Concept of mind in non-human primates

Abstract

Research concerning animal cognition explores the abilities and capacity of animals to perceive, think and conceive. As an extension of this, researchers have tried to ascertain the concept of animal minds. The field has been a matter of great debate as it has brought into question the uniqueness of the human mind. This dissertation will review the various areas of research that have contributed to our understanding of animal minds, with a specific focus on non-human primates. The term ‘theory of mind’ was originally proposed by Premack and Woodruff in 1978. The ability entails a recognition and understanding of another's mental states. Recently, this term has included the cognition of seeing. Throughout this article, the important distinction between theory of mind capabilities and complex behavioural analysis is emphasized. It is important to consider how various primates represent entities in their environment, including their own image. In particular reference to this latter point, self-recognition could act as a first step towards understanding others. With this ability, other individuals may then be understood and manipulated through deception, imitation and teaching. In addition to deception, pretend play and external representation are proposed as another dimension of understanding false representations. Decisions about the evolutionary point at which theory of mind may have developed will depend on interpretations of the evidence for these abilities in non-human primates and whether indeed theory of mind is underlying them. Since the conception of the term ‘theory of mind’, the issue may have evolved beyond whether or not there is theory of mind in non-human primates to a more sophisticated appreciation that the concept of mind has many facets and some of these may exist in non-human primates while others may not.

theory of mind, cognition, chimpanzee

Introduction

Many advanced behaviours have been observed in non-human primates including learning, tool use and forms of communication. This review will discuss how the existence of certain behaviours may allude to understanding of another's mind, with an aim to evaluate the possibility that non-human primates have a theory of mind.

First, this article shall clarify what is meant by ‘theory of mind’, although as will become evident that there are several different interpretations. Premack and Woodruff¹ first introduced the term in their seminal paper on the subject as an inferential capacity concerning mental constructs such as intentions and beliefs. The term ‘theory’ in particular is appropriate as such mental constructs are not directly observable and the ability enables prediction of another's behaviour.¹ More recently, the term ‘mental states’ has incorporated the cognition of ‘seeing’. This article will assume the definition, put forward by Penn and Povinelli² in 2007, of theory of mind as any process, which informs the subject about constructs of another's mind, allowing for behavioural prediction.

Throughout this review, it will be considered how this mental attribution/understanding differs from complex behavioural reasoning, although some researchers² have suggested that the two are closely related. Indeed, the difficulty has arisen to determine if and when experiments clearly distinguish the two.

It has been suggested that social environment may provide both pressure and context for the evolution of a higher intelligence.³ It is, therefore, not surprising that it is in the great apes who present complex social systems that we find the most evidence for theory of mind. It is these species, more specifically chimpanzees, gorillas and orangutans, that this article will focus on.

Representation

Perner⁴ broke down theory of mind into three components (Fig. 1): primary representation (understanding reality), secondary representation (using concepts that can be detached from reality, allowing the formation of hypothetical simulations) and meta-representations (allowing for the comprehension of misrepresentations or false beliefs).

Figure 1.

Schematic representation of Perner's⁴ theory of mind components with an example to explain the process through the components.

At the level of primary representation, individuals are unlikely to have a concept of how the world is represented to themselves or others; the world is taken at face value. At the stage of secondary representation, it is possible that individuals are starting to have a theory of their own minds. This stage is thought to enable, or at least parallel, the envisaging of future scenarios/means-end reasoning, recognition of one's self in a mirror and understanding external representation (e.g. pictures).⁴ It is meta-representation (encompassing the concept of another's mind as distinct from ones own) that is seen to isolate human capacity of theory of mind relative to other primates. As will be discussed, debate lies in which behaviours demonstrate this meta-representation. An analysis of how primates represent their worlds may help to establish the extent to which they have a theory of mind.

Back to the future

Secondary representation can be assessed using the extent to which primates can envisage scenarios outside the current context. Kohler⁵ stated ‘the time in which a chimpanzee lives is limited in past and future’. In actuality, the time frame in which primates perceive tools as relevant is unclear. Mulcahy and Call,⁶ having taught bonobos and orangutans to use tools to obtain rewards from a particular apparatus, then subjected the primates to various trials where the appropriate tool had to be selected for later use. Suitable tools were chosen significantly more than unsuitable ones even when the period between tool selection and reward was extended to overnight (for a couple of selected, best-performing subjects), and when the reward apparatus was not visible until after tool selection.

To determine that this application of objects was due to more than just learned associations, a group of monkeys were also presented with a scenario whereby reward was given to the subjects when they returned with the correct tool without having to use it. The authors suggested that poorer performance observed in this task eliminates the possibility that subjects were simply linking tool with reward. However, the authors also suggested that the performance could be due to the fact that the group participating in this control were naïve, having not participated in the other conditions. Although success at these experiments may indicate a concept of the future scenarios, it does not necessarily suggest an understanding of future mental states. For instance, in Mulcahy and Call's experiments detailed above, the motivation of the subjects (to desire the reward) is likely to have been constant. Therefore, as emphasized by other authors, there was no prediction of future need, only satisfaction of current need/state.⁷

As well as hypothetical scenarios based in a different time frame, external and false representation in the present may also indicate secondary representation. These are considered as separate from false mental representation, which will be discussed later. Several examples of non-human primate understanding of these concepts have been documented. Suddendorf and Whiten⁸ identified studies that have shown primates responding to television⁹ and magazines,¹⁰ by signing appropriately in recognition of the images displayed, including naming characters and indicating likes/dislikes. However, chimpanzee Viki made gestures towards photographs as if they were real objects,¹¹ indicating that these primates may not be recognizing the symbolic nature of the images.

This difference between exact representation (a copy) and symbolic representation is a crucial one. Children with autism who lack theory of mind can draw but only to reproduce images that they have had previous experience with. Indeed, when asked to produce fictitious entities, they were significantly worse in doing so than children without the disorder.¹² From an evolutionary perspective, emergence of pictorial representations of fictitious entities, evident from 30 000 years ago, is thought to demonstrate recognition of mental concepts and the presence of theory of mind at this time.¹³

More convincing evidence may come from the subjects' own use of objects as false representations such as through pretend play. Patterson and Linde¹⁴ (as cited by Whiten and Byrne¹⁵) recorded signing that reportedly indicated that the chimp Koko was pretending to be an elephant using a rubber tube as a trunk. Although the signing dialogue is open to interpretation, several other accounts of pretence have occurred in captivity and the wild.⁸ Although conclusions regarding subjects understanding of pretend play can only be drawn from rigorous experimental manipulations, if these provide evidence consistent with the observational records, non-human primates may possess an ability which precedes understanding other mental states.¹⁶

Another important representation is that of the self.

‘Mirror, mirror on the wall… ’¹⁷

When presented with a mirror, many species react with social behaviours, as if they were seeing another individual.¹⁸ However, in contrast, the great apes have been shown to display an understanding of self, originally observed by Gallup¹⁸ as subjects' self-directed behaviours such as grooming with particular reference to areas of their body they cannot see. Gallup produced a testable method for this assumption, the mark test whereby an anaesthetized subject is marked in a place on the head not directly observable to the subject and then its behaviours (when conscious) are observed. It was found that the number of times the mark was touched with the presence of a mirror was significantly higher, compared with without. Since its conception, all great apes have shown success at this task.⁸ Criticisms, including the saliency of the mark and residual effects of anaesthesia, have been discredited¹⁹ and the results of the mark test are accepted quite widely as evidence that the great apes understand, at least their physical, self. This is also supported by reports that chimpanzees can recognize photographs of themselves.²⁰ Gallup et al.²¹ state that an understanding of self may be requisite for other mentalistic understandings. Indeed, in human development, empathy and perspective taking skills become evident around the same time as self-recognition.²²

The neural basis for such distinction may be ‘mirror neurons’, a collection of neurons that showed similar activation patterns when a monkey was either observing or carrying out a particular arm movement.²³ Since this discovery, possible equivalent ‘mirror’ regions have been proposed in human brain.²⁴ Platek et al.²⁵ found, using functional magnetic resonance imaging, correlated activity in the right hemisphere of subjects during presentation of their own faces compared with faces of familiar others. Although this study was conducted on a relatively small number of participants and recognition of another (in terms of physical attributes) does not equate to an appreciation of another's mental states, it does reveal a potential neural basis for the self-other distinction. In humans²⁶ and monkeys,²⁷ it has been shown that these neurons can not only respond to surface behaviour but also respond differently depending on the intent behind the perceived action. This mapping of experiences of other organisms on to our own could be the basis for shared experiencing related to an understanding of the mind. In relation to this, human individuals on the autism spectrum show deficits in these proposed mirror regions correlated with their poor theory of mind capacities.²⁸ These mirror components may also allow for neural replication of subjective experiences,²⁴ which could aid skills such as empathy, teaching and imitation.

Me, myself and imitation

Imitation can elucidate primates' understanding of their own behaviour relative to another.²⁹ To ensure that subjects are forming a representation of the behaviour to be imitated, not just presenting an immediate reaction, Zentall argues that deferred imitation must be introduced, with a delay in the period between observation and performance.²⁹ This is yet to be demonstrated in primates. Deducing that chimpanzees understand the nature of imitation is evidenced further by virtue of the fact that they can understand when a human is imitating them³⁰ but Suddendorf and Whiten⁸ suggest that this could be systematically tested with ‘imitation games’ such as those employed with children. Great apes have demonstrated the ability to understand commands to imitate, to the extent of employing them towards a human experimenter,³¹ indicating that their understanding of imitation is more than simply copying.

Evidence of imitation is often evident through transmission (teaching) of particular skills. Whiten et al.³² demonstrated the transmission of acts among groups by introducing individuals, who had learnt one of two particular behaviours using a stick, into separate groups. Within a week, the two groups exclusively demonstrated the specific behaviour that had been introduced, indicating conformity as well as social learning. However, this may be simply copying of an overt behaviour, which as discussed, does not require an understanding of mental state. The more complex dimension of teaching requires an active interpretation of what needs to be conveyed and how, using a concept of what another knows. Vauclair highlights Boesch's reports of this more active teaching process including provision of more appropriate tools or correcting actions.^{33, 34}

Although copying behaviours may not indicate mentalistic reasoning, examples of understanding the concept of imitation and adaptation of teaching strategies may indicate this intentional capacity.

Intent

Intentions can be thought of as representations, which can be different between individuals. An understanding of minds as separate entities between individuals allows for manipulation,³⁵ as in deception. Byrne and Whiten³⁶ use the term tactical deception to highlight purposeful acts (as opposed to passive conditions such as camouflage) which aim to achieve a goal through an act, which may instil deceit. Within this category, there are reports of savannah baboons giving out a predator alarm call when being chased by adults³⁷ (as cited by Tomasello and Call³⁸) and bonnet macaques doing the same to avoid attack/chastisement from more dominant adults.³⁹ However, in one example, a subordinate male continued to give such a call although descending from a tree (which would not be done if a predator was present), indicating a lack of understanding that all actions, not just vocalisations, convey information.³⁵

Withholding information is as much an act of deception as offering false information. For example, deception can be achieved by emotional concealment. Demonstration of fear can often be disadvantageous as it may identify an individual displaying such emotion as weak and a potential target for rivals. In two examples, this fear was concealed by manually regulating facial expression (covering a grin by repositioning the lips with the hands)⁴⁰ (as cited by Whiten and Byrne⁴¹) and through using a distracting action (chewing on a twig).³⁵

In terms of physical concealment, Kummer⁴² observed baboons hiding behind a rock to conceal the forbidden behaviour of mating that excluded the alpha male. The interpretation that this was an intentional removal from sight of the alpha male is supported by the observation that the female curbed her vocalizations and returned to the alpha male between these incidents.⁴² Sinha reported the concealment of similar prohibited behaviours in bonnet macaques.³⁵

In response to the question of whether the primate is trying to instil a belief or simply using behavioural strategies which it has learnt will initiate certain responses in certain scenarios, Sinha points out that the chances for the primates to observe and learn deception are rare.³⁵ Furthermore, if learning did occur through this route, it would be expected that the diversity of deceptive acts and their application would be limited, which was not the case. Also, acts of deception occurred across a range of spatial and temporal contexts.

However, in tactical deception, the primate may still not understand why the particular act has achieved its goal. Whiten highlights that he has been misunderstood in stating that non-human primates intentionally deceive and in fact states that he has found no evidence of meta-representation akin to the attribution of false beliefs.⁸ False belief is a belief held by an individual that does not coincide with the actual state of affairs. For example, one person might believe it to be sunny outside whereas in fact it has started raining since they last looked. Understanding that one individual can have a belief which is different from your own is the foundation of grasping false belief. Typically, this is tested with children and other primates by getting a subject to watch as an item is concealed and then moved to an alternative location. This is conducted with a third party observing the initial placement but not the relocation. The subject must then predict where the observer will search for the item. The subject must understand that the observer will seek in the place they last saw it, not in its new location—where the subject knows it to be. The observer has a false belief about the location of the item. Although deception may be considered as one form of false-belief understanding, no non-human primate has passed the traditional false-belief test (Povinelli, reported by Pennisi, 2006)³⁰ even when controlling for the cognitive demands of the task.⁴³ Although it has been suggested that such current false-belief paradigms are still inappropriate for use with non-verbal primates (Hare [no date] reported by Pennisi, 2006).³⁰

Other experimental groups have seemingly demonstrated primate success on tasks assessing similar analysis of another's mental state. Premack and Woodruff's experiments can, to a certain extent, be considered as starting the theory of mind debate with primates. They set up several problems for a chimpanzee subject to solve which had to involve an assessment of the need and intention of a human experiment (feeling cold, wanting a certain object out of reach).¹ Povinelli et al.⁴⁴ later apparently demonstrated that chimpanzees could identify a knowledgeable experimenter from and ignorant one. They placed two human experimenters in front of a chimpanzee subject. One experimenter remained in the room while an item was concealed; the other left the room and did not observe the item's placement. Being present or absent during item placement resulted in different knowledge in each experimenter. Selection of the correct experimenter by the primate to seek the reward was thought to demonstrate the chimpanzees understanding of these different mental perspectives. However both these tests to establish primate subjects understanding of intent¹ and knowledge states⁴⁴ took many trials for subjects to show success which led researchers to believe that the ability was merely a demonstration of learned associations. Povinelli and Vonk⁴⁵ have suggested that demonstration of mentalistic understanding must be shown on the first few attempts at any such paradigm (novel to the subject) in order to rule out the possibility of learnt associations. In accordance with this, in Povinelli et al.'s (1990) experiment, apes were unable to transfer the inference to an experimenter who was ignorant for a different reason to that first tested (experimenter facing away), emphasizing the specific nature of the cue that was learnt.⁴⁴

Cooperation

Within hunting groups where cooperation is assumed to be present, primates could have independent goals that happen to coincide or the group may be inadvertently formed by attraction towards a dominant male.⁴⁶ Cooperation instead requires an understanding of another's goal, recognizing it as the same as ones own. In another example of cooperation, two chimpanzees took different roles to help each other up a tree.⁴⁷ Such division of labour could only be achieved with an understanding of a shared end point.

Across theory of mind research, experimenters try to look for evidence for primate subjects not just performing an act but also understanding it. In a study by Melis et al.,⁴⁸ chimpanzees could not only perform tasks in cooperation with another but also establish and recruit the most cooperative partners for a further trial. However, the apes may simply have been employing association between partner and most reward not necessarily understanding the mental state or characteristics of the other (partnering with conspecific A led to better results, not conspecific A is more helpful).

One variation of cooperation is altruism. This again requires an assessment of another's need. Warneken and Tomasello demonstrated experimental setups that showed chimpanzees⁴⁹ providing assistance to experimenters in terms of handing them out of reach objects, showing an understanding of another's perspective and related limitations in obtaining the desired object. Although competitive paradigms (two primate subjects after the same goal/reward) have been suggested to increase the context validity of some experiments into social cognition,⁵⁰ it is also thought that the same competition may change the motivation of the subjects and therefore obscure behaviours such as helping and cooperation. By using a human experimenter and non-food items in their experiment, Warneken and Tomasello believed that competitive drive was reduced.⁴⁹ Subjects were more likely to hand an item to a peer if they themselves do not desire it and do not see the peer as a competitor, which possibly explains why this manifestation of helping has not been shown previously. However, in contrast to human infants, where help was demonstrated across four varying tasks, chimpanzees only offered help in one condition. The authors themselves concede that success in this one (out-of reach) task could have been due to the saliency of the cue for help. As such, this experiment does not demonstrate that the primates are using anything beyond physical and behavioural cues to infer when another individual needs help.

In fact, in general, paradigms apparently demonstrating an understanding of empathy/altruism do not require an understanding of another's distress or mental state but rather just an analysis of the situation, e.g. conflict. Although such social analysis is important, the absence of mentalistic analysis limits the support that empathy trials lend to a primate theory of mind.

Empathy may be more accurately perceived as perspective taking (or cognitive empathy) which may have direct implications for interpretation of the mind. This may have specific benefits for the organism's capability to learn. For instance, chimpanzees have been shown to assume new roles with no deficit in performance following the interaction with a partner in that role.⁵¹ In contrast, this ability was not shown by rhesus macaques⁵² leading to support of species differences in cognitions correlated with their capability differences in self-recognition. Unfortunately, no other studies implementing larger sample sizes have been conducted since.

Seeing is believing

Seeing has been recognized as a mental state and relates to a change in the knowledge held by another.² Both observational and experimental studies have explored the extent that primates understand what others can and cannot see.

Primates have been shown to follow the gaze of both conspecifics⁵³ and humans.⁵⁴ However, it has been suggested that this may be occurring only as a result of superficial cues. Some research has argued that understanding eye movements in particular may demonstrate, more robustly, an understanding of seeing, but results on this have been mixed and this too may be as a result of physical indications. Appreciating the consequence of seeing may be a better indicator of this comprehension, which may be shown through the ability to follow a line of sight. This has been demonstrated through chimpanzees' behaviour of following gaze around barriers.⁵⁵ Also some deceptive acts, as mentioned previously, make use of this line of sight. As another example, Hare et al.⁵⁶ observed chimps approaching a food item from behind a barrier in order to beat the human experimenter to the reward. This behaviour was evident from the first few trials and was a novel situation for the subjects. The authors stated that the subjects were understanding and subsequently intentionally manipulating the view of the experimenter. Adaptation of action towards seeing or non-seeing targets would also suggest this understanding. Hattori et al.⁵⁷ showed that comprehension of eye direction did not equate to differentiation of pointing in an experimental setting. However, such differentiation of gesturing towards different partners has been shown by two experimental groups.^{53, 58} These however are open to the initial criticism that the subject is only responding to behavioural cues.

With an attempt to resolve the controversy in this area, Hare et al.⁵⁹ conceived a new paradigm (Fig. 2); the results of which they believed demonstrated a clear understanding of mentalistic concepts, namely seeing.

Figure 2.

Diagrammatic representation of Hare et al.'s⁵⁹ experimental setup. Subordinate (S) and dominant (D) chimpanzees are positioned either side of an open area where there are two food items, one obscured from view from the dominant by a barrier (bold line). Both chimpanzees could see through to, but were separated from, the middle area by barriers. The chimpanzees are released into the centre to obtain the food.

On each occasion, the subordinate chose the reward obscured from the view of the dominant (reward behind the barrier). The experimenters declared that the subordinate appreciated what the dominant could see and chose accordingly, although they did elaborate that the use of seeing and its implications could develop through experience. They distinguished this explanation from purely associative learning models as, through learning, mentalistic representation was achieved which could be flexibly applied across situations.⁵⁹

Several control conditions were performed to rule out alternative explanations. First, to ensure that the monkey was not reacting to the orientation of the dominant in the first place, a delay was introduced to the release of the dominant subject. This also prevents intimidation from the dominant, but Povinelli and Vonk argue if the subject cannot see the dominant's behaviours how is it to establish whether the dominant can see?⁴⁵

The explanation of the subordinate preferring food next to barriers was disproved as they showed no such preference when no competitor was present.⁵⁹ Finally, the subordinate chimp did not believe that the barrier impeded the dominant from accessing the food as the subordinate stayed clear of the food when the barrier was present but transparent. However, in an experiment by Karen-D'Arcy and Povinelli,⁶⁰ it was found that one of their subjects actually showed a preference towards this transparent barrier. Also across all their subjects, a comparable preference was made towards the ‘side’ condition (Fig. 3) where dominant's sight was not blocked.

Figure 3.

Karen-D'Arcy and Povinelli's⁶⁰ setup. Subordinate (S) and dominant (D) chimpanzees have view of central food items. The barrier (bold line) does not obscure the top food item from the dominant.

As a result, Povinelli and Vonk⁴⁵ disagree that Hare et al.'s experiments in 2000 demonstrate an understanding of the concept of seeing and Penn and Povinelli² say that current paradigms only demonstrate abilities where the existence of theory of mind is superfluous to analysis of behavioural cues. Povinelli and Vonk argue that to remove any possibility that the subject could be acting in reaction to behavioural cues, a setup that is completely novel to the organism must be used.⁴⁵ Two paradigms have been subsequently suggested to better evidence theory of mind.

The first paradigm suggested by Penn and Povinelli² is an adaptation of Hare's competitive food protocol⁵⁹ (mentioned above). First, without the presence of the dominant, the chimpanzee learns to retrieve the more desirable reward. Failure, therefore, cannot be attributed to not understanding the pragmatic aspects of the tasks. Secondly, several different conditions are presented involving variations in the manipulation of the visual access of the dominant ruling out rules such as ‘go toward the food if the dominant has already orientated toward it’ (this was also achieved by Hare et al.'s⁵⁹ control condition) or ‘always retrieve the less desirable reward when there is a dominant present’. By different successes and failures, it could be determined which process was being used by chimpanzees. However, it is important to note that among these conditions are trials such as invisible displacement that chimpanzees have been shown to have difficulties with.⁴³ Also with the number of choices and computations, the task may have been too complex for the subjects.² The authors suggest that this would be reduced with a theory of mind but this is an assumption, one which has its basis in superimposing human methods of cognition which they have criticized other researchers for doing.

Another paradigm, the opaque visor experiment,² involves the subject testing two visors—one that they can see through and the other that they cannot. The visors, distinguished by a superficial attribute such as colour, are then worn by a human experimenter. The authors suggest that if subjects can infer that wearing the opaque visor instils that state of not being able to see, as experienced by themselves, fewer begging gestures would be directed at them. This is different from experiments of preferential begging towards, for instance, an individual facing them as the subjects will not have had experience with the visors. Furthermore, success on this requires the transference of first to third person experience.⁴⁵ Though this experiment has been criticized with being low in ecological validity, the authors retort that it is no less valid than previous experiments of false belief and indeed the begging gesture itself is not entrained but rather appears naturally in wild and captive chimpanzees.⁴⁵ Chimpanzee subjects reportedly showed no aversion to the visors.² Finally, in response to Andrews's⁶¹ criticism that the monkeys learn that they cannot do anything while wearing the visor and therefore will assume the same of the experimenter, Penn and Povinelli² argue that on the contrary, the chimpanzees learn that they can do many things but not see.

Infants succeed at a similar task from the age of 18 months⁶² (as cited by Meltzoff⁶³) but this is much before it was originally considered that infants had a theory of mind (emerging around 3/4 years), which again brings light to the question: what is theory of mind and are these paradigms testing it?

The purpose of these two paradigms as emphasized by Penn and Povinelli² is not that these are so far superior that successes should be valued above all other previous experimental protocols or that failure indicates a clear absence of theory of mind but that there needs to be a paradigm shift towards ones that require the explicit implementation of mentalistic reasoning (over behavioural inferences).

The ability to understand ‘seeing’ is interesting as it intersects the line between cognition and overt action. Are the experiments that apparently show an understanding of ‘seeing’ actually demonstrating an understanding of ‘looking’? In each of the experiments, primates will be implementing one of the following two explanatory strategies: Povinelli and Vonk argue that visual perception is the ideal combination of understanding observable and unobservable states⁴⁵ but do any paradigms make this distinction? Even so, seeing may not equate to believing. Indeed, autistic children who are considered to have deficits in theory of mind still succeed in tasks that involve the understanding of what another person sees.⁶⁴

• (target) A is looking at me → A will respond OR

• A is looking at me → A can see me → A will respond.

Evolutionary significance

As previously discussed, having a theory of mind may convey numerous advantages to a species and facilitate abilities such as learning and communication. Humphrey⁶⁵ was first to suggest that social cognition in particular (apart from other mental facets/abilities) requires more brain capacity, which Whiten (reported by Pennisi, 2006³⁰) later extended as an explanation for the additional brain volumes among the primates relative to other species.

By establishing which extant species have aspects of theory of mind, researchers have attempted to isolate when these may have emerged on the evolutionary timeline. With the controversy around the results of particular studies and the definitions of what actually constitutes a theory of mind, estimations vary according to attribution of theory of mind to current species; current estimates range from 35 million years to just 30 000 years ago.¹³

This review will not attempt to produce such a figure but evidently readers' opinions on when theory of mind developed would be based upon interpretation of the existence of such a construct in our present primate relatives.

Discussion

Theory of mind, as the ability to interpret mental states in others, was originally thought of as a characteristic to distinguish humans from other primates. More recently, however, the area of research has tried to examine how primates think and the abilities they possess which may indicate at least some aspects of theory of mind.

There have been many reports of non-human primates having the capacity for representation. Although the extent of these is not clear, Suddendorf and Whiten⁸ do conclude that the great apes do have abilities of secondary representation. Understanding a representation of the self (through mirror recognition) has been a focus since Gallup's conception of the mark test. All great apes show this recognition, and it has been posited that this may form the basis of the self-other distinction required for theory of mind. Mirror neurons in humans and primates have been suggested as the neural basis for this, which would most likely facilitate imitation and teaching, both of which have been demonstrated in non-human primates. Further understanding of intentions may be shown through deception. Deception is often cited as evidence for false-belief understanding, however, this is controversial. Experimental examination of understanding intent and knowledge state in others has only been shown after several trials, leading researchers to believe subjects are simply learning associations.

Empathy has also been used in an attempt to gauge another's intention, although observations do not exclude the possibility of contextual analysis. A more viable test may be through perspective taking or seeing. Although at first look Hare et al.'s⁵⁹ experiment seems to highlight significant evidence for primate's concept of seeing, Karen-D'Arcy and Povinelli's⁶⁰ failure to replicate that this sheds significant doubt on the findings. The two alternative paradigms suggested by Penn and Povinelli are not perfect: the competitive food paradigm may be confounded by the complexities of the task demands and although the visor experiment may demonstrate an understanding (or lack) of seeing, seeing may not be a viable measure of theory of mind.

The main problems in this area are definition and testing of this construct. Paradigms that are truly novel to subjects may provide a context for applying a theory of mind but as Hare (reported by Pennisi, 2006³⁰) points out, theory of mind may consist of a whole host of abilities and classification of possession of this may be on a scale rather than an all-or-none rule. Sinha describes the primate theory of mind as incomplete.³⁵ This assumes a deficit in comparison with humans but this may be inaccurate. Indeed, chimpanzees outperform human infants on areas of tool use.⁶⁶ As such, the area of research should continue in search of a multifaceted pattern of abilities, some of which are shared between humans and other primates, some which are not.

Author biography

Sarah studied Neuroscience and Psychology at Keele University, and particularly enjoyed the behavioural and developmental modules. She graduated in July 2008 and has since undertaken various care roles to gain practical experience. In September 2009, Sarah began a new role as an assistant psychologist and support worker with young people (aged 10–19 years) with severe learning disabilities, working within a charity in Buckinghamshire. She hopes to develop her skills further and would like to progress to clinical psychology.

Author notes

Supervisors: Janet Hoole and Peter Chevins, Huxley Building, Keele University, Keele, Staffordshire ST5 5BG, UK.