Previously On: Prime Time Serials and the Mechanics of Memory
As is typical for me at the end of the school year, my to-do list has a pile of publishing projects that I’ve put off to the last minute. So I’ve spent the last month knocking things off the list with general success – I revised an essay on Lostpedia that will be coming out in the next issue of Transformative Works and Cultures, and contributed a short piece to a roundtable on genre for Mediascape.
But my main writing has been focused on an essay for an anthology called Intermediality & Storytelling co-edited by Marie-Laure Ryan and Marina Grishakova (now published!). I proposed to adapt my presentation given at last summer’s Society for the Cognitive Study of the Moving Image conference; as I blogged last year, my presentation explored how American prime time television copes with the challenges of cuing viewers’ long-term memories, which often catalog years of story material. Alas my presentation was oral/slide only, so I spent the past couple of weeks converting it to essay form.
Beneath the fold I’ve included the entire essay, and would appreciate any comments, as I’ll have a chance to revise before publication. I’m particularly interested if my examples, which include moments from Lost, Battlestar Galactica, Six Feet Under, Veronica Mars, Arrested Development, The Wire, and many more shows, make any sense to readers who haven’t seen the relevant programs. And as it’s written for an international anthology not primarily focused on television, I’ve included a bit of industrial and technological context that will might be fairly redundant for anyone reading a blog called Just TV.
[And a technical note for software geeks: OpenOffice drags & drops formatted text directly into WordPress! I've never had an easier time migrating a long document into this blog - OO FTW!]
Thanks in advance for any comments, and must note that unlike the rest of this blog’s CreativeCommons license, this post is ©2009 by Jason Mittell.
Previously On: Prime Time Serials and the Mechanics of Memory
In recent years, American television has embraced a model of narrative complexity that has proven to be both artistically innovative and financially lucrative. Dozens of series across genres, from comedies like Seinfeld and Arrested Development to dramas like Buffy the Vampire Slayer and 24, have explored serialized forms and non-conventional storytelling strategies such as intertwined flashbacks and shifting narrative perspectives that had previously been quite rare within mainstream American television. Serialized television has emerged as a vibrant artistic form that many critics suggest rival previous models of long-form narrative, such as 19th century novels.
Television’s poetics of narrative complexity are wide ranging. Series embrace a balance between episodic and serial form, allowing for partial closure within episodes while maintaining broad narrative arcs across episodes and even seasons. Such programs also embrace more elaborate storytelling techniques, such as temporal play, shifting perspectives and focalization, repetition, and overt experimentation with genre and narrative norms. Many contemporary programs are more reflexive in their narration, embracing an operational aesthetic encouraging viewers to pay attention to the level of narrative discourse as well as the storyworld. In all of these instances, narratively complex television programs both demand that viewers pay attention more closely than typical for the medium, and allow for viewers to experience more confusion in their process of narrative comprehension. In short, television has become more difficult to understand, requiring viewers to engage more fully as attentive viewers (see Mittell 2006).
In this essay, I want to explore how complex serials strategically trigger, confound, and play with viewers’ memories, considering how television storytelling strategies fit with our understanding of the cognitive mechanics of memory and highlighting the poetic techniques that programs use to engage viewers and enable long-term comprehension. The television medium employs specific strategies distinct from other narrative media. For instance, cinematic narratives typically engage a viewer’s short-term memory, cuing and obscuring moments from within the controlled unfolding of a two-hour feature film, while literature designs its stories to be consumed at the reader’s own pace and control, allowing for an on-demand return to previous pages as needed. The typical model of television consumption, divided into weekly episodes and annual seasons, constrains producers interested in telling stories that transcend individual installments, as any viewer’s memory of previous episodes is quite variable, with a significant number of viewers having missed numerous episodes altogether. These constraints have helped lead to a specific set of storytelling conventions and poetic possibilities that distinguish television as a narrative medium.
Before exploring television’s mechanics of memory, it is important to understand the numerous reasons why it has taken 50 years for television to broadly adopt such complex poetic possibilities. The commercial television industry in the United States avoids risks in search of economic stability, embracing a strategy of imitation and formula that often results in a model of “least objectionable content.” For decades, the commercial television industry was immensely profitable producing programming with minimal formal variety outside the conventional genre norms of sitcoms and procedural dramas. Serial narratives were primarily confined to the lowbrow arena of daytime soap operas, with more prestigious primetime offerings avoiding continuing storylines in lieu of episodic closure and limited continuity.
Economic strategies privileged the episodic form―in large part, serialized content posed problems for the production industry’s cash cow, syndication. Reruns distributed by syndicators could be aired in any order, making complex continuing storylines an obstacle to the lucrative aftermarket. Additionally, network research departments believed that even the biggest hit series could be guaranteed a consistent carryover audience of no more than 1/3 from week-to-week, meaning that the majority of viewers would not be sufficiently aware of a series’s backstory to follow continuing storylines. Coupled with the general risk-averse attitude of networks and the ongoing success of episodic programming, there was little economic rationale for television producers to undertake the risks necessary to embark on experiments in more serialized and complex storytelling.
These industrial conditions have changed over the last twenty years, with the much-reported erosion of broadcast networks’ audience size, and increased competition from cable, satellite, online video, and other media. In the wake of decreased audience shares, the industry found itself in a situation where programming risks could be justified as an attempt to discover a new model of popular programming―audience shares that would once be considered fringe or cult now qualify as mainstream hits. Cable channels could find a lucrative dedicated audience by creating programming that demanded regular viewing, and in the case of HBO and Showtime, help justify monthly subscription fees. Additionally, new technologies of recording and playback, from DVDs to DVRs to online streaming, all allow viewers more opportunities to catch-up on missed episodes. Thus the underlying contexts of television programming have transformed sufficiently to allow a mode of narrative complexity to flourish since the late-1990s (see Lotz, 2007).
As with any popular narrative mode, specific formal strategies have emerged to manage audience comprehension. These strategies have been formed through a mixture of industrial conventions and norms, and creative innovations that have shifted dominant practices. Even though the rise of cable and decline in network domination has resulted in greater risk-taking and innovation, there are still crucial assumptions that shape television storytelling. The industry still adheres to the view that viewers are rarely dedicated enough to consistently watch every episode in sequence. Thus producers are encouraged to develop strategies to fill-in narrative gaps and catch up erratic viewers. Additionally, the television industry is understandably reluctant to program a series whose narrative is so densely constructed that it is impossible for new viewers to leap in mid-series.
Along with this industrial conventional wisdom, viewers bring their own assumptions as to what they expect from narrative television. Even though other sites of evidence suggests that viewers are more likely to watch a series consistently than the industry assumes, most viewers still want to be able to miss an episode or start a series midway through a season without being alienated or confused. They also assume that there will be some means available to them to catch-up on necessary backstory, whether within the show itself or through some external site. Finally, they come to complex narratives with the expectation that mysteries and enigmas created within a series will eventually be revealed, hopefully with a satisfying resolution.
Television’s mechanisms of storytelling also provide some important constraints on how stories can be told. More than almost any other medium, commercial television has a highly-restrictive structured delivery system: weekly episodes of prescribed lengths, often with required breaks for advertisements. A given season will have a specific number of episodes, with variable scheduling for how long breaks between episodes might be―often producers cannot plan on precisely when a series will be aired or even in some extreme cases, in what order episodes might appear. Additionally, the series is consumed as it is still being produced, meaning that adjustments are often made midstream due to unexpected circumstances. Such adjustments can be due to casting constraints, as in an actor’s pregnancy, illness, or death, or feedback from networks, sponsors, or audience in reaction to an emerging storyline. Constraints like these make television storytelling distinct from nearly every other medium—a parallel would be if literature demanded the exact same page count for every chapter of every novel, regardless of genre, style or author.
Finally, a successful television series typically lacks a crucial element that has long been hailed as of supreme importance for a well-told story: an ending. Unlike nearly every other narrative medium, American commercial television operates on what might be termed the “infinite model” of storytelling―a series is deemed a success only as long as it keeps going. While other national television systems might end a successful series after a year or two, American series generally keep running as long as they are generating decent ratings. This becomes a significant issue for storytellers, who must design narrative worlds that are able to sustain themselves for years rather than closed narratives plans created for a specific run. Not surprisingly, this need to accommodate an infinite run privileges episodic content with little continuity and long-term story development, with recyclable characters and interchanging situations typical of police dramas and sitcoms.
Despite all of these constraints and norms, American television has developed a new mode of narrative complexity that pushes back against many of these limitations. One of the specific challenges that this mode has faced, with its emphasis on storyworld continuity and non-linear narration, is managing the memories of viewers. If the characters and events in the storyworld have internal coherence and continuity, then viewers need to follow along with the expansive narrative universe. When it concerns a series that is told over a period of months and years, this becomes a challenge for the mechanics of memory.
These challenges pull storytellers in a number of directions. Even though technologies and distribution systems have made it more possible to catch up on missed episodes, television producers still need to provide opportunities to fill-in gaps for viewers who may have missed an episode or two. However, they cannot be so redundant as to bore or annoy diehard fans who watch every episode, or DVD viewers who might be watching an entire season in a marathon binge. Likewise on some complex shows like Lost or Battlestar Galactica, dedicated fans might fill-in gaps between weekly episodes by reading online recaps and commentaries, keeping the events of previous weeks fresh in their minds. Thus writers need to balance between the needs of erratic and dedicated viewers.
Similarly, individual episodes need to manage the short-term memory of events that roll out over the course of the episode along with the longer term serialized recall from weeks, months, or even years beforehand. While the stereotype of the distracted television glance is less relevant today, especially concerning demanding narratives like The Wire or Mad Men, producers still need to create programs for a domestic environment that is more prone to split attention and accommodate viewers multitasking than for other media. Over the course of an episode, television narratives embed minor redundancies that remind viewers of key story information, ranging from establishing visuals locating a scene’s setting to subtle repetition of characters’ names and relationships. The entire process of narration in a television series needs to constantly reinforce story information and remind viewers of what they need to know to comprehend the next event.
Television producers have always erred toward redundancy and repetition in their narrative strategies, a tendency that was established in earlier modes of serial narrative. Before the last two decades, the primary model of serial television in America was the daytime soap opera, which developed its own conventions and norms for managing memory. As Robert Allen has explored, soaps embrace a poetics of redundancy―instead of treating repetition into a necessary evil, soaps raise it to an art form. Allen suggests that soap operas, which were designed both for dedicated fans as well as distracted and erratic viewers, derive their narrative pleasures less from the forward-moving plot of new events and developments, but more from the ripple effects of an event across the community of relationships within the drama (Allen, 1985).
The redundant narration of soap operas depends on a device that both facilitates viewer recall and the pleasures of watching character reactions to past events: diegetic retelling. Typically, a soap opera might portray a key event on-screen, although if it is a spectacle that would require a high-budget production like a car accident or disaster, it might occur off-screen. The event itself becomes less narratively important in its initial portrayal than in the chain of on-screen conversations about the event. Thus any single event might be retold through the dialogue-heavy conventions of the genre, as each character reacts to the news of hearing about the event and we witness the impact each moment of retelling has upon the characters and their web of relationships. Through this convention of recall, we are both repeatedly reminded of what happened with our attention focused on the characters and their emotional lives, making redundancy an active pleasure of the genre.
Soap operas use diegetic retelling within particular episodic structures. Typically an episode features between four and six separate storylines intercut throughout the hour, selecting between dozens of potential ongoing stories on a series. At the beginning of the episode, each storyline gets one scene to set up that day’s conversation, typically with the characters talking about some recent event and revealing some new information about how that event impacts their relationship or situation. These initial scenes are highly focused on retelling, reminding and catching viewers up about every element in the scene―previous events, relationships, settings, and even character names. As the episode progresses, the process of retelling continues, especially to remind viewers as each scene cycles back from a commercial break, but advances the plot by highlighting the new story elements rippling out from past events. At end of the episode, each scene concludes with a moment of uncertainty to prompt a series of retellings when the next episode featuring that storyline airs. While a viewer closely focused on every episode may find the level of repetition frustrating, more erratic, casual, and distracted viewers learn to use the redundancy as both a means of following the plot and enjoying the relationship-driven storytelling.
Prime time serials are far less dependent on the dialogue-based practice of diegetic retelling as a core narrative pleasure than daytime soaps, but still frequently use this traditional technique. Characters call each other by name and reference their relationships more frequently than in everyday life, using dialogue as a way to keep crucial character information active in our minds. Often past events are retold to new characters both to update them on the status of a situation and to remind us of what we have already seen. For a typical instance, early in Lost‘s fourth season episode “Cabin Fever,” a scene shows mercenary leader Keamy arriving via helicopter on a freighter with an injured man. The ship’s doctor asks, “What did this to him?” Keamy replies, “A black pillar of smoke threw him 50 feet in the air… ripped his guts out,” retelling an event spectacularly portrayed two episodes earlier (“The Shape of Things to Come”). While anyone who saw the previous episode was unlikely to have forgotten the source of the injury, this diegetic retelling reminds us of the events via naturalistic dialogue and reinforces what we have already previously seen.
This example points to an important concept in the way that viewers make sense of ongoing serials. At this point in Lost‘s run, a dedicated viewer will have watched 79 episodes over the course of four years, creating a vast array of narrative information to retain and recall. Even the most attentive and intent viewer could not possibly have all of that narrative information active in her operative working memory―most of the story information she has retained would be archived in long-term memory. When a character’s dialogue uses diegetic retelling, the viewer activates that bit of story information into working memory, making it part of her immediate narrative comprehension.1 While certainly some viewers might have been actively thinking about the smoke monster’s attack from two weeks earlier when starting “Cabin Fever,” this diegetic retelling ensures that everyone has this context active in working memory while watching the rest of the episode, as subsequent events build upon this past event to motivate Keamy’s actions to find his betrayer and return to the island.
The use of dialog to recall previous events does not have to necessary be motivated toward clarity; diegetic retelling can also work to purposefully create a sense of confusion or curiosity. As a series, The Wire tends to avoid redundancy, favoring a naturalistic mode of long-term storytelling in which viewers are often confused as to who is who and how it everything fits together. Eventually over the course of a season, the characters, roles, and systems become clear, making the process of discovery part of the show’s narrative pleasure. However, some elements are left perpetually vague; for instance, Lieutenant Cedric Daniels is introduced in the first season as an upright, career-driven police officer. This perception is undercut when an FBI officer tells Detective McNulty that there’s some dirt in Daniels’s past, and that the police department covered it up. This information is never directly addressed or fully clarified in the first season, serving as backstory on the otherwise ethical Daniels. In the fourth season, Daniels is promoted up the ranks by the new mayor with possibilities of rising to Commissioner, prompting current Commissioner Burrell to mention to his confidants, “I happen to know he’s less the saint than he pretends to be.” This casual mention is the only direct reference to the scandal until the fifth season, when Daniels’s still-vague past crimes prevent him from taking the Commissioner job. For Wire fans, the casual reference to the dirt on Daniels rewards our long-term memories from years before, but prompts the continued curiosity into the character’s enigmatic past that is never fully revealed in detail.
Diegetic retelling typically uses dialogue as a means to activate past events into working memory, but more subtle cues can also serve a similar function. As television is both an audio and visual medium, visual cues can serve the function to activate long-term memories. For instance, in the third season Battlestar Galactica episode “Maelstrom,” pilot Kara Thrace gives Admiral William Adama a figurine of a goddess to use as a masthead for his model ship; at the end of that episode, Adama destroys the model out of grief when Thrace’s ship appears to be destroyed in a fatal crash. In the next season’s episode “Six of One,” Adama is shown rebuilding the model after Thrace has seemingly returned from the grave. Lingering shots of the figurine and ship activate memories of the earlier episode, adding resonance to these character’s relationship and the mysterious circumstances of Thrace’s survival, but without the explicit expository function of dialog. Typically, visual cues are more subtle than dialog, functioning less to catch-up viewers who might have missed an episode than integrating more directly into a naturalistic style of moving image storytelling.
Producers of long-form stories constantly need to balance the needs of forward narrative momentum with the ability to keep audience’s memories activated for relevant story information from previous episodes. While diegetic strategies of dialog and visual cues are a primary means for activating viewer memories, many programs use non-naturalistic techniques to trigger recall. The use of voiceover is a common way to convey story information via a more self-conscious mode of narration. While many writers condemn voiceover as overly literary and a lazy tool for film and television, it can be used effectively in certain genres like detective shows, serving to both guide viewers within the narrative world and offer a distinctive personality to the storytelling.
The film noir infused teen drama Veronica Mars uses often sarcastic first-person voiceover narration by the titular character to both keep viewers on track with the complex story and convey the character’s perspective on narrative events. For instance, in the first season episode “Silence of the Lamb,” Veronica is helping her friend Mac grapple with the discovery that she was switched at birth with another baby. Veronica’s voiceover narration intones, “I could tell Mac I know how she feels, but the truth is, I don’t. When I had the opportunity to learn my paternity, I chose blissful ignorance with a side of gnawing doubt.” This reference to Veronica’s paternity refers to an event from two episodes earlier, as Veronica discovered that her mother had been unfaithful and she ordered a kit to test her father’s DNA, but decided not to go through with the test. While Veronica’s mysterious parentage does not become a significant plot point until later in the season, recalling this previous event helps viewers draw parallels between Mac and Veronica, and colors with way that Veronica and her father interact later in the episode.
Less commonly for television, voiceover narration can resemble the more literary model of third-person omniscient storytellers. Such narrators typically act only to frame a story, as in Rod Serling’s opening and closing narration on The Twilight Zone, but some recent series have played with third-person voiceover narration as a self-conscious device. Pushing Daisies uses the voice of Jim Dale, recognizable as the reader of the Harry Potter audiobooks, as an omniscient narrator to both present new story information and remind us of past events. In episode seven, “Smell of Success,” the narrator comments, “Chuck continued to keep the secret ingredient of her pies secret. Not even Olive Snook knew the baked secret she delivered contained homeopathic mood-enhancers meant to pry Chuck’s aunts out of their funk.” This voiceover reminds us of a plot development introduced four episodes earlier and that continued to run through the season; the reminder helps viewers remember both what is happening and who knows what about the secret pie ingredient. Given Pushing Daisies‘ highly elaborate narrative mechanics and fanciful storyworld, the omniscient narrator’s storybook style, reinforced by the intertextual link to Harry Potter, functions both to manage memories and promote a self-conscious playful tone.
An even more farcical use of third-person voiceover can be found in the sitcom Arrested Development, where producer Ron Howard narrates the action about a dysfunctional wealthy family as if he is providing deadpan commentary within a nature documentary (see Thompson, 2007). Howard’s narration constantly fills in gaps and moves the story forward, allowing the fast-paced show to cover an astounding amount of story ground in a half-hour. The narrator frequently provides a clarifying reference to a previous episode―in the second season episode, “The One Where They Build a House,” Tobias appears with blue paint on his ear, leading Howard to clarify, “Tobias had recently auditioned as an understudy for the silent performance-art trio, the Blue Man Group,” an event that occurred in the previous episode. Howard’s deadpan narration often serves to humorously undercut or comment on the character’s action, providing narrative momentum, clarifying recall, and comedic density.
Arrested Development‘s narration highlights how moving image media rely on more than just language to convey meaning―often the narrator’s comments are accompanied by images and scenes to further trigger memories and move the narrative forward. Following the comment on Tobias and the Blue Man Group, the scene shifts to a flashback of Tobias auditioning for the part. While this references an event that happened over the course of the previous episode, this scene was never shown, making it a flashback within the storyworld but adding new narrative information beyond just triggering recall. Arrested Development uses more than flashback scenes to retell past events, relying on a number of pseudo-documentary techniques for comedic effect. Later in the same episode, Michael and his son are talking about how he is no longer in charge of the family company. Howard reminds us of another event from the previous episode: “In fact, since Michael’s father escaped from prison, his brother G.O.B. had been made president.” The visuals cut to a shot of a newspaper reporting both the prison escape (complete with still photo taken from the previous episode) and the leadership succession. The scene then shifts to a conversation between Michael and G.O.B., in which they recount the events that led to G.O.B.’s presidency and the accompanying criminal investigation, all framed with the running gag of Michael disingenuously saying, “I have no problem with that,” which is even quoted in the newspaper. The effect of these narrative strategies is to combine a range of ways to prompt viewer recall while maintaining a humorous tone and self-conscious style.
Since narration is not necessarily verbal in moving image media like film and television, other techniques can be used to retell information aside from voiceover. Flashbacks are a more common technique than voiceover to incorporate previous events into an episode, and like voiceover, they can follow first or third-person focalization. A first-person subjective flashback is more common, presenting a character’s memories as cued by suggestive close-ups, subjective visuals, and special effects. For instance, in the season four Battlestar Galactica episode, “Guess What’s Coming To Dinner,” Cylon leader Natalie tells a group of humans that being rescued by Kara Thrace was their destiny. Kara watches the speech as the image begins to blur and break-up, leading into a subjective flashback of Kara being told that she is the “harbinger of death” in the previous episode. While this was an important prophecy that viewers are likely to recall, the explicit flashback both activates the memory and highlights its importance to Kara in imagining her own role in the battle between humans and Cylons. Reinforcing this line by re-showing the scene via flashback makes it more prominent in the show’s long-term mythology, which proves to be a central narrative concern in the show’s final season. Such glimpses of character’s memories via flashbacks are a common cue to trigger a viewer’s own memories, promote empathy with a main character, and frame our comprehension of an upcoming set of events.
Flashbacks can be paired with voiceover narration as a way of visualizing a narrator’s memories. Veronica Mars frequently uses this device, as we see bits from Veronica’s memories and clues about a lengthy mystery, often that we witness multiple times throughout a season. Comedies can use a similar technique, such as on My Name Is Earl, where Earl will reference a minor character we’ve met previously, and narrate a flashback comprised of earlier appearances and footage. In these instances, the voiceover typically serves as a determining thread of knowledge, framing previous scenes and cuing the relevant memories of earlier events and relationships as needed to advance the ongoing story.
Flashbacks presented from a more objective third-person perspective, or what we might call replays, are more commonly used as a way to fill in backstory rather than triggering memories―series like Lost, Jack and Bobby, and Boomtown use atemporal storytelling to craft their complex narratives, but flashbacks are rarely used to trigger memories rather than present new narrative material on such programs. Flashbacks of previously-seen events that are not framed as character memories are quite uncommon. Crime shows like C.S.I. often use replays in the context of retelling the previously-seen crime scene, but present new narrative information in the retelling, making the flashback less about memory than gap-filling. Damages and The Nine both use complex atemporal structures, portraying previous events repeatedly throughout the season and adding more information each time to string together a new storythread―again, this model of repetition is more about filling in gaps in multiple timelines rather than reminding us what we might have forgotten. Matt Hills discusses such objective flashbacks in the most recent version of the British show Doctor Who, but suggests that they function more to invite new viewers into the complex narrative rather refresh the memories of long-term fans (Hills, 2009).
A more common place to see such replays is reality television. Subjective flashbacks are a rarity, but it is quite common for a reality show replay earlier scenes and moments to refresh our memories of previous events. This technique is more consistent with the documentary style of reality television, as a subjective flashback would feel out-of-place, and the replays can be motivated as coming from the more omniscient documentary gaze. Reality television also uses more short-term replays, often returning from a commercial break by repeating the final few moments from the previous segment, or similarly starting a new episode with the final scene of the last. Scripted television sometimes adopts this technique between weeks as well, starting the new episode by replaying an earlier cliffhanger moment a bit to regain momentum and refresh viewers’ memories.
The most common examples of objective replays triggering memories within American narrative television might be within comedies. One recent trend has been the rise of the cutaway aside as a comedic technique, commonly found in animated series like Family Guy or single-camera sitcoms like Scrubs. Such asides frequently cut from the main action to an often random vignette to offset or comment on whatever just happened in the story. These asides can be fantasy sequences, unknown moments from a character’s past, or replays from past episodes. An example of the latter comes from “Kidney Now,” a third season episode of 30 Rock. Tracy tells Kenneth that he never cries, which cuts away to a montage of six moments from previous episodes showing Tracy crying. The sequence is certainly functions as a comedic aside, but builds upon our memories of Tracy’s frequent crying jags that counters Tracy’s own statement. However, the paucity of relevant examples suggests that replays are a comparatively less utilized strategy to promote memory recall.
Thus far, the strategies of triggering memories I have discussed all occur within the diegetic narration of serial television. However, television has also adopted a number of strategies outside of the core storytelling text to help manage memories. Most notably, most contemporary serials air a one-to-two-minute recap before each episode to summarize key events “previously on” the series. These recaps are generally crafted by producers, choosing key moments that they believe vital to refresh viewers’ memories for upcoming storylines and to enable new viewers to get on board with the series. While they are designed for the weekly original airings, recaps often do get included on DVDs, with some series offering the option of viewing each episode with or without recaps, while others leave them integrated into the core episode. The presence or absence of recaps can drastically change the way episodes are consumed and comprehended.
Most recaps highlight the most pertinent narrative information for the upcoming episode. For instance, the Veronica Mars episode “Silence of the Lamb,” replays three brief scenes in the recap, drawing from three different episodes, ranging back over nine weeks. The scenes capture highly expository moments―first is a two-line exchange between Veronica’s father and former sheriff Keith Mars and his successor Lamb, discussing the controversial murder case of Lily Kane, the basis of the main season-long arc. Next is the scene where Veronica and Mac meet, setting up Mac’s role in this episode’s primary plot. Finally, shots of Veronica investigating her mother’s past are overlaid with a voice-over explaining the contested paternity, which sets up the secondary plot of this episode. In just 30 seconds, the show triggers which long-term plotlines need to be activated into working memory to comprehend this episode’s developments. However, these clips would mean almost nothing to someone who had not seen most of the previous episodes, as the snippets are far too minimal to actually provide adequate exposition for new viewers. Just as notable is what the recap omits, with no reference to major characters Logan and Duncan—these characters do not appear in this episode, and thus can stay archived in long-term memory.
Recaps can serve more expository roles, especially early in a series run. The second episode of Dexter features a two-minute recap, culled exclusively from the 52-minute pilot. This recap functions as a true summary of the pilot, providing glimpses of each main character, highlighting the core narrative scenario, and establishing the ongoing arc of Dexter’s ludic pursuit of another serial killer. While it might be a bit confusing, it would certainly be possible to watch the series without viewing the pilot, filling in narrative gaps solely from this recap and other internal redundancies. For viewers who had seen the pilot, the recap seems quite redundant, offering little to cue memories aside from character names―the core narrative situation of a serial killer working as a forensic investigator is sufficiently memorable as to not need refreshing, as simply thinking about the name of the show would likely activate that basic narrative memory.
The recap from Dexter‘s first season finale is much more in keeping with the memory-refreshing role typically found later in a season. The 1:45 recap contains clips from many of the previous 11 episodes, and presents them in such quick succession that they would be incomprehensible to a new viewer. For ongoing viewers, however, the flashes of clues remind us of how far Dexter had gotten in his pursuit of the Ice Truck Killer, and the final shots of his sister in peril refreshes the cliffhanger from the previous episode. The recap also focuses on the stabbing of police office Angel from episode 10, which becomes a major plot point in the finale. More than anything, recaps like this one serve to filter the vast amount of story information that an ongoing viewer accrues, activating the most crucial bits of narrative into working memory while allowing other moments that will not become relevant in the upcoming episode to continue to reside in the archives of long-term memory.
Recaps can often trigger long dormant memories which might work to foreshadow upcoming narrative events. For instance, the recap before the fourth season Battlestar Galactica episode “Escape Velocity,” includes a scene from the third season episode “Exodus (part II)” featuring the death of Ellen Tigh. The gap between the original airdates of these episodes was over 18 months, marking this scene’s presence in the recap as unusual―at the time I first saw it, I hypothesized that the inclusion of Ellen’s death in the recap must mean that she’ll reappear in some fashion in the episode. That prediction proved correct, as Sol Tigh begins to hallucinate visions of Ellen, a connection that proves to be even more significant in the series mythology later in the season. The recap effectively reminded me about Ellen, who had certainly receded from my active memory, but also made her reappearance more predictable than it would have been within the diegetic narrative without the recap. Viewers watching the series on DVD or DVR might choose to fast-forward through the recaps, which might make Ellen’s reappearance prompt confusion or surprise, two reactions mitigated by her presence in the recap.
The strategies of recall prompted by recaps can run counter to one of the core narrative pleasures of most genres of storytelling: surprise. Within many complex long-term narratives, the deep mythology of the storyworld can be confusing and hard-to-follow without recaps to active working memories and remind us of deep-seated backstory. However, seeing a character or event in a recap can effectively “spoil” a surprise appearance or twist, undercutting the narrative effects that creators might have been hoping to produce. Clearly recaps need to balance between the dual demands of activating memories for comprehension and avoiding foreshadowing to allow for surprise to register for viewers without being confusing. Creators have devised a number of strategies for avoiding such recap spoilers.
One option is the use of diegetic flashbacks to serve as embedded recaps for viewers in the moment of the surprise itself. “Daybreak,” the series finale of Battlestar Galactica offers a good (if convoluted) example. Five characters agree to share in a complicated technological process that will share their memories with each other to facilitate a peace agreement between the warring Cyclons and humans. Prior to the procedure, Tory mentions that they may discover things that are shameful in their pasts, a protest that another character brushes aside. During the procedure, we glimpse memories in the form of flashbacks of some key moments from each character. Among these events, we see Tory confronting Cally, the late wife of Galen, another character in the memory exchange. Galen starts to focus on these memories, and we see Tory’s murder of Cally, prompting him to break from the procedure and strangle Tory. The flashbacks are to another season four episode, “The Ties That Bind,” which had originally aired 11 months before “Daybreak.”
Producer Ron Moore stated in his podcast that they intentionally “buried” the storyline of Cally’s murder, waiting for the climactic moment to payoff Galen’s revenge with high narrative stakes in the finale.2 Notably, the recap for “Daybreak” contains no reference to Cally or the murder, allowing the viewer to experience the memory along with Galen’s realization. While a dedicated viewer certainly could have recalled that Tory had murdered Cally without getting caught, it was far from active memory after 11 months and many subsequent plot machinations. For me and other viewers I spoke with, the revelation prompted a gradual surprising realization that Galen will witness his wife’s murder and the shock of his reaction. Had the recap reminded us about the murder, we would have been better able to anticipate the result of the memory meld, defusing a moment of high drama. The effect of such revelations might be called surprise memory, or the moment of being surprised by story information that you already know, but don’t have within working memory.
Surprise memory does not need to be triggered by a flashback. In the fourth season episode of Lost, “Cabin Fever,” which notably aired without a “previously on Lost” recap, Claire awakens in the jungle to discover that her infant son is not with her. She looks around for him, and we see Christian Shephard holding him. Claire looks at him with confusion and says, “Dad?” right before we cut to commercial. It had been revealed that Christian, who was introduced as Jack’s father in the first season, was also Claire’s father in the third season episode “Par Avion,” but that relationship had not been actively referenced for over 10 months. While it is surprising enough to see Christian in the woods (especially given that he is dead and previously had only appeared as an apparition for Jack), the average viewer would not likely have the relationship of him as Claire’s father in working memory until she calls him Dad and prompts this moment of surprise memory.
The practice of surprise memory highlights the importance of working memory for storytelling practice. When a long-term viewer has accrued a large amount of story information, a storyteller can guide emotional reactions based on what is in working memory―a show might highlight particular relationships and connections within working memory, or prompt surprise or suspense via elements buried in long-term memory. The feeling of being surprised through the act of remembering is quite pleasurable, rewarding the long-term viewer’s knowledge base while provoking the flood of recognition stemming from the activation of such memories. Such pleasures are hard to imagine working in non-serialized formats, as the shorter-term forms of cinema or novels do not allow sufficient time over the course of narrative consumption to enable the process of archiving and activation needed to create surprise memory.
One additional source of memory within television episodes can be the credit sequence. While such sequences vary greatly, from just a brief title card as on Lost or Breaking Bad, to a 2-minute sequence as on Deadwood or Veronica Mars. Some title sequences use footage outside the narrative, as with Tony’s drive from New York to his house on The Sopranos, with the sequence working to emphasize the setting and milieu of the show, or Dexter‘s visually stylized images of the title character preparing to go to work, highlighting the theme of finding the gruesome within the mundane. Many longer title sequences include images from the series itself, which for both episodic and serialized shows can evoke fond character moments, as with Friends or Buffy the Vampire Slayer.
One series whose credit sequence adds to mechanics of memory is The Wire. Each season offers a new montage of images of Baltimore and from narrative moments of the series, most of which have little explicit resonance within story. But some images do trigger particular memories. For instance, season four’s credits includes a brief close-up of a man putting a lollypop into his pocket. For the first four episodes, this image bears no real meaning, and seems out of place next to images of criminals, cops, and kids on the street. In the season’s fourth episode, “Refugees,” we see the image in context, as crime boss Marlo pickpockets the lollypop in an act of petty crime aimed to openly mock a security guard, who is later killed for daring to challenge Marlo. For the rest of the season, this repeated image in the credits serves as a reminder of Marlo’s arrogance and cold-blooded lust for power, highlighting how he might do anything to climb the ranks of Baltimore’s drug game and build his reputation. Through this repetition and constant reminder, we keep this minor action in working memory, consistently shading Marlo’s character.
The process of consuming television narratives plays out in a broader context than the television text itself. The television industry has devised a number of extra-textual means of helping manage viewer memory. One long-standing tradition that has been declining are reruns—for decades, networks typically played each episode of a season twice throughout the year, filling in off-times with earlier episodes. These network reruns have become less common in the 2000s, especially with DVD, DVR, and online video as methods for viewers to rewatch or catch-up on missed episodes. For instance, Lost aired with reruns over the summer and during breaks from new episodes in its first two seasons, but ABC ceased this practice in later seasons. Instead, Lost and other network shows have taken a page from cable channels, showing the same episode multiple times throughout the week, a scheduling practice that allows viewers to refresh their memories or take a closer look at an episode during the week’s gap, or catch-up on missed material.
Lost has take advantage of these multiple airings by offering what they call “enhanced versions” of episodes after the initial airing—these versions add caption annotations to the action, clarifying references and previous events. For instance, in “Something Nice Back Home” when Claire encounters Christian, the captions read: “Christian Shepard is also Claire’s father, making Jack and Claire half-brother and sister, though neither one of them know it.” Such comments certainly help refresh memories for viewers, but most diehard fans report dissatisfaction with the “enhanced” experience for being too obvious and literal in its annotations.
More commonly, serialized programs have created stand-alone texts designed to refresh memories and initiate new viewers. Lost has aired twelve hour-long compilation shows over the course of its first five seasons, with each show replaying key moments from the series along with voice-over narration retelling the narrative. Similar recap compilation shows have been used by Battlestar Galactica and The Wire, among others, often airing before the start of a new season to refresh viewer memories and invite new viewers. Compilation shows, like recaps, are quite strategic in their summaries, selecting plot threads with continued relevancy while ignoring storylines that have been resolved and made dormant within the ongoing narrative.
The rise of online video has enabled a number of other strategies for recapping. Some networks, channels, and programs have created “mini-sodes” briefly summing up previous episodes, such as NBC’s online-only “2 Minute Replays” or Rescue Me‘s “3 Minute Replays” that can be seen both online and on cable channel FX. Such replays probably function more to allow viewers who missed episodes to fill gaps, but they could also serve as memory refreshers like pre-show recaps; however, such replays are more designed to retell the entire episode rather than strategically present key story information for the upcoming episode.
More notably, a trend emerged online in 2007 with the popular YouTube video “The Seven Minute Sopranos.”3 A highly rapid recap of the previous five and a half seasons in advance of the final episodes, the humorous but affectionate fan-created video garnered over a million views and successfully promoted the final season. Producers took note of the success, and enlisted marketers to create similarly glib online recaps, such as “Lost in 8:15” and “What the frak is going on?” for Battlestar Galactica. These humorous recaps are designed for long-term fans as affectionate parodies, but they also function to effectively remind viewers of key events and highlight patterns and repetitions across the series, such as the numerous times that Carmela Soprano “gets pissed” at Tony, captured by the repeated visual of her throwing his luggage at him down the stairs.
Online recaps can be written as well presented as edited video. Network websites typically provide episode summaries for many series, but fan-created sites can serve as encyclopedic repositories of information for a complex long-form narrative. The fan-generated wiki Lostpedia is best known, but nearly every television series has a wiki where fans compile summaries and catalog events and characters, as well as using broader platforms such as Wikipedia and IMDB. The effect of this array of online media is that nearly any question a fan might have about a serialized television program can be answered by a quick Google search or perusing the show’s most active fansites, making these long-form storyworlds effectively searchable and highly documented (see Mittell, 2009).
While the primary aspect of narrative memory involves the events and characters within the storyworld, television fictions also rely on and play with viewer memory for how stories are told. As I have discussed elsewhere, narratively complex television plays with storytelling form in a range of ways, prompting an “operational aesthetic” in which viewers are simultaneously concerned with the storyworld and its telling, or in narratological terms, both story and discourse (Mittell, 2006). For viewers attending to the storytelling patterns that a series uses to convey their narrative worlds, formal memory helps frame the intrinsic norms that a series follows, and establishes storytelling expectations that can be relied upon or thwarted by creators.
One facet of formal memory involves the use stylistic cues. While television has been much more aesthetically adventurous in narrative than visual style, some series do use particular stylistic strategies as part of a show’s long-term reservoir of memory. Battlestar Galactica offers one such usage in the third season—for the first time in the series, we are taken inside a Cylon basestar as human Gaius Baltar is taken in custody by the Cylons. The scenes inside the basestar are edited with layered dissolves between shots, creating a dreamy and unreal quality to the setting. Across numerous episodes, the use of this formal pattern triggers memories of the setting and its previous events, serving as a means to create both a distinct sense of place and reinforce longer narrative arcs about the Cylon’s home.
A prime example of an intrinsic storytelling norm being used to play with viewer memories and expectations is Lost‘s use of flashbacks. Over the first three seasons, nearly every episode features a series of flashbacks focused on a single character (or couple, in the case of Sun and Jin), offering glimpses into lives before arriving on the island. These flashbacks are cued by a number of formal norms: typically starting with a close-up of the key character, followed by a “whooosh” sound and a straight-cut to the flashback. While these are not framed as character memories explicitly, they function as subjective narrative, providing story information that only one character on the island knows. As David Bordwell explores for film narration, learned intrinsic norms or schemas of a text help viewers make hypothesis, fill-in gaps, and anticipate actions, drawing upon earlier experiences to make sense of an ongoing narrative (Bordwell, 1985). For an ongoing serial, intrinsic norms are more long-term, requiring active engagement with memory. Thus when Lost viewers see and hear these cues, we draw upon our formal memories to comprehend the upcoming sequence as a flashback.
In “Through the Looking Glass,” the third season finale, Lost strategically plays with our memories to invite viewers to make a faulty assumption and prompt a spectacular surprise. The episode cues its off-island storyline as a flashback for the heroic character Jack, portraying him as drug-addled and despondent in Los Angeles. The formal devices used to present these sequences follow the norms we expect for flashbacks, leading viewers to assume that this is set in Jack’s pre-island past by activating the well-established formal memories of how Lost tells its stories. But in the episode’s final scene, we see Jack converse with fellow crash-survivor Kate about possibly returning to the island, establishing that what we’ve been seeing was actually a flash-forward. The only reason why this “narrative special effect” (Mittell, 2006) works is because of our activated memories of the show’s intrinsic norms, established over dozens of episodes throughout three seasons, highlighting the important role that formal memory plays within serial narrative.
Another more elaborated example of the role of formal memory can be found throughout the series Six Feet Under. The opening of every SFU episode portrays a death, typically of a previously-unknown character, who will be treated by family-run funeral home around which the series is centered. The first season typically portrays these deaths fairly straightforwardly, introducing the victim and showing the context of their demise. By the second season, the producers took a more baroque approach to the weekly deaths—knowing that viewers would remember the formal pattern of each episode, the show played with our expectations by delaying and feinting the deaths. For instance, the second season premiere, “In the Game,” starts with a young woman undressing in her bedroom seemingly preparing for a romantic evening, but is interrupted by a masked knife-wielding maniac breaking into her home. The scene then cuts to a movie theater, revealing that the previous actions were on-screen during a horror film premiere. We then follow the actress playing the film’s victim, as she binges on cocaine and dies of an overdose in the bathroom of the post-premiere party. Our memories of previous episodes set the expectation that this character will die, and the producers play with the “how” in spite of our learned norms.
Other episodes play on our expectations for who will die. In season two’s “Out, Out, Brief Candle,” the opening scene shows a college football practice in blistering heat, with one character collapsing from heatstroke. While the team attends to him, another player collapses and dies, confounding our narrative expectations. An even more baroque example is from season four’s “In Case of Rapture,” as a pair of young men fill inflatable sex dolls with helium to deliver to a porn awards party. Their truck almost hits a kid on a skateboard, causing the payload to come loose and rise into the air. We cut to a middle-aged woman driving a car with a bumper sticker, reading “I Brake for the Rapture”—when she sees the dolls floating up to the heavens, she exits her car and runs into traffic, yelling “I’m ready Jesus!” only to be hit by a car. The show can also offer minimalistic deaths to counter our heightened memories of previous ornate openings; season four’s “Coming and Going” starts with an old man driving his car to the funeral home, parking in the driveway, and dying simply. The broad range of variations on the show’s established formal pattern builds on established norms and offers a set of playful pleasures that depend on the long-term commitments and memories of viewers unique to serialized storytelling.
Clearly, prime time serials use a range of narrative strategies to trigger and play with viewer memories. The significance of this poetic catalog of techniques is to highlight the importance of underlying cognitive processes in the seemingly simple act of narrative comprehension. Managing a multi-year narrative universe is difficult enough for television writers, but they also face significant challenges to ensure that viewers can follow the action without falling into either confusion or boredom from redundancy. As serialized television as evolved into a robust and unique art form over the past decade, producing some of this century’s most compelling stories regardless of medium, it is vital that we recognize television’s unique narrative techniques and highlight the innovative strategies it employs that help make it a distinctive and aesthetically valid medium. If we want to understand the potential ways that long-form narratives can be constructed and consumed, we need to remember how television has offered compelling solutions for mastering the mechanics of memory.
Robert C. Allen, Speaking of Soap Operas (Chapel Hill: University of North Carolina Press, 1985).
David Bordwell, Narration in the Fiction Film (Madison: University of Wisconsin Press, 1985).
David Bordwell, “Cognition and Comprehension: Viewing and Forgetting in Mildred Pierce,” in Poetics of Cinema (New York: Routledge, 2008), 135-150.
Matt Hills, “Absent Epic, Implied Story Arcs, and Variation on a Narrative Theme: Doctor Who (2005-2008) as Cult/Mainstream Television,” in Third Person: Authoring and Exploring Vast Narratives, ed. Pat Harrigan and Noah Wardrip-Fruin (Cambridge: MIT Press, 2009), 333-342.
Amanda D. Lotz, The Television Will Be Revolutionized (New York: New York University Press, 2007).
Jason Mittell, ‘Narrative Complexity in Contemporary American Television,’ The Velvet Light Trap, no. 58 (2006): 29-40.
Jason Mittell, “Sites of Participation: Wiki Fandom and the Case of Lostpedia,” Transformative Works and Cultures 3 (forthcoming Fall 2009), http://journal.transformativeworks.org.
Henry L. Roediger, Yadin Dudai, and Susan M. Fitzpatrick, eds., Science of Memory: Concepts (Oxford: Oxford University Press, 2007).
Ethan Thompson, “Comedy Verité? The Observational Documentary Meets the Televisual Sitcom,” The Velvet Light Trap, no. 60 (Fall 2007): 63-72.
1For an overview of the cognitive understanding of memory, see Roediger et. al., 2007; for an application of cognitive theories of memory to moving image storytelling, see Bordwell, 2008.
Filed under: Academia, Media Studies, Narrative, Television, TV Shows, Viewers | 12 Comments
Tags: Arrested Development, Battlestar, Lost, memory, Six Feet Under, The Wire, Veronica Mars