Mahr & Csibra (M&C) argue that the function of episodic memory is to enable a subject to persuade others to endorse the subject's descriptions of past events. Although the authors build an impressive case for this communicative account, it turns out to be committed to a counterintuitive claim, namely, that episodic memory is systematically misleading. Other accounts, including the future-oriented account (e.g., Schacter & Addis Reference Schacter and Addis2007), likewise turn out to be committed to this misleadingness claim. The future-oriented account sees episodic memory, along with episodic future thought (Szpunar Reference Szpunar2010), as a form of mental time travel (MTT) (Suddendorf & Corballis Reference Suddendorf and Corballis1997), with future-oriented MTT or episodic future thought being primary, in the sense that the function of the MTT system is to enable the subject to imagine future events, while the ability to engage in forms of past-oriented MTT, including episodic memory, emerges as a by-product. Although they differ on the question of the function of the memory or MTT system, the future-oriented account and the communicative account agree on two claims that together imply the misleadingness claim: (1) that episodic memories necessarily involve autonoesis (the autonoesis claim) and (2) that episodic remembering is necessarily a constructive process (the construction claim).
The autonoesis claim: M&C understand autonoesis in metarepresentational terms (cf. Dokic Reference Dokic2014; Fernández Reference Fernández2016), characterizing the content of a retrieved memory as having two components: a first-order component informing a subject about an event and a second-order component informing him or her that the information provided by the first-order component originates in the subject's own experience of the event. If retrieved memories are indeed metarepresentational, then, when retrieval results in the formation of a belief, the subject believes not simply that such-and-such an event occurred but rather that he or she knows that such-and-such an event occurred because of having experienced its occurrence. Crucially, the second-order component of a memory belief might be inaccurate – and hence the belief as a whole might be false – even if the first-order component is accurate, simply because there are sources of accurate information about an event other than one's own experience.
The autonoesis claim is essential to the communicative account: In making a memory claim, a subject claims epistemic authority over the event in question, and autonoesis is normally the subject's only ground for doing so. The claim might not, strictly speaking, be essential to the future-oriented account: Because autonoesis may not play a role in episodic future thinking (Perrin Reference Perrin, Michaelian, Klein and Szpunar2016), the future-oriented account might replace it with a weaker claim, namely, that although autonoesis typically plays a role in episodic remembering, it is not a necessary feature of retrieved memories (Michaelian Reference Michaelian2016b). Even this weakened claim is, however, sufficient to commit the future-oriented account to the misleadingness claim.
The construction claim: M&C understand construction as occurring through Bayesian prediction of features of past events based on evidence provided by both episodic traces and semantic information (De Brigard Reference De Brigard2014a). Alternative understandings are available (Michaelian Reference Michaelian2016b), but they concur that, at least in typical cases, not all of the content of a given retrieved memory originates in the subject's experience of the remembered event. This, in turn, implies that retrieved memories will often be, to some degree, inaccurate with respect to remembered events. But construction does not make inaccuracy inevitable: The incorporation of nonexperiential information into a retrieved memory representation, in particular, does not necessarily imply inaccuracy, simply because incorporated information may itself be accurate (Michaelian Reference Michaelian2013).
The construction claim is essential to the communicative account: If the point of making memory claims is not to convey accurate descriptions of past events but rather to convey descriptions that the subject wants an audience to endorse, a constructive memory process is needed to enable the subject to generate suitable representations of events. The claim is likewise essential to the future-oriented account: The MTT system must be able to constructively recombine and modify information from various sources in order to generate representations of possible events in episodic future thinking, and if episodic remembering is carried out by the same system, it is bound to be constructive in the same sense.
The misleadingness claim: Together, the (weakened) autonoesis claim and the construction claim imply the misleadingness claim. If the autonoesis claim is right, a memory might be false even if the event that it represents occurred exactly as the belief represents it as having occurred. In particular, the belief will be false in cases in which its first-order content originates at least in part in a source other than the subject's own experience of the event. If the construction claim is right, such cases occur frequently. Indeed, because, as M&C acknowledge, episodic remembering is driven as much by current beliefs as by episodic traces, they are the rule rather than the exception. Therefore, the second-order component of a memory belief – and the belief as a whole – will frequently be false. In short, both the communicative account and the future-oriented account are committed to the claim that episodic memory beliefs are frequently false, not because construction results in inaccurate representations of events, but rather because autonoesis results in inaccurate metarepresentations of the relationship between representations and the sources in which they originate, both where events are represented inaccurately and where they are represented accurately.
We might, in principle, attempt to avoid the misleadingness claim by rejecting either the construction claim or the autonoesis claim, but we have good reason to accept both of these claims. We might also attempt to avoid it by modifying the metarepresentational understanding of autonoesis so that the autonoesis claim says that the second-order component of a retrieved memory informs a subject only that part of the first-order component of the memory, as opposed to the first-order component as a whole, originates in the subject's experience of the event, but it is unclear whether this is compatible with the roles assigned to autonoesis by the communicative and future-oriented accounts. We may thus be forced to accept the counterintuitive conclusion that episodic memory is indeed systematically misleading.
You have
Access
Mahr & Csibra (M&C) argue that the function of episodic memory is to enable a subject to persuade others to endorse the subject's descriptions of past events. Although the authors build an impressive case for this communicative account, it turns out to be committed to a counterintuitive claim, namely, that episodic memory is systematically misleading. Other accounts, including the future-oriented account (e.g., Schacter & Addis Reference Schacter and Addis2007), likewise turn out to be committed to this misleadingness claim. The future-oriented account sees episodic memory, along with episodic future thought (Szpunar Reference Szpunar2010), as a form of mental time travel (MTT) (Suddendorf & Corballis Reference Suddendorf and Corballis1997), with future-oriented MTT or episodic future thought being primary, in the sense that the function of the MTT system is to enable the subject to imagine future events, while the ability to engage in forms of past-oriented MTT, including episodic memory, emerges as a by-product. Although they differ on the question of the function of the memory or MTT system, the future-oriented account and the communicative account agree on two claims that together imply the misleadingness claim: (1) that episodic memories necessarily involve autonoesis (the autonoesis claim) and (2) that episodic remembering is necessarily a constructive process (the construction claim).
The autonoesis claim: M&C understand autonoesis in metarepresentational terms (cf. Dokic Reference Dokic2014; Fernández Reference Fernández2016), characterizing the content of a retrieved memory as having two components: a first-order component informing a subject about an event and a second-order component informing him or her that the information provided by the first-order component originates in the subject's own experience of the event. If retrieved memories are indeed metarepresentational, then, when retrieval results in the formation of a belief, the subject believes not simply that such-and-such an event occurred but rather that he or she knows that such-and-such an event occurred because of having experienced its occurrence. Crucially, the second-order component of a memory belief might be inaccurate – and hence the belief as a whole might be false – even if the first-order component is accurate, simply because there are sources of accurate information about an event other than one's own experience.
The autonoesis claim is essential to the communicative account: In making a memory claim, a subject claims epistemic authority over the event in question, and autonoesis is normally the subject's only ground for doing so. The claim might not, strictly speaking, be essential to the future-oriented account: Because autonoesis may not play a role in episodic future thinking (Perrin Reference Perrin, Michaelian, Klein and Szpunar2016), the future-oriented account might replace it with a weaker claim, namely, that although autonoesis typically plays a role in episodic remembering, it is not a necessary feature of retrieved memories (Michaelian Reference Michaelian2016b). Even this weakened claim is, however, sufficient to commit the future-oriented account to the misleadingness claim.
The construction claim: M&C understand construction as occurring through Bayesian prediction of features of past events based on evidence provided by both episodic traces and semantic information (De Brigard Reference De Brigard2014a). Alternative understandings are available (Michaelian Reference Michaelian2016b), but they concur that, at least in typical cases, not all of the content of a given retrieved memory originates in the subject's experience of the remembered event. This, in turn, implies that retrieved memories will often be, to some degree, inaccurate with respect to remembered events. But construction does not make inaccuracy inevitable: The incorporation of nonexperiential information into a retrieved memory representation, in particular, does not necessarily imply inaccuracy, simply because incorporated information may itself be accurate (Michaelian Reference Michaelian2013).
The construction claim is essential to the communicative account: If the point of making memory claims is not to convey accurate descriptions of past events but rather to convey descriptions that the subject wants an audience to endorse, a constructive memory process is needed to enable the subject to generate suitable representations of events. The claim is likewise essential to the future-oriented account: The MTT system must be able to constructively recombine and modify information from various sources in order to generate representations of possible events in episodic future thinking, and if episodic remembering is carried out by the same system, it is bound to be constructive in the same sense.
The misleadingness claim: Together, the (weakened) autonoesis claim and the construction claim imply the misleadingness claim. If the autonoesis claim is right, a memory might be false even if the event that it represents occurred exactly as the belief represents it as having occurred. In particular, the belief will be false in cases in which its first-order content originates at least in part in a source other than the subject's own experience of the event. If the construction claim is right, such cases occur frequently. Indeed, because, as M&C acknowledge, episodic remembering is driven as much by current beliefs as by episodic traces, they are the rule rather than the exception. Therefore, the second-order component of a memory belief – and the belief as a whole – will frequently be false. In short, both the communicative account and the future-oriented account are committed to the claim that episodic memory beliefs are frequently false, not because construction results in inaccurate representations of events, but rather because autonoesis results in inaccurate metarepresentations of the relationship between representations and the sources in which they originate, both where events are represented inaccurately and where they are represented accurately.
We might, in principle, attempt to avoid the misleadingness claim by rejecting either the construction claim or the autonoesis claim, but we have good reason to accept both of these claims. We might also attempt to avoid it by modifying the metarepresentational understanding of autonoesis so that the autonoesis claim says that the second-order component of a retrieved memory informs a subject only that part of the first-order component of the memory, as opposed to the first-order component as a whole, originates in the subject's experience of the event, but it is unclear whether this is compatible with the roles assigned to autonoesis by the communicative and future-oriented accounts. We may thus be forced to accept the counterintuitive conclusion that episodic memory is indeed systematically misleading.