2. False Memories

One might argue that episodic memories of sexual assault might stem from false memories of the sort described by Elizabeth Loftus (Reference Loftus1997, Reference Loftus2003). According to Loftus, it is possible to implant “false memories” about one’s personal past. In a “false memory,” an individual seems to remember an event that did not occur in their personal past, and the individual acquires the relevant information from the suggestions of others (as opposed to associative processes in the mind) (Otgaar et al. Reference Otgaar, Howe and Patihis2021). Loftus (Reference Loftus1997, 74) considers the extent to which imagining a possible past event will increase individuals’ confidence that the event happened in their past. Thirty-eight college students participated in a three-staged study and were asked about the likelihood of 40 possible events happening in their past (Garry et al. Reference Garry, Manning and Loftus1996). Two weeks later, participants were split into two groups and randomly assigned a set of events they were instructed to imagine and a set of events that they were not instructed to imagine, with the latter set acting as a control. In the last phase of the experiment, participants were asked questions about the specific events they were to consider in the study and were instructed to report their estimate of the likelihood that the event in question happened in their past. For many scenarios there was an increase in confidence (ibid., 212), but for events among the set of those imagined, “24% of subjects who imagined the event increased their subjective confidence that it had happened to them, while the corresponding figure for subjects who had not imagined it was only 12%” (ibid., 211). ^{Footnote 8}

In addition to pioneering studies into the psychological nature of false memories, Loftus proposed three conditions for implanting a false memory. First, an individual must become convinced of the plausibility of the relevant event. Second, the individual must become convinced that the event was personally experienced. Third, an individual must imagine the event (typically in a guided way). When these conditions are met, it’s likely that a rich but false episodic memory of the event will develop (Loftus Reference Loftus2003, 871).

One might argue that individuals who claim to be victims of sexual assault are likely to be influenced by others (e.g., therapists) in ways that intensify and distort real or imagined events or even lead to the development of false memories. In support of this claim, they might appeal to studies involving sexual assault in children with supposedly repressed memory (Otgaar et al. Reference Otgaar, Howe and Patihis2021, 3). However, it’s unclear how far findings on false memory development that only concerned relatively nontraumatic events generalize to cases of traumatic memories of sexual assault. Because these findings do not involve traumatic memories it’s not clear that they would generalize to traumatic experiences that involve greater emotional valence and normally different ways of processing (ibid., 2). Moreover, for any given case, it would be unreasonable to suspect a victim’s memory of being false in the absence of positive evidence that Loftus’s conditions for implanting a false memory have been met. ^{Footnote 9}

3. Stress and Memory

A different way one might try to defend P2 of the TUA is by appealing to findings which appear to show that victims who suffer from PTSD after a traumatic experience perform worse overall on episodic memory assessments (Forest and Blanchette Reference Forest and Blanchette2018; Zlomuzica et al. Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018).

However, victims of PTSD often remember the traumatic event veridically and highly vividly (Zlomuzica et al. Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018). First, although it has been reported that individuals suffering from PTSD are overall worse at episodically remembering (Forest and Blanchette Reference Forest and Blanchette2018; Zlomuzica et al. Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018), most studies show only a “small-to-moderate” effect in memory deficits related to PTSD and that the problem seems to lie in the acquisition of information rather than retention (an issue we’ll get into in what follows) (Brewin et al. Reference Brewin, Kleiner, Vasterling and Field2007; Samuelson Reference Samuelson2011, 347). What’s more, it has also been shown that individuals suffering from PTSD perform just as well as healthy subjects when it comes to remembering negative or trauma-related affective or emotional content (Forest and Blanchette Reference Forest and Blanchette2018, 6). Individuals suffering from PTSD in particular exhibited higher arousal when it came to negative stimuli than healthy subjects which speaks in favor of them being better able to remember negative events than happy or neutral events because higher arousal aids encoding (Zlomuzica et al. Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018; Forest and Blanchette Reference Forest and Blanchette2018). This suggests that there may not be any general memory deficit for sexual assault victims because, although negative arousal might not bear on the veridicality of episodic memory, it facilitates the consolidation of arousing aspects of the event in episodic memory that enables a more vivid and possibly detailed recollection. Moreover, when we assess the correctness of the TUA, it is precisely memories of the traumatic experience that we need to consider. While it is sometimes reported that these memories are fragmented, lack information, or lack a narrative structure (e.g., Brewin Reference Brewin2011, Reference Brewin2016), these findings have not been replicated (Rubin et al. Reference Rubin, Deffler, Ogle, Dowell, Graesser and Beckham2016; Engelhard Reference Engelhard, McNally and van Schie2019, 92). And other findings point in quite the other direction: It has been reported that episodic memorires of trauma are often more vivid and clear than other memories (Tulving Reference Tulving2001, Reference Tulving2002; Brewin Reference Brewin2015; Zlomuzica Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018, 1). Indeed, a study by Forest and Blanchette (Reference Forest and Blanchette2018) found that while there was poorer episodic memory of affectively neutral content, “[v]ictims did not show any impairment in memory for trauma-related content” (Forest and Blanchette Reference Forest and Blanchette2018, 6). In part for this reason, Forest and Blanchette (Reference Forest and Blanchette2018) warn that “impairment and deficit words” with respect to the memories of PTSD victims “must be used with caution” because “[they] should not be understood to mean that there is a pathological level of memory function” (Forest and Blanchette Reference Forest and Blanchette2018, 7).

How can victims of trauma be worse at episodically remembering neutral events but proficient at episodically remembering emotionally valenced aspects of a traumatic event? First, our memory capacity is enhanced when we are in an especially emotional situation. This is so because the stress of traumatic events boosts the activation of the amygdala and leads to stronger emotional memory (Stevens Reference Stevens, Renuka Reddy, van Rooij, Ely, Hamann, Ressler and Jovanovic2018, 653). Furthermore, our memory is enhanced especially for negative memories. Various studies show that we retrieve episodic memories attached to negative emotions with greater detail and accuracy and with less reconstructive memory-errors (Bless and Schwarz Reference Bless, Schwarz, Chaiken and Trope1999; Kensinger Reference Kensinger2007; Kensinger et al. Reference Kensinger, Garoff-Eaton and Schacter2007). This can be explained by higher activation of the amygdala during encoding and/or consolidation. For instance, the activity in the right amygdala in particular correlates with enhanced visual details of memories. Indeed, the “strong correlation between the amount of activity in the right amygdala and in the right fusiform gyrus during the encoding of negative items later remembered with specific visual detail suggests that interactions between these regions may underlie this enhancement” of episodic memory (Kensinger et al. Reference Kensinger, Garoff-Eaton and Schacter2007, 1883).

Traumatic memory is connected to high negative arousal. As it has been observed that PTSD patients are just as good as healthy subjects at remembering negative content, it is not surprising that the former’s traumatic memory is intact (Forest and Blanchette Reference Forest and Blanchette2018, 6; Zlomuzica et al. Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018; Stevens et al. Reference Stevens, Renuka Reddy, van Rooij, Ely, Hamann, Ressler and Jovanovic2018, 356).

In sum, studies appear to confirm the consistency and veridicality of traumatic memories. They also suggest a mechanistic explanation for why traumatic experiences might even be better remembered (as we will argue presently).

4. Gaps and Inconsistencies

A proponent of the TUA might argue that evidence of missing information and inconsistency with respect to the narrative structure of the remembered traumatic experience is reason to be skeptical of an individual’s testimony about that experience. Even if an individual did experience a trauma, they may not be able to recount it accurately. Moreover, there are neurophysiological explanations of the likelihood of missing information and inconsistency in the narrative of the traumatic experience, namely, the impact of traumatic stress on hippocampal function.

However, the claim that gappy memory cannot be trusted to be veridical threatens all memory because, due to a variety of constraints, we cannot possibly remember every detail of an experience. With respect to episodic memory, we typically remember only what we consider important “central” details and not other “peripheral” details. Gappiness is a feature of both nontraumatic and traumatic memory. The TUA is motivated by the intuition that traumatic memories differ from other types of memory in terms of veridicality. If traumatic memories differ from nontraumatic ones, this cannot be a function of the former’s gappiness. The proponent of the TUA must look elsewhere to impugn traumatic memories or else risk losing memory as an epistemic source altogether.

In addition, the neurophysiological explanations of the likelihood of missing information and inconsistency in the narrative of the traumatic experience need to be considered in the context of a broader neurological picture of traumatic memory. Both the hippocampus and amygdala are involved in encoding, storing, and retrieving remembered events. The hippocampus is responsible for explicit memory and coherent memory structure. Its job is to put a remembered experience into chronological order and perspective. Activation of the amygdala is quick, working on the order of tens of milliseconds and enabling instantaneous responses to danger and rapid association of stimuli with potential threats. It catalogs past sensory experiences (threats, anger) as implicit memories. By contrast, the hippocampus works more slowly when consolidating memories, on the order of minutes, hours, days, and even weeks. And unconsolidated episodic memory at early stages can be disrupted and is generally fragile, but after some time the memory becomes resilient. Studies have shown that increased cortisol levels due to stress correlates with higher activation of the amygdala during a traumatic event, and that higher stress levels degrade the function of the hippocampus (Samuelson Reference Samuelson2011; Stevens et al. Reference Stevens, Renuka Reddy, van Rooij, Ely, Hamann, Ressler and Jovanovic2018; Haskell and Randall Reference Haskell and Randall2019). This provides a plausible explanation of why victims, shortly after a trauma, will not be able to retrieve all of the information which has been encoded as well as why traumatic memories are often fragmented, incomplete, or nonlinear. However, after a few days, the hippocampus is often able to complete encoding, and a traumatic memory often becomes more fully accessible. Traumatic memories also often manifest as “flashbulb memories”—memories that are especially vivid (Haskell and Randall Reference Haskell and Randall2019). Stress can enhance memory storage such that victims of traumatic events often have full and vivid memories about the beginning of a sexual assault. Increased adrenaline is correlated with increased memory intensity and memory enhancement—memories of stressful or traumatic events are sometimes said to be “burnt” into our minds.

Another reason victims may remember the beginning of a sexual assault more accurately than other moments is that the brain encodes what one pays attention to, and one coping strategy for dealing with a sexually traumatic experience as it is occurring is distraction. Distracting oneself during the traumatic experience may explain why memories of the beginning of the experience are especially vivid, while memories of the middle or end are not. This may also help to explain why memories about sexual traumas tend to be fragmented. The upshot is that sexual and other types of trauma are likely to produce highly veridical memories of the beginnings of a traumatic experience, and the veridicality of such memories is consistent with the fact that victims may struggle to remember other parts of the assault in detail or coherently. This is also consistent with victims remembering the central facts about the full event (ibid., 21). ^{Footnote 10}

It appears, then, that the neurophysiological mechanisms underlying the encoding and retrieval of traumatic memories can explain inconsistency and spottiness in those memories in a way that does not impugn their veridicality.

5. Two Suggestions for Gauging Traumatic Memories

The neurophysiological evidence we’ve so far considered suggests that victims of trauma are not in general worse at remembering traumatic events. Not only can victims of trauma later remember new trauma-related content in the same way as nonvictims but some studies (Forest & Blanchette Reference Forest and Blanchette2018; Zlomuzica et al. Reference Zlomuzica, Woud, Alla Machulska, Lisa Dietrich, Wolf, Huston, Angelica De Souza Silva and Margraf2018) suggest that they can coherently and clearly remember their traumatic experiences as well. This suffices to undercut P2. There’s no reason to think that it is true and, in fact, some reason to think that it’s false. So, the neuropsychological evidence does not warrant skepticism about a victim’s capacity to remember traumatic events. It does, however, warrant care in the way we go about gathering information from victims about traumatic events.

First, interviewing victims immediately after a traumatic experience could undermine a victim’s capacity to receive justice in court. Due to the effects of increased amygdala activity on the hippocampus or of the time required to consolidate, what a victim can recall immediately after a trauma is likely not to be the same as what she can recall after a few days. As a result, reports taken immediately after a traumatic event may fail to match the report given in court and this may arouse suspicion in a jury (Haskell and Randall Reference Haskell and Randall2019). ^{Footnote 11} At the same time, reports taken much later are subject to the forgetting curve, severely reducing the details recalled. Ideally, reports should be taken after consolidation has occurred but before too much has been forgotten. Alternatively, and especially in the case of traumatic memories, it may be advantageous to take several reports at later intervals to see which details are retained as central. Changes in the report over time should not, by themselves, be seen as undermining the veridicality of the memory. Like gappiness, changes in detail, loss of detail, and retention only of central details over time are germane to nontraumatic memories.

Second, some questions may be better probes than others when it comes to assessing what happened on the basis of a traumatic memory. For example, during a sexual assault trial victims are often asked what they were wearing or how they behaved in the hours before the assault. This information would not necessarily be stored as trauma-related content but instead as every day or even positively valenced content. According to Forest and Blanchette (Reference Forest and Blanchette2018), traumatized victims may be worse at remembering positively and neutrally valenced content. But inability to remember such details does not impugn the veridicality of a victim's memory of the more central details of the event nor that of the details of the event that are especially negatively valenced. ^{Footnote 12,Footnote 13}