test cats' causal understanding using a different modality. Cats have a remarkable hearing ability (Bradshaw, 2013). For example, they have more than 20 muscles in their ear (see Tabor, 1983), and the audible range extends from 48 Hz to 85 kHz (Heffner and Heffner, 1985). We propose that the auditory modality may be more suitable test focus.

Here we present 2 experiments that investigated whether cats could show causal-logical understanding about the existence of an object inside an opaque container when they observe the container being moved accompanied by a rattling sound. We tested cats in 3 conditions in each of the 2 experiments. In experiment 1, we ran “contingent noise,” “irrelevant noise,” and “no noise” conditions. The experimenter shook the container repeatedly while the cats watched. In contingent noise condition, a block of wood inside the container made a rattling noise as it moved. In irrelevant noise condition, white noise was played during the movement of the container. In no noise condition, the experimenter shook the empty container. In experiment 2, “noncontingent noise” condition replaced “irrelevant noise” condition. In this new condition, the rattling noise was not synchronized with the motion of the container. We hypothesized that if cats form a representation of an invisible object from auditory stimuli, they would pay more attention to the contingent noise condition than to any other conditions.