I remember seeing the Grand Canyon for the first time. What's the point of remembering what that experience was like, over and above simply remembering that I've seen it? According to Mahr & Csibra (M&C), the autonoetic character of episodic memory is a signal of one's epistemic authority about the experience. It indicates credibility, which can be conveyed to others by asserting “I remember.”
M&C's proclaimed ability to account for the autonoetic character of episodic memory is exciting, and gives their account an advantage over “future-oriented mental time travel” (sect. 2.1) accounts. Future-oriented accounts are popular because of their ability to explain the persistence and pervasiveness of errors in episodic remembering. Despite their popularity, M&C rightly note that these accounts fail to make sense of episodic memory's first-person perspective. If M&C's communicative account is to supplant them, then they must offer their own explanation of memory errors. In this commentary I question whether M&C's account of errors is successful, focusing specifically on confabulation. Confabulation has an autonoetic character, yet damages one's epistemic authority.
Future-oriented accounts of episodic memory explain memory errors by downplaying or outright denying the importance of accuracy in the memory system. M&C's approach must be different. They view episodic memory as a source of epistemic authority; episodic memory must be (by and large) accurate in order to play this role. Still, they acknowledge that episodic errors are possible. M&C focus their discussion on the relationship between memory errors and a person's overall credibility. They begin by noting the bidirectional influence between episodic memory and belief. Sometimes an episodic memory forces a change in one's beliefs; other times beliefs alter a memory's content. The direction of influence is determined by the situational demands of epistemic credibility. They cite studies indicating that people alter their memories of a past event to be consistent with public reports when doing so maintains or improves their status as a witness. In other cases, where one's credibility depends on immunity to social influence, M&C note that participants often resist alterations. M&C also offer a prediction about when we should expect people to be at their most vigilant about misinformation: “If the costs of being found wrong are high, or our audience can monitor our assertions effectively, we ourselves should be more skeptical toward the outputs of our own construction system … and consequently be less likely to form a false memory” (sect. 3.1.3, para. 4).
To evaluate M&C's explanation, it helps to distinguish between different types of memory error. Elsewhere, I have argued for a distinction between misremembering and confabulation (Robins Reference Robins2016a; Reference Robins2017; see also Michaelian 2016b for an expanded alternative taxonomy). Misremembering is the distortion of retained information; confabulation is an entirely fabricated “memory.” The errors M&C discuss are misremembering errors. They focus on cases where one swaps or adds details to the report of a past event. And it seems right that these edits, even when they result in less accurate memories, will have a minimal influence on a person's epistemic credibility overall.
Confabulation errors are different. M&C discuss Loftus's misinformation studies, which provide evidence of misremembering but do not mention other suggestibility studies by Loftus and colleagues, which show that participants can be led, somewhat easily, to confabulate past experiences. Loftus and Pickrell (Reference Loftus and Pickrell1995), for example, found that many participants formed false memories of childhood experiences – events like spilling the punch at a family wedding or being lost in a shopping mall – in response to repeated interviews by researchers. Most importantly, these fabricated memories have an autonoetic character. These studies are intriguing, and disturbing, because participants report details of events that never happened, and do so from a first-person perspective.
The existence of confabulation errors is difficult to reconcile with M&C's communicative account of episodic memory. Feeling as if one remembers a confabulated event would seem to present a serious threat to one's epistemic authority. How is this tendency to generate false memories compatible with the view that the feeling of remembering is a signal of epistemic certainty?
M&C do not discuss confabulation errors. This oversight is likely due to their focus on memory's role in communication. As they note, “when we argue about the past, we often do not contest whether the event in question happened, but rather in what way it did” (sect. 3.1.1, para. 8). It is true that we do not often argue about the occurrence of past events, but this would only seem to make people's willingness to invent past events all the more troubling – and damning for the confabulator's epistemic credibility if discovered. At the very least, participants' willingness to endorse these false reports indicates a lack of the sort of epistemic vigilance M&C predict that people would display in such circumstances.
One might respond, on M&C's behalf, that the experimental situation entices participants to comply in order to appear credible. Perhaps. But it seems equally plausible to assume that participants would be on their best epistemic behavior while monitored by researchers. And even if participants feel compelled to comply, it would seem most prudent for them to accept the information without the first-person perspective, forming an event memory rather than an episodic one. By M&C's own lights, adding autonoetic character is optional, and it is mysterious why it should appear in these cases.
Confabulation has its limits. Few people can be led to “remember” being abducted by aliens, for instance (Clancy et al. Reference Clancy, McNally, Schacter, Lenzenweger and Pittman2002). But the repeated empirical demonstration that mundane confabulations are somewhat easy to induce is a threat to our epistemic authority over past events and to M&C's account of episodic memory formed on its basis.
I remember seeing the Grand Canyon for the first time. What's the point of remembering what that experience was like, over and above simply remembering that I've seen it? According to Mahr & Csibra (M&C), the autonoetic character of episodic memory is a signal of one's epistemic authority about the experience. It indicates credibility, which can be conveyed to others by asserting “I remember.”
M&C's proclaimed ability to account for the autonoetic character of episodic memory is exciting, and gives their account an advantage over “future-oriented mental time travel” (sect. 2.1) accounts. Future-oriented accounts are popular because of their ability to explain the persistence and pervasiveness of errors in episodic remembering. Despite their popularity, M&C rightly note that these accounts fail to make sense of episodic memory's first-person perspective. If M&C's communicative account is to supplant them, then they must offer their own explanation of memory errors. In this commentary I question whether M&C's account of errors is successful, focusing specifically on confabulation. Confabulation has an autonoetic character, yet damages one's epistemic authority.
Future-oriented accounts of episodic memory explain memory errors by downplaying or outright denying the importance of accuracy in the memory system. M&C's approach must be different. They view episodic memory as a source of epistemic authority; episodic memory must be (by and large) accurate in order to play this role. Still, they acknowledge that episodic errors are possible. M&C focus their discussion on the relationship between memory errors and a person's overall credibility. They begin by noting the bidirectional influence between episodic memory and belief. Sometimes an episodic memory forces a change in one's beliefs; other times beliefs alter a memory's content. The direction of influence is determined by the situational demands of epistemic credibility. They cite studies indicating that people alter their memories of a past event to be consistent with public reports when doing so maintains or improves their status as a witness. In other cases, where one's credibility depends on immunity to social influence, M&C note that participants often resist alterations. M&C also offer a prediction about when we should expect people to be at their most vigilant about misinformation: “If the costs of being found wrong are high, or our audience can monitor our assertions effectively, we ourselves should be more skeptical toward the outputs of our own construction system … and consequently be less likely to form a false memory” (sect. 3.1.3, para. 4).
To evaluate M&C's explanation, it helps to distinguish between different types of memory error. Elsewhere, I have argued for a distinction between misremembering and confabulation (Robins Reference Robins2016a; Reference Robins2017; see also Michaelian 2016b for an expanded alternative taxonomy). Misremembering is the distortion of retained information; confabulation is an entirely fabricated “memory.” The errors M&C discuss are misremembering errors. They focus on cases where one swaps or adds details to the report of a past event. And it seems right that these edits, even when they result in less accurate memories, will have a minimal influence on a person's epistemic credibility overall.
Confabulation errors are different. M&C discuss Loftus's misinformation studies, which provide evidence of misremembering but do not mention other suggestibility studies by Loftus and colleagues, which show that participants can be led, somewhat easily, to confabulate past experiences. Loftus and Pickrell (Reference Loftus and Pickrell1995), for example, found that many participants formed false memories of childhood experiences – events like spilling the punch at a family wedding or being lost in a shopping mall – in response to repeated interviews by researchers. Most importantly, these fabricated memories have an autonoetic character. These studies are intriguing, and disturbing, because participants report details of events that never happened, and do so from a first-person perspective.
The existence of confabulation errors is difficult to reconcile with M&C's communicative account of episodic memory. Feeling as if one remembers a confabulated event would seem to present a serious threat to one's epistemic authority. How is this tendency to generate false memories compatible with the view that the feeling of remembering is a signal of epistemic certainty?
M&C do not discuss confabulation errors. This oversight is likely due to their focus on memory's role in communication. As they note, “when we argue about the past, we often do not contest whether the event in question happened, but rather in what way it did” (sect. 3.1.1, para. 8). It is true that we do not often argue about the occurrence of past events, but this would only seem to make people's willingness to invent past events all the more troubling – and damning for the confabulator's epistemic credibility if discovered. At the very least, participants' willingness to endorse these false reports indicates a lack of the sort of epistemic vigilance M&C predict that people would display in such circumstances.
One might respond, on M&C's behalf, that the experimental situation entices participants to comply in order to appear credible. Perhaps. But it seems equally plausible to assume that participants would be on their best epistemic behavior while monitored by researchers. And even if participants feel compelled to comply, it would seem most prudent for them to accept the information without the first-person perspective, forming an event memory rather than an episodic one. By M&C's own lights, adding autonoetic character is optional, and it is mysterious why it should appear in these cases.
Confabulation has its limits. Few people can be led to “remember” being abducted by aliens, for instance (Clancy et al. Reference Clancy, McNally, Schacter, Lenzenweger and Pittman2002). But the repeated empirical demonstration that mundane confabulations are somewhat easy to induce is a threat to our epistemic authority over past events and to M&C's account of episodic memory formed on its basis.