“What is mental time travel, and is it unique to humans?” As the focus for research on cognitive evolution, this question is doubly flawed. Not only can we never know what other species' mental experiences are like, searching for a yes or no answer to a question about human uniqueness is not a productive way to proceed with research in comparative cognition. This path leads to a quest for an existence proof (just one animal with “it” is enough), followed by endless disputes over whether “it” really was demonstrated. The history of attempts to teach forms of human language to apes (Shettleworth Reference Shettleworth1998) is evidence enough for this. Far more productive as well as consistent with evolutionary thinking is to ask something like, “What components of process X are and are not shared among what species and why, in evolutionary, functional, and perhaps neurological terms?” Progress in answering such questions is most likely to be made when the process being compared across species is already well understood within at least one of the species in question, human or other animal.

In various places, Suddendorf & Corballis (S&C) do address the difficulty of drawing conclusions about animals' mental experiences, for example, when discussing episodic-like memory in section 3.1. But this is not merely a matter of not having enough evidence yet: Because we can only observe behavior, we can never know if animals are mentally time traveling anywhere. Researchers can productively seek no more than functional similarities between human and animal behaviors in analogous situations. A pattern of brain activity (S&C, sect. 2.3) is just another item of function similarity, and one that is a priori denied to species with very different brains from humans.

An instructive example of what is meant by documenting functional similarity of behaviors is research on metacognition in rhesus macaques. Contrary to what S&C suggest in section 4.5, there is a now a rich body of convergent evidence from several laboratories and testing paradigms consistent with these monkeys' being able to monitor both the certainty of their perceptual judgments and the strength of their memories. For example, when memory is poor, monkeys choose to escape tests of memory more often; but when they choose to take the test, they continue to perform well, better than when forced to take it (Hampton Reference Hampton2001; see also Smith & Washburn Reference Smith and Washburn2005). Monkeys show this functional relationship immediately when memory is taxed, further indicating that their behavior does not reflect learning to use cues like length of the retention interval. Monkeys' reports of metacognitive certainty transfer across tasks (Kornell et al. Reference Kornell, Son and Terrace2007), and they behave as if they are using it to control information seeking in naturalistic (Hampton et al. Reference Hampton, Zivin and Murray2004) and operant tasks (Kornell et al. Reference Kornell, Son and Terrace2007). The availability of multiple convergent measures and more than one testing paradigm also renders meaningful repeated failures to find similar patterns of data in pigeons (Sutton & Shettleworth, in press).

One reason this body of work is so compelling is that there is reasonable theoretical agreement on relevant procedures and predicted patterns of data. The tests for animals are also closely analogous to those used with people, and indeed data from people and monkeys have sometimes been reported in the same article (see Smith et al. Reference Smith, Shields and Washburn2003). As alluded to by S&C, in some of these situations, rather than explicitly “reporting on” memory strength, the monkeys may be discriminating some self-generated correlate of memory strength such as latency to select an option; but the same may be true of people (Terrace & Metcalfe Reference Terrace and Metcalfe2005). Such possibilities remain for further study. Nevertheless, among attempts to document behaviors in animals functionally similar to those accompanied by distinct kinds of awareness in humans, research on monkey metacognition is unique in richness and diversity of convergent data. Even so, it still cannot reveal the monkeys' subjective states.

S&C do try to specify (in sect. 2.1) what nonverbal behavior would count as evidence of “future mental time travel.” It should be a novel behavior or combination of behaviors, the outcome of which later satisfies a need not present at the time of performance. They then suggest an experiment very similar to one recently reported by Raby et al. (Reference Raby, Alexis, Dickinson and Clayton2007) with positive results. But, as S&C say, an additional requirement that follows from the folk-psychological notion of human planning is that the ability revealed in such a test be flexible, that is, applicable in more than one context. It remains to be seen whether this new study with scrub jays has tapped a generalized ability or a hitherto unrecognized learning adaptation in the food-caching system (Shettleworth Reference Shettleworth2007).

S&C also propose separable components of mental time travel that might be sought in developing children and/or other species. This approach to comparative research is undoubtedly useful in general. For example, the comparative study of numerical cognition has made real progress since researchers moved beyond asking, “Do animals count?” (see Shettleworth Reference Shettleworth1998) to the comparative analysis of components of numerical competence such as exact perception of small numerosities and Weber's law–based discrimination among larger ones (Feigenson et al. Reference Feigenson, Dehaene and Spelke2004). These may be widely shared among vertebrates, whereas exact appreciation of quantities larger than about 5 is normally confined to numerate humans (Pica et al. Reference Pica, Lemer, Izard and Dehaene2004). In contrast, the components of mental time travel proposed by S&C seem too many and too vaguely metaphorical for similar progress in this area to be expected any time soon. It is abundantly clear (Roberts Reference Roberts2002; Shettleworth Reference Shettleworth1998) that all animals are superbly equipped with learning and timing mechanisms allowing them to use their own and their ancestors' past to determine future behavior. Interpreting what they do as revealing mental time travel is one example of a rather widespread return of anthropomorphism, which appeals to the public without necessarily advancing understanding (Wynne Reference Wynne2007). Perhaps there should be a moratorium on studies of animal “mental time travel” until its psychological and behavioral components have been better specified and more carefully dissected in developmental studies.