Anda di halaman 1dari 6

ANTHROPOMORPHIC INTERPRETATIONS AND ETHOLOGICAL DESCRIPTIONS OF DOG AND CAT BEHAVIOR BY LAY PEOPLE

Zana Bahlig-Pieren and Dennis C. Turner


Zoology Institute, University of Zurich-Irchel, Zurich, Switzerland

ABSTRACT
Unlike ethologists and veterinarians, lay people supposedly use their own unstructured observations to interpret their companion animals behavior, often in anthropomorphic terms. Recently, anthropomorphism has evoked new interest amongst scientists as a result of provocative publications concerning animal mind, that have prompted investigations in the field of cognitive ethology, and much speculation about the emotions of animals. The purpose of the research reported here was to determine the concurrency and plausibility of anthropomorphic interpretations and ethological descriptions of dog and cat behavior and facial expressions, taking into account the observers experience with the companion animal species. Independent sets of pet owners and non-owners (n=128) were presented with selected still photographs and edited video sequences to interpret, first spontaneously, then in a multiple choice mode. These were programmed on an interactive CD-Rom. Nested ANOVA showed that the choice of different kinds of multiple choice answers was influenced by their level of plausibility, by the personal experience which the people had had with the species, and by the species shown in stills or movie sequences. Scenes which were interpreted with high agreement of over 80% either showed key elements, a consummatory act, or the context in which the behavior took place. They generally represented scenes relevant to the relationship between an owner and an animal. The facial expressions of fear and curiosity in dog scenes and stress in cat scenes were also interpreted with agreement of over 80%.

INTRODUCTION
here is a considerable gap between the way ethologists, veterinarians, and lay people describe and interpret behavioral sequences and events in the lives of companion animals, especially dogs and cats. Nevertheless, these three groups must communicate with each other in the case of common interest in the animals being studied, treated or housed privately. Humans and animals often communicate with each other on an intuitive level. We observe people talking to or stroking their companion animals and the latter reciprocating with an appreciative bark or wag of the tail. Human-animal communication can be broken down into two categories: vocal and non-vocal. In the category of non-vocal communication, facial expressions, body language, touch and visual signals are certainly involved, but we might also include empathy as a form of non-vocal communication. Empathy has been

Address correspondence to: Dr. D. C. Turner, c/o I.E.T., P.O. Box 32, CH8816 Hirzel, Switzerland. Ph: +41 - 1 - 729 92 27; Fax: +41 - 1 - 729 92 86; e-mail: dennis@turner-iet.ch

defined as the capacity of a person or an animal to experience the needs and feelings of others as if they were his or her own (Levinson 1983). Of course this implies that the empathizer has a theory of mind about the subject of their empathy, an issue about which there is much debate at the moment. The tendency to ascribe human mental experiences to non-human animals is known as anthropomorphism. According to Kennedy (1992) and Mithen (1996), the capacity for anthropomorphic thinking may be a product of natural selection, probably because it proved to be useful to our ancestors for predicting and controlling the behavior of animals. While Goodall (1986), Masson and McCarthy (1994) and Midgely (1992) argue that it is useful to describe animal behavior in terms of intentions, emotions and cognition, Kennedy (1992) and McFarland (1981) are highly critical of the use of anthropomorphic language in the scientific field. The ordinary lay person nevertheless readily believes that his/her dog or cat feels (Masson and McCarthy 1994). There have been numerous surveys (e.g. Ganster and Voith 1983; Lee and Lee 1992; and Hills 1995) indi-

Bahlig-Pieren & Turner

ANTHROZOS, 12(4), 1999

205

cating that over 90% of dog and/or cat owners believe that they can sense the mood of their companion animals, and that their animals can sense and react to changes in their own moods. The purpose of our research was to determine the concurrency and plausibility of anthropomorphic interpretations and ethological descriptions of dog and cat behavior (animalanimal interaction and animal-human interaction) as well as facial expressions in different edited video scenes and still photographs. We also took into account the observers experience with the companion animal species and asked the following questions: 1. How do pet owners and non-pet owners interpret animal behavior anthropomorphically or ethologically? 2. When do they interpret anthropomorphically and show a high level of agreement (concordance)? 3. Are dog and cat pictures and film sequences interpreted in the same way and equally well? 4. Do people agree in their interpretation of the facial expressions associated with different emotional states such as fear, anger, contentment, jealousy, curiosity and affection? 5. Can behavior be correctly predicted? 6. Which factors influence the level of agreement? context experience with the species occurrence of key elements occurrence of a consummatory act amount of information presented, e.g., number of stills shown. Key elements are those most easily ascribed to one motivational state; consummatory acts are defined as those ending a natural sequence of behavior and lowering motivation in that functional area (see Immelmann 1982).

tographs each of dogs and cats; two series of dog stills representing two different behavioral scenes and four series of cat stills representing four different behavioral scenes; 16 dog video sequences and 11 cat video sequences as well as one interrupted movie of each species. For interpretation of the still photographs, the test persons were first asked to do this spontaneously and then to choose from a set of multiple choice answers. Of the multiple choice answers, three were anthropomorphic (A) and three ethological (E), and for both categories the three answers were rated from 13: very plausible (1), plausible (2) or implausible (3). This rating was defined by three ethologists, who interpreted the material independently of each other and came to over 95% agreement. For interpretation of the video sequences only multiple choice answers could be chosen. Validity of interpretations was tested through a series of successive photographs describing a variety of behavioral sequences. After each picture the participants were asked to indicate spontaneously the intentions of the animal portrayed. Validity was also tested with a short film sequence that was interrupted halfway by asking the individual to predict the outcome of the sequence. We tested the degree of agreement between the different answers using the index of concordance (Bortz, Lienert and Boehmke 1990).

RESULTS
Nested ANOVA showed significant effects of experience on the different kinds of multiple choice answers selected. There were significant differences between the inexperienced and the cat- and dog-experienced groups, as well as between the cat-experienced and the cat- and dog-experienced groups. Cat- and dog-experienced people interpreted more ethologically very plausible (E1) than the totally inexperienced and the cat-experienced people (F=11.56, p=0.000). In addition, cat scenes were generally less often interpreted anthropomorphically (A1 and A2) than dog scenes (F=51.98, p=0.000). Looking at the distribution of answers given by the experienced and inexperienced people (Figure 1, pg. 207), a similar pattern was found. In both groups the very plausible answer, anthropomorphic as well as ethological (A1 and E1), was chosen most frequently, followed by the plausible (A2 and E2) and the implausible answers (A3 and E3). Taking only the spontaneous answers into account, we found an even more profound difference between the experienced and inexpe-

METHODS
Volunteers were solicited through advertisements in local newspapers and animal magazines. A set of 128 lay people, 19 people without any experience with either species, 45 cat-experienced only, 12 dog-experienced only and 52 dog- and cat-experienced people, were presented still photographs and video sequences. The questionnaire was programmed on an interactive CD-Rom, which enabled the participants to work at their own pace. By using a CD-Rom, the video sequences could also be easily watched, interrupted, forwarded and reversed as often as the participants liked. The material presented consisted of: 10 still pho-

206

ANTHROZOS, 12(4), 1999

Bahlig-Pieren & Turner

Figure 1. Distribution of Different Categories of Cat and Dog Answers

Figure 2. Number of Correct Answers after Each Still

(A = anthropomorphic, E = ethological; rated from 1= very plausible, 2 = plausible to 3 = implausible). Shown are medians, upper and lower quartiles, and maximum and minimum values.

rienced test persons. The inexperienced group showed significantly more incorrect answers than the experienced ones (WILCOXON, n=9, T=5.5, p=0.025 for dog answers; n=10, T=0, p<0.005 for cat answers). The same result was found in predicting an interrupted movie of one cat and dog sequence each. Again the predictions of the inexperienced persons were significantly more often incorrect than those of the experienced participants (Dog sequence: 2=8.1, p=0.005; Cat sequence: 2=7.44, p=0.001). In order to test the hypothesis whether the correct interpretation of behavior is dependent on a key element, sight of a consummatory act, or on the amount of the behavioral sequence shown, the participants were shown several series of photographs each depicting different behavioral sequences: Dominance behavior and play behavior were the dog sequences, and sexual behavior, food expectation, attention-seeking and territoriality were the cat sequences. After each picture the test persons were asked to indicate any intentions of the portrayed animal. We then determined at which point in each sequence the correct answer was given.

Qualitative comparison between four different series of cat and dog scenes.

There was a significant correlation between the number of stills shown in all four of the progressing sequences of cat behavior and correct interpretations (PAGE Trend test L(4,4)=117.5, p<0.01). However, this was not the case in the progressing sequences of dog behavior. We also analyzed these results qualitatively by looking at the distribution of correct answers for each progressing sequence, four of which are illustrated in Figure 2. In the case of cat sexual behavior, we found a gradual increase of correct answers with increasing number of scenes the people had seen, although there was

Bahlig-Pieren & Turner

ANTHROZOS, 12(4), 1999

207

a distinct gap between the fourth and the last scene (see Figure 2). This indicates that the consummatory act, in this case mounting, was the cue for a correct answer. In the sequence of cat feeding (Food Expectation in Figure 2), there was a gap between the first and second still. Looking at the photographs, a food bowl becomes visible in the second picture, which can easily be associated with feeding behavior. In this case we have to conclude that the bowl (or context) presented the cue for the correct answer. In the sequence on cat territoriality, a gap is once again found between the first and second stills (Figure 2). People were able to recognize a hiss with a defense posture of the cat in the second scene of the photographs. The hiss was apparently the key element for a correct interpretation. In the sequence of play behavior of dogs, the general pattern of answers given seems to differ greatly from those found for all the other progressing sequences (see Figure 2). Here the test persons already answered correctly after the first still, which showed the dog pawing and can be regarded as a key element for an intent to play. After subsequent stills, the participants obviously became uncertain about their first answer and changed their interpretation. If we consider that play behavior consists of different elements from various motivational and functional states, this result is not surprising. These findings suggest that key elements as well as consummatory acts play an important role in correct interpretation, and that the number of stills of a behavioral sequence (total amount of information) shown doesnt necessarily lead to a correct interpretation. Of particular interest was the question whether the volunteers could recognize an emotional state of an animal without being shown the context in which the behavior took place. We used a film sequence of disappointment behavior in the dog for this purpose: disappointment was stimulated by preparing the dog for a walk, then abandoning this intention and ignoring the dog for a few minutes. In the scene of disappointment without context, the test persons were only shown the dog towards the end of the sequence. The answers on this film sequence were significantly more plausible when the context in which the behavior took place was recognizable (McNEMAR 2=17.98, p<0.001) (see Figure 3). Lastly, we were able to confirm by concordance test that all categories of test persons showed high agreement in their interpretations

Figure 3. Influence of Context on Interpretation in Dog-disappointment Behavior.

McNemar 2=17.98, p<0.001.

(dog-experienced: n=64, =35.7, p<0.001; catexperienced: n=97, =46.9, p<0.001; inexperienced: n=19, [dog] =9.7, p<0.001; [cat] =8.7, p<0.001). Qualitatively speaking, scenes with high agreement of over 80% either showed a consummatory act, e.g. pouncing on prey (cat hunting behavior), mounting (cat sexual behavior), or ritualized behavior, e.g. cat defensive display and dog dominance/submission, or represented important scenes for the relationship between an owner and an animal, e.g. dog greeting behavior, dog pawing (attention-seeking) and cat greeting behavior (see Serpell 1983). Only the scenes showing facial expressions associated with fear and curiosity in dogs and stress in cats were interpreted with high agreement (>80%). Other scenes, showing facial expressions associated with contentment, anger and affection (in both dogs and cats) or disappointment and jealousy (in dogs) were not.

DISCUSSION
In a similar, but less extensive study, Fidler, Light and Costall (1996) found that pet owners tend to describe dog-human interactions more mentalistically (using our terminology, anthropomorphically) than non-pet owners. We could not confirm this result, but this may have been due to the different methods used in the two studies. We did find, however, that anthropomorphic interpretations depended more on the species shown in the stills and movies. Dog scenes were generally interpreted more anthropomorphically than cat scenes. This may be explained by the fact that dogs faces are more richly endowed with facial muscles that are reportedly used to express emotions in ways similar to those of humans. This obviously leads people to interpret dogs facial expressions as if they were looking at another person (Serpell 1983,

208

ANTHROZOS, 12(4), 1999

Bahlig-Pieren & Turner

Feddersen-Petersen and Ohl 1995). Moreover, anthropomorphic interpretations might also have depended on the content of the various scenes shown: A typical dominant/submissive behavior of dogs was often described as such, whereas pawing was often interpreted anthropomorphically as seeking attention. With the scene of disappointment behavior of the dog we were able to show that the context in which the behavior took place had to be visible for a correct interpretation. Of course, context might have been important in some of the other scenes as well, but they were not analyzed in this connection. Nevertheless, there were at least two scenes with facial expressions and absolutely no context which were interpreted with high agreement of over 80% (fear in dogs and stress in cats). Needless to say, these expressions represent emotional states, which scientists normally find difficult to define but which are important to an animals welfare. The fact that high agreement amongst interpreters was found may have consequences for this field. To assess ability to predict behavior, only one interrupted movie per species was utilized in this study. By presenting a variety of movies in this way (which were not available to us), one could examine the hypothesis that people who mainly anthropomorphize are more successful in predicting companion animal behavior than those who do not. Hebb (1946) demonstrated this for the interpretation of chimpanzee behavior. We were, of course, aware of the problems associated with multiple choice testing, especially of influencing the choice of answers. For this reason, spontaneous answers were first requested in the interactive CD-Rom and ana-

lyzed separately. Furthermore, the number of test persons in each category differed and may have had an influence on the results. Few people without any experience with companion animals took part in the test and the number of male participants was also low. Additionally, many people who knew the work of the second author volunteered, and these persons had more knowledge of cat behavior. They generally chose the ethological descriptions of the scenes as opposed to the anthropomorphic answers. In summary, our findings clearly showed that there was a difference in the interpretations of the experienced and inexperienced groups, even though the inexperienced people answered surprisingly plausibly and with high agreement. This result may offer some support for Kennedys (1992) and Mithens (1996) notion that humans have been selected for an ability to predict and interpret the behavior and emotional states of other non-human species. Empathy and anthropomorphism play an important role in this capacity, as it presumably also did in the original process of animal domestication.

ACKNOWLEDGMENTS
This study was financed by the Swiss National Science Foundation, Grant Nr. 31.49388.96 to Turner. We thank Jeannine Traber for providing one of the video sequences, Karin StammbachGeering and Sonja Doll-Sonderegger for their assistance in interpreting the scenes and videosequences (ethologically and anthropomorphically) and James Serpell for improving the English manuscript. Lukas Meyer (University TV Dept.) was extremely helpful in programming the interactive CD-Rom. Discussions with all members of the Zurich companion animal ethology team were most fruitful.

REFERENCES
Bortz, J., Lienert, G. and Boehmke, K. 1990. Verteilungsfreie Methoden in der Biostatistik. Berlin: Springer Verlag. Feddersen-Petersen, D. and Ohl, F. 1995. Ausdrucksverhalten beim Hund. Gustav Fischer Verlag, Vet special series, Jena. Fidler, M., Light, P. and Costall, A. 1996. Describing dog behavior psychologically: pet owners versus non-owners. Anthrozos 9(4): 196200. Ganster, D. and Voith, V. L. 1983. Attitudes of cat owners toward their cats. Feline Practice 13(2): 2129. Goodall J. 1986. The Chimpanzees of Gombe. Cambridge, MA: Harvard University Press. Hebb, D. O. 1946. Emotion in man and animal: an analysis of the intuitive processes of recognition. Psychological Review 53: 88106. Hills, A. M. 1995. Empathy and belief in the mental experience of animals. Anthrozos 8(3): 132142. Immelmann, K. 1982. Wrterbuch der Verhaltsforschung. Berlin: Verlag Paul Parey. Kennedy, J. S. 1992. The New Anthropomorphism. Cambridge, UK: Cambridge University Press. Lee, L. and Lee, M. 1992. Absent Friend:

Bahlig-Pieren & Turner

ANTHROZOS, 12(4), 1999

209

Coping with the Loss of a Treasured Pet. Bucks, UK: Henston. Levinson, B. 1983. The future of research into relationships between people and their animal companions. In New Perspectives on Our Lives with Companion Animals, 536550, ed. A. H. Katcher and A. M. Beck. University of Pennsylvania Press, Philadelphia. Masson, J. and McCarthy, S. 1994. When Elephants Weep: The Emotional Lives of Animals. London: Vintage. McFarland, D. 1981. The Oxford Companion to Animal Behaviour. Oxford: Oxford University Press. Midgely, M. 1992. The Mixed Community. In

the Animal Rights / Environmental Ethics Debate. New York: SUNY Press. Mithen, S. 1996. The Prehistory of the Mind: A Search for the Origins of Art, Religion and Science. London: Thames & Hudson. Serpell, J. 1983. The personality of the dog and its influence on the pet-owner bond. In New Perspectives on Our Lives with Companion Animals, 5763, ed. A. H. Katcher and A. M. Beck. Philadelphia: University of Pennsylvania Press.

210

ANTHROZOS, 12(4), 1999

Bahlig-Pieren & Turner

Anda mungkin juga menyukai