The possible effects of the COVID-19 pandemic on the contents and organization of autobiographical memory: A Transition-Theory perspective

Abstract

The COVID-19 Pandemic is unique in its near universal scope and in the way that it has changed our lives. These facts suggest that it might also be unique in its effects on memory. A framework outlined in this article, Transition Theory, is used to explicate the mnemonically relevant ways in which the onset of the Pandemic differs from other personal and collective transitions and how the Pandemic Period might differ from other personally-defined and historically-defined autobiographical periods. Transition Theory also provides the basis for several predictions. Specifically, it predicts (a) a COVID bump (an increase in availability of event memories at the outset of the Pandemic) followed by (b) a lockdown dip (a decrease in availability of event memories from lockdown periods compared to other stable periods). It also predicts that (c) people may consider the Pandemic an important chapter in their life stories, but only when there is little continuity between their pre-Pandemic and post-Pandemic lives. Time will tell whether these predictions pan out. However, it is not too soon to highlight those aspects of the COVID-19 Pandemic that are likely to shape our personal and collective memories of this very unusual historical period.

There is the world B.C. — Before Corona — and the world A.C. — After Corona.

Thomas Friedman, New York Times, March 17, 2020

The evidence of calamity is overwhelmingly of absence…

Geoff Dyer, The New Yorker, April 13, 2020

At this point (December 2020), it is clear, at least intuitively clear, that the COVID-19 virus has created a “new normal”, a Pandemic Period. What is less clear is how well events from this period will be remembered, whether it will stand out as a major chapter in our lives, and whether its effect on memory will differ across regions, between groups or from one individual to another. It is too soon to address some of these issues empirically. However, given what we now know about autobiographical memory, it is possible to speculate about them in an informed manner. To this end, I first outline a framework, Transition Theory (Brown, 2016; Brown, Hansen, Lee, Vanderveen, & Conrad, 2012; Brown, Schweickart, & Svob, 2016), that explains how transitional events affect the contents and organization of autobiographical memory and then use this framework to discuss the ways in which the Pandemic appears to differ from other historically significant public events and to consider how the Pandemic is likely to affect our memories.

1. Transition Theory: an overview

Transition Theory provides an account of autobiographical memory that takes the Event Component (EC), as its basic unit. ECs are those concrete, identifiable elements of our lives that can be captured in a verbal description of an event (Barsalou, 1988; Linton, 1986; Morton, Hammersley, & Bekerian, 1985; Norman & Bobrow, 1979). Familiar people, locations, activities and objects are all considered ECs. Transition Theory assumes that: (a) ECs are represented in memory as individual units, (b) each memorable experience is represented as a bound set of ECs and indexed by them (Barsalou, 1988; Brewin, Dalgleish, & Joseph, 1996; Conway, 2009; Morton et al., 1985; Norman & Bobrow, 1979; Rubin, Boals, & Berntsen, 2008; Shimamura, 2014), and (c) experiences also function as Hebbian learning trials, creating and strengthening associations between co-occurring and contiguous ECs (Hebb, 1949; McClelland, McNaughton, & O'Reilly, 1995; Munakata & Pfaffly, 2004; Nelson & Shiffrin, 2013; Smith & DeCoster, 2000).

During stable periods, we spend much of our time engaged in mundane and repetitive activities (e.g., commuting, shopping, preparing meals, socializing; Kahneman, Krueger, Schkade, Schwarz, & Stone, 2004; White & Dolan, 2009). As a result, Hebbian learning processes give rise to robust, richly interconnected networks, which link frequently encountered ECs to one another. These EC networks provide the mnemonic foundation for lifetime periods, where a lifetime period is defined as a high-level memory structure that integrates period-specific knowledge (e.g., ECs) and subsumes thematically related event memories (Conway, 2005; Kubovy, 2020; Linton, 1986; Neisser, 1986; Thomsen, 2015). On this view, transitions mark the beginnings and ends of lifetime periods by causing the synchronized termination of one set of familiar ECs and/or by creating conditions that establish encounters with a new set of ECs. Over time and with repeated exposure, some of these new elements become (highly) familiar and hence come to define a new post-transitional period.

The focus on ECs enables us to think about autobiographical memory in a quantitative manner. In particular, it is useful to consider two theoretical values: The Stability Ratio (Eq. (1)) and the Continuity Ratio (Eqs. (2a) and (2b)).

$$\text{Stability}{\text{Ratio}}_{\mathrm{T}}=\frac{\#\text{Familiar}{\mathrm{EC}}_{\mathrm{T}}}{\#\text{Novel}{\mathrm{EC}}_{\mathrm{T}}+\#\text{Familiar}{\mathrm{EC}}_{\mathrm{T}}}$$ (1)

Underlying the Stability Ratio (Eq. 1) are three observations. First, at any point in time, some ECs are familiar (e.g., one's spouse, one's bicycle) and others are novel (e.g., a new colleague, a just-opened café); second, the mix of familiar and novel ECs fluctuates overtime; and third, there are times when this mix is dominated by familiar ECs and others when it is not.

Obviously, EC familiarity is continuous. However for simplicity's sake, it is assumed that familiar ECs are those that have been encountered, say, 50 times or more and novel ECs are those that have been encountered, say, l0 times or less. Also, to keep things simple, each ratio is computed over some standard unit of time (e.g., a week, a month). The Stability Ratio at Time_T approaches 1.0 during stable lifetime periods, periods during which an individual spends most of his or her time following a well-established routine and hence encounters many familiar ECs and few novel ones. In contrast, the ratio approaches 0.0 immediately following major transitions (e.g., relocation to a foreign country), when everything seems new – when novel ECs are much more common than familiar ones.

$$\text{Continuity Ratio}=1-\frac{\#\text{Abandoned}{\mathrm{EC}}_{\mathrm{pre}}+\#\mathrm{New}{\mathrm{EC}}_{\text{post}}}{\#\text{Abandoned}{\mathrm{EC}}_{\mathrm{pre}}+\#\text{Retained}{\mathrm{EC}}_{\mathrm{pre}}+\#\mathrm{New}{\mathrm{EC}}_{\text{post}}}$$ (2a)

$$\text{Continuity Ratio}=\frac{\#\text{Retained}{\mathrm{EC}}_{\mathrm{pre}}}{\#{\mathrm{EC}}_{\mathrm{pre}\cup \text{post}}}$$ (2b)

Where:

• EC_pre⋃post = EC_pre ⋃ EC_post

  • [PRE ⋃ POST – the set of all ECs from the Pre- and Post-transition periods]
• Retained EC_pre = EC_pre ⋂ EC_post

  • [PRE ⋂ POST: the set of ECs present during both the pre- and post-transition periods]
• Abandoned EC_pre = EC_pre – Retained EC_pre

  • [PRE ~ POST: the set of ECs present ONLY during the pre-transition period]
• New EC_post = EC_post – Retained EC_pre

  • [POST~ PRE: the set of ECs present ONLY during the post-transition period]

• Period_pre < Period_post

The Continuity Ratio, a simplified version of Tversky's ratio model (Tversky, 1977, p. 333), reflects the degree to which life following some identified transition (i.e., Period_post) is similar to life during some prior period (Period_pre). It does so by assessing the degree to which the ECs present during one period are also present during the other. The intuition here is that two periods will be experienced as very similar (a) when the two share many of the same ECs and (b) when neither includes many ECs that are unique to one period but not the other. When these conditions are not met, the two periods will be experienced as very different.

More specifically, the Continuity Ratio recognizes that some familiar ECs from Period_pre will be retained across a transition (Retained EC_pre)_, that others will not (Abandoned EC_pre), and that the post-transitional period should eventually give rise to new familiar ECs (New EC_post). This ratio captures two aspects of our understanding of transitions (Brown et al., 2012). First, transitions are graded; hence, the ratio is large when a post-transitional period (Period_post) is very similar to a pre-transitional period (Period_pre) i.e., when #Retained EC_pre is relatively large and the #Abandoned EC_pre and/or #New EC_post are relatively small. Conversely, this ratio is small when the two periods are very different, i.e., when #Retained EC_pre is relatively small and #Abandoned EC_pre and/or #New EC_post are relatively large. Second, transitions can affect lives through a process of addition (yielding a larger #New EC_post), through the process of subtraction (yielding a smaller #Retained EC_pre, and a larger #Abandoned EC_pre) or both. Note, although the ratio, as presented, is defined over adjacent periods, the Continuity Ratio can be computed between any two points in time.

In principle, an individual's Stability Ratio can be plotted over time to produce a Stability Profile. Fig. 1 provides an example. This profile fragment begins with a period of stability (e.g., a post-doctoral position in the US), followed by a major transition that involves both a change of country and a new role (e.g., a professorship in Canada), which in turn leads to a new stable period. In this figure, this second stable period is punctuated twice, first by a vacation in Mexico, and then by the birth of a child. The changes in the shape (from square to circle) and color (red, dark blue, light blue) of the symbols represent different sets of familiar ECs and hence give an indication of the continuity between periods. This figure illustrates the fact that there is little continuity between the first and second stable periods (resulting in a relatively small Continuity Ratio), and that the birth of a child served as a transition but not a radical one (resulting in a relative large Continuity Ratio). The Mexican vacation provides an example of a brief interlude, defined as a span of time dominated by novel ECs but which has no lasting effect on the set of familiar ECs; when the vacation is over, things return to normal; here, the Continuity Ratio computed across the vacation would be at ceiling.

Fig. 1.

An example of a stability profile presenting changes in the Stability Ratio over time.

Note: See text for explanation.

Broadly speaking, the temporal distribution of memorable personal events and the Stability Profile should be reverse images of one another. On the one hand, steep declines in the Stability Profile and long stretches of instability should give rise to memory bumps of one sort or another (e.g., the reminiscence bump, the upheaval bump, the immigration bump; Brown et al., 2016; Koppel & Berntsen, 2015; Shi & Brown, 2016). On the other, relatively few memories should be available during stable periods, i.e., stretches of time during which the Stability Ratio is high and remains steady. These points are developed further below.

2. The COVID-19 pandemic: implications for autobiographical memory

2.1. How the COVID transition differs from other transitions

Important transitions typically bring to an abrupt end one way of life and usher in another (Brown, 2016; Gu, Tse, & Brown, 2017; Shi & Brown, 2016; Uzer & Brown, 2015). In other words, they produce: (a) a sharp decline and gradual increase in the Stability Ratio and (b) a small Continuity Ratio. This general pattern recognizes that life altering transitions typically involve an adjustment phase, during which people necessarily encounter many novel ECs (i.e., #Novel EC > > # Familiar EC). With time, however, some novel ECs become familiar, resulting in a large-scale replacement of ECs across the transition (i.e., #Abandoned EC_Pre ≈ #New EC_Post; see Fig. 2 , Panel A). Consider, for example, relocation from one city to another. A transition of this sort would involve, at minimum, the replacement of one dwelling for another, one set of neighbors for and another, and one set of shops for another.

Fig. 2.

Venn diagrams representing a transition-by-replacement (Panel A) and a transition-by-omission (Panel B).

Note. In each panel, the horizontally-lined region represents the set of Abandoned Event Components (ECs), the region marked by the square grids represents the set of Retained ECs, and the vertically-lined regions represent the set of New ECs.

One of the unusual features of the COVID Pandemic is that it appears to have changed our lives by narrowing them. Simply put, at least during lockdown – and by definition – the new normal differs from the old, largely in what we cannot do, where we cannot go, and whom we cannot meet. As a result, we are experiencing a transition-by-omission (i.e., #Abandoned EC_Pre > > #New EC_Post; see Fig. 2, Panel B). A recent study by Heanoy, Shi, and Brown (2020) provides empirical support for the notion that the onset of the COVID Pandemic was experienced as a transition-by-omission. The study itself is based on a large (n = 1215) North American, web-based survey conducted in late March 2020, some two weeks after the World Health Organization declared COVID-19 a global pandemic. Of direct relevance were responses to a set of questions drawn from the Transitional Impact Scale (TIS, Svob, Brown, Reddon, Uzer, & Lee, 2014). These questions provided information about the degree of change and the type of change people experienced as a result of the COVID Pandemic. Two findings are of particular interest. First, the Cronbach's Alpha for this scale is usually in the 0.80 to 0.90 range (Svob et al., 2014; Uzer, 2020) indicating that major transitions (e.g., relocation) typically change many aspects of a person's life in a synchronized manner. In contrast, the Pandemic produced a Cronbach's Alpha of 0.60; this implies that some elements were changing and others were not. The second finding clarifies this point. Specifically, it turned out that only two of five items received moderately high scores; these concerned the activities people engaged in (mean = 3.95 on a 1 to 5 scale) and the locations where they spent their time (mean = 3.33). The other items, which concerned people, possessions and general material circumstances, produced ratings at or below the midpoint of the scale. One way to interpret these findings is to recognize that, at the outset of the Pandemic, many areas went into lockdown and many schools and businesses were moved online. As a result, people were not going where they had previously gone and were not doing what they had previously done. They were, however, still living in the same location where they had been before the Pandemic and spending time with a subset of people with whom they were familiar (e.g., family members). Thus, these data indicate that the fabric of daily life did change for many people, but not in a wholesale manner.

From a Transition-Theory perspective, then, for many of us1 , the onset of the Pandemic can be characterized by a relatively shallow decline and then a moderately steep increase in the Stability Profile (i.e., a relatively brief and shallow adjustment period). The decline was caused by the introduction of a modest number of novel ECs (e.g., online video meetings, masks, hand sanitizer) in combination with a substantial decrease in the number of familiar ECs (e.g., in-person classes, restaurant meals, social gatherings). The Stability Ratio was expected to increase rapidly because the life under lockdown tends to be highly repetitive. As a result, newly encountered ECs should lose their novelty rapidly. As for the Continuity Ratio, it should be less than 1.0 – yes, the Pandemic has been disruptive – but considerably greater than the Continuity Ratio produced by a major life-changing event (e.g., relocation to a foreign country). This moderate degree of continuity reflects the fact that, for many of us, the Pandemic has introduced a relatively small number of new EC_post; it has caused us to abandon, at least temporally, a fairly large number of EC_pre, but has not resulted in the large-scale replacement of one set of ECs for another.

2.2. Predictions: A COVID bump and a lockdown dip

This situation has implications for both the contents and organization of autobiographical memory. First, consider how the Pandemic might affect the temporal distribution of memorable personal events. Prior research has demonstrated that memorable personal events tend to “pile up” around impactful transitions (Enz, Pillemer, & Johnson, 2016; Kurbat, Shevell, & Rips, 1998; Pillemer, Goldsmith, Panther, & White, 1988; Shi & Brown, 2016; Thomsen and Berntsen, 2005) and occur more often during unstable periods (e.g., wars and natural disasters) than during stable ones (Bohn & Habermas, 2016; Brown et al., 2016; Gu et al., 2017; Zebian & Brown, 2014). This happens because people in transition (or living in unstable times) have left their old routines and are taking part in many one-off events and having many first-time experiences (Robinson, 1992; Rubin, Rahhal, & Poon, 1998). We know that the episodic distinctiveness conferred by an unusual experience predicts memorability (Hunt, 2006; Hunt & Rawson, 2011; Wagenaar, 1986). Thus, it makes sense that more events are remembered, per unit time, from the transitional/adjustment periods that separate stable periods than from the stable periods themselves. Given that we have all had a number of COVID-related first-time experiences (e.g., online teaching, home-schooling, shopping trips involving cleaning routines, etc.), it seems likely that we will observe a COVID bump.

In addition to predicting a COVID bump, there are reasons for predicting a lockdown dip – a decrease in availability of specific event memories for a lockdown period, compared to other stable periods (Uzer & Brown, 2015). First, during lockdowns, people are restricted in their activities and movements, and as a result, they have fewer opportunities to do “interesting” things. Hence, relative to stable periods, individuals living under lockdown conditions are less likely to have memorable experiences. Second, even activities introduced by or required as a result of the Pandemic (e.g., online meetings) might not be well remembered as individual events. This is because these activities are likely to be frequently repeated and as a result, rapidly schematized (Barsalou, 1988; Conrad, Brown, & Cashman, 1998; Linton, 1982; Neisser, 1981, Neisser, 1986; Rumelhart, Smolensky, McClelland, & Hinton, 1986; Schank, 1999).

One way to test these predictions is to collect two sets of freely recalled autobiographical memories, one from 2020 and the other from a subsequent “normal” year. More concretely, consider a study that requires participants, in this case first-year undergraduates in their second term of university, to recall a dozen memorable events from the previous calendar year. This method is known to produce a calendar effect (Kurbat et al., 1998; Pillemer et al., 1988; Pillemer, Rhinehart, & White, 1986; Robinson, 1986), which is the tendency to recall more events from the beginning and the end of the academic year than from other times2 .

Using this design, the expectation is that data collected from the “normal” year should display the typical calendar effect, producing pronounced peaks for June and September. The temporal distribution of event memories from 2020 is predicted to look very different, though the exact locations of the COVID bump and the lockdown dip should differ depending on when the Pandemic hit and for how long a population was locked down. More concretely, consider the situation in Edmonton, Alberta, Canada (the author's residence). Here, schools went online and non-essential businesses closed in mid-March (approximately two weeks after the World Health Organization declared the COVID-19 outbreak to be a global Pandemic) and remained closed until mid-June. Thus, compared to a typical distribution of event memories, the 2020 distribution should display an increase in the frequency of memories from March 2020, the COVID bump, and a decrease in memories from April and May, the lockdown dip. It also seems possible that the June and September peaks – the defining characteristics of the calendar effect – will be attenuated in 2020, compared to those observed following a “normal” year. This prediction recognizes that the Pandemic caused the cancellation of many distinctive, graduation-related activities and that incoming first-year students were denied a normal (and hence memorable) “freshman experience” in September 2020 because all university classes were conducted online and because incoming students did not have the opportunity to relocate to campus.

This general method – a method that elicits personal event memories from 2020 – can be extended to non-university samples drawn from different regions. The expectation is that the location of the COVID bump should reflect the arrival of the Pandemic in a particular region and that dips should occur during different months in different regions depending on when those regions were locked down. Finally, it should be noted that for some people (e.g., frontline workers, people who have lost their livelihood), the Pandemic has been a highly emotional and eventful time. These people are likely to produce a relatively large COVID bump in this event-recall task and should also produce evidence indicating that they represent the Pandemic as a historically-defined autobiographical period (see below).

2.3. The pandemic period – H-DAP or extended interlude?

Public events sometimes define lifetime periods (Bohn & Habermas, 2016: Brown, 2016; Brown & Lee, 2010; Brown et al., 2009; Brown et al., 2012; Gu et al., 2017; Zebian & Brown, 2014). These historically-defined autobiographical periods (H-DAPs) are uncommon, occurring only when a population has undergone a major collective transition, a transition that has imposed rapid, fundamental, and enduring changes in the lives of all people affected by it. Typically, then, and in contrast to the COVID Pandemic (see above), H-DAP formation involves (a) a sharp decrease in the stability Ratio brought on by some identifiable public event or process. This is followed by (b) an extended span of instability – the H-DAP itself – which in turn leads to (c) the gradual establishment of a personally-defined post-transitional period, a period that is often very different from (i.e., discontinuous with) either the earlier personally-defined autobiographical period or the more recently experienced H-DAP. Where H-DAPs do exist, they are considered important chapters in a person's life story (Gu et al., 2017), are often conveyed from one generation to next (Gu, Tse, & Brown, 2020; Svob & Brown, 2012) and frequently serve as temporal landmarks (Bohn & Habermas, 2016; Brown et al., 2009, Brown et al., 2016; Zebian & Brown, 2014).

At present, several billion people are in (or have recently been in) lockdown, self-isolation, or quarantine; all these people are living (or have been living) lives that are to some degree different from the ones they lived before the Pandemic. At least in a numerical sense, the arrival of the Novel Coronavirus has brought about the largest collective transition humanity has experienced. Does it follow that all these people will inevitably represent the Pandemic as an important historically-defined autobiographical period? In other words, in the future, will we frequently mention the Pandemic as a temporal landmark when dating unrelated events (Brown et al., 2016; Friedman, 1993; Shum, 1998)? Will we list it as an important event in our lives and include it as a major chapter in our life stories (Habermas & Bluck, 2000; McAdams, 2001; Thomsen, 2015)?

In order to predict how the COVID Pandemic might affect the organization of autobiographical memory, it is necessary to consider the degree to which it has produced a fundamental and enduring change in people's lives3 . As noted above, it appears that we have arrived at a new normal, a state of affairs that is, at least, moderately stable and somewhat different from life pre-COVID. It is also clear that despite its near universal scope, the immediate impact of the Pandemic has differed greatly from place to place (e.g., Italy with 116.5 COVID deaths per 100,000 people vs Indonesia with 7.6 COVID deaths per 100,000; Johns Hopkins Coronavirus Resource Center, 2020, December 24), from individual to individual (e.g., a bankrupted business owner vs a professor working from home), and from time to time (life during and after lockdown). What is less clear is the degree to which the next new normal – life after the Pandemic – will differ from the pre-COVID period. This is an important issue because only some transitional events are considered important and only some identifiable periods become chapters in the life story. In general, in retrospect, transitions are considered important and personal periods treated as life-story chapters only when they have shaped the trajectory of a person's life (Gu et al., 2017, Gu et al., 2020; Habermas & Bluck, 2000; McAdams, 2001; Thomsen et al., 2015, Thomsen and Berntsen, 2008). Thus, if there is a high degree of continuity between the two periods, the Pandemic is likely to have the status of an extended interlude 4 . In other words, a person whose post-Pandemic life is much the same as his/her pre-Pandemic life may well remember the Pandemic as a special time, but should not, in retrospect, rate it as being particularly important nor consider it to be a major chapter in the life story.

It will be possible to test these ideas directly by combining the TIS (Svob et al., 2014) with standard methods used to extract life stories. Again, the TIS can be used to identify those aspects of a person's life that have changed because of the COVID Pandemic. In a research setting, life stories are elicited by asking participants to provide an open-ended narrative of their lives (McAdams, 2001; Thomsen, 2009), by asking them to divide their lives into chapters (Thomsen, Pillemer, & Ivcevic, 2011) or by asking for a listing of most-important life events (Glück & Bluck, 2007). In each case, it will be necessary to collect the life-story events first and then to assess the transitional impact of the Pandemic using the TIS.

It is reasonable to expect that some people will mention the Pandemic in their life stories and others will not. If so, for reasons developed above, (a) TIS scores should be high when people include the Pandemic in their life stories and (b) people who include the Pandemic in their life stories will provide higher TIS ratings than those who do not. (Gu et al., 2017; Shi & Brown, 2016). This approach can be extended in a quasi-experimental manner to assess predictable between-group differences in the Pandemic's mnemonic impact. For example, one could compare people who have lost their homes and their jobs to the Pandemic with those who have not.

Lastly, as mentioned above, the COVID-19 Pandemic can be seen as the most extensive collective transition in history. Yet Transition Theory treats collective transitions and personal transitions in the same way. Thus the theory predicts that the Pandemic, like any other transition, will come to organize autobiographical memory only if it has changed a person's life in fundamental ways and that this should be true regardless of the Pandemic's scope, its cultural prominence or its historical importance (Brown, 2016; Brown et al., 2012; Gu et al., 2017). It will be interesting, going forward, to determine whether these social aspects of the Pandemic will also affect its mnemonic impact. Given its scope and scale, it is possible that “life during the Pandemic” will be a lasting topic of conversation and that, by convention, people, even those unscathed by it, will claim the Pandemic as an important chapter in their lives (Cohn, Mehl, & Pennebaker, 2004; Hirst & Echterhoff, 2012). If so, the collective and pervasive nature of the Pandemic may override or at least augment the basic memory processes that are thought to give rise to H-DAPs. This in turn might weaken or eliminate the predicted relation between the Pandemic's long-term transitional impact and its inclusion in the life story. This possibility can be tested by focusing on people whose lives were not radically altered the Pandemic. If social factors play an important role in determining the inclusion of a COVID chapter in the life story, then, other things being equal (i.e., holding the transitional impact constant), these chapters should be common in hard-hit areas and uncommon in areas that were relatively unaffected by the Pandemic.

3. General discussion

In many ways, the COVID Pandemic is unique; it is unique in its near universal scope; it is unique in the way that it has changed our lives; and it is likely to be unique in the way that we remember it. Transition Theory provides a language to describe at least some of this uniqueness. From a Transition-Theory perspective, the onset of the COVID-19 Pandemic can be characterized as a transition-by-omission, one that has produced a relatively shallow decline and then a steep increase in the Stability Profile leading to a moderately stable Pandemic Period. Transition Theory has also provided the basis for predicting how the Pandemic might affect autobiographical memory. Specifically, it seems likely that that the Pandemic will produce a COVID bump and, depending on location, one or more lockdown dips. It is also likely that Pandemic will create an identifiable period in autobiographical memory. Whether this period will play a central role in a person's life story will depend on its long-term impact. The prediction here is that people will consider the Pandemic to be an important chapter in their life stories only if there is little continuity between their pre-Pandemic and post-Pandemic lives. Otherwise, the Pandemic will be treated as an extended interlude.

Two final points. The first concerns the quantitative approach outlined above. In this article, the Stability Ratio, the Stability Profile, and the Continuity Ratio were introduced to highlight the special properties of the COVID Pandemic. However, these notions are grounded in what appear to be fundamental aspects of experience, aspects that have, to some extent, been underappreciated by autobiographical memory researchers, but surely play an important role in determining what we remember about our lives and how we organize what we do remember. Specifically, (a) some time spans are dominated by predictable and repetitive activities, others are filled with novel experiences and new challenges – this aspect is captured by the Stability Ratio; (b) some transitions change many facets of our lives and cause a sharp break between the past and present, others alter one or two facets, but are not particularly disruptive – differences in transitional impact are reflected in the Stability Profile and the Continuity Ratio; finally, on a larger scale, (c) some lives are tumultuous and chaotic; others placid and orderly – these different ways of being yield markedly different Stability Profiles.

In brief, these tools – the Stability Ratio, the Continuity Ratio, and the Stability Profile –make it possible to describe the texture of life at a given point in time, to assess the similarity of periods across time, and to plot the contours of a life over time. We know that change and stability, also distinctiveness and repetition, the concepts that underlie the current approach, affect the organization and content of autobiographical memory (Brown, 2016; McAdams, 2001; Thomsen, 2015); we also know that these factors affect people's mental health and physical well-being (Hobson & Delunas, 2001; Holmes & Rahe, 1967; Lundberg, Theorell, & Lind, 1975; Tennant, 2002; Wheaton, 1990). Thus, an approach like the present one that quantifies these aspects of a person's life is likely to have a wide range of research applications.

Admittedly, moving from theory to application will be a challenge – this is the second point. The challenge lies in the requirement to catalogue and classify ECs accurately, at various points in time. Fortunately, methods are available to help with this task. These include traditional approaches like social-network elicitation (McCallister & Fischer, 1978), diary keeping (Thompson, Skowronski, Larsen, & Betz, 1996), and experience sampling (Larson & Csikszentmihalyi, 2014). In addition, there exist computer technologies that capture people's activities, locations and interactions as they happen (e.g., Greaves et al., 2015; Hodges et al., 2006). There are also techniques for extracting information about people's lives from their email, scheduling software and social media (Gemmell, Bell, & Lueder, 2006; Vianna, Yong, Xia, Marian, & Nguyen, 2014). The existence of this suite of methods indicates that it is feasible to create individualized EC catalogues and to update them over time, and hence that it is realistic to believe we can acquire the data required to compute Stability and Continuity Ratios and to construct Stability Profiles.

4. Conclusion

There is no doubt that the Pandemic will be seen as an extremely important historical event, and there is no doubt that it is permanently altering the lives of a vast number of people and temporarily altering the lives of many more. And it seems equally certain that the Pandemic will leave its mark on autobiographical memory. What is less certain is how the Pandemic will be remembered. This article provides a theoretical characterization of the COVID-19 Pandemic and develops a set of predictions based on this characterization. Although it is too soon to assess the accuracy of all of these predictions, it would seem that there is much to gain by considering the memory issues associated with this world-changing historical event as it unfolds rather than in retrospect. On the one hand, the unusual nature of the Pandemic has provided the impetus for extending Transition Theory itself (cf. Brown, 2016); on the other, predictions derived from this theory call for a longitudinal approach, one grounded in an accurate and nuanced understanding of the Pandemic experience. Thus, it is clear that the time to begin studying the long-term effects of the COVID Pandemic is now.