According to a recent study, prairie voles console one another when stressed. J. P. Burkett and his colleagues separated a pair of voles and subjected one of them (the “demonstrator”) to light foot shocks.1 The demonstrator was then reunited with the other vole (the “observer”), who had not witnessed the shocks. Burkett and colleagues found that the observer engaged in significantly longer grooming behavior directed toward the demonstrator than in the control situations in which the demonstrator was not shocked. The grooming, which Burkett and colleagues also call “consolation behavior,” was limited to familiar voles; observers would not groom or console strangers. The New York Times reported on this experiment under the headline, “A Furry Shoulder to Cry On.”2 The Times report begins with the claim that “prairie voles console one another when distressed.” There is a noticeable gap between the claims made by the scientists – that the observer grooms the stressed demonstrator – and those made by The New York Times – that the observer provides a shoulder for the distressed demonstrator to cry on. The claims made by The Times are evocative, suggesting a sympathetic response characterized by empathy and fellow feeling. The reader is practically invited to imagine the observer reaching out to the demonstrator, saying “I feel your pain.” Such a scenario is clearly not experimentally supported. Any attribution of human-like experience to the prairie vole would be premature. Or would it? What warrants an attribution of human-like experience to a nonhuman animal? What evidence might we gather to support the claim that a nonhuman animal has thoughts and experiences similar to those that humans have?
In what follows, I suggest that the comparative evidence (gathered mostly from primate studies) suggests that even though some animals may have an active mental life that includes thoughts and goals, it is unlikely that the nonhuman animal’s experience of its mental life is similar to human beings’ experience of their mental life. Human mental experience relies on a concept of a self, which evidence suggests is missing in many nonhuman animals. To see why such a concept of the self is important, I start by looking some more at the ways in which prairie vole consolation behavior is similar to empathetic behavior found in human beings. Then I turn to remarkable planning behavior by a chimpanzee. Careful consideration of this behavior and the nature of the evidence it provides for the understanding of chimpanzee minds indicates the role a concept of the self plays in human mental life.
Burkett and colleagues uncover several details about prairie vole consolation behavior that suggest parallels to human empathic response. First, like a human being in an empathetic state, the prairie vole observers apparently experience a stressful state similar to the one the demonstrator is in, as indicated by their engaging in similar fear and stress-related behaviors, as well as by increases in hormones associated with stress. These hormone levels decreased if the observer was allowed to interact with and groom the demonstrator, but remained elevated if the observer was not allowed to groom the demonstrator. The observer voles also were more likely to groom demonstrators who were familiar to them, much like empathetic human beings, who are more likely to demonstrate empathy toward familiars. The voles also demonstrated an awareness of the distinction between self and other: even though both demonstrators and observers exhibited elevated levels of stress, only the demonstrator received comfort; demonstrators did not engage in increased grooming of the observers. Moreover, in human beings, empathetic behavior has been linked to the presence of oxytocin, so Burkett’s team injected observers with an oxytocin antagonist in order to ascertain the effects of oxytocin on prairie vole empathetic behavior. Those prairie voles who had been injected with the oxytocin antagonist engaged in no consolation behavior subsequent to that injection.
In both human beings and prairie voles, consolation behavior is triggered by stress. Consolation behavior involves the empathetic increased sensation of stress in the consoling individual; involves a recognition that, of the two individuals experiencing stress, the originally stressed individual is the one who receives consoling; and is closely related to the presence of oxytocin in the system. Why not think that the prairie vole is experiencing a state much like that of a human being engaged in empathetic behavior?
We can ask a similar question about Santino, a chimpanzee at the Furuvik Zoo in Sweden. As reported by Osvath (2009),3 Santino exhibited some remarkable aggressive behavior towards zoo visitors. Captive chimpanzees will often express themselves by throwing objects at visitors. Often these objects are whatever is at hand, including feces. However, Santino adopted a different strategy: he cached stones early in the day, before there were visitors at the zoo. Then, later, when visitors came to the zoo, he would recover the hidden stones and throw them at the visitors. Initially, his ammunition consisted of rocks found in the waterbed adjacent to his enclosure, cached along the shoreline closest to where the visitors would walk. Later, however, he added chunks of concrete to his ammunition, even going so far as finding where in his enclosure the concrete was likely to be weak and then breaking it off to use as ammunition. As zoo officials would remove his caches of stones, he took to manufacturing hiding places from hay that he found in his enclosure, and then concealing the stones and concrete under those new hiding places.4 Santino was calm when he planted the ammunition in caches and angry later when he recovered the ammunition and threw it at the visitors. These details tempt one to conclude that Santino was engaging in advance planning. He knew that he would be angered by the presence of visitors, that he would want to demonstrate that anger, and that unless he prepared ahead of time, he would not be able to demonstrate that anger – and so he prepared ahead of time. This interpretation attributes to Santino several complex mental attitudes: a sense of self, knowledge about his current states of mind and how they may differ from his future states of mind, and an ability to take steps so that his future self will be satisfied, even at the cost of energy expenditure which is useless for his present self. Is this interpretation of Santino’s behavior justified? Or is it an example of an unwarranted projection of human-like psychological attributes onto another animal?5
As these questions suggest, there are two obvious alternative accounts of Santino’s behavior: the deflationary account that we can explain Santino’s behavior in ways that don’t involve attributing human-like psychological attributes to him (in the case of the prairie voles, this would amount to asserting that what they are experiencing is not compassion), and the inflationary account that Santino is truly engaged in planning for the future, that he is truly taking steps to ensure that his future self is able to achieve its goals (or, in the case of the prairie voles, that they are truly empathetic).6 Several ethologists argue for the deflationary answer. For example, Roberts and Feeney (2010) argue that Santino’s behavior does not demonstrate advanced planning. The fact that he does not stockpile ammunition during the off-season (during which the zoo is closed), but instead waits until visitors start arriving at the zoo in the spring to cache stones suggests that the first visitors “served as a contextual stimulus for storing rocks … without the anticipation of doing so at any specific time in the future” (Roberts and Feeney 2010: 53).7 Suddendorf and Corballis (2010) argue that there are alternative explanations for Santino’s behavior; the data collected by Osvath underdetermines the conclusion that Santino was planning for future dominance displays. Shettleworth (2010) repeats Suddendorf and Corballis’s analysis, adding in her discussion that often apparently complex behavior can be explained by means of “elementary mechanisms.” More generally, Balter (2012), reporting in Science, describes a general attitude among scientists that when there is a “simpler” explanation available (where cognitive explanations are viewed as more complex, and non-cognitive explanations are thought to be simpler), that simpler explanation is preferable to an explanation citing cognitive processes. Thus, the deflationary strategy: look for non-cognitive (or less cognitive) explanations of Santino’s behavior and argue that because such explanations exist, the behavior does not merit a fully cognitive explanation.8 We should not conclude on the basis of this evidence alone that Santino is planning for the future, nor that he is taking steps to ensure that his future self is able to achieve his goals, perhaps at the expense of his present goals.
This argument apparently relies on Morgan’s Canon, the principle that we should avoid appealing to “higher” processes when we can explain behavior as a consequence of “lower” processes alone: “In no case is an animal activity to be interpreted in terms of higher psychological processes if it can be fairly interpreted in terms of processes which stand lower in the scale of psychological evolution and development.”9 However, Morgan’s Canon has come under attack recently.10 It relies on several questionable assumptions, for example, that there is a higher and a lower process when applied to psychology, and that evolution somehow makes a distinction between higher and lower.
Perhaps the deflationists need not rely on Morgan’s Canon. Povinelli and Vonk (2003) argue that when it comes to attribution of a theory of mind (that is, thoughts that other individuals have minds), chimpanzee behavior can always be explained more simply. That is, when a chimpanzee’s behavior is putatively caused by its representation of another animal’s mental state, it must also be mediated by an abstract representation of the other animal’s behavior. But, Povinelli and Vonk argue, the behavior we want to explain can be explained simply by reference to that mediating abstract representation of behavior. No further representation of the other animal’s mental state is necessary. Rather than depending on a spurious distinction between higher and lower, Povinelli and Vonk are making an appeal to parsimony: the mindreading hypothesis requires chimpanzees to have representations of animals’ mental states and of their behavior, while the behavior-reading hypothesis requires representations only of animals’ behavior. Unless the evidence specifically requires both kinds of representation, we should prefer the explanation that postulates fewer intervening representations.11
Can this reasoning be applied to the debate about Santino? The key to Povinelli and Vonk’s argument is that they find two structures that must be present for the mindreading explanation of the chimpanzee’s behavior, but only one of those structures must be present for the behavior-reading explanation of the same behavior. The deflationists can argue that in order for the proper explanation for Santino’s behavior to be that he is planning ahead, then that explanation must assert that he is both responding to cues in his current environment and engaging in mental states that involve projecting himself into the future. On the other hand, if the proper explanation for Santino’s behavior does not involve advance planning, then the explanation need assert only that he is responding to cues in his current environment. Unless there is evidence that explaining Santino’s behavior specifically requires both projecting himself into the future and responding to cues in the environment, we should prefer the deflationary view. This is the argument pattern that Roberts and Feeney (2010) seem to rely on when they argue that there is an explanation of Santino’s behavior that relies only on contextual clues and not on any planning for the future.12
The problem with this argument pattern is that just as it could be easily generalized from issues having to do with theory of mind to issues having to do with advance planning, so too can it be easily generalized to any situation in which a cognitive explanation vies with a non-cognitive explanation. In each of these situations, the cognitive hypothesis will compete with a non-cognitive hypothesis that is largely the same, except that it does not postulate the existence of some extra mechanism. The problem is that every instance of behavior might have a non-cognitive explanation. That includes even paradigms of cognitive behavior such as speech; there are non-cognitive speech acts, e.g., parroting expressions without understanding them. If we follow Povinelli and Vonk’s logic, no behavior requires us to adopt the cognitive explanation. Despite this, it seems apparent to us that some behavior is appropriately explained cognitively. Human beings engage in mindreading. Human beings plan for the future. Human beings engage in meaningful speech acts. But, of course, our mindreading behavior can also be explained by simply postulating behavior-reading mechanisms. Since any act of mindreading involves behavior-reading as well, doesn’t Povinelli and Vonk’s argument entail that we should never attribute mindreading to human beings? This suggests that something has gone wrong here; Povinelli and Vonk’s conclusion is too strong.
I think Povinelli and Vonk go wrong by focusing on individual behaviors in isolation. (Morgan’s Canon, at least as it is typically interpreted, makes this mistake as well.) When we look at patterns of behavior, and at different behaviors across different situations, we can observe evidence that supports (or fails to support) cognitive explanations. If Santino fails to engage in behavior suggestive of advance planning in other situations in which advance planning would be beneficial, then we might be justly hesitant to attribute advance planning to him in this situation. But if he engages in behaviors suggestive of advance planning in other situations, or in behaviors suggesting that he has a sense of a self who exists in the future, then the advance planning explanation of these behaviors looks much more plausible. We see the same in human behavior: you may be able to explain my reaction to a particular political candidate’s name as a mere associative reaction, but when you observe other behaviors related to that candidate, each of which would require a different association, you may prefer to hypothesize that my behavior is better explained by citing my beliefs about that candidate.
The preceding argument is, at best, a strategy for a negative argument against the deflationary explanation of Santino’s behavior. Is there positive reason to prefer the inflationary explanation? Several details of Santino’s behavior are worth noting in this regard. First, while Santino initially hid his ammunition behind naturally occurring bodies (such as logs) that prevented the zoo visitors from seeing the piled rocks, he eventually manufactured hiding places from hay he found elsewhere in his enclosure, coming to prefer the manufactured hiding places.13
Second, Santino initially stockpiled stones that he found in the waterbed surrounding his enclosure. In addition to those stones, he would occasionally find bits of concrete that he also used as ammunition. The concrete bits were present as a consequence of cycles of freezing and thawing of the water that had seeped into the concrete structures of the zoo. Santino learned that when he knocked on the concrete rocks in his enclosure, he could determine by the sound which areas were damaged. He would then strike these areas more forcefully to detach more concrete to add to his stockpile.14 Third, when Santino approached the zoo visitors aggressively, holding his ammunition and ready to throw it, the zoo guide would direct the visitors away from Santino, so that they would be out of his range. Consequently, Santino learned to approach without engaging in dominance behavior, with his ammunition hidden. In this way, he was more likely to be successful in throwing the stones at the visitors.15 I have argued elsewhere that the ability to learn relative to a goal indicates that an organism’s behavior (relative to that goal) is best explained by appeal to that organism’s psychological states, such as beliefs and desires.16 The reasoning supporting that conclusion is that an organism is only able to abandon an ineffective means of reaching a goal and adopt a different, perhaps more effective, means of reaching the goal if it has distinct representations of the means to the goal and the goal itself. Such distinct representations are necessitated by the structure of belief-desire explanations of behavior. Furthermore, if the organism is able to continue to learn relative to that goal, then there is evidence that the organism is not modifying its behavior according to what we might call “programming,” prestructured by evolution. This suggests that the organism is actually sensitive to its goals and to the various means for reaching those goals that are at its disposal. Santino nicely fits this model. First he learns to collect and throw rocks, then he learns to collect the rocks ahead of time so that he is prepared to throw the rocks, then he learns to harvest concrete ammunition, and then, finally, he learns to better hide his ammunition. He continues to modify his behavior, all while maintaining the same goal, to throw rocks at the zoo visitors. This strongly suggests that Santino has the goal of throwing rocks at the zoo visitors and has representations of his environment that guide him in achieving that goal.
Does this argument indicate that Santino’s stockpiling behavior is mediated by plans? Is he stockpiling ammunition because he recognizes that his future self will want to throw rocks and because he wants to make that future self happy? If so, Santino would have the ability to reflect on his current mental states, and his behavior would be guided by a self-concept. Do Santino’s mental states include (1) representations of his mental states, and (2) beliefs (or other representations) about himself? Before addressing this question, let’s pause and review how we got here. Often nonhuman animals exhibit behaviors that suggest that they have thoughts and feelings similar to those had by human beings. Examples of this occur in the consolation behavior of the prairie voles and in the planning behavior of Santino. There are two tempting responses to these behaviors. According to the deflationary response, careful scientific examination of the behavior reveals that it can be explained without recourse to concepts such as representation, empathy, goals, and distress; and that when the behavior is described as involving “consolation” and “planning,” those descriptions should not be taken literally. However, according to the inflationary response, there is more to some nonhuman animal behaviors than a superficial similarity to human behavior. The prairie voles have an internal chemical response that is remarkably similar to the internal chemical response found in human beings when they demonstrate empathy. I have argued that Santino’s behavior exhibits the plasticity that is concomitant with distinct representations of one’s goals and one’s means to achieve those goals. Thus, there is good reason to conclude that Santino does in fact have thoughts about what he wants to achieve and about how to achieve those goals. Does this mean that Santino is genuinely planning? Planning of this type – sacrificing in the present in order that one’s future self is happier – requires more than distinct representations of one’s goals and means to those goals. This sort of planning requires that one be able to represent one’s own mental states, that one have a concept of one’s self, inasmuch as one has a special relationship to the organism that benefits in the future as a consequence of one’s present sacrifices. Thus, the question I am asking here: we have evidence that Santino has thoughts about how to best satisfy his goal of throwing rocks at the zoo’s visitors, does he also have thoughts about how he will feel when he has satisfied this goal?
The data relevant to answering this question are equivocal at best. Several experiments seem to show that some primates (including chimpanzees) are capable of monitoring their internal mental states. In one paradigm, the subject is trained to respond to a forced-choice situation in which, if the subject makes the wrong choice, the reward is significantly worse than the reward for making the correct choice. For example, Smith et al. (1997) trained rhesus monkeys to respond differentially to a pixel display. When the display was dense (more dots), the monkeys learned to make one response; when it was sparse (fewer dots), the monkeys made a different response. Correct responses led to a reward, while incorrect responses led to a time-out, during which the monkey was unable to take the test and thus unable to secure the reward. The subjects became quite adept, except for displays that were near the border of sparse and dense. Once subjects learn the differentiation task, they are given an opt-out alternative. Choosing not to take the test guarantees a better reward than taking the test and failing, but not as good a reward as taking the test and passing. Thus, if the subjects were capable of monitoring their confidence level, we would expect them to opt out of taking the test when they are not confident that they will pass the test. In several experiments, this is exactly what happens.17 Variations of this paradigm that involve betting (subjects bet more if they are more confident) or testing one’s knowledge (subjects check the stimulus again if they are less confident) seem to show the same result: many primates are able to monitor their own confidence levels.18 Of course, Santino’s planning behavior does not involve monitoring his confidence levels, but these results do suggest that some animals, including chimpanzees, are able to think about their own states of mind and act on those thoughts.
Before we are swayed by these impressive results and we conclude that Santino does have thoughts about his state of mind, it is worth noting that these experiments have been subject to incisive critiques.19 Both Peter Carruthers (2008) and Josef Perner (2012) argue that there are ways to interpret these experiments that do not require attributing metacognitive abilities to the subjects. Perner, for example, points out that the density test can be passed if the subjects merely create a third category of response for moderately dense displays. Rather than opting out, subjects are indicating that the display fits in this third category. Carruthers adopts a strategy that relies on Morgan’s Canon. First, for each supposed metacognitive result, there is an explanation that postulates only the mediating factor of belief strength. If the subject has a weak belief that the pattern is strong and a strong belief that pressing the opt-out button will lead to a reward, the strong belief can overwhelm the weak belief; it can trigger action without the animal being self-aware. Since self-awareness is not necessary to explain the animal’s behavior, Morgan’s Canon suggests that we not appeal to self-awareness in our explanations of the results of the self-monitoring experiments.
The results of false-belief tasks also cast doubt on the conclusion that Santino has thoughts about his own states of mind. In these tasks, one of the standards for measuring metacognitive ability, subjects are asked to differentiate between their own knowledge of a situation and another’s false belief about the same situation. Chimpanzees typically fail false-belief tasks.20 For example, in a competition task in which a subdominant chimpanzee is asked to predict where a dominant will look for food, the subdominant fails to treat the case in which the dominant is ignorant of the location of the food as different than the case in which the dominant is wrong about the location of the food. (The consequence of failing to make this distinction is that the subdominant chimpanzees reap fewer rewards than they would otherwise.) This and other failures strongly suggest that chimpanzees do not have an understanding of false beliefs.
As I remarked above, the data are not univocal when it comes to chimpanzee self-concepts. Some data, such as the opt-out experiments, leave open the possibility that chimpanzees have internal representations of their own mental states. Other data, such as provided by the false-belief tasks, suggest that they do not have internal representations of the states of mind of other chimpanzees. How might we resolve this apparent impasse? I think two lines of argument suggest that chimpanzees are unable to represent their own states of mind. The first has to do with what we might call a meta-analysis of the experimental results. The second has to do with a careful consideration of what we are looking for when we look at these results.
First, recall that Carruthers explicitly appeals to Morgan’s Canon in his deflationary arguments that opt-out tasks do not demonstrate that the animal subjects are relying on an introspective sense of self. His reasoning is that metacognition is more complex than first-order cognitive processes. Since we can explain the opt-out behavior by appeals only to first-order cognitive processes, we ought to prefer those explanations. I suggest above that there are good reasons to be suspicious of appeals to Morgan’s Canon. However, I also suggest that when we broaden our scope of inquiry to look for patterns that appear in different domains, these patterns suggest what sort of cognitive processes can be legitimately appealed to in order to explain an organism’s behavior. In this case, the fact that we can explain chimpanzee behavior on opt-out tasks without appealing to higher-order metacognitive states, coupled with poor chimpanzee performance on false-belief tasks, strongly suggests that chimpanzees do not have an understanding of belief states. These considerations indicate that we should prefer an explanation of Santino’s behavior that does not rely on his beliefs about his mental states.
Second, we are looking for evidence that Santino is able to represent his own mental states. However, it is not enough that he represent his own mental states; the task facing Santino, if he is indeed legitimately planning for the future, is to represent future mental states, states that will differ from his present mental states, and to represent those future mental states as being of special interest to him (as opposed to the mental states had by some other chimpanzee or other organism). The first of these tasks is comparable to the task facing chimpanzees in the false-belief task: represent (and act on) mental states had by another organism, which are different than one’s own. If chimpanzees are unable to do this, then Santino is likely to be unable to represent his own future mental states when they differ from his current mental states. Whether his current mental states are available to him or not, his future thoughts, when they differ from his current thoughts, are as opaque to him as are those thoughts of his conspecifics that differ from his own.
Nor is it clear that chimpanzees are capable of generating representations of their selves as being of special interest. The evidence relevant to chimpanzee self-representation comes from mirror self-recognition tests. In these tests, individuals first learn about mirrors (by interacting with them) and are later marked (with, for example, dye on the forehead) in a way that cannot be observed except by looking in the mirror. If the animal examines its mark in the mirror significantly more than it had examined that previously unmarked region, then experimenters conclude that it recognizes itself in the mirror. Such recognition involves a self-concept, according to the standard reasoning,21 because in order to examine the mark, the animal must have both a stable self-image that is different than the image the animal now sees, and the animal must identify the image it now sees with itself. Without such identification, there would be no reason for the animal to examine its own forehead upon seeing the marked forehead in the mirror. No representation of the self; no mirror self-recognition. Mirror self-recognition thus implies a representation of the self. Since chimpanzees recognize themselves in the mirror, they must have a representation of themselves, right? Unfortunately these data are equivocal as well.22 While the consensus seems to be that chimpanzees do recognize themselves in the mirror, and that such recognition is mediated by a self-concept, the data are not as clear as this consensus would suggest. A majority of chimpanzees tested actually fail the mark test for self-recognition; according to a review by Thornton and Lukas (2012), only 40% of the chimpanzees tested pass the mark test. Moreover, there have been few studies with the goal of illuminating the nature of this self-concept. Provocatively enough, one study (de Veer et al. 2002), the only study I know of which returned years later to reexamine the performance of chimpanzees who had passed the test, found that of nine chimpanzees that had passed the mark test in an earlier study, only six chimpanzees passed in the later study. It seems that whatever self-concept those chimpanzees relied on to pass the test was lost in the intervening time. At the very least, this suggests that any self-concept possessed by chimpanzees is not like the human self-concept; when chimpanzees and humans recognize themselves in the mirror, they seem to be doing different things.
We might think that Santino’s behavior shows us that he has something of an understanding of his own current and future mental states, and that he is choosing to help satisfy the goals of his future self. But the studies I cite here point in a different direction. It seems unlikely that Santino is reflecting on his current mental states. These studies suggest that the sort of self-concept necessary for an understanding of one’s own mental states is likely to be missing in the chimpanzee’s mental economy. What, then, of the prairie voles, who not only console their familiars, but who also show many of the same physiological concomitants of consolation behavior as do human beings? Aren’t the physiological data evidence that there is a strong similarity between the causes (and effects) of human and prairie vole consolation behavior? Yes, but our inclination to attribute to prairie voles an empathetic fellow feeling, a recognition that the observer is reaching out to the demonstrator as if to say “I feel your pain,” may well be a symptom of something that is added on to our oxytocin-mediated empathetic response: human beings are capable of both empathetic behavior and the awareness that we are engaging in empathetic behavior (and that such behavior causes in us certain experiences). It may well be that a difference between us and Santino, and between us and the prairie voles, does not lie in the behavior, but in our awareness of the behavior.23
1 Burkett et al. (2016).
2 Bhanoo (2016).
3 See also Andrews (2015).
4 Osvath and Karvonen (2012).
5 My answer to these questions will be somewhere between these poles. I think Santino’s behavior is evidence of human-like mental characteristics in some ways, but not in other ways. For this reason, I focus more on inflationary versus deflationary explanations than on the issues of anthropomorphic and anthropectic explanations.
6 There is another position one might adopt: Santino’s behavior is unusual and not good fodder for induction to conclusions about the cognitive skills of typical, wild chimpanzees because it is an artifact of the training Santino may have received as a consequence of his prior interactions with his human caretakers. I am sympathetic to this response (see Saidel 2016), but I will not explore these issues here.
7 Roberts and Feeney’s main focus is that Santino’s planning, inasmuch as he is engaged in planning, involves semantic memory rather than episodic memory. That focus, while important, is not strictly relevant to the issues I am raising here.
8 What do “cognitive,” “non-cognitive,” and “fully cognitive” mean in this context? I’ll admit that these are fuzzy terms, and, as will become apparent below, we should be suspicious of them in the context of Morgan’s Canon. However, the intuitive idea is clear enough: don’t give an explanation citing complex mental states when an explanation that does not cite complex mental states is available. My ultimate goal has to do with explanations that cite a concept of the self. So, a clearer version of a principle like Morgan’s Canon is available to me: don’t give explanations that rely on a concept of the self when explanations that do not rely on a concept of the self are available.
9 Morgan (1894).
10 See, for example, Sober (1998) and Fitzpatrick (2008), and see the essays in this volume by Buckner (2018, Chapter 39), Fitzpatrick (2018, Chapter 42), and Dacey (2018, Chapter 40).
11 On mindreading and Povinelli and Vonk, see the essays in this volume by Halina (2018, Chapter 22) and Lurz (2018, Chapter 21).
12 Again, the issues raised in the literature in which Santino figures have to do with different memory systems. Osvath believes that the best explanations of Santino’s behavior appeal to episodic memory, while Suddendorf believes that the behavior may be explicable without appeal to episodic memory. Since both Osvath and Suddendorf agree on the other structures present in Santino’s cognitive economy, if Suddendorf is correct, appeals to episodic memory would be unwarranted additions, doing no explanatory work that isn’t already being done. Suddendorf can make this argument without appealing to Morgan’s Canon.
13 Osvath and Karvonen (2012).
14 Osvath (2009).
15 Osvath and Karvonen (2012).
16 See, for example, Saidel (1998, 2009). In Saidel (1998), I argue that the only way an organism can exhibit the sort of goal-oriented plasticity Santino exhibits is if it has distinct representations of its goals and means to achieve those goals. I understand “beliefs and desires” relatively thinly here; a better description might be “representations of the environment” and “motivational or goal states.”
17 For an overview of these experiments, see Terrace and Son (2009). There are many versions of these experiments. One might start with Hampton (2001) (rhesus monkeys) and Rosati and Hare (2011) (chimpanzees and bonobos). Even rats have been tested: Foote and Crystal (2007). For discussion, see Griffin (2001).
18 For betting, see Kornell, Son, and Terrace (2007); for knowledge testing, see Call (2010).
19 Proust (2018, Chapter 13 in this volume) discusses these experiments in much greater depth. She concludes that they provide evidential support for a kind of nonhuman animal metacognition, one that is based in evaluative, non-cognitive, sensing of affordances. This fits with the conclusion I draw here, that these experiments do not support claims that these animals have thoughts about their mental states. Proust’s conclusion is also in line with my own; she argues that metacognitive nonhuman animals “probably don’t think about themselves the way that we do.”
20 See Call and Tomasello (1999); Kaminski, Call and Tomasello (2008); and Call and Tomasello (2008).
21 For the mark test for mirror self-recognition, see Gallup (1970). For arguments that a self-concept is involved in recognizing oneself in the mirror, see Boccia (1994) and Mitchell (1995).
22 See Saidel (2016) for a more in-depth discussion of these and other issues raised by the mirror self-recognition tests.
23 I am indebted to Mark Engelbert, Dan Hicks, Aleta Quinn, and, especially, Kristin Andrews for comments on a previous draft of this essay.
I recommend starting with the chapters from this volume cited in the text. Sober (1998) and Fitzpatrick (2008) are excellent places to start for critiques of Morgan’s Canon. Thornton and Lukas (2012) provide an indispensible overview of not just mirror self-recognition experiments, but also raise questions about the sorts of conclusions we are tempted to draw from limited experimentation. Also recommended reading are K. Andrews, Do Apes Read Minds? Toward a New Folk Psychology (Cambridge, MA: MIT Press 2012) and the papers collected in R. Lurz (ed.), Animal Minds (Cambridge: Cambridge University Press 2009).
Andrews, K. (2015) The Animal Mind New York: Routledge.
Balter, M. (2012) “‘Killjoys’ challenge claims of clever animals” Science 335(6072) pp. 1036–8.
Bhanoo, S. N. (2016) “A furry shoulder to cry on” New York Times January 26, 2016, p. D2.
Boccia, M. (1994) “Mirror behavior in macaques” in S. T. Parker, R. Mitchell, and M. L. Boccia (eds.) Self-Awareness in Animals and Humans: Developmental Perspectives. Cambridge: Cambridge University Press.
Buckner, C. (2018) “Understanding associative and cognitive explanations in comparative psychology” in K. Andrews and J. Beck (eds.) Routledge Handbook of Philosophy of Animal Minds. London and New York: Routledge.
Burkett, J. P., Andari, E., Johnson, Z. V., Curry, D. C., de Waal, F. B. M., and Young, L. J. (2016) “Oxytocin-dependent consolation behavior in rodents” Science 351:6271 pp. 375–8.
Call, J. (2010) “Do apes know they could be wrong?” Animal Cognition 13(5) pp. 689–700.
Call, J., and Tomasello, M. (1999) “A nonverbal false belief task: The performance of children and great apes” Child Development 70 pp. 381–95.
——— (2008) “Does the chimpanzee have a theory of mind? 30 years later” Trends in Cognitive Sciences 12(5) pp. 187–92.
Carruthers, P. (2008) “Meta-cognition in animals: A skeptical look” Mind & Language 23(1) pp. 58–89.
Dacey, M. (2018) “A new view of association and associative models” in K. Andrews and J. Beck (eds.) Routledge Handbook of Philosophy of Animal Minds. London and New York: Routledge.
de Veer, M., Gallup, G., Theall, L., van den Bos, R., Povinelli, D. (2002) “An 8-year longitudinal study of mirror self-recognition in chimpanzees (Pan troglodytes)” Neuropsychologia 1493 pp. 1–6.
Fitzpatrick, S. (2008) “Doing away with Morgan’s Canon” Mind & Language 23(2) pp. 224–46.
——— (2018) “Against Morgan’s Canon” in K. Andrews and J. Beck (eds.) Routledge Handbook of Philosophy of Animal Minds. London and New York: Routledge.
Foote, A. L., and Crystal, J. D. (2007) “Metacognition in the rat” Current Biology 17(6) pp. 551–5.
Gallup, G. (1970) “Chimpanzees: Self-recognition” Science, New Series 167(3914) pp. 86–87.
Griffin, D. R. (2001) “Animals know more than we used to think” Proceedings of the National Academy of Sciences 98(9) pp. 4833–34.
Halina, M. (2018) “What apes know about seeing” in K. Andrews and J. Beck (eds.) Routledge Handbook of Philosophy of Animal Minds. London and New York: Routledge.
Hampton, R. R. (2001) “Rhesus monkeys know when they remember” Proceedings of the National Academy of Sciences 98(9) pp. 5359–62.
Kaminski, J., Call, J., and Tomasello, M. (2008) “Chimpanzees know what others know, but not what they believe” Cognition 109(2) pp. 224–34.
Kornell, N., Son, L. K., and Terrace, H. S. (2007) “Transfer of metacognitive skills and hint seeking in monkeys” Psychological Science 18 pp. 64–71.
Lurz, R. (2018) “Animal mindreading: The problem and how it can be solved” in K. Andrews and J. Beck (eds.) Routledge Handbook of Philosophy of Animal Minds. London and New York: Routledge.
Mitchell, R. (1995) “Evidence of dolphin self-recognition and the difficulties of interpretation” Consciousness and Cognition 4 pp. 229–34.
Morgan, C. L. (1894) An Introduction to Comparative Psychology London: Walter Scott.
Osvath, M. (2009) “Spontaneous planning for future stone throwing by a male chimpanzee” Current Biology 19(5) pp. R190–1.
Osvath, K., and Karvonen, E. (2012) “Spontaneous innovation for future deception in a male chimpanzee” PLoS ONE 7(5) p. e367782.
Perner, J. (2012) “Minimeta: In search of minimal criteria for metacognition” in M. J. Beran, J. L. Brandl, J. Perner, and J. Proust (eds.) Foundations of Metacognition Oxford: Oxford University Press.
Povinelli, D. J., and Vonk, J. (2003) “Chimpanzee minds: Suspiciously human?” Trends in Cognitive Sciences 7(4) pp. 157–60.
Proust, J. (2018) “Metacognition and self-awareness in nonhumans” in K. Andrews and J. Beck (eds.) Routledge Handbook of Philosophy of Animal Minds. London and New York: Routledge.
Roberts, W. A., and Feeney, M. C. (2010) “Temporal sequencing is essential to future planning: Response to Osvath, Raby and Clayton” Trends in Cognitive Sciences 14(2) pp. 52–53.
Rosati, A. G., and Hare, B. (2011) “Chimpanzees and bonobos distinguish between risk and ambiguity” Biology Letters, 7 pp. 15–18.
Saidel, E. (1998) “Beliefs, desires, and the ability to learn” American Philosophical Quarterly January, 1998 pp. 21–37.
——— (2009) “Animal thoughts,” in R. Lurz (ed.) Animal Minds Cambridge: Cambridge University Press.
——— (2016) “Through the looking glass, and what we (don’t) find there” Biology and Philosophy 31(3) pp. 335–52.
Shettleworth, S. J. (2010) “Clever animals and killjoy explanations in comparative psychology” Trends in Cognitive Sciences 14(11) pp. 477–81.
Smith, J. D., Shields, W. E., Schull, J., and Washburn, D. A. (1997) “The uncertain response in humans and animals” Cognition 62(1) pp. 75–97.
Sober, E. (1998) “Morgan’s canon” in D. D. Cummins and C. Allen (eds.) The Evolution of Mind New York: Oxford University Press.
Suddendorf, T., and Corballis, M. C. (2010) “Behavioural evidence for mental time travel in nonhuman animals” Behavioural Brain Research 215 pp. 292–98.
Terrace, H. S., and Son, L. K. (2009) “Comparative metacognition” Current Opinion in Neurobiology 19 pp. 67–74.
Thornton, A., and Lukas, D. (2012) “Individual variation in cognitive performance: Developmental and evolutionary perspectives” Philosophical Transactions of the Royal Society B 367 pp. 2773–83.