Can we infer mental time travel in nonhuman animals?

This essay was written in early 2018 and is primarily an investigation into animal cognition. Specifically, it looks at how we typically infer cognitive capacities in animals and the limits of such inferences. Enjoy!


1. Introduction

In 1997, Thomas Suddendorf and Michael Corballis began discussion on the capabilities of nonhuman animals (hereafter, ‘animals’) to experience mental time travel (MTT), which they describe being “the mental reconstruction of personal events from the past […] and the mental construction of possible events in the future” (Suddendorf & Corballis, 1997, p.133). Twenty years on, there remains considerable debate as to whether or not humans are unique in this ability (Clayton, 2017, p.227). The aim of this paper, therefore, shall be to assess the current landscape of MTT research in animals, and to investigate the method by which researchers infer the contents of animal minds.

In §2 of this paper I shall provide an explanation of what MTT is, particularly within the context of current MTT research on animals, and I shall discuss some of the attributing factors for its nonlinguistic inference. This will be necessary for §3, in which I shall present research on the capabilities of animal minds to experience it. MTT is, I shall argue, a functional capacity of the mind, so to make an accurate assessment of the mental abilities in other species we must consider all relevant evidence that counts toward the MTT inference. My evaluation of behavioural evidence, as the most studied aspect of MTT research in animals, shall be split into two sections; §3.1.1 & §3.1.2, to study both past and future-oriented thinking (respectively). The second and third forms of evidence that count towards the MTT inference are neurological and evolutionary, which I shall present and defend in §3.2. §4 shall conclude by discussing some of the relevant methodological considerations on MTT research and its associated terminology, in which I shall propose directions for future theory of mind research in animals.

2. What is MTT?

MTT refers to the distinctive human ability of being able to create and experience ersatz mental simulations of personal memory, and future speculation (Boyer, 2008, p.219). There are therefore two distinguishable aspects of MTT, which are the conscious recollections of past events in one’s life, and the anticipatory imaginings of possible future events. Declared necessary to MTT is the ability to mentally project the subjective self in time and space, which comprises the autonoetic (self-aware) quality of remembrances and preremembrances (Clayton, 2017, p.227). Using MTT, our mind becomes a theatre and our self the “main actor in old (reconstructed past events), scheduled (anticipated events) or new plays (imaginary future scenarios)” (Dere et al., 2017, p.2).

Without these first-person projective qualities, memories about past and speculation on future events become semantic knowledge which, in contrast to episodic cognition, takes the form of general knowledge about the world and is not subjectively bound (insofar as it does not follow a specific viewpoint in space and time). For example, the statement ‘The University of Amsterdam offers a study course in the Philosophy of Time’ is a semantic fact about the world – but this fact does not take into account my personal episodic memories of attending the course, of which include my memories of completing text readings, attending lectures, and working on assignments (such as the one currently being read). This distinction will become important in §3 as we attempt to infer working episodic cognition in animals.

MTT can arguably best be understood as a functional capacity of the mind. Although certain authors suppose the human brain’s exclusivity in its ability to undergo MTT (see §3.2 of this paper), it is generally accepted that the experience of MTT is a functional state that arises from no specific brain configuration – such as how most identify pain experience in terms of the functional role it plays within an organism. This is further supported by empirical research finding close similarities in the performance of cognitive operations belonging to the brains of neuroanatomically very different species (for example; insects, birds, and mice – see Dere et al., 2017, p.2). In contrast with behaviourist and brain-state theories of the mind which identify mental states or capacities as behavioural dispositions and neuroanatomical states (respectively), the functionalist method of inference seeks to identify mental states by such evidence, since “similarities in the behaviour [and physical details] of two systems are at least a reason to suspect similarities in the[ir] functional organisation” (Putnam, 1975, p.437). This approach takes into account any relevant evidence that counts towards the hypothesis that is being considered, which in this case is the claim to MTT capacities in animals.

2.1. Episodic versus Episodic-like Memory

‘Episodic-like’ memory is a label for the kind of memory that closely resembles working episodic memory in humans for they contain many of the same features and are, from outside observation, seemingly indistinguishable from one another. The key difference between the two is that autonoetic consciousness and chronesthesia are sufficient components for episodic-like memory but are not necessary (as in the latter case) (Clayton, 2017, pp.227). Researchers prefer the term ‘episodic-like’ when analysing the results of studies on animal cognition because it is not yet possible to evaluate these phenomenological aspects empirically, due to limitations in communication and inference (ibid, pp.227-228).

Identifying episodic-like memory in nonlinguistic animals as opposed to semantic memory, however, can be done by observing the behavioural expressions of their specific internal states and capacities. These include ‘content’ which Clayton et al. (2003) describe as the encoded what, when, where (WWW) information of a memory episode, ‘structure’ as the glue for WWW information to “represent the same event by being bound together in an integrated representation”, and ‘flexibility’ to ensure that these memory episodes remain consciously accessible for recall and deployment in new situations (unlike nondeclarative memory which “is thought to include motor skill learning, priming and simple forms of conditioning”) (ibid, pp.686-687). All of these aspects can, under the right experimental conditions, be observed in the behaviour of animals possessing of them. Additionally, the term ‘episodic like’ seemingly leaves the door open for autonoetic consciousness and chronesthesia (were such behavioural markers to emerge in nonlinguistic animals).

2.2. Evolutionary Economics of MTT

Without the added ‘like’ suffix, regular episodic memory therefore refers to the structure and flexible deployment of encoded WWW memory as well as the “specific type of recall; that is, MTT into the past” or future (Suddendorf & Busby, 2003, p.391). Given this distinction, we may now ask ourselves what is the benefit to having conscious MTT in addition to encoded WWW memory? Because evolution can only select for traits which enhance the reproductive fitness of organisms, we can assume that the presence of private experience in humans (and possibly other animals) has evolutionary underpinnings (Suddendorf, 2013, p.151). Pascal Boyer (2008) argues that the answer “must lie in mental activity that is present in the former [MTT] but not in the latter [WWW]” (Boyer, 2008, p.219). In other words, the ability to consciously relive past experiences must offer some tangible fitness benefit that simple WWW encoding does not achieve by itself.

Perhaps the most immediate explanation for this is that reflections on one’s past promotes hindsight, allowing individuals to realize information that was previously unnoticed or deemed circumstantially irrelevant, or to use past events for case-based inferences on current and future events. Social interactions, as an example of this, “require […] case-based reasoning rather than (or as well as) rule-governed inferences”, and it is additionally possible that other socially-based uses of MTT would create a selective pressure for its evolution in the human brain (ibid, p.220). Being able to recognise the epistemic and economic reliability of others has a clear fitness benefit and was perhaps fundamental in our movement as a species toward social – rather than solitary – living. It is worth noting that this particular explanation likely only includes humans and other similarly developed primates in its scope.

Another possible reason for MTT into the past which Boyer considers is that it better allows us to engage in foresight and flexible planning, resulting from our accumulated store of memories of past events which can be combined and altered to aid future-thinking (bear in mind that an accurate recollection of the past is useful only to the extent that it is fitness enhancing, memories can and do change according to the fitness preferences of the belonging individual – see Suddendorf & Corballis, 2010, p.292). An alternative (though less compelling) explanation posits foresight as the selected cognitive capacity from which memory-based MTT arises as an adaptive design feature – a spandrel (ibid). But given the fitness benefit of episodic memory, this is probably not the case.

In any case, either of these explanations closely resembles the Janus hypothesis – “the idea that mentally conceiving of past events and imagining future events are closely linked in mind and brain” – which is further supported by empirical research showing the parallel rise and decline of past and future MTT abilities in children and elderly people, the corresponding impairment of both abilities in amnesiac patients, and the overlap in brain imaging studies that are associated with using either of these capacities (ibid, pp.292-293) (Clayton, 2017, p.228). Because of these similarities, and the lack of a formal set of criteria to assess the individual aspects of phenomenal consciousness in animals, evidence for one of these aspects of MTT may reasonably be used as evidence the other. This is important for the inference of human-like MTT in animals as specific studies may only show evidence for one of these.

3. Current Research into MTT in Nonhuman Animals

So far in the paper we have discussed the difference between future and past-oriented MTT, differences in memory and cognition which play a role in the experience of MTT, and the nature of MTT as a functional capacity of the mind with evolutionary underpinnings. The key question, however, asks us whether or not animals are capable of MTT such as humans are. Because animals cannot linguistically communicate the contents of their minds to us (nor are they perhaps cognitively capable of doing so), the evidence for our inference toward their having MTT capacities must adhere to a strict set of criteria so that any other plausible explanation may be ruled out (this pertains to behavioural evidence specifically – see Suddendorf & Corballis, 2010, pp.295-296). However, as outlined in §2a, mental states are not exclusive to any particular expression of their occurrence – they are functional states of the mind. The inferential evidence of their presence in nonlinguistic minds can take a variety of forms which include (but are not limited to); behavioural dispositions, neuroanatomical structures, physiological processes, and evolutionary inferences. While the former of these has been given the most research in MTT study to date, I shall consider each in turn.

3.1. Behavioural Evidence of MTT

3.1.1. Evidence of Episodic-like Memory

Research into episodic-like memory capacities in animals began in 1998 with the renowned scrub jay caching experiments, in which the birds demonstrated the ability to discriminate between WWW (content) memory knowledge, possibly indicating a kind of autonoetic recollection of past events (Clayton, 2017, p.228). The original experiments were conducted using two foods typically consumed by scrub jays – worms and peanuts – that were chosen for their varying degradative properties and degree of preference among the birds (the former, though ranked higher in preference, does not remain edible for as long as the latter). Two groups of scrub jays were used as subjects in the experiment and were appropriately labelled ‘replenish’ and ‘deplete’ according to the context of their testing environment. The group ‘replenish’ were raised to believe that cached worms stayed fresh even after long periods of delay, taught by replacing decayed cached worms with fresh worms before the birds were released to recover their stores. The ‘degrade’ group, by contrast, were given the chance to learn that the worms remained fresh after a short delay of four hours, but that they did not last for a long delay of 124 hours.

The outcome of this teaching process was consistent with the MTT hypothesis; after caching their food, the replenish group continued to search for worms after both short and long delays, whereas the degrade group elected to search for worms only after a short delay. These results show that the jays at least satisfy the content criterion of episodic-like memory; the ‘what’ evidenced by the degrade group differentiating between the cached worms and peanuts according to the time that had passed, ‘when’ by their ability to take into account the passage of time (reflected by their food choice) without reference to their internal circadian rhythm, and ‘where’ by the specified searching of individuals in either group in their respective caching locations, and even after the food was manually removed (so that visual/olfactory cues did not play a role in their searching ability) (ibid, pp.228-230).

wefw

Figure 1: From ‘Episodic-Like Memory and Mental Time Travel in Animals’ in APA Handbook of Comparative Psychology (ibid)

In later experiments with additional protocols, the jays “discriminated between similar episodes that occurred at different times and places, demonstrating that they formed integrated what, where, and when components”, and thereby satisfying the ‘structure’ criterion of episodic-like memory (ibid, p.230). Also, experiments were conducted to evaluate the jays’ ability to flexibly update their knowledge on WWW (content) memory of caching episodes. These tests showed that following their caching of foods, they could learn and adapt their food retrieval preferences according to new information that was presented on their specific perishability rates. Clayton notes that to her knowledge, “this is the only published demonstration of the declarative flexibility with which animals can update their information after the time of encoding”, making corvids the only animal class with evidence to satisfy all of the three proposed criteria of episodic-like memory.

It should also be noted that many non-corvids with remarkably different brain structures have been shown to satisfy the content (and in some cases the structural) criterion of episodic memory. Such animals include nonhuman primates, dogs, zebrafish, rodents, and notably; invertebrate species including bees (insects) and cuttlefish (cephalopods) (Dere et al., 2017, p.2). The discovery of MTT abilities in these species (even in its most basic form, and regardless of how closely it resembles human MTT) lends credence to the functional identification of MTT capacities in the mind and further supports the evolutionary inference to its presence in other species (I shall investigate this further in §3.2).

3.1.2. Evidence of Future Thinking Capabilities

While the outcomes of these experiments may serve as good evidence for episodic-like memory capacities in animals, they do not necessarily indicate foresight. It is, for instance, possible that during the caching process the jays were not conscious of their future need for food and acted instead from a psychological impulse or mechanism. This is seemingly consistent with the Bischof-Köhler hypothesis, which states that animals “anticipate future events only to the extent that the future relates to the satisfaction of current drives and needs” (Suddendorf & Corballis, 2010, p.294). But other studies suggest that scrub jays are capable of predicting their own future motivational states – which can be considered an aspect of foresight.

In these tests, the jays were able to take into account their current and future satiety of a specific food and plan accordingly by caching different foods which they expect to prefer at a later date, thereby “maximiz[ing] the satisfaction of their future motivational state over the current one” (and seemingly showing the presence of future-oriented MTT) (ibid). Some writers have criticised the fact that the jays only chose to reduce the amount of cached food that the birds had been prefed with, rather than increasing the cached amount of non-prefed food. But as Clayton (2017) responds to this critique, “this is precisely what [we] ought to expect given that they began on trial 1 by caching both food types. […] The prefeeding with one food type before retrieval reduced the value of that food type most, but also reduced the value of all foods through its effect on general satiety” (ibid, p.235). Other responses critique the methodology of this experiment for not adhering to a strict feeding schedule, but similar work studying the food satiety of Eurasian jays were able to replicate the findings of the original experiment, so it is unclear how this is relevant to the study’s findings.

Additional work with Eurasian jays show that as well as being able to predict their own future motivational states, they can do so for their mates, whom the males share food with during their breeding season (ibid). In the experiment, female jays were prefed with one of two foods, and later, males were given the opportunity to share either of these foods with their partner. In cases where the males could observe their mate being prefed with a specific food, they would choose to share the alternative food option, indicating outward future-thinking capacities and being able to predict the motivational states of others (this possibly shows the jay’s ability to consciously differentiate between themselves and other minds – although this claim cannot be empirically tested). This is further supported by evidence that in three to four-year-old (human) children, “episodic foresight for the self and other emerge at the same point in development”, and possibly gives rise to a new version of the Janus hypothesis in which these two distinguishing capacities are closely linked (Payne et al., 2015, p.680).

As an alternative explanation to the jays behaviour, it is possible that they instead engage in memory-mediated reinforcement of caching/retrieval activities. In this reasoning, the birds remain capable of having episodic-like memory and it is the acting upon recollections of caching episodes, followed by the rewarding experience of satisfying the jay’s current desire for the specific food it has retrieved, that causes memory-mediated reinforcement of the caching act to take place. The result “does not require the bird to envisage future motivational states”, only to associate current states with past episodes, which is not by itself a sufficient basis for foresight (Clayton et al., 2009, p.67).

Additional research in other animals – notably; nonhuman primates – suggests that they, too, might possess future-thinking capacities. If so, it should not be surprising given the evolutionary proximity of humans to other primates. In a study by Osvath and Osvath involving two chimpanzees and one orangutan, subjects were capable of selecting a straw to access a store of fruit soup that only became accessible 70 minutes after their tool selection (Suddendorf & Corballis, 2010, p.295). In addition to the straw, subjects were presented with a choice between several non-functional tools and, in a second experiment, a piece of their preferred fruit (Clayton, 2017, p.237). Even with the option of an immediate food reward subjects continued to choose the straw tool, plausibly showing that they anticipated their need for the straw using future-oriented MTT. Another explanation posits that their selection of the straw was memory-mediated or based on a prior association or conditioning of the straw with the pleasurable experience of drinking fruit soup. Subjects were given an opportunity to learn about the function of the straw prior to the experiment taking place whereas control subjects did not receive this learning experience and performed poorer, creating a possible basis for this association.

In attempt to circumvent the possible problem of tool association, a third experiment was conducted in which subjects were given two choices which were between (1) a functional straw tool or one of several alternative non-functioning tools, and then (2) which contained the same options as (1), but with the addition of a fruit piece. If the subjects were able to disassociate their own current and future motivational states, then we would expect them to choose the straw for their first choice and the fruit for their second choice, which is just what they did, suggesting that the tool was recognised in terms of its future use and not in terms of its prior reinforcement (or else the subjects would have chosen it for both choices) (ibid).

Suddendorf and Corballis (2010) claim that there are difficulties associated with the control conditions of these experiments, since (as they argue on a separate occasion) “it is not clear that desire for the fruit and desire for the fruit soup are qualitatively different motivational states” (ibid, p.295) (Clayton, 2017, p.238). But as Clayton responds, “[i]f the apes were indeed preexperiencing a future encounter with the apparatus containing fruit soup, would anticipation of the need for the tool not be sufficient to be considered a case of episodic foresight?” (ibid). This criticism thus seems to rely more upon the ambiguity of the Bischof-Köhler hypothesis than it does on a tangible fault in the experimental conditions or results.

3.2. Neurological and Evolutionary Evidence of MTT

Although neurological evidence toward the mental states of animals presents a less compelling case for inference than behavioural evidence (because of the non-exclusivity of mental states to specific neuroanatomical configurations), it does nonetheless present a case, meaning that relevant neurological evidence can be considered in conjunction with other forms of evidence (Dere et al., 2017, p.2). Earlier in §2.1 I made the distinction between episodic and episodic-like memory, the former containing behaviourally inexpressible elements (such as autonoetic consciousness and chronesthesia) which may or may not be present in the latter. If these subjective elements correspond with specific neuroanatomical and/or neurophysiological states in humans (given that humans are the only species we know to be capable of MTT), comparisons between these states with the neurological states of animals may reasonably count as evidence for similarities in their experienced mental states.

Recent work in neuro-imaging has found that the ‘default mode network’, which in humans is closely associated with autonomous mental activity and MTT, exists to some degree in the brains of rats (Corballis, 2013, p.5). Additionally, rats possess hippocampal structures that are seemingly capable of constructing and activating mental maps of environments that they have visited, achieved through the firing of individual cells within that cognitive region. Researchers have observed this activity in the context of a rat-maze traversal during and then following completion of the maze, finding that in short-wave ripples (SWR) (either during slow wave sleep or sedentary rest) the same hippocampal cells as activated in the maze would discharge, sometimes in reverse order, and sometimes corresponding to a path that the rat had not taken (suggesting a mental replay of events and speculation of possible future scenarios within the maze). Because of the hippocampus’ importance in the human default mode network, and because human MTT depends on its activation in the brain, “it may not be unreasonable to suppose that SWR activity in the rat brain is more than a process of consolidation, but that it also represents ongoing mental processes […] tantamount to mental time travel” (ibid).

In §2.2 I considered several evolutionary reasons for MTT’s development in humans, which included primarily social functions like keeping track of the epistemic and economic vigilance and accountability of social partners, or using hindsight, foresight, and case based inferences to make judgements on past and present situations (Boyer, 2008, p.220). While the former two can be considered specific to animals that live complex social lives (including humans and other primates), the latter of these can also be construed in other contexts – possibly including the case of maze rats, who seem to (during their runs) make active judgements according to their own past experience, and also (following their runs) reflect and speculate on their navigatory performance within the maze. Corballis (2013) argues that the neurophysiological evidence of the rats suggests capacities for MTT, but that this ability is probably limited in its complexity (in comparison to humans). The difference is therefore one of token rather than type, “one of degree and not of kind”, by virtue of MTT’s evolutionary continuity between species (ibid, p.6).

Suddendorf’s response to this claim is that rats “lack certain components critical to the human capacity” of MTT – for example; recursive thought, the capacity to “flexibly combine and recombine basic elements to simulate almost any future situation” (for which there is, he claims, no evidence of in rats) (Suddendorf, 2013, p.151). Additionally, behavioural evidence ought to be given preference over neurological forms of evidence because of the ambiguity/indiscernibility of the specific function of animal cognitive faculties. Suddendorf grants that general neural processes and capacities can be evolutionarily continuous across species – including those that are associated with aspects of MTT – but “this need not entail that animals travel mentally in time as humans do” (ibid). Thus, the difference (he claims) really is one of type rather than token.

In a second response, Corballis (2013b) rejects Suddendorf’s strictly behavioural approach and argues that we can make comparative inferences about the functions of cognitive faculties in animals to how we understand their function in our own species (note that this disqualifies certain species like scrub jays who satisfy the behavioural criteria for episodic-like memory, but “do not have a structure homologous to the mammalian cortex” – see Clayton et al., 2003, p.690). On the basis of neuroanatomical comparison, he claims (though not explicitly) that we are able to make neurophysiological inferences on the mental states and capacities of other species. Additionally, to ask whether the evidence of MTT in rats amounts to a fundamental discontinuity depends on how precisely one wishes to define ‘MTT’; ought we consider human MTT the gold standard, or merely an expression of an evolutionary continuous capacity? Whatever the case, evolutionary parallels in the behaviour and neurological function of humans and other species would make it unsurprising to discover that there are parallels in the mental experience of each. On this basis, Corballis asks that we “assume differences in degree rather than in kind as the default position, and seek evidence that this is not the case” (Corballis, 2013b, p.152).

4. Closing Remarks

To conclude, I would like to address some of the methodological issues surrounding the use of behavioural evidence in animal MTT research. While evidence in some animal behaviour is consistent with their possessing MTT capacities, its limitations in not being able to assess relevant aspects of phenomenal consciousness (to some) put research at a standstill. Certain authors (for example; Suddendorf in the previous section, among many other sceptics of animal theory of mind) hold that such evidence must be more than consistent – it must be such than any other plausible explanation for behaviour can be ruled out. To achieve this, the notion of MTT and its associated elements must be formally described and tested. It is here that the research is lacking.

Consider for instance the difference between episodic and episodic-like memory, and how this difference (while useful as distinctive terminology) does not meaningfully further the MTT debate in animals (insofar as it does not answer the question ‘do animals, like humans, think episodically?’). Two solutions have been proposed by Clayton et al.; to continue marking this difference and accept that the claim to MTT in animals cannot be empirically proven (because of limits in phenomenal inference), or treat Clayton et al.’s (2003) earlier definition of content, structure, and flexibility in WWW memory as criteria for inferring regular episodic memory (Suddendorf & Busby, 2003, p.392). Accepting the first of these options would seemingly shut out all future behavioural research into animal MTT, while the latter allows for the possibility that more than one cognitive capacity is responsible for one behaviour at a given point – unless one develops sufficiently closed experimental conditions to disqualify an explanation from an alternative cognitive capacity. This seemingly has not yet been achieved (despite great effort), though if it were it would not count as definitive proof of a particular mental state’s occurrence due to unforeseeable circumstances in animal cognition.

The presentation of either of these options is, I argue, misguided, as it assumes a widespread understanding of the term ‘episodic memory’ has already been reached – it has not. Not only is it unclear as to whether ‘episodic’ and ‘episodic-like’ denote the same mental experience, but it is also questionable whether certain qualifying aspects of episodic memory really are sufficient, and thus, whether they ought to be included in its definition. In 2013, Cameron Buckner wrote a scalding critique against theory of mind research in animals, in which he notes three aspects of anthropocentric bias; methodological, evaluative, and semantic (Buckner, 2013, p.858). In our case it is the latter – semantic anthropocentrism, which “involves precisifying vaguely-defined psychological terms to human-level ability” – that most succinctly addresses the problem of MTT research in animals.

The term ‘episodic memory’ demonstrates this bias, in which the qualifying condition of autonoetic consciousness refers to the specific human experience of conscious recollection, while other non-qualifying conditions such as those specific to the cognitive niches of other animals are discarded. More so, Buckner reminds us that “human remembering is largely constructive and frequently confabulatory. [T]he sceptical position [thus] rests on a tenuous semantic premise”, in which episodic memory (by its definition) can only include human cognitive capacities in its scope (ibid, pp.860-861). Suddendorf (2013) is guilty of this bias, as he openly discounts episodic MTT evidence in rats for “lack[ing] certain components critical to the human capacity” (ibid, p.151). Buckner (2013) proposes the term ‘anthropofabulation’ to describe these semantic and exaggeratory biases, which result in unfair comparisons between animal and human data (to discount the former). Likely, it would be better not to engage in comparative research between the cognitive capacities of humans and other animals if such comparisons are made with the bias of one being exemplary for the other. Rather, future attempts at study ought to try and understand the cognitive abilities of animals with respect to the relevant aspects of their own particular species (and not our own).

In closing, although there remains considerable disagreement on whether animals possess MTT capacities that are similar to human capacities, this paper has brought to attention some of the problems associated with MTT discourse and recommendations for future research in the field. As Clayton (2017) remarks; “[a]bsence of evidence is not evidence of absence” – the normative implications of MTT research must therefore also account for the possibility of impassable research, by which a lack of available evidence cannot be used to justify the mistreatment of animals on the basis of their supposed lack of mental capacities (ibid, p.235).

Bibliography

Boyer, P., 2008, ‘Evolutionary economics of mental time travel?’, Trends in Cognitive Sciences, vol. 12, no. 6, pp. 219-224.

Buckner, C., 2013, ‘Morgan’s Canon, meet Hume’s Dictum: avoiding anthropofabulation in cross-species comparisons’, Biology & Philosophy, vol. 28, no. 5, pp. 853–871.

Clayton, N. S., Bussey, T. J. & Dickenson, A., 2003, ‘Can animals recall the past and plan for the future?’, Nature Reviews Neuroscience, vol. 4, pp. 685-691.

Clayton, N. S., Russell, J. & Dickinson, A., 2009, ‘Are Animals Stuck in Time or Are They Chronesthetic Creatures?’, Topics in Cognitive Science, vol. 1, no. 1, pp. 59-71.

Clayton, N. S., 2017, ‘Episodic-Like Memory and Mental Time Travel in Animals’ in J Call, (ed), APA Handbook of Comparative Psychology: Vol.2, pp. 227-243. American Psychological Association: Washington.

Corballis, M., 2013, ‘Mental time travel: a case for evolutionary continuity’, Trends in Cognitive Sciences, vol. 17, no. 1, pp. 5-6.

Corballis, M., 2013b, ‘The wandering rat: response to Suddendorf’, Trends in Cognitive Sciences, vol. 17, no. 4, p. 152.

Dere, E., Dere, D., Silva, M. A., Huston, J. P. & Zlomuzika, A., 2017, ‘Fellow travellers: Working memory and mental time travel in rodents’, Behavioural Brain Research, [In Press].

Payne, G., Taylor, R., Hayne, H. & Scarf, D., 2015, ‘Mental time travel for self and other in three- and four-year-old children’, Memory, vol. 23, no. 5, pp. 675-682.

Suddendorf, T., 2013, ‘Mental time travel: Continuities and discontinuities’, Trends in Cognitive Sciences, vol. 17, no. 4, pp. 151-152.

Suddendorf, T. & Busby, J., 2003, ‘Mental Time Travel in Animals?’, Trends in Cognitive Sciences, vol. 7, no. 9, pp. 391-396.

Suddendorf, T. & Corballis, M., 2010, ‘Behavioural evidence for mental time travel in nonhuman animals’, Behavioural Brain Research, vol. 215, no. 2, pp. 292-298.

Citation style in accordance with the University of Western Australia, Harvard

Word count: 5011

Scroll to Top