University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Faculty Publications, Department of Psychology Psychology, Department of

11-2010

The challenges of understanding animal minds

Jeffrey R. Stevens University of Nebraska-Lincoln, [email protected]

Follow this and additional works at: https://digitalcommons.unl.edu/psychfacpub

Part of the Psychiatry and Psychology Commons

Stevens, Jeffrey R., "The challenges of understanding animal minds" (2010). Faculty Publications, Department of Psychology. 512. https://digitalcommons.unl.edu/psychfacpub/512

This Article is brought to you for free and open access by the Psychology, Department of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Faculty Publications, Department of Psychology by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. Specialty Grand Challenge Article published: 19 November 2010 doi: 10.3389/fpsyg.2010.00203 The challenges of understanding animal minds

Jeffrey R. Stevens*

Center for Adaptive Behavior and Cognition, Max Planck Institute for Human Development, Berlin, Germany *Correspondence: [email protected], [email protected]

Comparative psychology is by nature an • Can animals use referential benchmark for parsimonious explanations. interdisciplinary science that lies at the communication? If a simpler, learning account can explain crossroads of psychology and biology but • How do animals track time and the data, then we should not invoke a more also draws from other fields in the natural, number? complex explanation. This stringent bench- social, and cognitive sciences. The study of • Do animals maintain a cognitive map mark has set a high bar for researchers in the the psychology of animals has been labeled of their environment? cognitive tradition to reject simple learning animal cognition, comparative cognition, • Do animals attend to the well being of explanations. Occasionally, the bar is raised animal learning, animal psychology, and others? higher and higher, sometimes to the extent animal intelligence. Here, comparative • How do animals categorize objects in that the learning mechanism becomes quite psychology is used interchangeably with their world? complex. How complex does a parsimoni- these terms, encompassing all fields that • Which emotional and motivational fac- ous explanation need to be before we accept explore the psychological mechanisms tors underly animal behavior? a representational one? How do we define underlying animal behavior, including the complexity of a mechanism? One solu- human behavior. Studying the psychology of other species tion is to sidestep the parsimony/complexity The primary goal of comparative psy- is not easy. With rare exceptions, we can- question completely. Rather than resorting chology is to understand the cognitive, not ask directly about their psychological to a scale of psychological complexity, we emotional, and motivational processes of states. Instead, we must make inferences can develop and test more precise models the animal mind. How do other animals about the psychology of animals based on of the phenomena of interest. perceive, learn about, and make decisions their behavior. Inferring internal mecha- Even after excluding simpler explana- in their worlds? From our pets to exotic nisms from external behavior results in a tions for phenomena, the question arises, animals portrayed in nature documenta- number of challenges for comparative psy- under what circumstances do animals ries, we are inherently curious about other chologists. Though we have met a number use the more complex form of cognition? animals. Comparative psychology both pro- of these challenges, several remain, slow- Possessing a particular psychological ability vides a window into their minds, as well as ing our progress in advancing comparative does not imply using it at every opportu- offers a unique perspective on the human psychology. To push forward, we must meet nity. In many cases, animals may use simpler mind. Which aspects of our psychology do these conceptual and practical challenges mechanisms, reserving the more complex animals share? Human uniqueness is con- head on. mechanisms for situations in which the sim- stantly challenged as we learn more about pler mechanisms do not work. The circum- the psychology of animal minds. Once dis- The complexity of parsimony stances under which animals use different tinctive human abilities – such as tool use, The difficulties of making inferences about mechanisms remains an open question in language, and mental time travel – appear, internal mechanisms has spawned two comparative psychology. at least to a degree, in other species. Though general approaches to studying cognitive other species exhibit elements of these abili- aspects of animal behavior. The animal Individual differences ties, the central question for comparing learning approach emphasizes the general A hallmark of data in comparative psychol- humans and animals remains, do humans learning principles, such as instrumental, ogy is variation across individuals. Despite and other animals share the same psycho- and Pavlovian conditioning, espoused by this, we have very little understanding of logical mechanisms? Hull, Spence, Tolman, and Skinner. The the sources of variation in psychological Comparative psychology explores many cognitive approach examines other forms mechanisms. Evolutionary models predict of the same topics as human psychology. of cognition such as perception, attention, individual differences in behavior (Wolf From learning and memory to commu- memory, categorization, navigation, timing, et al., 2007), but few models explore why nication and decision making, the field number, communication, decision mak- cognition and emotion should vary across investigates a number of key questions, for ing, and social cognition. Though learning individuals. A number of questions arise example: mechanisms often are considered simpler when developing a theoretical understand- explanations, some cognitive mechanisms ing of individual differences. Is there really • How do animals understand causal are more complex, requiring an organism variation in animal psychological abilities relationships in their environments? to generate a mental representation. Since or only in the expression of these abilities? • Can animals represent the perceptions, Morgan’s (1903) admonition to opt for the To what extent is this variation adaptive? To intentions, and beliefs of others? most parsimonious explanation supported what extent does the variation result from • Do animals plan for the future? by the data, the learning tradition has set the environmental influences? Are there general

www.frontiersin.org November 2010 | Volume 1 | Article 203 | 1 Stevens Understanding animal minds

reasons for variation in psychology or do share similar psychological mechanisms? that tap natural behaviors and situations we need to consider each ability separately? What kinds of evolutionary pressures can yield more valid studies of compara- Researchers of animal behavior investigate shape psychological mechanisms across tive psychology. individual differences under the headings of species? Does the social environment play animal personality, behavioral syndromes, a special role in shaping psychology? How Strong inference and temperament. We must now extend this can we test whether species share actual Much of comparative psychology is based study of variation at the psychological level mechanisms rather than just general psy- on existence proofs: Do animals have theory and explore the sources of such variation. chological abilities? Comparisons across of mind, metacognition, episodic memory, species pose difficulties and require great empathy, or other-regarding preferences? To Putting the comparative back care regarding the species tested and the further mature, however, the field needs to into comparative psychology experimental methods used. Nevertheless, develop more sophisticated models of how Despite the name and origins of the field, the fruits of this labor will yield valu- psychological mechanisms work, not just comparative psychology lacked many com- able insights into the nature of animal whether they are present and what influ- parative studies during much of the twenti- psychology. ences them. Weisman (2008) offers an eth century. Only a handful of species acted important challenge to the young research- as the workhorses of the field: the pigeon, Ecological relevance ers in behavioral and cognitive science: the rat, and the rhesus macaque. Focusing Simon (1990) argued that a serious study embrace Platt’s (1964) notion of strong on a few species offers a more in-depth and of cognition must explore both the mind inference. Strong inference emphasizes the cumulative method for understanding the and the environment in which the mind development of multiple hypotheses and psychology of those species. This restric- interacts. He likened this to two blades the design of critical experiments allowing tion, however, limits our broader under- in a pair of scissors that cannot function for alternative outcomes that will exclude standing of psychological mechanisms in unless both blades are present. This pro- some of the hypotheses. Currently, many animals. Comparative psychologists have vides an important lesson for comparative behavioral scientists rely on testing a single lamented this lack of diversity for decades psychology as well. Studies of comparative hypothesis against a null hypothesis, despite (Beach, 1950; Cook, 1993; Shettleworth, psychology frequently face criticism for the pitfalls of this approach (Marewski and 1993), and, more recently, researchers have overly artificial task environments. Using Olsson, 2009). Rather than only testing for tested a number of interesting questions in a simple stimuli in impoverished apparatuses the presence or absence of a psychological wide variety of species, ranging from insects offers clear advantages in terms of reduc- ability or for factors that influence an abil- and octopods to tortoises and manatees. We ing the likelihood of confounding variables. ity, we need to explore how the processes should continue to meet the challenge of Unfortunately, this gain in internal valid- work. To this end, testing a series of for- casting a wide taxonomic net. ity comes at the price of reducing external mal process-based models allows a precise Cataloging the abilities of various species validity. Striking a balance between careful understanding of psychological mecha- provides a necessary first step in understand- control over the experimental situation and nisms. Though formal models are common ing the psychology of animals. However, eliciting meaningful responses is critical to in the animal learning tradition (e.g., mod- comparative does not mean simply collect- understanding animal psychology (Cook, els of timing), other areas of comparative ing information on a broad range of spe- 1993). The right balance starts with asking psychology have yet to use these models to cies; it also implies active comparison across ecologically relevant questions. What kinds their full potential. species (Shettleworth, 1993). The implicit of mechanisms do animals need to navigate Developing and testing multiple hypoth- comparison in the early days of compara- through their physical and social worlds? eses yields more careful inferences because tive psychology was between a single ani- The next step is to develop experimen- more powerful statistical techniques can mal species and humans. Comparisons have tal stimuli and tasks that tap the natural be used. Instead of relying on the weak become more explicit recently, borrowing abilities of animals. Though simple artifi- null hypothesis testing paradigm, we can methods directly from fields such as cogni- cial stimuli offer complete control over the use competitive model testing techniques tive psychology, developmental psychology, features presented to animal subjects, Cook to discriminate between hypotheses. social psychology, and economics. When (1993) argues that more naturalistic stimuli Incorporating process models and competi- testing questions of human psychology of intermediate complexity can offer a rea- tive model selection into the strong infer- in animals, we should strive to make the sonable balance between control and eco- ence approach can only raise the level of experiments as comparable as possible logical relevance. In addition, Hare (2001) rigor for comparative psychology. across species. argues that we must consider the natural Comparing not just to humans but ecology of the species to develop appro- Replication and reproducible across animal species opens a host of priate experimental tasks. When placed in research new questions to investigation. In par- unnatural situations, animals may not dem- Around the turn of the twentieth century, ticular, the comparative method allows onstrate the relevant psychological mecha- a horse named Clever Hans captivated first the testing of evolutionary questions nism. Though not easy to conduct, field Germany and then the world with amazing regarding the origins of and ecological experiments can offer the most naturalis- feats of arithmetic skill. Alas, the even more pressures on psychological mechanisms. tic circumstances for testing psychological clever Pfungst (1911) debunked the horse Do ­phylogenetically closely related species mechanisms in animals. Developing tasks by demonstrating that Hans simply used

Frontiers in Psychology | Comparative Psychology November 2010 | Volume 1 | Article 203 | 2 Stevens Understanding animal minds

inadvertent cues from his owner or others Allowing free access to data and data questions and continually be mindful of the during the questioning. This story is often analyses is equally important in advancing most rigorous means to achieve our goal of the first cautionary tale learned by all com- comparative psychology because it allows delving into the minds of animals. parative psychologists as they enter the field. for reproducible research (de Leeuw, It reminds us that the wishes and biases of 2001). With the available electronic tech- Acknowledgments the observer can influence the observation. nologies, scientists can include data files I wish to thank Lael Schooler, Sara It also has encouraged meticulous method- either as supplementary materials pub- Shettleworth, and Alison Stevens for com- ologies for researching animal psychology. lished on a journal’s website or posted on ments on this paper. Our field could improve, however, in foster- the author’s website. In addition to pub- ing replication and reproducible research. lishing the data, granting access to how References A core principle of experimental the data were analyzed can be useful for Beach, F. A. (1950). The Snark was a Boojum. Am. Psychol. 5, 115–124. research methods is to replicate studies. other scientists. Some statistical software Cook, R. G. (1993). The experimental analysis of cogni- This is common practice in human psy- programs, including the freely available tion in animals. Psychol. Sci. 4, 174–178. doi: 10.1111/ chology experiments, perhaps because a R software for statistical computing2, j.1467-9280.1993.tb00483.x. new pool of participants is available each allow the reporting of not only statistical de Leeuw, J. (2001). Reproducible Research: The Bottom semester at the university. The typically results but also the code used to generate Line. Department of Statistics Papers, University of California, Los Angeles. small sample sizes used in comparative the results. Making this code freely avail- Hare, B. (2001). Can competitive paradigms increase the psychology can make replication difficult. able facilitates both error checking and validity of experiments on primate social cognition? Nevertheless, we should encourage the gen- direct reproduction of data analysis. Thus, Anim. Cogn. 4, 269–280. doi: 10.1007/s100710100084. eral practice of replicating our work and posting experimental protocols, example Marewski, J. N., and Olsson, H. (2009). Beyond the publishing these replications, even if they videos, data files, and analysis code main- null ritual: formal modeling of psychological proc- esses. J. Psychol. 217, 49–60. doi: 10.1027/0044- result in different findings. As one solu- tains the transparent nature of science. In 3409.217.1.49. tion to facilitating replications, we can the spirit of the mission of open access, Morgan, C. L. (1903). Introduction to Comparative grant free access to experimental methods. we at Frontiers in Comparative Psychology Psychology, 2nd Edn. London: Walter Scott. The Comparative Mind Database1 offers a encourage the publication of methods, Pfungst, O. (1911). Clever Hans: (The Horse of Mr. Von Osten.) A Contribution to Experimental Animal and systematic attempt to maintain free access data, and data analysis along with scien- Human Psychology. New York: Henry Holt. to experimental methods in comparative tific articles. Platt, J. R. (1964). Strong inference. Science 146, psychology by archiving experimental 347–353. protocols and video clips of experimental Looking ahead Shettleworth, S. J. (1993). Where is the comparison in sessions. This database facilitates compara- Comparative psychology explores the fasci- comparative cognition? Alternative research pro- grams. Psychol. Sci. 4, 179–184. doi: 10.1111/j.1467- tive studies by standardizing experimental nating topic of the animal mind. The field 9280.1993.tb00484.x. procedures such that researchers may use has made great strides in the recent years, Simon, H. A. (1990). Invariants of human behavior. similar protocols across species. In addi- though more work remains. We face excit- Annu. Rev. Psychol. 41, 1–20. doi: 10.1146/annurev. tion to exact replications of methods, we ing conceptual challenges such as grappling ps.41.020190.000245. need to replicate our studies using differ- with issues of parsimony and complexity, Weisman, R. G. (2008). Advice to young behavioral and cognitive scientists. Behav. Processes 77, 142–148. doi: ent methods. As scientists, we often become exploring individual differences, and pos- 10.1016/j.beproc.2007.09.004. enamored with our methodologies. Yet, the ing careful comparative questions of psy- Wolf, M., van Doorn, G. S., Leimar, O., and Weissing, use of a single method to test a hypothesis chological mechanism. We also share more F. J. (2007). Life-history trade-offs favour the evolu- can result in a biased view of the phenom- practical challenges with other areas in tion of animal personalities. Nature 447, 581–584. enon of interest. Returning to Simon’s scis- behavioral science and beyond, including doi: 10.1038/nature05835. sors, cognition cannot be understood fully developing more naturalistic experimental Received: 15 September 2010; accepted: 28 October 2010; without accounting for the environment, paradigms, implementing strong inference published online: 19 November 2010. and ensuring that our findings generalize techniques, and producing reproducible Citation: Stevens JR (2010) The challenges of understand- across multiple experimental environments and transparent research. No single piece ing animal minds. Front. Psychology 1:203. doi: 10.3389/ fpsyg.2010.00203 is crucial. Replicating our work within lab- of scientific work will likely address all of This article was submitted to Frontiers in Comparative oratories, between laboratories, and across these challenges. Nevertheless, to advance Psychology, a specialty of Frontiers in Psychology. methodologies strengthens the integrity of comparative psychology into the twenty- Copyright © 2010 Stevens. This is an open-access article our conclusions. first century, we must carefully craft our subject to an exclusive license agreement between the authors and the Frontiers Research Foundation, which permits unre- stricted use, distribution, and reproduction in any medium, 1http://www.cmdbase.org/ 2http://www.r-project.org/ provided the original authors and source are credited.

www.frontiersin.org November 2010 | Volume 1 | Article 203 | 3