Human knowledge of past particular events is enabled not only by episodic memory, but also by semantic memory and by inference: for example, in detective work from currently perceivable evidence. With various ways of knowing past events, it is an interesting question why we have autonoesis, the self-reflexive consciousness that is applied to the past in episodic memory. Mahr & Csibra (M&C) propose that its function is communicative: We need to keep track of whether judgments about past events originate in firsthand experience in order to share these judgments most effectively with others. M&C may be right to connect consciousness and communication, but some details of their argument are debatable.
According to M&C, autonoesis is “the capacity that enables us to distinguish between cases in which we can assert epistemic authority for our own testimony and cases in which we cannot” (sect. 3.1.1, para. 4), where asserting epistemic authority is a matter of staking one's credibility on the truth of what one is asserting. However, one's credibility is at stake in any assertion, whether it is grounded in present or past firsthand experience, inference, or testimony. M&C are right that one may choose to assert that another person said something, endorsing only the fact that this person made a statement (“Jane said that Bill was fired”), but one may also assert the bare proposition learned through testimony (“Bill was fired”), taking a source's word for it and risking one's own reputation in spreading the gossip. One's reputation for reliability is at stake even if one explicitly marks the assertion with an evidential like “apparently” to indicate a testimonial basis, or an epistemic modal (“Bill must have been fired”) to indicate inference; indeed, claims made with such markers are not inherently weaker than their directly grounded counterparts (Von Fintel & Gillies Reference Von Fintel and Gillies2010).
Rather than seeing epistemic authority as something binary, either staked or not, one might see it as a matter of degree, with the greatest confidence vested in what we have seen for ourselves. However, this would leave it unclear just why autonoesis was needed alongside the simpler quantitative sensibilities that already regulate the reliability of our reports from semantic memory (Goldsmith et al. Reference Goldsmith, Koriat and Weinberg-Eliezer2002). More qualitative ways of understanding epistemic authority are perhaps more promising. Perhaps autonoesis tells us which facts we know will be most informative for others to hear. Semantic memory encodes similar facts and regularities across a community: Indeed, the background beliefs in which one has greatest individual confidence are most likely to be shared already by others (Koriat Reference Koriat2008). By contrast, episodic memory provides a unique record of particular events not widely known to others, together with information on who else witnessed these events. Knowing whether others know about a past event may indeed be useful in arguments, but it could also be valuable to the species for more cooperative purposes in communication, just as the conscious availability of subjective confidence enables pooling of current perceptual judgments to increase accuracy in joint decision making (Bahrami et al. Reference Bahrami, Olsen, Latham, Roepstorff, Rees and Frith2010).
Autonoesis can fail to accompany the capacity to retrieve events, although this is rare: Klein (Reference Klein2015a) reports finding only three such cases in the literature. One describes a recently amnesiac patient who was able, when questioned, to narrate many details of his older brother's death in an automobile accident roughly a year before. When complimented on his recall, the patient looked puzzled and insisted that it was the interviewer who had just told him about what had happened to his brother (Talland Reference Talland1964). Despite his confusion between another's testimony and his own recall, the patient was not reported as lacking confidence in the facts surrounding the automobile accident: He expressed no uncertainty as to whether these were real or merely imagined events. If this patient had detailed knowledge of events in his personal past and the capacity to narrate these events to others, what was he missing? Talland notes that recall in this case was secured only by a series of prodding questions, rather than flowing spontaneously. Autonoesis enables recognition of what one's audience will not already know: The intuitive sense of private access to one's personal past enables recognition that description of these events will be informative to others; lacking that sense leaves one unmotivated to share one's knowledge.
M&C's qualitative characterization of the epistemic role of episodic memory focuses on argument: They define episodic memory as “an epistemic attitude taken toward the simulation of a specific past event, which serves to justify a belief about the occurrence of this event” (sect. 1.3, para. 1). However, one may take an epistemic attitude toward a simulation of a past event without remembering the event in question; for example, when visualizing how an animal might have freed itself from the broken trap one is looking at, or when reconstructing what one must have done in the course of last night's drunken stupor. Beliefs about past events may be well justified by imaginative simulation informed by current perceptual input, and these simulations can make us aware of why we believe what we do, supporting argumentation without constituting moments of episodic memory.
M&C take the importance of having and enforcing human social commitments as a reason for the development of episodic memory. I agree that commitments and autonoesis are closely linked, but I wonder about the direction of explanation: It seems to me that in order to enter into a social commitment binding my future self, or to feel bound by my past commitments, I must already have the sort of sense of self, enduring in time, that is enabled by autonoesis. M&C would be on firmer ground if proto-humans with mere event memory could already make social commitments, where the emergence of autonoesis could then offer a way to strengthen and enforce those commitments. I wonder whether the power to make social commitments might not be just one of the many adaptive consequences of the human sense of self (Metzinger Reference Metzinger2004), so that autonoesis would be better explained by its contribution to that larger construct, with its diverse adaptive advantages.
Human knowledge of past particular events is enabled not only by episodic memory, but also by semantic memory and by inference: for example, in detective work from currently perceivable evidence. With various ways of knowing past events, it is an interesting question why we have autonoesis, the self-reflexive consciousness that is applied to the past in episodic memory. Mahr & Csibra (M&C) propose that its function is communicative: We need to keep track of whether judgments about past events originate in firsthand experience in order to share these judgments most effectively with others. M&C may be right to connect consciousness and communication, but some details of their argument are debatable.
According to M&C, autonoesis is “the capacity that enables us to distinguish between cases in which we can assert epistemic authority for our own testimony and cases in which we cannot” (sect. 3.1.1, para. 4), where asserting epistemic authority is a matter of staking one's credibility on the truth of what one is asserting. However, one's credibility is at stake in any assertion, whether it is grounded in present or past firsthand experience, inference, or testimony. M&C are right that one may choose to assert that another person said something, endorsing only the fact that this person made a statement (“Jane said that Bill was fired”), but one may also assert the bare proposition learned through testimony (“Bill was fired”), taking a source's word for it and risking one's own reputation in spreading the gossip. One's reputation for reliability is at stake even if one explicitly marks the assertion with an evidential like “apparently” to indicate a testimonial basis, or an epistemic modal (“Bill must have been fired”) to indicate inference; indeed, claims made with such markers are not inherently weaker than their directly grounded counterparts (Von Fintel & Gillies Reference Von Fintel and Gillies2010).
Rather than seeing epistemic authority as something binary, either staked or not, one might see it as a matter of degree, with the greatest confidence vested in what we have seen for ourselves. However, this would leave it unclear just why autonoesis was needed alongside the simpler quantitative sensibilities that already regulate the reliability of our reports from semantic memory (Goldsmith et al. Reference Goldsmith, Koriat and Weinberg-Eliezer2002). More qualitative ways of understanding epistemic authority are perhaps more promising. Perhaps autonoesis tells us which facts we know will be most informative for others to hear. Semantic memory encodes similar facts and regularities across a community: Indeed, the background beliefs in which one has greatest individual confidence are most likely to be shared already by others (Koriat Reference Koriat2008). By contrast, episodic memory provides a unique record of particular events not widely known to others, together with information on who else witnessed these events. Knowing whether others know about a past event may indeed be useful in arguments, but it could also be valuable to the species for more cooperative purposes in communication, just as the conscious availability of subjective confidence enables pooling of current perceptual judgments to increase accuracy in joint decision making (Bahrami et al. Reference Bahrami, Olsen, Latham, Roepstorff, Rees and Frith2010).
Autonoesis can fail to accompany the capacity to retrieve events, although this is rare: Klein (Reference Klein2015a) reports finding only three such cases in the literature. One describes a recently amnesiac patient who was able, when questioned, to narrate many details of his older brother's death in an automobile accident roughly a year before. When complimented on his recall, the patient looked puzzled and insisted that it was the interviewer who had just told him about what had happened to his brother (Talland Reference Talland1964). Despite his confusion between another's testimony and his own recall, the patient was not reported as lacking confidence in the facts surrounding the automobile accident: He expressed no uncertainty as to whether these were real or merely imagined events. If this patient had detailed knowledge of events in his personal past and the capacity to narrate these events to others, what was he missing? Talland notes that recall in this case was secured only by a series of prodding questions, rather than flowing spontaneously. Autonoesis enables recognition of what one's audience will not already know: The intuitive sense of private access to one's personal past enables recognition that description of these events will be informative to others; lacking that sense leaves one unmotivated to share one's knowledge.
M&C's qualitative characterization of the epistemic role of episodic memory focuses on argument: They define episodic memory as “an epistemic attitude taken toward the simulation of a specific past event, which serves to justify a belief about the occurrence of this event” (sect. 1.3, para. 1). However, one may take an epistemic attitude toward a simulation of a past event without remembering the event in question; for example, when visualizing how an animal might have freed itself from the broken trap one is looking at, or when reconstructing what one must have done in the course of last night's drunken stupor. Beliefs about past events may be well justified by imaginative simulation informed by current perceptual input, and these simulations can make us aware of why we believe what we do, supporting argumentation without constituting moments of episodic memory.
M&C take the importance of having and enforcing human social commitments as a reason for the development of episodic memory. I agree that commitments and autonoesis are closely linked, but I wonder about the direction of explanation: It seems to me that in order to enter into a social commitment binding my future self, or to feel bound by my past commitments, I must already have the sort of sense of self, enduring in time, that is enabled by autonoesis. M&C would be on firmer ground if proto-humans with mere event memory could already make social commitments, where the emergence of autonoesis could then offer a way to strengthen and enforce those commitments. I wonder whether the power to make social commitments might not be just one of the many adaptive consequences of the human sense of self (Metzinger Reference Metzinger2004), so that autonoesis would be better explained by its contribution to that larger construct, with its diverse adaptive advantages.