A Brief History of Visual Novels

In any discussion of visual novels, there’s a good chance that the same misconceptions are going to emerge. One of the more obvious examples is conflating visual novels with dating sims, although I’m not too concerned about that one. There are enough games in both genres to prove a distinction, and more experienced fans are already careful to separate them anyway. What catches my attention is how people separate visual novels from other genres. They define these games through a very narrow set of conventions, like those Patrick Galbraith uses in this Game Studies article:

Backgrounds are static and change when [the protagonist] changes locations; they are often recycled. Onscreen text describes the place and situation.[…] [The character speaking with the protagonist] is mostly static, alternating between different poses based on the onscreen text and her reactions to the situation. As with the backgrounds, these posed character images are also recycled.

While his description isn’t inaccurate by any means (at least for the games he describes), it doesn’t capture the whole picture. Visual novels aren’t a static object. They’ve changed a lot since their inception, so limiting them to just one definition risks overlooking important historical developments. In fact, when we examine visual novel history in greater detail, we see the genre operating on at least two distinct definitions. Where earlier visual novels told their stories primarily through manga-like composition, the modern visual novel (where character portraits tell the story through theatrical presentation) was a later development. Each type had their own unique origins, and each had a significant impact on how visual novels conveyed narrative.

The early years: Adventure games and Manga

Mystery House, 1982; The Portopia Serial Murder Case, 1983

Any history of visual novels must begin with The Portopia Serial Murder Case, and for good reason: many people credit the game with inventing the genre as a whole. However, if we accept that premise as true, then we must trace the genre’s origins to early American adventure games. Yuji Horii (the game’s designer) would say as much. In an interview with Retro Gamer, he remembers adventure games as the dominant genre for narrative in games at the time. So he created Portopia as an attempt to introduce American adventure games to Japanese audiences. That’s why the game takes inspiration from games like King’s Quest and Sam’s Spade (and perhaps Mystery House). Subsequent visual novels would make the link much more apparent: in these games, you’d look at objects, talk to people, use items from your inventory, go to specific areas, etc. They were more concerned with individual puzzles than they were the story that connected them.

Although that’s not to say these games could partition off elements of their design. Because adventure games were basically a string of puzzles for the player to solve, their design necessitated a holistic approach where every aspect of the game worked toward the same end. The narrative, for instance, had to contextualize whatever puzzles the player encountered, defining the logic in a consistent enough way that the player could reasonably solve them.

Yet the most important area of influence was the visuals. The challenge in adventure games came from observing static environments and applying knowledge about the narrative in just the right way to advance the. From a gameplay standpoint, this meant that the composition would have to hide key details from the player enough that they wouldn’t notice those details right away, but not so much that the player could never find them. And from a narrative standpoint, the visuals not only had to make sense within whatever story the game was telling, but further that story by adding to the player’s understanding of it. Otherwise, the player wouldn’t be equipped to handle the challenges they encounter later in the game. Unsurprisingly, this design shows up throughout Portopia, where the player analyzes a crime scene for clues that will advance the narrative.

While visual novels would slowly fade out puzzle-oriented design in favor of narrative, all of their other structures remained intact, including their visuals. Admittedly, games like Portopia, Famicom Tantei Club, and Mindseeker used the same “sprite on background” approach that visual novels are known for today. However, they didn’t emphasize this approach like later visual novels would; the sprites wouldn’t shift across the backgrounds, and they wouldn’t move a certain way to imply action. It wouldn’t be until Tokimeki Memorial’s release (which I’ll address in the next section) that that strategy really took off. So until that time, early visual novels largely employed a strategy that eschewed sprites altogether. They told their stories not through a collection of art assets that were constructed independently of each other, but through individual panels. Like manga, each panel was composed as a whole, its angles and shapes and other visual elements coming together to create a very specific scene or convey a particular mood.

Metal Slader Glory, 1991; Snatcher, 1994

This scene from Snatcher, for example, is drawn to feel tense and exciting. The dynamic angle we view it from connotes action not only by placing emphasis on the gun, but also by leading to an imagined target just outside our sight. Gillian Seed’s stern expression only reinforces these feelings. He looks so certain about what he’s about to do that it’s as if he’s already shot the gun. Meanwhile, Metal Slader Glory emphasizes the scenery and tells the viewer to enjoy it. Yoshimiru Hoshi (the main designer behind the game) separates the major architectural features with a lot of negative space, giving each one relatively equal focus without any of them cluttering one another. He also gives Azusa an active posture connoting the same joy he expects the viewer to feel. Finally, the angle implies that the viewer is looking up at both Azusa and the scenery, as if to say that what they’re seeing is larger than life.

In any case, the art for both of these games wouldn’t allow the narrative strategies that later visual novels employ. The artist couldn’t swap characters between scenes, and even if they tried to, they’d encounter obvious limits. The character would be at the wrong angle, or they might be the wrong color for that environment, or key details (facial features, certain body parts, etc.) might not be revealed. The same would also apply to swapping backgrounds between shots. It might not match the characters’ size, or it wouldn’t make sense with whatever pose the character was in.

So rather than swap assets between scenes like later visual novels would, early visual novels instead relate individual shots to each other in sequential order, much like manga. I don’t make that comparison lightly. In fact, comic theorist Scott McCloud considered this kind of sequential storytelling to be one of the defining features of comic books/manga. He theorized that the key to understanding those media was to look at how they related panels to one another, and what those relations (“panel transitions”, as he calls them) told the reader about the story. Some of these transitions include moment-to-moment (time is seen passing between panels), subject-to-subject (the focus switches from one character to another), and aspect-to-aspect (the focus switches from one aspect of the environment to another). But the individual categories aren’t that important. What matters is the broader insight that these kinds of transitions are vital to how these two mediums tell their stories.

And along those lines, we might add early visual novels. Like manga and comic books, visual novels also rely on a combination of text and visuals to convey narrative. Therefore some of the storytelling had to take place within the visuals, since text alone wouldn’t provide sufficient context for the plot. While this doesn’t necessarily lead early visual novels toward manga-esque narrative strategies, the art these games used certainly does. Because their art was so rigid, it only made sense to relate them to each other like manga panels. So when we look at early visual novels, we see them employing the same kinds of panel transitions that comics and manga had been comfortable using for years. Examples include:

  • Jesus: Kyoufu no Bio-Monster: an aspect-to-aspect transition that switches from an establishing shot of the ship to the characters piloting it.
  • Moving Adventure Psy-O-Blade: several scene-to-scene transitions to show the character moving throughout the world.
  • Policenauts: a subject-to-subject transition reflecting the shift in conversation.

Panel transitions aren’t the only strategy early visual novels borrow from manga. Subjective motion lines show up in Dennou Tenshi at the bottom, and both manga and visual novels represent movement in similar ways, albeit for different reasons. Yet out of all the techniques early visual novels borrowed from manga, panel transitions have proven the most useful.

Akira, 1988; Famicom Tantei Club Part II: Ushiro ni Tatsu Shoujo, 1989; Hikaru Genji, 1989
Misty Blue, 1990; Metal Slader Glory, 1991; Princess Tomato in the Salad Kingdom, 1991
Phantasy Star II Text Adventures: Shilka’s Adventure, 1991; Lunatic Fantasy, 1992; Dennou Tenshi: Digital Ange, 1993

(I’d also recommend this Snatcher gallery at VGMuseum. Some of the screenshots there are fantastic.)

Dating Sims and Theatrical Storytelling

No Ri Ko, 1988; Tokimeki Memorial, 1994

Despite dating sims and visual novels being distinct terms, it’s hard to deny the influence the former has had on the latter. Unfortunately, this doesn’t mean it’s easy to trace the history of dating sims. For all the writing I’ve been able to find on the structure of dating sims and their place in a Japanese cultural/economic context, finding a straightforward history of the genre has proven more difficult. Part of this may come down to a reluctance to research a genre with strong pornographic roots. However, it may also have something to do with dating sims lacking a central history. A variety of different games, like strategy games and early visual novels, all contributed to the dating sim in some important way. Yet there’s one game that fans agree put dating sims on the map: Konami’s Tokimeki Memorial. True, it wasn’t the first game of its kind. No Ri Ko foreshadowed much of its design about five and a half years before Tokimeki’s release, and it’s possible that Tokimeki may have modeled itself after No Ri Ko.

Whatever the case may be, Tokimeki Memorial not only popularized the genre, but it also became the standard that future dating sims would look toward. Some of its influences include:

  • A focus on foregrounding ordinary, relatable settings and events over fantastic scenarios.
  • The idea of managing a collection of romantic relationships through daily activity.
  • Gameplay that consists of explicit choices, such as where to go or what to say from a set of given options.
  • Those choices determining which girl your character becomes linked with romantically, and thus which narrative thread the player follows.

The most important influence was how that gameplay affected Tokimeki’s visuals, and how those visuals would influence the art in later visual novels. Because the game lets the player take any girl to (theoretically) any location in the game, the old strategy of composing shots for specific scenes just wasn’t feasible. While Tokimeki does utilize this strategy from time to time (like here), it’s far from a common occurrence. And for good reason: designing an entire game around them would have proved cumbersome and would have likely exhausted the available memory limitations. (Keep in mind that Metal Slader Glory pushed the Famicom to its limits with this kind of design.) Narrative workarounds were outright impossible. All the game’s characters attend the same high school, and some of them even share classes with the protagonist.

So unable to create distinct images for every situation the player could encounter, Konami instead decided to compile generic assets that could be swapped out depending on the situation. Backgrounds represent generic locations like “school library” or “aquarium”, rather than any specific parts of those locations. Likewise, the characters do not inhabit these spaces, but instead stand against a backdrop. All they can do is change their expression and, on a few occasions, their poses. Although this style couldn’t present an organic world, it suited the game’s needs nonetheless. Not only could any character theoretically inhabit any space in the game, but they could also express themselves in ways they couldn’t before. Where previous visual novels could only muster a few different expressions per shot, Tokimeki’s sprites allowed Konami to create as many as six expressions per character (see: Megumi, Shiori, Yuko, Nozomi).

It is this model — portraits sitting on top of backgrounds — that visual novels have become known for. Many writers have compared this to watching anime or reading manga. Those writers aren’t wrong, but I think visual novels are better understood as stage productions. As Omar Elaasar details in “Theatre, Artifice, and the Flawed Emulation of Cinema”, theatrical games don’t try to represent their events as absolute reality, like many AAA games do. Instead, they go in the other direction. They abstract; they give the player the essence or idea of something. They’re a collaborative effort between game and player, where the former supplies a performance that the latter is expected to complete with their active imagination. For Elaasar, games like Outrun 2 and Final Fantasy IX and Black Knight Sword qualify as theatrical.

And if we return to the Galbraith quote from the beginning, we see that visual novels would also qualify. Very rarely do we see anything directly represented in these games. Actions are implied through basic gestures (if at all), and environments only represent the idea of the setting, not the setting itself. The trade-off from this arrangement is that while visual novels can no long frame a scene to create a specific effect, they encourage the player to approach the narrative in a way early visual novels didn’t. It’s because things are only suggested that the player has the space they need to put the story together as they see fit. Far from watching a narrative passively unfold before them, the player becomes an active participant in it, filling in the gaps that either the story or the visuals leave blank.

Metal Slader Glory, 1991; Katawa Shoujo, 2012

Let’s consider Metal Slader Glory and Katawa Shoujo as our examples. Granted, the former’s an extreme example; Metal Slader Glory pushed the NES to its limits, so few games would attempt the kind of animation it was capable of. Nevertheless, the strategies it employs for representing action speak to how visual novels would represent visual action. There was little abstraction at play; the visuals more or less represented exactly what they depicted. While it may only take the form of a still shot, there was never any ambiguity what was going on in that shot. In addition, the art was designed specifically around that action, meaning it had to accommodate for any action that wasn’t just a still shot. Take note of how Azusa’s head pops into the negative space between characters, for example.

These options are less available for Katawa Shoujo. Instead, it implies actions in simple ways that have come to characterize visual novels. Characters rotate toward the bottom of the screen to indicate they’re falling, or swap between portraits to indicate that they’re rustling their hair. (That’s what’s happening in that GIF, by the way. She’s rustling her hair.) These actions coincide with some line of text that supplies the context about the current situation. In any case, the game only hints at what the action is, and as Katawa Shoujo demonstrates, those hints can be too abstract to decipher on their own. It’s up to the player to use the combined text and visuals to interpret what they’re seeing. Granted, Katawa Shoujo may be just as extreme an example as Metal Slader Glory was. Where the girls of Yamaku High School are very expressive, the characters in Ame no Marginal and Amnesia: Memories are stiff in comparison. Yet all three games treat their characters’ actions in the same way: imply that the action happened. Don’t represent it directly.

This isn’t to say that later visual novels haven’t experimented with their form. Indeed, some visual novels really do feel like watching an anime or reading a novel. Sakura Taisen and Revolutionary Girl Utena: Itsuka Kakumeisareru Monogatari, two Saturn visual novels developed by Sega, go out of their way to present themselves like anime. The first achieves this by framing shots to look like they’re being watched on a television set; the second by cutting between shots to mimic the show’s cinematography. And outside these examples, we find many visual novels using the same manga-esque tactics their forerunners were known for. Where theatrics won’t suffice, they instead depict a very specific scene, framed and angled in such a way that they achieve a particular effect. Yet both within individual games and in the genre at large, manga shots are the exception, not the rule. Theatrics remain the dominant mode of communication for the modern visual novel.

Sakura Taisen 2, 1998; Revolutionary Girl Utena: Itsuka Kakumeisareru Monogatari, 1998; Ever17: The Out of Infinity, 2002
The Question, 2006; True Remembrance, 2008; The Answer, 2008
Nine Hours, Nine Persons, Nine Doors, 2010; Katawa Shoujo, 2012; Ame no Marginal, 2015

Conclusion

What I’ve said so far shouldn’t be interpreted as the definitive text on visual novels. Exceptions to these trends are inevitable, especially considering how many non-visual novel games employ visual novel-esque visual techniques (Shin Megami Tensei: Persona 3, Trauma Center: Second Opinion, Fire Emblem: Awakening, etc.). Instead, it’s better to look at this as a useful framework for understanding the genre. At the very least, what I’ve written should open up new possibilities for what visual novels can do.