Defining Dialogue SystemsBy Brent Ellison
[In an in-depth Gamasutra analysis piece, Ellison looks at the universe and history of player-NPC dialogue interaction in games, analyzing titles from Mass Effect through Facade to The Sims and beyond.]
In 1966, Joseph Weizenbaum created ELIZA (pictured below), a computer program designed to emulate interaction between the user and an artificial therapist. Ever since, designers of interactive entertainment have attempted to incorporate meaningful interactions with virtual characters in order to aid immersion.
The most popular Western RPGs, like the Baldur's Gate and Fallout series, live and die by the strength of their dialogue and the player's ability to influence the non-player characters. Japanese romance games such as Konami's Tokimeki Memorial and visual novels like the Phoenix Wright series revolve almost entirely around character interactions.
Even action titles like Grand Theft Auto: San Andreas incorporate elements of non-violent character interaction to make the world come to life. However, while a great deal has been written about the process of creating game characters and writing for games, very little literature has addressed the mechanics behind character interaction in games.
While likeable design and well-written dialogue are among the most vital aspects in engaging a player with in-game characters, the systems of interaction certainly play an important role as well.
NPC interaction gameplay is a part of nearly every modern game, in the form of dialogue, barking orders at followers, or just choosing who to talk to in town. However, the western games industry has always found it difficult to create successful games based solely around character interactions, even though such things are popular in other media (television, novels, etc.).
The difficulties in interactive conversation lie in giving the player the illusion of freedom while still feeling natural and driving the story forward along interesting paths. Finding the most interesting and engaging way for a player to interact with game characters and develop relationships potentially opens up a wide array of game concepts and themes not typically explored by western game developers.
NPC interaction gameplay has seen numerous permutations throughout the years, but most can be separated into a few categories based on common design patterns. Gameplay varies with regard to level of interactivity, interface, and how the game presents the player with his potential responses. What follows is an overview of the gameplay of character interaction to date.
In games using non-branching dialogue, the simplest form of interaction, the player walks up to an NPC and initiates conversation. The NPC delivers his or her lines and the conversation ends. Alternately, initiating a conversation with an NPC triggers a cutscene where the player's avatar and the NPC have a non-interactive dialogue.
If the player talks to the same NPC again after certain events, the NPC may have different things to say, but the player never has any control over the conversation. The only decision-making involved is whether or not to talk to the NPC.
This type of interaction is extremely common and easy to implement, and can therefore be seen in almost any game featuring non-hostile NPCs, such as the original Final Fantasy and The Legend of Zelda: Ocarina of Time. While non-branching dialogue is still a regular facet of modern games, due to the absence of gameplay it is included in this article only for reference.
Game dialogue becomes more interactive when conversations can take different paths. The player reads dialogue and chooses their response from a limited set of choices available to them. Conversation typically moves forward such that the player cannot go back to previous topics or responses.
From this basic framework, NPC interaction can be as simple as the player answering a yes or no question in a three-line conversation with a random NPC in town, or as complex as the relationship-building simulations in Japanese dating games like Tokimeki Memorial.
In games where branching dialogue is the primary gameplay focus, the player's choices often affect the NPCs' attitudes toward the player one way or another, with the player attempting to guess the "best" response in order to maximize NPC disposition.
One common technique employed to give the player a greater illusion of freedom is to have multiple responses lead to the same path. This is usually done as an attempt to limit the quantity of dialogue that must be produced for the game. Therefore, branching dialogue usually curves back in on itself such that while an individual choice may immediately produce a unique response, the rest of the conversation is typically not unique to that choice.
In games where the goal of conversation is to improve the player's relationship with the NPC, however, while every choice may not change the course of the dialogue, each choice may have a different number of "mood points" associated with it, and thus the player must still carefully consider his options at every turn.
An interesting variation in branching dialogue, found in the game Culpa Innata, sees the player choosing one of several tactics at the beginning of the conversation. For example, while interviewing a potential witness in a crime, the player chooses to use either a formal, casual, or accusatory manner.
This decision affects the tone of the conversation, the options, and, ultimately, the information gleaned from the interviewee. Some tacts are more successful than others, and the player cannot immediately go back and try a different method. In order to approach the conversation differently, the player must come back another day.
The interface used by the player in branching dialogue varies significantly from game to game. The most obvious method is to present the player's possible responses word-for-word as his or her avatar would say them. The player has an infinite amount of time from which to make his decision, and the NPC gives his or her reaction as soon as the player makes his choice.
This is the case with most dating simulations and many western RPGs like Planescape: Torment. There is no ambiguity in the player's decision, but reading all the possible responses takes time and brings conversation flow to a halt.
While this is not necessarily a problem in games featuring dialogue presented entirely in text, modern games typically present all dialogue with full voice acting. As a result, the menu navigation and long pauses while the player chooses their next response can negatively impact immersion.
Games have attempted to address this issue by presenting responses in a different manner, or controlling the speed at which the player responds. In some cases, designers choose to present full responses alongside symbols or text that quickly sum up the general gist of the response.
An example of this technique can be seen in Desperate Housewives: The Game (Liquid Entertainment 2006). The game presents the player's options as a series of stylized faces depicting various emotions associated with the response, such as happy, sad, or angry. When the player moves their mouse over one of the faces, the full response appears in a text bubble for the player to read if they wish.
If the player knows they want to respond to a particular comment in a negative manner, they can quickly filter out the responses they do not want and just read the angry choices, thereby reducing the load on the player and speeding up dialogue to a more natural level.
One game notable for its aspirations for cinematic-quality dialogue scenes is Indigo Prophecy, which eliminates full responses entirely. The player sees his choices pared down to their essence, such as "Lie", "Avoid the question", or "Ask about murder weapon."
Once the player decides, his avatar delivers the full line related to the player's choice. In addition, the player has only a limited time in which to make a decision after the NPC finishes speaking. If the player fails to make a decision in that time, the game chooses a response for them, in a deliberate attempt to keep the conversation moving at a more natural pace.
Quantic Dream/Atari's Indigo Prophecy
The recent Mass Effect makes similar attempts at simplifying the presentation of the player's choices, but rather than limit the player's response time, it gives the player his options before the NPC finishes speaking. In this manner, the player makes his decision and the avatar delivers a response with little to no pause in the conversation.
Thus, both Indigo Prophecy and Mass Effect attempt to make conversations more natural by reducing the amount of time and effort the player spends considering their next response.
Although the heavily scripted nature of branching dialogue allows designers and writers to craft natural, flowing conversations, the limited nature of interactivity is very transparent to the player.
It is easy for players to see that they are simply choosing from paths laid out for them, rather than creating their own story. Further, players may be frustrated that they must follow such a straightforward path and cannot choose to inquire about certain topics. The Hub-and-Spokes Dialogue method addresses some of these problems.
Though ultimately a variation of the previous method, Hub-and-Spokes Dialogue creates a very different conversation flow compared to basic Branching Dialogue. The player listens to the NPC's lines and then chooses their response from the main "hub" of the conversation.
After hearing the NPC's response, the player either returns to the main hub, from which they can ask the same question again or inquire about another topic, or enters a deeper hub with more options to choose from.
The player can typically always find their way back to any hub by navigating through their responses, and thus can explore the dialogue in any order they wish. In this manner, a player can exhaust a conversation by trying every possible option at their disposal (with no penalty), and the interaction only ends when the player chooses the "goodbye" option.
Most conversations in Mass Effect and other BioWare titles take this form, with occasional basic Branching Dialogue implemented when the player has to make an important decision that may affect quest outcomes or the NPC's disposition towards the player.
Hub-and-Spokes Dialogue gives the player more freedom and control over conversation and often allows them to interrogate NPCs to find out every last piece of information about them. However, this method of dialogue tends to create conversations strongly divorced from reality.
The NPC usually has infinite patience for the player's strange inquisitions, and every dialogue plays out like an interrogation as the player keeps pressing the NPC for info. Furthermore, the player hears a lot of the same lines over and over as he navigates between hubs, potentially breaking immersion.
While the original conversation simulator, ELIZA, used a text parser to get player input, this method is relatively rare today. Some experimental titles like Façade, however, still explore its possibilities.
In a parser-driven dialogue, players type their exact response and the system attempts to parse the input in a way it can understand. The NPC then replies with one of a number of pre-set responses, or builds a response based around the words used by the player in combination with pre-set phrases. In many cases, the player directly controls the flow of conversation, veering wildly off-topic whenever they wish without eliciting much surprise from the NPC.
Façade's design, however, always attempts to keep the player on track. In this game, the player plays an active part in dealing with marital troubles as a third-party helping a couple work through their problems.
Although input from the player comes from a text parser, Façade tries to keep conversations as natural as possible by choosing a response based on not only the player's immediate input, but also their behavior thus far and the current state of the dialogue.
Procedural Arts' Façade
In Façade, the conversation always moves forward, despite the player's best attempts to change subjects, as the NPCs always try to return to the subject at hand. The conversation can go many different ways, and a message after the game ends encourages the player to replay the game many times.
Parser-Driven Dialogues are rare in modern games for two reasons. The first is that the freedom they provide is extremely time-consuming to produce. The system needs hundreds of potential responses to accurately simulate a single, short conversation.
The second reason is that even the most robust parsers frequently misinterpret the player's input. In Façade, an innocent inquiry can send the NPCs into shock, horrified by what they thought the player just said. These misunderstandings ruin virtual relationships and frustrate the player, while at the same time exposing the program's failings and distracting the user from the interaction.
Some games do not feature full conversation systems, but merely simple means for interacting with characters that pass by. Often, interacting with NPCs in this way is part of the same gameplay used for exploration, combat, etc. The game does not change into "conversation mode" when it comes to NPC interaction.
In many cases, the player chooses a gesture or attitude to take when initiating interaction with an NPC. The NPC responds briefly, and sometimes the player has the opportunity to choose another gesture or attitude.
Often, the player rarely has more than two interactions with a given NPC in one of these encounters, which are typically carried out for the purpose of an immediate reward, such as a health boost, like in Bully or a temporary bodyguard, such as in Grand Theft Auto: San Andreas.
It is possible for these interactions to have meaning over the long term, however. In Fable, romances are handled by this same sort of system.
The player builds up their relationship with a potential partner by returning to them again and again over the course of the game, choosing the appropriate gestures and giving gifts.
Eventually the player reaches the point where he can propose to the NPC, at which point the NPC sometimes gives gifts to the player.
The transparent nature of Systemic Interactions is the most obvious problem with this method, since the player receives constant reminders of the exact nature and purpose of his interactions, making immersion difficult to achieve.
Another common issue seen in games using this method, though not necessarily inherent to the system, is that NPCs, even the romantic interests in Fable, typically have no real character. Interactions with them are superficial not only due to the nature of the interaction, but because there is nothing to learn about them.
None of the characters have unique responses; every NPC with the same character model has the same voice and dialogue. Bully has several unique characters that express their personality through their responses, but interactions remain shallow.
The implementation of Systemic Interactions in a game is usually made in order to allow the player to deal with a great many NPCs, so making those interactions meaningful is a difficult challenge. The Sims is an example of a system similar to this type that has had great success in this area.
Some NPC interaction gameplay is unique, or difficult to categorize. The Sims is a notable example as the best-selling PC game of all time. While The Sims' basic interactions are very similar to Systemic Interactions, the context is significantly different due to the way the player has control over multiple characters and can control both sides of a relationship.
In The Sims, the player chooses which character to control, and then issues simple commands like "compliment," "brag," or "flirt."
Characters in the game speak nonsensical gibberish, however, so their dialogue reflects only their mood or their emotional response to the topic and the character they converse with; it conveys no other information to the player.
However, using a fictional language avoids the issue seen in many games where the player must endure countless repetition of the same few lines of dialogue.
While the player eventually comes to recognize repeated snippets of gibberish, their nonsensical nature makes them much more palatable.
Conversations in The Sims serve to improve or worsen relationships between characters, with new options opening up on either end of the social spectrum. The Sims also has a more organic form of the Time Scheduling system (described in the next section), as relationships with other Sims worsen when the player ignores them, and the player must determine how they want their player to spend their time.
There are no scripted relationship arcs in The Sims, so the player has complete control over their development and outcome. In this manner, the developers save hundreds of man-hours in writing dialogue while still engaging players with the characters.
The popularity of The Sims and the emotional attachment some players develop may indicate that the freedom of the game leads to engagement, but players tend to make these attachments to characters under their direct control, and therefore these characters are not necessarily NPCs.
Another method of character interaction in games involves the use of minigames to simulate certain aspects of conversation. In The Elder Scrolls IV: Oblivion and Nancy Drew: The Deadly Secret of Olde World Park, the player plays a minigame to improve the NPC's disposition, but any dialogue produced during the minigame consists of stock phrases that simply serve as feedback to the player's performance rather than giving any insight into the character.
These minigames do affect the course of conversation, however, by giving the NPC new responses to topics or opening up new avenues depending on the result of the minigame.
In contrast, the minigames in Leisure Suit Larry: Magna Cum Laude take place during a conversation, and dynamically affect its course. If the player makes a mistake in the minigame, his avatar says something inappropriate that upsets the NPC, whereas if the player continually performs well, Larry (the player character) impresses the NPC, improving their relationship.
Of course, playing a minigame makes it difficult for the player to pay full attention to the dialogue in this case, and makes the ultimate result (pass or fail) relatively transparent.
While not a method of direct NPC interaction, Time Scheduling systems are relevant to this article both due to their importance in games centered around character interaction, such as Tokimeki Memorial, and because of the way they provide context and relevance to NPC interactions.
In most games, the player chooses when and if to interact with NPCs, but some games build the entire framework of their game around managing time spent with different characters. Typically, these games take place over a defined period of time, during which the player chooses their activities. For example, in a given day, the player may choose to go to work at the office in the morning, and spend time in the library building their academic skills in the afternoon.
These decisions are important because of the limited time scale. Usually, developing relationships plays a pivotal role in these games. After initially meeting or learning about an NPC, the player has the option to spend time with them.
Each time the player chooses to visit an NPC, their relationship progresses. Often these encounters play out with Branching Dialogue, giving the player an opportunity to deepen their relationship with the NPC further.
The character interaction activities usually culminate in some ultimate point in the relationship where the scripted sequences run out and the player is considered to have "completed" the relationship, as in the RPG title, Shin Megami Tensei: Persona 3, or dialogue begins to repeat at the maximum disposition, such as in the American-developed dating sim, Brooktown High.
Atlus' Shin Megami Tensei: Persona 3
Time Scheduling is effective in engaging a player with an NPC because the act of repeatedly choosing the characters to spend time with creates intrinsic value to the interactions. In the player's mind, if they chose to interact with a given character, it must be because the character is interesting. The player sees value in the interaction before it even starts.
The downside of this method is that it requires a large number of NPCs to choose from in order for the player's choice to be relevant, so developers must create NPCs and dialogues that the player does not necessarily encounter during the course of the game.
Often scripted sequences play out the same regardless of how much game time has passed between them, so it can be strange for an NPC to refer to an event that may have happened months ago (the last time the player chose to interact with them) as if it were just yesterday, exposing the limits of interactivity.
Considering the potential market for games that play like interactive television dramas or sitcoms, it is likely only a matter of time until the Western games industry hits upon a recipe for a wildly successful game based almost entirely around character interactions.
Microsoft/BioWare's Mass Effect
Games like Mass Effect and Star Wars: Knights of the Old Republic have proven that there is definitely an interest in extensive NPC interaction gameplay, and the surprise success of the Phoenix Wright series has shown that gamers are even ready for the visual novel.
By taking cues from past techniques, NPC relationships in the games of tomorrow have the opportunity to be deeper and more immersive than ever before.
ELIZA screenshot courtesy of Wikipedia.
Return to the full version of this article
Copyright © UBM Tech, All rights reserved