Long live the King! Beginnings loom larger than endings of past and recurrent events

Highlights

• Events are temporal figures, asymmetrically bounded by beginnings and endings.

• Lay readers show a preference for beginnings of historical epochs and events.

• Beginnings claim more attention and are viewed as more important and interesting.

• Transitional events become more salient framed as beginnings than endings.

• The ‘beginning advantage’ also applies to recurring events (the yearly seasons)

Abstract

Events are temporal “figures”, which can be defined as identifiable segments in time, bounded by beginnings and endings. But the functions and importance of these two boundaries differ. We argue that beginnings loom larger than endings by attracting more attention, being judged as more important and interesting, warranting more explanation, and having more causal power. This difference follows from a lay notion that additions (the introduction of something new) imply more change and demand more effort than do subtractions (returning to a previous state of affairs). This “beginning advantage” is demonstrated in eight studies of people’s representations of epochs and events on a historical timeline as well as in cyclical change in the annual seasons. People think it is more important to know when wars and reigns started than when they ended, and are more interested in reading about beginnings than endings of historical movements. Transitional events (such as elections and passages from one season to the next) claim more interest and grow in importance when framed as beginnings of what follows than as conclusions of what came before. As beginnings are often identified in retrospect, the beginning advantage may distort and exaggerate their actual historical importance.

Introduction

Time, in the eyes of a human perceiver, is not continuous and seamless. Prehistoric time is divided into geological periods, historical time into ages, eras, or dynasties, calendars chop it up in months, weeks, and days, and tragedies unfold on the stage in acts and scenes. Most people, looking back upon their pasts, find it natural to describe their life stories as a sequence of distinct lifetime periods (Conway & Pleydell-Pearce, 2000), or chapters (Thomsen, Pillemer, & Ivcevic, 2011). Within each chapter, they remember distinguishable episodes, often referred to as events. The changes between lifetime periods can themselves be described as transitional events (Brown, 2016). Recently, Rubin and Umanath (2015) have suggested a theory of event memory as an alternative to episodic memory for the recall of mentally constructed single scenes. In cognitive psychology, a field of event perception has emerged, particularly concerned with segmentation and identification of action episodes of relatively brief durations (from seconds to minutes) that are perceived or witnessed directly rather than being read or talked about (see Radvansky & Zacks, 2014, for an overview).

In the present article, we use events more broadly as a general label for all identifiable segments of time, from historical epochs to more specific happenings nested within the larger ones. Some of these segments are, or appear to be, objectively defined, like a journey that starts when the travelers leave home, and ends when they arrive at their point of destination. Others are more clearly the result of human observers’ attempts to make sense and impose a structure upon a temporal sequence, as for instance with historical categories such as the Age of enlightenment (Withers, 2007) and the Cognitive revolution (Baars, 1986, Leahey, 2001), whose nature, boundaries, and even claims to existence strongly depend on the perspective of the narrator.

In contrast to studies of event perception, we are in this research primarily concerned with people’s representations of temporally extended events that have taken place in the past rather than being observed in the present. Such events play an important role in structuring not only our personal histories but also the landscape of our collective history that is continually updated, changed, or reinforced by public narratives (Zerubavel, 2003).

Philosophers have suggested that events serve a similar function in the temporal domain as objects do in the spatial domain, namely as units with their own identity and their own boundaries, by which they can be distinguished from the surrounding field (Vendler, 1967). Objects in space have physical boundaries that separate them from their surroundings. Tables have sides, pictures have frames, and figures have contours that seem to “belong” to the figure, rather than to the ground (Rubin, 1915). Similarly, the protoypical historical event can be viewed as a figure, standing out against the general backdrop of a “normal”, less remarkable state of affairs (Bruckmüller et al., in press). While ordinary physical objects are supposed to have relatively crisp spatial boundaries and vague temporal boundaries, events, by contrast, are supposed to have relatively vague spatial boundaries and crisp temporal boundaries (Casati & Varzi, 2015).

Following this analysis, events may be regarded as figures in time that can be separated from what came before and what happened later. Indeed, Zacks and Tversky (2001) suggest that otherwise divergent philosophical and psychological analyses of an event converge on one basic idea, namely that all events have a beginning and an end, and that anything that has a beginning and an end in time can be regarded as an event. By this definition, we can describe a party as an event starting with the arrival of the guests and ending when they leave, or a war as an event starting with an assault or a declaration of war and ending with a victory or a proclamation of peace. Even more arbitrary partitions of time, such as the successive seasons of a year, can be described by cues marking their emergence and their disappearance. In short, these happenings would not be described as event entities unless they came into being at a specific point in time and were concluded at another, later, occasion.

Events have parts that in themselves can be described as subordinate or micro-events, and are included in more comprehensive macro-events extended over larger time spans, forming hierarchically structured “partonomies” (Hard, Tversky, & Lang, 2006). Thus, a war can be described as a fairly comprehensive event including part events like troop movements, individual battles, and peace negotiations, each with a structure of its own.¹ Beginnings and endings belong to the structure of any event, but may in turn be viewed as subordinate events in their own right, which implies, in Churchill’s (1943) words, that we can have “a beginning of the end” as well as “an end of the beginning”. In the present studies, we do not set upper or lower limits to the scope and extension of an event, but use this term to encompass all temporally defined happenings, from episodes of short duration, like the shots in Sarajevo on 28 June 1914, to long term epochs like wars and monarchs’ reigns spanning several years. Similarly, we regard the four seasons of the year as annual macro-events, which encompass more circumscribed, culturally or climatically defined events such as summer vacation, harvest, and school start. With adjacent events, such as successive reigns or the passage from one season to another, the transition itself is sometimes conceived as an event, or alternatively framed as the end of one epoch or the beginning of a new one, as expressed by the epigrammatic announcement: The king is dead. Long live the king! This traditional proclamation, used in several countries to mark the end of one (male) monarch’s reign and the beginning of a new reign, suggests that these two phases nevertheless belong together in one single constitutional act.

For a graphical illustration of events as separated or adjacent “figures”, see Fig. 1. The events might in both cases be historical or natural, and their “contours” (the beginnings and endings) can be well defined or more poorly defined, naturally given or arbitrarily imposed.

The temporal boundaries of events differ from the spatial contours of objects in that they appear in a fixed sequential order. Specifically, both boundaries of an event describe transitions, or changes. The beginning marks a transition from absence to presence of the target event, as in the announcement of the new king (who is elevated to monarch from his passive status as heir to the throne), whereas endings tell us that something has passed out of existence (literally, in the case of the deceased king). Even if both transitions may be of comparable scope and magnitude, we claim that beginnings suggest more of a contrast with the default state of affairs, than endings do, which sometimes simply imply a return “back to normal”. In other words, the “step up” from non-existence to existence implied by a beginning of an event may loom larger than the “step down” for something that simply has ceased to exist. Analogous asymmetries have been observed in other areas, as with the action/inaction asymmetry (Kahneman & Tversky, 1982) and the omission bias in decision making (Baron & Ritov, 2004), which both assume that people are more affected by what they do than by what they abstain from doing. Rozin, Fischler, and Shields-Argelès (2009) showed that additions change the nature of a product more than subtractions, suggesting a principle of “additivity dominance”. In analogy, journal editors seem to think that adding a study to a submitted manuscript would constitute a major revision, whereas removing one would only be a minor revision.

From the principles of contrast and additivity dominance, several predictions can be derived:

Beginnings will attract more attention, and often be regarded as more important than endings. Within history, we predict that beginnings of wars will focus attention more than their terminations, and that the introduction of a new cultural product (a style of dress, a school of art) will appear as more striking than the same product going out of fashion. As a result, beginnings will be given more coverage in historical accounts. Similarly, we suggest that the same event, framed as a beginning, will capture the reader’s attention more than the same event, framed as an ending (e.g., a new law introduced vs. an old law repealed).

Beginnings will also be considered more interesting than endings. This follows from theories of curiosity (Berlyne, 1960, Silvia, 2008), which see interest as related to novelty and amount of surprise. Levels of surprise are associated with the extent to which an event contrasts with the default, expected alternative (Teigen & Keren, 2003). Unexpectedness, novelty, and importance can make beginnings more vivid and memorable, and also make them beg for explanations more than endings do (Bruckmüller et al., in press). Endings may be perceived as flowing more naturally from the event itself, whereas beginnings appear to spring from causes situated outside of the events they begin.

Relatedly, people may think of beginnings as active causal forces that are more powerful than endings in bringing about further developments. Some beginnings are claimed to have “changed the course of history”, whereas endings by their very nature conclude rather than open a chapter in history. Even when conclusions allow new developments to occur, they need not be perceived as actual causal forces in their own right. When we say: “winter gives way to the spring”, spring still appears more causal than winter does.

The advantage of beginnings over endings bears some similarity to the well-known primacy effects in studies of learning (Bolhuis & Bateson, 1990), memory (Murdock, 1962), persuasion (Hovland, 1957), and impression formation (Jones & Goethals, 1971), where first items in a list are recalled better than later ones, and first arguments and first encounters are shown to have a stronger impact on subsequent thoughts and behaviour than later experiences of similar kinds. These effects typically refer to situations where individual facts or arguments are presented and experienced sequentially by the same individual. Weakened encoding of later items have been explained as an outcome of processes such as interference and neural fatigue (Tulving, 2008). Similarly, autobiographical memories cluster at the beginnings of life ‘chapters’ (Thomsen et al., 2011, Fig. 6, Fig. 7), and college students recall many more events from the beginning of the term than from later in the academic year (Pillemer, 1998). Going beyond such findings we examine events whose beginnings and endings are not personally relevant but historically important, and examine if people also prioritize historical transitions to a greater extent when they are framed as beginnings than as endings.

Of course, we do not claim that beginnings are inherently the most important part of any event, as a time period could have its peak moments early, in the middle, or late, depending on circumstances. For instance, a specific war could be remembered for the moment the tide of the battle turned, and the summer season could be remembered for its hottest day. Task requirements (e.g., a writing assignment) might activate a search for representative aspects of an event, which perhaps are more often found in the middle than in the beginning. In other cases, specific motives, such as self-relevance or an interest in outcomes, might direct attention towards features more strongly related to the end. For instance, people seem to be more accurate in timing endings than beginnings when watching brief action episodes (Lu, Harter, & Graesser, 2009), suggestive of an “end-state bias hypothesis”. We will return to the question of boundary conditions for the beginning advantage in the General Discussion.

Our proposal is more general, namely, that people, in the absence of specific case information, expect beginnings to be more memorable and intriguing by virtue of their temporal position than they would have been otherwise. This postulated beginning advantage² could manifest itself in two ways: (1) As a preference for information about the beginning rather than about the ending of the same event. For instance, people might prefer reading about why a war started to reading about why it came to an end, and they might have more to say about the beginning of a season than about its termination. (2) As a preference for the same transition framed as a beginning of a new event rather than as the ending of a previous one. For instance, 1914 should more likely be described as the year World War I began than “the year peace ended” (MacMillan, 2013), and December 31 is better described as New Years’ Eve than as Old Years’ Night.

The current studies investigate segmentation of time in two different domains, namely in lay people’s representations of historical epochs and annual seasons. The concept of an event holds a central role in both domains, both with respect to its identity (Was there really a revolution? Did we have summer this year?) and to its boundaries (When did it start?).³ There are also important differences. Historical events are typically presented chronologically and exemplify a linear conception of time, with individual particular happenings stringed in a non-repeating sequence from more remote to more recent events. In contrast, people’s calendars illustrate a cyclical conception of time, where weekly and yearly events recur in a periodical fashion. Moreover, historical events are to a large extent believed to be a product of human activities, whereas seasonal events are considered natural.

We report eight studies where the proposed preference for beginnings over endings was put to the test. In doing so, our primary endeavor has been to establish the beginning advantage as an empirical phenomenon that can be demonstrated in several domains, rather than to exhaustively account for all factors that might be responsible for it. The first five studies are devoted to cognitions about relatively unique events in human history and the last three concern naturally reoccurring events (annual seasons).

We begin the studies on historical events by examining milestones displayed in popular historical timelines (Study 1), expecting beginnings to be more frequently highlighted than endings in such lists. People’s thoughts about the impact of historical beginnings versus endings can be studied in two ways: By comparing beginnings and endings of a particular event or a specific historical period, as indicated by B₁ and E₁ (or B₂ and E₂) in the upper panel of Fig. 1. Alternatively, the same transition between two periods can be construed as an ending of one period or as the beginning of another, allowing us to compare E₁ with B₂ as in the lower panel of the figure. The first of these approaches is pursued in Studies 2 and 3, where participants (non-historians) were asked which dates of wars and reigns of monarchs are important to remember, and which part of a historical period they would rather read about. The second approach is explored in Studies 4 and 5, where people were given lists of transitional events (US presidential elections and milestones in European history) that can be framed as endings or as beginnings.

Beginnings are sometimes preferred for what they promise or inaugurate. To control for a positivity bias all studies included both positively and negatively valenced events. Historical beginnings may further be preferred to endings because of their continued relevance for the present. Self-relevance is a central determinant in autobiographical memory (Conway, 2005), availability (e.g., Gregory, Cialdini, & Carpenter, 1982), and attribution (e.g., Ross & Sicoly, 1979). In historical accounts self-relevance manifests itself as presentism, where those aspects of the past that fit one’s current picture of the world are retained and emphasized, and those that do not are downplayed or ignored (Fischer, 1970). In line with this, collective memories of historical events are biased towards events in the participants’ own country and recent events (Liu et al., 2005). Despite the fact that endings of events are, by definition, more recent than beginnings of those same events, endings describe states of affairs that may be less self-relevant than beginnings because they refer to events that have been replaced by other, still more recent events. Beginnings, on the other hand, might have inaugurated societal changes and practices that are still in force in the present. To control for this alternative explanation, the studies included events that are remote in time and of minor relevance for the present.

The last set of studies informs questions about order effects by investigating the beginning advantage in the way people talk and think about the recurring seasons of the year. Seasons are recurrent events where endings are not final, but happen every year, and the ending of one season may be coextensive with the beginning of the next, balancing potential order effects. We predicted that people have more to say about the beginning of a new season than about the ending of the previous one (Study 6), that they find beginnings more causal than endings (Study 7), and finally, that they think of the beginning as defining the season more strongly than the ending, by preferring to speak of beginnings rather than endings as taking place surprisingly “early” or “late” (Study 8). An overview of the studies reported in the paper is given in Table 1.