The Pluralistic Evolution of Factive Mindreading
Gavin Robert Foster
Purdue University
The following is a significantly abridged version of a much-longer paper. As such, there are certain qualification and nuances that are lacking. For those interested in reading the current draft, please send me an email at foste362 at purdue dot edu. Feedback and questions are more than welcome!
Most research on social cognition focuses on how we predict and explain the minds of others by attributing to them mental states such as beliefs or desires. Such states are non-factive: their contents need not obtain. Recently, however, researchers have begun to consider the role of factive mindreading in social cognition, which involves attribution of states such as knowledge, sight, and awareness. Essential to factive states is their factivity: their contents must obtain. This also means that sensitivity to factive states does not require an understanding of the difference between appearance and reality. Due to this comparative simplicity, many researchers have adopted the knowledge-first view of mindreading: phylogenetically and ontogenetically, knowledge-attribution arose before belief-attribution. I refer to such proponents as KF-theorists (e.g., Nagel 2013, 2017; Pavese 2024; Phillips & Norby 2021; Phillips et al. 2021; Westra & Nagel 2021). However, I use “knowledge-sensitivity” rather than “knowledge-attribution” to remain neutral about what mindreading amounts to. Regardless of whether one is a ‘romantic’ (mindreading is the conscious inference of an unobservable mental state) or ‘killjoy’ (mindreading can be achieved implicitly and through less-sophisticated cognitive processes) about mindreading, the pluralistic argument I present holds.
In my paper, I argue that current KF-theorists go awry in adopting, usually implicitly, a general account of knowledge-sensitivity. On their view, a general ability to track knowledge states in other creatures evolved earlier than belief-attribution because it is highly adaptive. Yet, this would not be adaptive for most non-human species where the ecological challenges that they face are very narrow. Instead, I argue that what would first be selected for is a plurality of narrower or domain-specific knowledge-sensitivity abilities, such as the ability to track knowledge-where in other creatures based on their visual access to the world, the ability to track knowledge-who in particular social contexts, or the ability to track knowledge-how when judging competence.
To determine what kind of knowledge states a species would evolve to be sensitive to, one must first consider its ecological background. Broadly speaking, cognitive abilities emerge as responses to ecological challenges that confer greater fitness on the creature. To outweigh the cognitive costs involved in employing such an ability, the sort of knowledge being tracked should enable the species to do something that is fitness-conferring. Usually, what is required is a sensitivity to a very narrow slice of knowledge states another creature has. Evolving a general knowledge-sensitivity ability to solve a narrow ecological challenge would be akin to getting the entire toolbox when all you need is a hammer. This is the basis of what I term the evolutionary argument, and one such way of illustrating it is to consider the food-caching behaviour of western scrub jays (and other corvids).
The relevant ecological challenge that this ability evolved to solve is the reduction of pilfered caches, which significantly impacts their caloric intake (Vander Wall & Jenkins 2003).
Scrub jays will wait for potential pilferers to either leave or be distracted before caching the food, or else may even move the storage-location if cached while a potential pilferer watched (Dally et al. 2005). So, we have grounds to attribute to such birds a knowledge-where sensitivity. However, according to the KF-theorist, if they possess this ability they must also possess the more general knowledge-sensitivity ability. On the basis of their ability to track knowledge-where, should we also attribute to scrub-jays the ability to track knowledge-why? The ability to track knowledge-why would be something akin to tracking someone else’s causal understanding of the world. Certainly humans can do this, which is why we go to the doctor; however, we have no evidence of such an ability in scrub jays, nor should we expect there to be an ecological challenge it would solve.
In my paper, I present multiple predictions of this account. Here I focus on one in particular: if the pluralistic account of factive mindreading is correct, then there should be evidence of creatures that are capable of tracking knowledge states of certain types while failing to track others. The scrub-jay case illustrates the absence of a sort of broader knowledge-tracking. But the same point can be made even within a single kind of knowledge-sensitivity. However, this point is also illustrated by one species tracking a particular kind of knowledge within a particular sensory modality. An example of this is the knowledge-where tracking of great apes compared to the more phylogenetically distant ringtail lemurs (a strepsirrhine). As outlined more fully in the paper, tracking knowledge-where often involves a sensitivity to behavioural cues such as eye-gaze, head-direction, chest posture, and so on (Reaux et al. 1999). A sensitivity to knowledge-where states thus means adequately picking out these sorts of behavioral-patterns in a creature.
In Hare and colleagues’ landmark study (2000), it was demonstrated that submissive chimpanzees will avoid food that dominant chimpanzees have seen behind a transparent barrier, instead opting for food that is outside the dominant’s line-of-sight. Numerous experiments have extended this research, suggesting chimpanzees will deliberately modify the knowledge-states of other creatures to their advantage (Karg et al. 2015). Knowledge-where is not only limited to the visual domain, as both monkeys and apes have demonstrated sensitivity to know-where on the basis of auditory access (Melis et al. 2006; Santos et al. 2006). However, in the slightly more phylogenetically distant ringtail lemurs (a strepsirrhine), lemurs are able to track knowledge-where on the basis of behavior patterns connected to seeing but not to hearing (MacLean et al. 2013; Bray et al. 2014). As Bray and colleagues note, “it may be that lemurs are not under significant pressure to exploit the auditory perception of conspecifics” (2014, 743).
The evidence described above only hints at the importance of exploring which particular knowledge-detection abilities species have across phylogenetic space. The existing fragmentation also suggests at minimum that human beings are unique in our ability to attribute an extensive array of knowledge types. I suggest that this the framework outlined here offers a new way of designing experiments to test for the mindreading abilities of non-human animals. Further, future research should be directed at explaining the ease with which humans track and exploit the factive states of other agents in a domain-general way. What event in the history of hominin evolution led to this generalized capacity that KF-theorists have so far concerned themselves with? More narrowly, what sorts of ecological challenges would have driven this generalization process, and what other cognitive abilities are required for the adaptation to evolve?
I want to conclude by briefly thanking all of those involved in organizing SSPP this year, which (in my opinion) was a tremendous success. Beyond the audience of this paper at SSPP, I would also like to thank audiences at the International Comparative Cognition Conference in Montreal, as well as the attendees of the Joint Workshop for Animal Minds in Bochum. Special thanks go to the many people who have helped inform and refine this paper, including (but not limited to): Evan Westra, Chris Krupenye, Myrto Mylopoulos, Mike Dacey, Noam Miller, Brett Karlan, Colin Allen, Luke Townrow, Robert Lurz, Ty Henley, Amalia Bastos, Albert Newen, Theo Camp, and Nina Van Rooy.
Bibliography
Bray, Joel, Christopher Krupenye, and Brian Hare. 2014. “Ring-tailed lemurs (Lemur catta) exploit information about what others see but not what they can hear.” Animal Cognition 17 (3): 735-44.
Dally, Joanna M., Nathan J. Emery, and Nicola S. Clayton. 2005. “Cache protection strategies by western scrub-jays, Aphelocoma californica: implications for social cognition.” Animal Behaviour 70 (6): 1251-63.
Hare, Brian, Josep Call, Bryan Agnetta, and Michael Tomasello. 2000. “Chimpanzees know what conspecifics do and do not see.” Animal Behaviour 59 (4): 771-85.
Karg, Katja, Martin Schmelz, Josep Call, and Michael Tomasello. 2015. “Chimpanzees strategically manipulate what others can see.” Animal Cognition 18 (5): 1069-76.
MacLean, Evan L., Aaron A. Sandel, Joel Bray, Ricki E. Oldenkamp, Rachna B. Reddy, and Brian A. Hare. 2013. “Group Size Predicts Social but Not Nonsocial Cognition in Lemurs.” PLoS ONE 8: e66359.
Melis, Alicia P., Josep Call, and Michael Tomasello. 2006. “Chimpanzees (Pan troglodytes) conceal visual and auditory information from others.” Journal of Comparative Psychology 120 (2): 154-62.
Nagel, Jennifer. 2013. “Knowledge as a Mental State.” In Oxford Studies in Epistemology Volume 4, 272-308. Edited by Tamar Szabó Gendler & John Hawthorne. Oxford University Press.
Nagel, Jennifer. 2017. “Factive and nonfactive mental state attribution.” Mind & Language 32: 525-44.
Pavese, Carlotta. 2024. “Factive Mindreading in the Folk Psychology of Action.” In Putting Knowledge to Work: New Directions for Knowledge-First Epistemology, 145-69. Edited by Artūrs Login & Jacques-Henri Vollet. Oxford University Press.
Phillips, Jonathan, and Aaron Norby. 2021. “Factive theory of mind.” Mind & Language 36 (1): 3-26.
Phillips, Jonathan, Wesley Buckwalter, Fiery Cushman, Ori Friedman, Alia Martin, John Turri, Laurie Santos, and Joshua Knobe. 2021. “Knowledge before belief.” Behavioral and Brain Sciences 44 (e140): 1-75.
Santos, Laurie R., Aaron G. Nissen, and Jonathan A. Ferrugia. 2006. “Rhesus monkeys, Macaca mulatta, know what others can and cannot hear.” Animal Behaviour 71 (5): 1175-81.
Vander Wall, Stephen B., and Stephen H. Jenkins. 2003. “Reciprocal pilferage and the evolution of food-hoarding behavior.” Behavioral Ecology 14 (5): 656-67.
Westra, Evan, and Jennifer Nagel. 2021. “Mindreading in conversation.” Cognition 210: 104618.