Hoerl & McCormack (H&M) let our study on raven planning exemplify studies measuring mental time travel (MTT) to the future. However, we did not investigate MTT, but conducted functional behavioural comparisons between apes and ravens, without directly inferring neurocognitive mechanisms (Kabadayi & Osvath Reference Kabadayi and Osvath2017; Osvath & Kabadayi Reference Osvath and Kabadayi2018). Such operational definitions are useful when comparing taxa vastly separated in phylogeny and brain constitution. It follows that we remained agnostic about whether ravens are temporal reasoners sensu H&M. It is perhaps not necessary to ascribe ravens’ temporal reasoning based on this study, but H&M's claims are oversimplified. This reflects the conceptual narrowness of the two systems: updaters and reasoners. The distinction risks throwing out a whole kindergarten with the bath water.
H&M assert that our study failed to show that the ravens selected items for a situation they represented as distinct from the present, and that the delays played no role in their reasoning behind the selection. However, this is not true. H&M ignore a pivotal part of the study: the delay-based control, which teased apart whether the ravens selected the tool because it was immediately rewarding or because it carried value in relation to the non-present apparatus event. Two intertemporal choice experiments showed that the ravens exerted significantly more self-control in selecting the tool over an immediate reward if the apparatus event was closer in time (less than a minute compared with 15 minutes). This revealed, contrary to H&M's claims, that the delay played a crucial role in how the ravens selected and also that they represented the apparatus event as distinct from the current situation. In H&M's account, the ravens’ behaviours are governed by what they call primitive goal representations. They illustrate this by a thought bubble in their Figure 2, which purportedly depicts the raven's reasoning: “The apparatus is somewhere in the environment. The stone operates the apparatus. Choose the stone” (Fig. 2, Temporal updating account). As the raven's representation does not include delays, it will produce the same thoughts regardless of any delays. Hence, the raven would select the item equally often despite differences in delays. Indeed, one would expect that the ravens would never even select the immediate reward, as it is smaller than the future one: In the mind of the raven there are, according to H&M, only two “immediate” rewards. The empirical results spell out something other than the thought bubble. So, what are the ravens doing?
We have already excluded primitive goal representations. The other option given for an updater by H&M is temporal sensitivity, which triggers a behaviour when a phase timing system is in a certain state. However, this does not work either, as there were no phases to time: two identical selection events, that could be offered anytime during the day, where only the upcoming delay differed. We are left with “elapsed-time sensitivity,” which H&M used to explain how an animal senses that a certain time has elapsed. The problem here is that according to H&M, such sensitivity is experienced in the ongoing present; it is not represented afterwards. So, the selection event could not cue any memories of the different sensations that relate to the two different delays.
The explanation of the ravens’ behaviours boils down to three possibilities: They did not do what they did; they are temporal reasoners; or they did something not captured by H&M's theory. As for the third alternative, one can think of many cognitively rich ways in which some animals could implicitly represent time, without representing time as such. If animals can be time sensitive, why could some not remember such sensations? If you have means-end reasoning why can you not represent change? (H&M seem to think that representing change is the same as representing it as change.) If they can represent locations in their inner map, why can they not represent distances? “Here” is not “there,” and “there1” is not “there2,” and reaching any of these locations requires different work effort, often related to time. (Many animals fancy shortcuts.) All sorts of memories of actions, sensations, or distances could become embodied representations, which are intrinsically related to time, without having a “pure” representation of the dimension itself. Such representations may differ widely between species.
Millions of species, with half a billion years of evolutionary history as animals, are lumped together as updaters, whereas only one is identified as a reasoner. H&M think that this distinction is more helpful than the dichotomy that often surfaces in MTT debates. Nevertheless, not only do they argue that MTT requires temporal reasoning; they still put forward a strong dichotomy. This is unhelpful for the study of the evolution of cognition. Ravens and great apes pass planning tasks that monkeys and young children do not. According to H&M, all these animals are incapable of representing time, which might be true, but then how is the distinction helpful in explaining striking differences between updating species? Relatedly, there is no account on how one evolves from an updater to a reasoner. The distinction is untimely when animal cognition research is increasingly directed at cognitive evolution beyond the very thin human lineage. It is not uninteresting whether some animals reason about time, but it is only one among a host of equally interesting questions. This rehashed “stuck in time hypothesis” comes across as stuck in the past.
As an endnote, one may ask whether this truly is a division between two distinct and conceptually equivalent systems. It seems, rather, that the authors use temporal updating just to explain the virtues of temporal reasoning. There appear to be many ways in which one can be cognitive about time without representing time, but only one in which one can be a temporal reasoner. This relationship is too unbalanced to warrant two categories. Moreover, temporal reasoning likely depends on the ability of abstract conceptualisation, which in humans is greatly aided by linguistic minds, and such conceptualisations pertain to a myriad of phenomena, where “time” is only one. This is not a temporal cognition system per se.
Hoerl & McCormack (H&M) let our study on raven planning exemplify studies measuring mental time travel (MTT) to the future. However, we did not investigate MTT, but conducted functional behavioural comparisons between apes and ravens, without directly inferring neurocognitive mechanisms (Kabadayi & Osvath Reference Kabadayi and Osvath2017; Osvath & Kabadayi Reference Osvath and Kabadayi2018). Such operational definitions are useful when comparing taxa vastly separated in phylogeny and brain constitution. It follows that we remained agnostic about whether ravens are temporal reasoners sensu H&M. It is perhaps not necessary to ascribe ravens’ temporal reasoning based on this study, but H&M's claims are oversimplified. This reflects the conceptual narrowness of the two systems: updaters and reasoners. The distinction risks throwing out a whole kindergarten with the bath water.
H&M assert that our study failed to show that the ravens selected items for a situation they represented as distinct from the present, and that the delays played no role in their reasoning behind the selection. However, this is not true. H&M ignore a pivotal part of the study: the delay-based control, which teased apart whether the ravens selected the tool because it was immediately rewarding or because it carried value in relation to the non-present apparatus event. Two intertemporal choice experiments showed that the ravens exerted significantly more self-control in selecting the tool over an immediate reward if the apparatus event was closer in time (less than a minute compared with 15 minutes). This revealed, contrary to H&M's claims, that the delay played a crucial role in how the ravens selected and also that they represented the apparatus event as distinct from the current situation. In H&M's account, the ravens’ behaviours are governed by what they call primitive goal representations. They illustrate this by a thought bubble in their Figure 2, which purportedly depicts the raven's reasoning: “The apparatus is somewhere in the environment. The stone operates the apparatus. Choose the stone” (Fig. 2, Temporal updating account). As the raven's representation does not include delays, it will produce the same thoughts regardless of any delays. Hence, the raven would select the item equally often despite differences in delays. Indeed, one would expect that the ravens would never even select the immediate reward, as it is smaller than the future one: In the mind of the raven there are, according to H&M, only two “immediate” rewards. The empirical results spell out something other than the thought bubble. So, what are the ravens doing?
We have already excluded primitive goal representations. The other option given for an updater by H&M is temporal sensitivity, which triggers a behaviour when a phase timing system is in a certain state. However, this does not work either, as there were no phases to time: two identical selection events, that could be offered anytime during the day, where only the upcoming delay differed. We are left with “elapsed-time sensitivity,” which H&M used to explain how an animal senses that a certain time has elapsed. The problem here is that according to H&M, such sensitivity is experienced in the ongoing present; it is not represented afterwards. So, the selection event could not cue any memories of the different sensations that relate to the two different delays.
The explanation of the ravens’ behaviours boils down to three possibilities: They did not do what they did; they are temporal reasoners; or they did something not captured by H&M's theory. As for the third alternative, one can think of many cognitively rich ways in which some animals could implicitly represent time, without representing time as such. If animals can be time sensitive, why could some not remember such sensations? If you have means-end reasoning why can you not represent change? (H&M seem to think that representing change is the same as representing it as change.) If they can represent locations in their inner map, why can they not represent distances? “Here” is not “there,” and “there1” is not “there2,” and reaching any of these locations requires different work effort, often related to time. (Many animals fancy shortcuts.) All sorts of memories of actions, sensations, or distances could become embodied representations, which are intrinsically related to time, without having a “pure” representation of the dimension itself. Such representations may differ widely between species.
Millions of species, with half a billion years of evolutionary history as animals, are lumped together as updaters, whereas only one is identified as a reasoner. H&M think that this distinction is more helpful than the dichotomy that often surfaces in MTT debates. Nevertheless, not only do they argue that MTT requires temporal reasoning; they still put forward a strong dichotomy. This is unhelpful for the study of the evolution of cognition. Ravens and great apes pass planning tasks that monkeys and young children do not. According to H&M, all these animals are incapable of representing time, which might be true, but then how is the distinction helpful in explaining striking differences between updating species? Relatedly, there is no account on how one evolves from an updater to a reasoner. The distinction is untimely when animal cognition research is increasingly directed at cognitive evolution beyond the very thin human lineage. It is not uninteresting whether some animals reason about time, but it is only one among a host of equally interesting questions. This rehashed “stuck in time hypothesis” comes across as stuck in the past.
As an endnote, one may ask whether this truly is a division between two distinct and conceptually equivalent systems. It seems, rather, that the authors use temporal updating just to explain the virtues of temporal reasoning. There appear to be many ways in which one can be cognitive about time without representing time, but only one in which one can be a temporal reasoner. This relationship is too unbalanced to warrant two categories. Moreover, temporal reasoning likely depends on the ability of abstract conceptualisation, which in humans is greatly aided by linguistic minds, and such conceptualisations pertain to a myriad of phenomena, where “time” is only one. This is not a temporal cognition system per se.