Time, Space and Interface

Reframing Dialogue Design for Better Flow

MA Thesis Submission

Thomas Martin


Though the ‘games vs narrative’ debate has cooled in recent years (Koenitz, 2018), a brief survey of current game dialogue design trends still reveals many sites disruptive to the concept of ‘flow’ proposed by Csikszentmihalyi (1990). In pursuit of a diagnosis, many academic categorisations of general game narrative are available, but - to date - few of these tools have been specifically applied to game dialogue systems. Two frameworks appear particularly productive for translation: Wei, Bizzocchi and Calvert’s categorisations of time and space (2010), and Jorgenson’s gameworld interface theory (2013). Through a broad assessment of the ‘standard dialogue meta’ and qualitative case studies of three contemporary narrative games, this paper investigates how the concepts of time, space and interface, when applied to game dialogue systems and sequences, can better facilitate flow.


Introduction - Games, Narrative and Disrupted Flow

We play games to be transported; to feel emotions different from those we find quotidian. One way for designers to engender this experience in players is narrative, a driving force for player involvement and self-identification in games (Murray, 1997). As a tool for contextualising rules (Juul, 2005), as a means for self-insertion and investment (Calleja, 2011), or as a way of ‘reading’ the entire experience of play (Fernandez-Vara, 2014), it is one of the most studied and debated elements of game design, with a rift between ‘narratologists’ and ‘ludologists’ emerging in the late nineties and early noughties (Koenitz, 2018). Prompted by articles such as Juul’s ‘Games telling stories’ (2001) and Aarseth’s ‘Computer game studies, year one’ (2001), ludic purists and narrative cheerleaders argued as to whether narrative enhanced gameplay, or disrupted it.

Today, the argument has cooled significantly. More specific definitions, as well as the broadening of game studies generally, have helped paper over academic disagreements. The market has spoken in narrative’s favour too, with previously narrative-sceptic genres now incorporating story - Fifa has an authored campaign mode (EA Sports, 2020); Forza has a writer’s room (Playground Games, 2021). The commercial and critical dominance of adventure games like The Last of Us 2 (Naughty Dog, 2020), the Uncharted series and Red Dead Redemption 2(Rockstar Studios, 2018) demonstrates developers looking to cinema to bridge the gap between narrative and gameplay, but Disco Elysium’s recent success shows that more literary games can strike a deeply popular chord (ZA/UM, 2019). Many of these works, however, still evince the issues that Juul et al were skeptical of at the turn of the century - crucially, that their narratives disrupt the ‘flow’ of their gameplay.

The psychological concept of ‘flow,’ posited by Mihael Csikszentmihalyi (1990), is a popular lens for analysing games. Flow refers to the state of ‘deep focus in which one is intensely engaged in an activity for its own sake,’ and the phenomenon has been clearly mapped onto game design pipelines and player experience (Chen 2007). If something were being disrupted by, say, narrative, it would be the player’s ‘flow-state’ while engaging with the game’s mechanics - their ‘ludic involvement’ (Calleja 2011). Conversely, if gameplay butted up against the sense of a story, the player’s ‘narrative involvement’ would be interrupted. But, to take the ludic side for a moment, what specifically about narrative could interrupt gameplay? Many (if not most) elements of game design can also be employed to deliver a sense of narrative. Ludic purists would hardly object to, say, the visual design of Chess, which communicates a clear narrative of hierarchy and medieval militarism (Koenitz, 2018), nor to audio design, which can contribute to narratives of place and progress. No, it is the written narrative - the game’s text - that for ludic purists most clearly interrupted the flow of gameplay.

Text & Text Types

A contemporary videogame will often feature a variety of different texts, from instructional writing to spoken dialogue, item descriptions to environmental graffiti (Bateman, 2007; Heussner et al, 2015).

Each type of text seems to have its own narrative mechanics to motivate readers. The text type structures the way in which the text is composed and processed.
(Bauer & Suter, 2021)

The narratological term ‘text type’ in the above quote refers to genres - detective, romance, thriller - but can just as easily be applied semantically to the different types of written content in games. The Game Narrative Toolbox(2015) identifies seven different ‘types’ of text common in games:

Text type Function Form
Instructional / UI text Tutor rules of game Clear language, short sentences
Item descriptions Increase narrative immersion, communicate purpose of item Flavourful, tonally heightened (for comic/atmospheric effect)
Investigated diegetic text (eg journal entries, audio logs) Increase narrative immersion at player’s discretion, break up pace Often more prosaically written
Environmental diegetic text (eg shop signs, directions, graffiti) Increase narrative immersion, communicate player position re: goals Context-dependent, but often humorous; short, ‘realistic’
Cutscene dialogue Increase narrative / emotional immersion, define goals Punchy, ‘cinematic’, clear
Interactive dialogue Increase narrative and emotional immersion, represent game mechanics, provide player with agency, act as ‘puzzles’ Highly game dependant, multi-voiced, dramatic, emotive, characterful, clear
‘Barks’ Communicate game-state, maintain narrative immersion Concise, ‘flavourful’, not overly differentiated

Table A: common game text types, functions and forms

While each of the above text types can be critical in forming the narrative of a game, it is clear that interactive dialogue has the most possible functions. It is of some surprise, therefore, how little academic or technical attention has been given to it. Narrative textbooks and resources focus on the difference between narrative, plot and story, or on effective worldbuilding, or on quest design, but pay little attention to the art of dialogue. Narrative Mechanics (Bauer et al, 2021) mentions the word 11 times in its 363 pages. The Game Narrative Toolbox (Heussner et al, 2015) defines dialogue’s purpose as ‘to keep the game moving.’ Academics have famously looked to theatrical theorists like Aristotle (Mateas, 2004), Stanislavsky (Fernandez-Manjon et al, 2013) and Boal (Frasca, 2004) to provide narratological frameworks more rooted in the stage than cinema or literature; none of these delve into the specific details of dialogue. Of the countless papers submitted to DiGRA, only 4 feature ‘dialogue’ as a keyword (contrast this with the 33 concerning ‘narrative’ alone) (DiGRA, 2021).

While a game’s written content may exist to outline gamestate-relevant information to the player (Heussner et al, 2015), dramatic dialogue serves a different purpose - communicating not only story, but also theme and emotion (Edgar, 2021). A well-written line of dialogue must tell the audience both what the character is saying, thinking and feeling. Unlike characters in a play or a film, however, the player’s decision of what to do next is rarely already made for them by the writer. Upon hearing / reading and understanding the story, the player must make the ‘narratively salient’ decision to move it onward (Tanenbaum & Tanenbaum, 2008), answering the question ‘what do I / what does my character do next?’ In service of this, game dialogue is often co-opted into the pedagogical modes of tutorial text or instruction manual. Additionally, players are already primed to search beneath the ‘flavour’ and ‘lore’ of investigated and environmental diegetic text types to uncover clues as to the game state. The extension of this detectorising ‘mode’ (coupled with the popularity of achievements for experiencing or ‘collecting’ every piece of content) has been lamented by independent game designers (Ingold, 2018; Yang, 2018) as one of the reasons game dialogue often disappoints - when much of a game’s written content exists to direct or inform, is it any wonder that players react to the dramatically written text in much the same way they do to item descriptions or UI text, searching beneath the ‘flavour’ for decision-relevant information? Or that this activity is necessarily both distracting from gameplay and counterproductive to narrative involvement? And have designers perhaps internalised this as the ‘standard’ way of designing dialogue?

The Standard Dialogue Metagame & Flow

Whether through design best practices, the accessibility of technology and assets, or the loop of consumer feedback, games have never looked - or played - more similar. Boluk and LeMieux (2017) condense this cultural codification of contemporary digital games into what they call the ‘standard metagame’: the generally accepted way that games feel, behave, react etc., including technical standards, UX, tropes and customs across both genre and design. This usefully broad model can be focused on specific elements, too. For example, the ‘standard dialogue metagame’ contains: voiced cutscenes and subtitles for AAA games; abstract response menus for open-world RPGs; scrolling text boxes for CRPGs; single-line text boxes for JRPGs. Popular innovations in the standard dialogue metagame are quickly adopted en masse: the floating button or trailing ellipses, signifying that a dialogue can be sped up by player input; UI-displayed 'cell phone’ chatboxes; exposed success percentages for dialogue options that contain risk; vividly animated ‘speech bubbles’ borrowed from Japanese manga, to name but a few.

Csikszentmihalyi (1990) defines eight factors that enable flow: a challenging activity requiring skill; a merging of action and awareness; clear goals; direct, immediate feedback; concentration on the task at hand; a sense of control; a loss of self-consciousness; an altered sense of time. In order to identify flow-disruptive tropes in contemporary dialogue design, I will map the ‘standard dialogue metagame’ to Csikszentmihalyi’s eight factors of flow.

Challenging activity requiring skill

Bauer and Suter maintain that, when engaging with written text, ‘the reader is permanently challenged at the level of the reading process’ (2021). Although this is evidently true - the experience of ‘getting lost in a good book’ is one of the ur-examples of Csikszentmihalyi’s flow - the difference in challenge between A: employing language skills to parse a sentence, and B: engaging in the complexities of even the simplest game mechanics must be considered. Some mental repositioning on the part of the player is always to be expected, and it is this repositioning - from dealing with rules just learnt (gameplay) to performing an action taught to most at the earliest age (reading) - that disrupts flow.

Additionally, the act of engaging with dialogue is often of the simplest nature. Following contemporary UX good practice, oftentimes any button can be used to ‘skip’ a piece of voiced dialogue; holding it down ‘prints’ out a textual line faster than its originally designed animation. Selecting responses is similarly simple, with radial or scrolling menus a popular choice. The continuous requirement to press ‘next’ or ‘select’ - in place of the skilled inputs required to progress in the rest of the game - might well be a barrier to flow.

Action and awareness

Shifting into a dialogue scene often signifies that a player’s actions won’t affect the wider gameworld, nor that the gameworld will affect her. In this way, the player’s relationship to both action and awareness abruptly changes. To effect a smoother refocus of this awareness (even if the sense of action is altered), designers make great use of environmental design and camera animation. In the environments of open world RPGs, for example, interactable NPCs are often found ‘nestled’ in corners of the environ, where conversation is ‘safe’; when an NPC is found out in the open, this often signifies a trap that will fling the player back into the action of the game. With cameras, designers program a gentle zooming to direct a player’s attention on a nearby NPC available for dialogic interaction (Legend of Zelda: Breath of the Wild - Nintendo EBD, 2017), or else subtly altering the frame of the image. Player avatars are even sometimes forced to lower their weapon to indicate a ‘safe zone for listening’ too (Half-Life 2 - Valve, 2004) - these types of intervention make players aware of a game’s limiting behaviours and design intentions, while reducing their capacity for action.


The goals of experiencing or interacting with game dialogue are sometimes nebulous. As discussed above, the subtext of much game dialogue is instructional; the goal of moving through that sort of text could be about learning what action the game ‘wants’ a player to do next. The Game Narrative Toolbox, with its assertion that dialogue’s purpose is to ‘keep the game moving,’ might even lead one to think that the player’s goal in engaging with dialogue is to return to ‘playing the game’ as soon as possible (Heussner et al, 2015). In RPGs, dialogue sequences that themselves contain goals (‘get Briony to give you a good deal on the pineapple’) are telegraphed by dialogic asides with other characters (‘we really need that pineapple, see if you can get a good deal’). With regards to ludic flow, dialogue goals are rarely as visceral or as juicy as the gameplay goals.

Direct, immediate feedback

When selecting dialogue options, game feedback is usually indirect, and less than immediate, often left to the next ‘line’ of dialogue: if voiced, the actor’s performance provides feedback; if animated, then the character model and camera do so. Notable exceptions are Fallout 3’s Pip-Boy ‘ker-ching’ on successful dialogue skill rolls (Bethesda Game Studios, 2008) and Disco Elysium’s high-octane UI colour splashes, but most dialogue interfaces are light on what Jayemanne terms ‘technical overstimulation’ (2019). If dialogue is voiced, response options will float like subtitles in front of the scene, or occupy a distinct UI panel; if written, text will be displayed in a readable font, and usually animated at a readable pace. The UI generally gets out of the game’s way when it comes to feedback, though more exceptions can be found in games with ‘diegetic’ dialogue interfaces, like ‘phones’ with messaging apps (Coming Out Simulator - Case, 2014) or ‘computers’ with faux-email and chat applications (Uplink - Introversion, 2001). These games mimic the fulsome feedback of the devices from which their UIs are drawn, and like those devices some are capable of holding multiple goals and tasks in hand - something designers are more commonly leaning into to increase flow.

Concentration on the task at hand

Depending on player persuasion, the definition of ‘task at hand’ could refer to either the gameplay or the narrative experience, but the academic consensus is that the two go hand in hand, and that players feel involved ludically and narratively to different degrees at different moments (Calleja, 2011; Koenitz, 2018). However, it must be conceded that the gear-shift between the two ‘tasks’ - as with the shifting between ludic and narrative goals - can be a site for the disruption of flow. This can be felt in interface design, with a tension appearing when ‘the game activity becomes centered on the overlay and not on the gameworld itself’ (Jorgenson, 2013).

Sense of control

Contemporary 3D games feature responsive, finely tuned control schemes for a player exploring an environment; these stand in contrast to the often simplistic level of interaction required for their dialogue navigation and selection. Indeed, dialogue sequences can be one of the few moments in which the player is not in control of how their character is moving or behaving.

The most popular AAA method of delivering dialogue is via cutscenes or else ‘played’ under the action, like an underscore. Depending on the game, these methods mostly disregard players’ immediate choice; a player may have performed an earlier action which triggered this particular cutscene, but during it they have no say on what their character says. During cutscenes, controls are generally locked - understandably the player loses control in these sequences. ‘Underscore’ or ‘overheard’ dialogues can give players a feeling of voyeurism and agency, although some provoke feelings of player impotence in worse ways than cutscenes: in the Assassin’s Creed series, players are often tasked with ‘stalking’ a target in order to overhear their conversation; if the player strays from ‘earshot,’ the mission fails and the process (and entire conversation) frustratingly repeated.

When dialogue selection is required, two control schemes are usually available: full sentence selection (what you see is what you say) and abstract response (paraphrased indicators of content or intent). Full sentence selection can entail a large amount of reading, and runs the risk of disrupting action and awareness. With abstract response systems, players gain a less time-consuming dialogue selection process, but lose the certainty of what their character would say. Games like Mass Effect (Bioware, 2007) and Alpha Protocol (Obsidian Entertainment, 2010) played into this uncertainty skillfully, while games like Fallout 4 (Bethesda Game Studios, 2015) received criticism over a loss of vocal character control. Unless a game tutors for NPCs reacting unexpectedly, abstract response dialogues can contribute to narrative disconnect - and a disruption in flow.

Loss of self-consciousness

Losing self-consciousness is a function of player involvement, be that narrative, ludic or emotional (Calleja, 2011), but dialogue systems work hardest to give meaning to the player’s narrative agency. When players encounter a ‘break’ in narrative sense - perhaps they’re offered a choice their character would never consider, perhaps the game offers them no choice - their immersion suffers. Though these can be seen as issues with the writing of game dialogue, Ingold has remarked on the popular hub-and-spoke approach to dialogue structure, and its propensity for breaking immersion (2018). The instructional nature of dialogue can also disrupt flow, especially if it becomes notably repetitive.

Additionally, in 3D RPGs like The Witcher 3 (CD Projekt Red, 2015), or narratively driven action games like Max Payne (Remedy Entertainment, 2001), the camera is usually floating some in-game feet behind and above the avatar. Dialogic interactions (whether interactive or cutscenes) are one of the few times the player is made aware of their character’s face, or their expressions, and this effect is of course even more pronounced in first-person games. Whether this reminder of avatar appearance and expression has any effect on player self-consciousness is something yet to be studied.

An altered sense of time

Real-time narrative games make various efforts to keep players’ experience of time as continuous as possible. When it doesn’t rely on cutscenes, the GTA and Assassin’s Creed series’ dialogue takes the form of a non-interactive scene played and animated while the player is busy with another action - driving in the case of GTA; following or eavesdropping in the case of AC. In games like The Walking Dead (Telltale Games, 2012) or Detroit: Being Human (Quantic Dream, 2018), more intense dialogue decisions are made under the pressure of a ticking clock or timer. Contemporary RPGs such as Baldur’s Gate 3 (Larian Studios, 2020) and The Witcher 3, while they do make use of cut-scenes, rely on dialogue trees accompanied by fully voiced and animated character models. These models are animated even during the player’s co-ordination time, giving the impression that action is continuing even when the gameworld is - for all mechanical intents and purposes - paused. This results in the uncanny experience of watching NPCs emotively ‘wait’ for a dialogue choice to be selected. Disco Elysium inverts this trend to some effect: the in-game timer advances by three minutes per dialogue selection, while traditionally time-advancing activities like walking, changing outfits or examining the environment take place in a state of suspension.

Generally speaking, though, games will either: suspend time in order to ‘enter’ a dialogue scene; maintain a turn-based time signature that lessens the interruptive feel, or else blend their dialogue ‘over’ or ‘under’ a player’s actions. These approaches can have various effects on a player’s perception of time.


The above (far from exhaustive) breakdown makes a few things clear. Firstly, the standard dialogue metagame still includes many sites prone to flow disruption. Secondly, developers have bent over backwards devising workarounds for these issues, and their design choices have been co-opted by other developers when faced with similar issues. Thirdly, given the continuing commercial growth of the games industry, consumers have grown to expect games to both feature these industry-standard design choices (Jayemanne, 2019), and to contain sites disruptive to flow. The general picture is of a medium well aware of the potential power and reach of its stories, but only now awakening to the implications of - and assumptions it has made about - the tools it has been using to tell them. Deepening this understanding must begin with an appraisal of the material in question - dialogue.

Dialogue - Finding A Working Definition

Dialogue can be defined as:

  • conversation between two or more people

  • an exchange of opinions on a particular subject; discussion

  • the lines spoken by characters in drama or fiction

  • a particular passage of conversation in a literary or dramatic work

  • a literary composition in the form of a dialogue

(Collins, 2021)

Most of these definitions describe an extant artifact - dialogue as a whole, or a chunk - and indeed this is helpful from the perspective of literary analysis. Books and poems have little variance in objective reader experience, different readers almost always being presented with the same words in the same order, experienced in real-time then reflected upon after completion (Iser, 1991). However, as games academics have found, analytical techniques borrowed from literature studies rarely account for things like different playstyles, session lengths, or even the sheer length of some games. Even less do they map to the player’s order and pace of experience - entire scenes, subplots or characters can be absent from one player’s playthrough, whereas they may be the most critical narrative moments of another’s (Herman’s ‘fuzzy temporality’ (2002)).

With all this in mind, it has proved difficult to categorise or analyse the design, writing or effect of a game’s dialogue beyond the surface level (a journalistic ‘is the writing good or bad?’). Yet games academics are frequently able to analyse game loops that take hundreds of hours and vary between players - why not so with text? Perhaps a different definition of dialogue could help?

In both the fields of semantics and computing, a dialogue (whether spoken, written or signed) is defined as a communication that takes place between two (or more) agents (Brusk & Bjork, 2009). Both of these agents - which we can assume to exist in both time and space - are involved in the forming and receiving of utterances, as well as negotiating and ‘playing’ the rules of a language-game using a common interface, which is usually a shared language (Austin, 1962; Searle, 1969). Considering dialogue as such a process of exchange cements interactivity into the definition, which is obviously useful from a games studies perspective. So we might arrive at a working definition of dialogue:

an interaction that takes place between agents, over time, in space, and through some kind of interface.

Some more clarity can be gained by separating dialogue into the two forms in which it is experienced: as a system and as a sequence. The system defines how the dialogue has been designed and how it is delivered; the sequence refers to a passage of time during which dialogue is experienced. In playwriting one might say that the system is the script as a whole; the sequence is the scene, or even the beat. This definition now allows for the analysis of how the dialogue has been designed, how it is experienced, and how it has been written. I will now focus on the first two considerations in order to better map concepts of dialogue onto Csikszentmihalyi’s elements of flow, using as lenses the concepts of time, space and interface.

Game Dialogue & Time

Zagal and Mateas (2007) separate time into four discrete categories: real-time (physical time / actual play-time); gameworld time (how and at what rate events within a game progress relative to player action); coordination time (the relationship between the actions of game agents and the player), and fictive time (what the progression of gameworld time means within the narrative of the game). Much has been built upon their work (Eskelenin, 2004; Tychson et al, 2009), and recently Jayemanne proposed an elegant trinary categorisation that allows for games’ temporal complexity to be addressed using comparative literary theory (2019). His diachronous time - the emotional deja vu of multiple playthroughs - is of particular use when considering dialogue, but it is Wei, Bizzocchi and Calvert’s thorough mapping of established concepts of narratological time onto digital game storytelling methods that I will apply to dialogue in this paper (2010).

Wei et al, like Zagal and Mateas, separate the notions of gameworld time and real-time, but apply the terms of literary theorists Bal (2009), Prince (2003) and Bridgeman (2007). In their unified framework, story time is the overarching term for what happens in a narrative - were it to be laid out on a timeline from beginning to end, as it were. Discourse time refers to the time in which a narrative is experienced - for the listener of an anecdote it is the time spent hearing; for the reader of a book, the time spent reading. In games, however, a player reads and acts at the same time;_discourse time_accounts for this entire process. The relationship between story time and discourse time is what defines a narrative, and can be divided into three categories: order, speed and frequency.

Order (linear, non-linear)

Order can be separated into linear and nonlinear. Experiencing dialogue linearly allows a player to build up assumptions about the role they are playing, advancing alongside their character through discourse time. Though many games make use of nonlinear story structures (flashbacks / time travel), few include non-linear dialogue - it is relatively rare for lines of dialogue to be presented out of order, and certainly for a player’s dialogue options to be disconnected from linear sense. Order can, however, be used to consider interruption - can a player break the natural ‘order’ of dialogue by interjecting? - and also the flexibility of the game’s dialogue systems. For example, Pendragon (Inkle, 2020) uses a storylet-style system to select the dialogue sequences, snippets and choices that are most relevant to the player at that moment in discourse time, taking into account choices they have made in this and previous rounds, other game agents that exist, their proximity to the endgame, and the type of dialogue sequences they have been shown thus far. While the order of individual ‘lines’ is still largely linear, that they are chosen procedurally rather than authorially lends a sense of reactive nonlinearity to the game’s narrative.

Wei et al also observe that ‘games have one distinct use of temporal order that is not found in other narrative forms, that is, to use order as the answer of a puzzle’ (2010). Whether or not the order of responses within a dialogue tree matters will alter a player’s relationship to a dialogue system, and the game’s discourse time, significantly.


The speed of narrative can be divided into five categories, and these terms can easily be applied to most game dialogue systems. Scene is the most common type, with story time and discourse time in unison - characters ‘say’ lines, they are either voiced entirely or the player reads them in real-time. When discourse time is shorter than story time, the dialogue can be described as a summary. This can be found in games like King of Dragon Pass (A Sharp, 1999) and lately the Crusader Kings series, in which description often contextualises a single dialogue choice - ‘off with her head!’ - in a scene that would normally take longer to experience; ‘token-based’ dialogue systems triggered by the use or exchange of inventory items (Bateman, 2007) - as in Monkey Island (Lucasfilm Games, 1990) - might also be described as summary. Stretch (in which discourse time is longer than story time) might include dialogue with additional details akin to stage directions or the intervention of an internal monologue - Disco Elysium is full of these tropes, and oscillates between scene and stretch. Ellipsis is an extended kind of summary, but with intermediate action removed. Moments of ellipsis are frequent in games, but rarely make up the bulk of dialogue systems; they can sometimes be found in MMOs like World of Warcraft (Blizzard Entertainment, 2004), where dialogue might represent the time it takes for an NPC crafting an item for the player, for example. Pause refers to both story time and discourse time being suspended. When a game’s dialogue is unvoiced, reading often feels like a pause; however, by reading the text the player is still experiencing discourse time. It is therefore more useful to associate pause with text types that provide solely ludic information - tutorial and menu text, for example.


Frequency can be conceived as singular, repetitious, or iterative. Singular dialogue sequences can be found in more cutscene-heavy AAA titles, in which - like a movie - the story is always moving forwards. The player can never repeat the ‘guitar scene’ in The Last of Us (Naughty Dog, 2013), for example, although it may be referenced diachronously later in the game. More narratively experimental independent games, like those of studio Inkle, feature an approach that could also be described as singular - conversations never (or at least very rarely) repeat, and players get one shot at decisions and responses (Ingold, 2018).

Repetitious dialogue - in which the same conversation can be accessed by the player at different times - is common in open-world RPGs, MMORPGs and CRPGs, although by altering the game state the player also encourages that dialogue to change - this might be termed iterative dialogue. For example, a seemingly repetitive NPC might give three dialogue options consistently, but only after a certain player action will their dialogue menu iterate into offering a fourth, or deleting the second. Generally speaking, the more iterative the dialogue, the more interactive the narrative.

Game Dialogue & Space

‘The defining element in computer games is spatiality.’ So holds Espen Aarseth in their article ‘Allegories of space’, continuing that ‘games are essentially concerned with spatial representation and negotiation; therefore the classification of a computer game can be based on how it represents or, perhaps, implements space’ (2007). Nitsche contends that the player’s experience is driven primarily by spatial structure, as - unlike in cinema - most game narrative elements are organised by space rather than time (2007). Different categorisations of game space and narrative space have been proposed by Jenkins (2004) and Wolf (2001), but I revisit Wei, Bizzocchi and Calvert here for their work on mapping narratological theories of space onto digital game storytelling.

Following Zolan (1984), they organise game narrative space into three types: topographical, operational and presentational. The topographical is the underlying structure - it is the spatial design as it exists before a player experiences it, without time. Worldbuilding, setting, character backstory, but also everything that might happen within a game. The operational is the story space reacting to the addition of the player’s experience and influence - time has been added, in other words. The plot as it develops during this specific playthrough; everything that does happen during a playthrough. The presentational refers to how the game manifests both the topographical and the operational - how it shapes and colours narrative through ‘visual, auditory, textual, haptic and other cues.’ To stretch these definitions over dialogue design, one might consider the topographical to be how dialogue sequences are shaped and structured during the design and writing process, and how they exist in the game’s files or mapped out in a wiki or guide; the operational to be how a player experiences and interacts with the dialogue during a single playthrough; and the presentational to be the method in which the dialogue is presented to the player. Wei et al are at pains to remind us that ‘it is [always] the combination of the three modes working together that make the design and experience of a game’s narrative space effective and satisfying’ (2010).

Topographical: Layouts

In the fields of narrative design, story space has been famously visualised in Murray’s maze and rhizome (1997), the ‘story cave’, and Short’s quality-based, salience-based and waypoint structures (2016). These layouts are applied both at the macro level (story structure, scene ordering) and the micro level (quest design, branching dialogue trees), and provide different affordances for both player agency and authorial control. Wei et al refer to Adams’ (2010) common patterns:open layout, network, combination, linear, parallel, ring and hub-and-spoke - the hub-and-spoke dialogue tree being recommended in The Game Narrative Toolbox(Heussner et al, 2015). While, unlike the map of a gameworld, the topographical layout of a game’s dialogue design is rarely something a player will get a bird’s eye look at while they play, both the number of interactive options and the frequency of sequences will often provide subliminal clues as to the true topographical layout, like the colour of water indicates its depth. Returning to the same dialogue options indicates a hub-and-spoke design (Fallout - Interplay Productions, 1997); single, repetitive lines might indicate a linear dialogue sequence set within an open story layout (Legend of Zelda: Breath of the Wild).

For a player, ‘where’ a dialogue sequence sits within a game’s story space layout also affects their experience of it. Sequences might be locked after, say, the player has entered a certain district, or require three in-game ‘years’ in order to trigger. The density of dialogue sequences as related to the system’s layout also creates meaning, quest-related dialogue sequences in denser groups at the more hub-style central acts, and sparser at the more networked peripheral areas (Star Wars: KOTOR - Bioware, 2003). Games tend to taper out the amount of interactive dialogue sequences in their end stages, in order to focus players towards one of (a handful of) predesigned endings.

Topographical: Spatial oppositions

Spatial opposition - the use of spatial contrast to create meaning - has a surprisingly wide application in dialogue design. Wei et al direct the reader towards ideas of ‘inside’ and ‘outside’, and these distinctions can be found in the dungeon-specific NPC dialogues of Baldur’s Gate II (Bioware, 2001). Stretching the definition, a line can be drawn to levels of internality. The ‘internal’ and ‘external’ dialogues of Disco Elysium or the ‘stage directions’ common in tabletop-inspired CRPGs, for example, create layers of narrative and ludic meaning that a more ‘realistically’ designed dialogue system, like The Walking Dead’s, cannot.

However, looking at the literal shape of a dialogue system’s writing can reveal much about the designers’ approach to spatial opposition. ‘Background’ NPCs will often be granted short, single lines of dialogue to compensate for the amount of times those lines will likely be activated. Players may only have a single available response, or an entirely non-interactive system may be used for triggering these lines, as in the comical streetwalking NPCs of the Grand Theft Auto or Saints Row series. Contrast this with the cinematic flair of these games’ ‘main’ dialogue sequences, complete with chunkier soliloquies for their villains, and one can observe as much difference in complexity and rhythm as between theatrical dialogue, monologue, and song (with Saints Row IV (Volition, 2013) indulging in the latter to memorable effect).

A game’s dialogue system can also engage with spatial opposition on the fly. Emily Short, in a talk on designing Animal Farm (Nerial, 2020), reveals a storylet system that alternates its bias between ‘major’ and ‘minor’ events, so as not to overwhelm the player with a series of story-critical - and usually more lengthily written - events (Short, 2021). One can easily imagine a dialogue system using a similar approach to vary line length in conversations.

Operational: Mobility of characters and objects

This classification would usually apply to the movement of agents within a gameworld; with that comes the consideration of whether the dialogue itself moves with the player across the gameworld? We have already discussed examples of this with the ‘underscore’ dialogue design of AAA games, but included in this consideration might also be the dialogue assigned to a player’s party in a CRPG like Pillars of Eternity (Obsidian Entertainment, 2015). In that game, most NPCs have a fixed position, both in the operation of the story and in the gameworld, usually only moving if engaged in combat. Their dialogue can be accessed only at their specific location. A number of ‘playable’ NPCs, however, will join the player on her adventure, and carry with them dialogue sequences triggered by specific gamestates - time spent with the player, for example, or the party reaching a specific level of ‘reputation.’ These dialogues could be said to be mobile.

Of further interest is whether player mobility is taken into consideration by the dialogue system. While not offering explicit dialogue ‘choices’ as such, the dialogue systems of Portal 2 (Valve, 2011), Firewatch (Campo Santo, 2016) and Half-Life: Alyx (Valve, 2020) interpret player position, movement and action as specifically as a dialogue tree. These games minimise the player’s ‘verbal’ agency while maximising their ‘physical’ agency.

Operational: Paths and axes

How does a player move through a dialogue sequence itself? Is there a single path in and out of a dialogue? In the Mass Effect series, while there is a single path through many of the dialogue sequences, a player might be said to have chosen either the Paragon or the Renegade axes. Dialogue ‘branching’ is one of the most widely employed design approaches, but can players ‘move’ between branches, across paths? Once they are on a path, is it unidirectional or bidirectional? Put glibly, is there a ‘back’ button - and if there is, what does it mean for the game’s temporality? The interventions of companion NPCs might provide short ‘detour’ paths within sequences; relying on a particular RPG skill or stat might literally move the player along an axis of ‘experience’, which represents both the player’s ludic progression and the narrative of their character’s growth.

Paths and axes within a dialogue system can also be explored through means other than dialogue choice. The much-lauded Hades (Supergiant Games, 2020) showcases how every one of a player’s in-game actions can be folded into dialogue paths and axes - its looping structure allows for multiple quest paths to be pursued simultaneously, with the events of each run biasing the player into different relationship axes with its cast of NPCs.

Less commonly, axes can be emotional or inflective as well as intentional - how a player chooses to ‘perform’ a dialogue choice. Background- or alignment-specific choices in RPGs might not provide different dialogue paths, but they can provide a flavourful axis into which a role-playing-savvy player can lean (Vampire: The Masquerade: Bloodlines- Troika Games, 2004).

Presentational: On-screen / off-screen

Voiceover and subtitling can sometimes reference an offscreen character (or, more often, a narrator), but this technique is rarely used in interactive dialogue sequences (‘overheard’ dialogue is covered in acoustic space). Dialogue does need to be presented to the player, after all, and games generally only have the medium of screen to do this. The ludic element of a dialogue system, however, might be said to exist ‘off-screen’; systems like NPC attitude and quest progression - though often displayed to the player in contemporary games - for many years were hidden. Whether the narrative or ludic consequences of dialogue choices are made available ‘on-screen’ is also a design consideration, as revealing or hinting at these consequences expands the story space, and can either complicate or oversimplify a player’s decision-making process (see: colour-coded choices for Paragon/Renegade).

Presentational: Acoustic space

The acoustic space, as it applies to dialogue sequences and systems, is most commonly connected to voice acting and cutscene underscore, but even before this became common, games were using sound cues to indicate that dialogue was being ‘spoken.’ The whirring or mumbling of text ‘loading’ will be familiar to players of JRPGs, and this approach remains common as a simple way for the dialogue system to offer the player feedback.

Entering dialogue sequences often subtly alters the acoustic space of the game, muffling environmental sounds to focus player attention. In voiced 3D games like Mass Effect, for example, the player experiences the introductory lines of a dialogue sequence through the same fully multidirectional sound space as the rest of the game. However, once the scene gets going (in Mass Effect’s case, when the visual aspect ratio changes and the dynamic camera system starts ‘directing’) the game opts for a more traditionally cinematic stereo recording.

Presentational: Perspective

Perspective is a term most associated in game design with camera perspective - first-person, third-person, isometric etc. - and can have a significant impact on the player’s involvement (Calleja, 2011). The first-person camera ‘reinforces the psychological perspective of the protagonist,’ while observing one’s character speak and act through a third-person perspective can encourage more objective, thematically focused involvement (Wei et al, 2010). The breaking of a player-focused third- or first-person perspective to a more cinematically familiar camera technique during dialogue will be familiar to anyone who has played recent AAA games, but the approach is hardly unique to blockbuster releases. Divinity: Original Sin’s mild zooming in might be one example (Larian Studios, 2014); Deus Ex - a first person game - enters a cinematic third person mode during dialogue sequences, complete with a narrowing of aspect ratio (Ion Storm, 2001). Baldur’s Gate I & II feature few perspective changes during dialogue (although the camera will sometimes ‘snap’ to above a ‘speaking’ character model), but Baldur’s Gate 3 makes use of an entire cinematics team to both procedurally- and hand-animate the camera movement of every dialogue scene.

A different interpretation of dialogic perspective might be found in the writing itself - here we could revisit the internal and external voices of Disco Elysium, but also the hard-boiled narration of Max Payne. Shifting interactive perspective (say, making choices as one character and then making choices as another) is relatively rare; though it does occur in games with multiple playable protagonists, with these characters sometimes talking to each other, switching mid-sequence is unheard of. Even strategic games like Frostpunk (11bit Studios, 2018), Humankind (Amplitude Studios, 2021) and Crusader Kings II (Paradox Interactive, 2012), with relatively simplistic dialogue systems, still maintain a unity of player perspective, by casting them as the leader of a post-apocalyptic community, pre-modern civilisation or feudal tribe respectively.

Presentational: Spatial segmentation

This category of operational space refers to the ‘division of the gameworld into different spaces that also partition gameplay’ (Zagal & Mateas, 2007). In most games, we would consider these spaces to be differently designed biomes (if topographically connected) or levels (if topographically separate), but applying it to dialogue throws up a number of interesting questions. Is the dialogue segmented from the gameworld? Does a player ‘enter’ this new subspace, and what happens when they do? We have already touched on shifting aspect ratios signifying the presentational transition into dialogue sequence space, but mechanical affordances can also segment space. The player knows, wheeling into one of Sunless Sea’s (Failbetter Games, 2015) ports, that the monsters and pirates of the Underzee won’t harm them, as their control scheme changes from real-time mouse-and-keyboard to turn-based point-and-click. And once inside a dialogue sequence, where are the barriers, the exit points? In some games, dialogue can be halted by moving away from an NPC model, blending the subspace transition; at the close of a conversation in Baldur’s Gate, the ‘End Dialogue’ button does the obvious, and in many games an option that ends a dialogue will be subtitled similarly.

A dialogue system may even allow for multiple segmented spaces within itself. Some games include ‘mini-games’ within dialogue, and the representation of ludic mechanics such as dice-rolling are often represented in new ‘spaces,’ or at least in different coloured text.Tyranny (Obsidian Entertainment, 2016) and Pillars of Eternity innovate with mouse-over, hyperlink-style text within their dialogue scenes, deepening lore, delivering details your character observes or, in one memorable case, representing an NPC communicating with you psychically.

Screen interface

Wei et al (2010) identify another category for presentational space - screen interface - but, given its overwhelming use as a dialogue delivery system, I have chosen to consider it separately.

Game Dialogue & Interface

From the earliest days of text-based games, interface has played an important role in how dialogue has been interpreted. Parser-based games like Colossal Cave Adventure (Crowther & Woods, 1977) and Zork (Infocom, 1980) were played within the context of early computers’ command-line interfaces. Point and click adventure games like The Secret of Monkey Island (Lucasfilm Games, 1990) or Discworld Noir (Perfect Entertainment, 1999), with their ‘token-based’ dialogue and inventory systems (Bateman, 2007), used a graphical UI to communicate the player’s context-sensitive options. Even in contemporary games with entirely voiced dialogue, lines are by default subtitled or superimposed in writing on the UI, or else paraphrased in flashing HUD-bound mission objective text.

Kristine Jorgenson, in her 2013 book Gameworld Interfaces, translates the ontological concept of ‘frames’ to the simulated world of digital games, and specifically user interfaces. The frames existing in game UI design, she offers, can be both interdependent and interposed. Jorgenson contends that ‘game-system information might not just be integrated [internal to the gameworld] or superimposed [part of an overlay]; it might also be fictional [diegetic] or ludic [connected to game systems] and emphatic [adding/highlighting information] or ecological [naturally containing information]; and the two elements of each category might be blended as well.’ Game dialogue interfaces make use of a wide range of these frames, often adding or removing during gameplay as necessary (this process is called up- or downkeying).

Interface Type Examples in dialogue design
Integrated, Fictional, Ecological Emails in Uplink / messages we see on a device that exists within the gameworld; Captain Blood’s spaceship bridge (Exxos, 1988)
Integrated, Fictional, Emphatic Text appearing above characters’ heads (Fallout)
Integrated, Ludic, Ecological Relatively rare in dialogue design!
Integrated, Ludic, Emphatic Exclamation point or similar above character (in gameworld) indicating expression or reaction
Superimposed, Fictional, Ecological Subtitled unvoiced dialogue, written and unvoiced dialogue within a UI text box, messages on a device we do not see through the character’s ‘eyes’ - GTA’s pager/cell phone
Superimposed, Fictional, Emphatic Subtitled voiced dialogue, first-person ‘diegetic’ HUDs (in-helmet text read-outs); animated character ‘headshots’ (Tyranny; visual novels)
Superimposed, Ludic, Ecological Descriptions of character actions in a text box (‘you attack the gnoll’)
Superimposed, Ludic, Emphatic Descriptions of dice rolls / outcomes or items / experience gained

Table B: examples of gameworld interface frames in dialogue designs

Dialogue - when it is represented within a game’s interface - is generally presented through superimposed subtitles, or else in a superimposed text box. A dialogue system consisting of textual communication over a phone interface, however, would be considered integrated. Dialogue is of course generally fictional, as it represents something diegetic to the game’s story, though when dialogue systems include skill thresholds or other mechanical signifiers, it can be said to be ludic. Jorgenson acknowledges the *liminality* between fictional and ludic frames, stating that ‘when props and other kinds of representation [which we can here take to include dialogue] support a fictionally coherent world, the primary framework for understanding these features is fiction, but when the representations break with fiction and focus on their own ludic value, the primary framework for understanding them is ludic.’ ‘Clementine will remember this’ is a cogent example of an otherwise superimposed and fictional dialogue interface providing ludic information (Telltale Games, 2012). A subtitling interface can be characterised as emphatic, as it highlights information that exists in the acoustic gameworld, but when the dialogue is primarily designed to be read it could be said to be ecological - indeed, given CRPGs historical design links to Choose Your Own Adventure novels, everything other than their written text might be said to be emphatic.

Of critical importance in dialogue interface design is its position relative to other elements of the gameworld and of game UI. If the entire dialogue system is superimposed, does it sit within the gameworld (World of Warcraft) or does it entirely frame it (Planescape: Torment - Black Isle Entertainment, 1999)? Is the dialogue interface constant (visual novels) or context-sensitive (The Elder Scrolls V: Skyrim - Bethesda Game Studios, 2011)? And while it is good practice to design a consistent and cohesive aesthetic for a game’s UI, it is important to consider with which other elements a player will associate a dialogue interface. When it is visually and mechanically indistinguishable from inventory, character customisation or menu UI, what is that teaching the player with regards to where a dialogue interface sits in relation to other forms of gameplay?

Case Studies

In the penultimate section of this paper I will outline three recent narrative games, and how the relationships of their dialogue systems to - variously -time, space and interface heighten narrative and ludic involvement by mitigating the disruption of flow. While time, space and interface are critical elements in the dialogic innovation of all three games, I have taken the liberty of aligning each example with the element it most strongly evinces in its design. The player’s experience, of course, is mediated through all three elements.

Time - Signs of the Sojourner

Signs of the Sojourner is a 2018 narrative deck-building game developed and published by Echodog Games. Set in a post-apocalyptic landscape, the player joins a caravan of merchants, encountering other merchants and artisans in the surrounding villages in the hopes of supplying the trading post of their recently deceased mother. Between encounters, dialogue interactivity is on the light side - the player is given the choice of whether or not to enter conversation with NPCs in the settlement; no dialogue options are given to the player; the NPC is the only character that ‘speaks’. But within an encounter, the dialogue system becomes more complex. Players still aren’t afforded dialogue ‘options’; rather, they interact with the NPC through a turn-based symbol-matching card game, the goal being to match a chain of symbols to create an ‘accord,’ and avoid mismatches that might create ‘discord.’ If three accords are created, the dialogue ends positively; otherwise, it ends acrimoniously. Regardless of the player’s performance in each encounter, they must replace one card in their 10-card deck with one of the NPC’s, both symbolically internalising something they’ve learned from that character and altering their relationship to the dialogue system going forward, encouraging the player to identify closely with their role, and experience a loss of self-consciousness. This connection between memory, character and deck-building means that it isn’t just the written dialogue that iterates through the different acts, it’s the dialogue system itself - the player’s ability to interact with the system is predicated on cards they bring from encounters. Some of these cards will last for only a handful of encounters, but some will form the core of a strategically formed deck over the course of an entire act, or the entire game, making the player reflect diachronously on the conversations and characters from which they took the cards they now use. This inversion of the usual action/dialogue interactive balance (gameworld navigation is as basic as inventory management) evolves dialogue interaction into a challenging activity requiring skill.

While the time signature of the written dialogue could be defined as summary, the player’s experience of a dialogue sequence is far more akin to stretch - the intense, dilated version of narrative time. During each suspended moment of dialogic negotiation (itself an altered sense of time), the player achieves a merging of action and awareness as they reflect on NPC animation, musical cues, and strategic considerations. Direct, immediate feedback is achieved through the matching process, and with the game’s reliance on imperfect information (Costikyan, 2013), the dramatic tension heightens with each new draw, each player action and each NPC card placement - encouraging concentration on the task at hand.

Understanding for how long a dialogue will continue helps players define a clear goal for the immediate moment, and Signs of the Sojourner’s card game mechanics help facilitate this. Conversation ‘rounds’ can last for a maximum of five turns; an entire dialogue sequence for a maximum of fifteen turns (to reach 3 accords). The clarity and predictability of this structure is repeated in the larger narrative, with the game organised into five ‘acts’; each act represents a different caravan journey starting from Bartow, proceeding along a route of the player’s choosing, and returning to Bartow. This ring-shaped narrative layout, coupled with the round- and turn-based time signature of encounters, awakens the player to the frequency of dialogue - the dialogue sequences might be seen as singular or iterative, depending on whether the player is able to make any headway with individual NPCs, but the route the player takes through the game’s five acts (ie whether they revisit locations) defines whether dialogue is repetitive. It’s possible to have an almost entirely singular experience, if the player refuses to revisit locations.

Signs of the Sojourner finds an elegant, flow-friendly way to reshape the discourse time of dialogue, and demonstrates how occupying a player with even mildly challenging dialogue mechanics can deepen narrative involvement.

Space - Oxenfree

Oxenfree (2016), developed by Night School Studios, is a 2D narrative adventure game set in a self-confessedly Spielbergian world of teenage investigators, mysteries, gadgets and ghosts. The player takes on the role of Alex, a young woman who, along with a gaggle of friends, explores a mysterious island. Dialogue is fully voiced, and the game uses an abstract response system for player dialogue choice, generally presenting between two and three responses. So far, so Mass Effect. However, it also allows the player to remain silent, to walk away from an interaction, to continue while moving, and to use any dialogue option as an interrupt. Unlike in traditional narrative-led games, in which NPCs react to the selection of content, Oxenfree’s cast will react differently depending on the player’s spatial and rhythmic choices. Not all dialogue sequences are guaranteed to appear in a given playthrough either; developers Night School Studios developed an AI system to select dialogue sequences salient to the player’s prior behaviour, as well as their position and the amount of time elapsed since the last dialogue sequence. The topographical layout of the dialogue system is thus far more open than might be first anticipated, and maps dynamically onto the game’s more networked environmental layout. By disconnecting some of the narrative layout from the environmental layout, Oxenfree even alleviates some of the position-locked predictability from which its point-and-click forebears suffer.

The player, in contrast to more hub-and-spoke-influenced dialogue designs, is locked into a unidirectional path - like in the films from which it draws inspiration, Oxenfree’s narrative progression is fast-paced, and there are few options to have directions or information repeated. However, players can lean into different axes with their dialogue choices, and accentuate those choices through interruption or silence.Oxenfree makes little differentiation between its dialogic acoustic space and its sonic environment (radio tuning mechanic aside), and this focus, coupled with the reduced friction between gameworld and dialogue interfaces, encourages concentration on the task at hand.

Perspective is locked to Alex, the player character, and the game almost never removes the player’s mobility, increasing their sense of control. With direct player control almost never interrupted, the game avoids the pitfalls of other games covering the spatial segmentation of dialogue and play - action and awareness are thus encouraged to merge. Being able to listen, ignore, move away etc during dialogue also affords moments of spatial opposition akin to the performed stage directions of a playscript. The invitation to perform spatially rather than just ‘vocally’ further embodies the player in their character, encouraging a loss of self-consciousness. Interrupting provides an immediate and fulsome sense of feedback, and while immersed in dialogue the player can also choose to carry out actions within the game world that would traditionally be separated from dialogue sequences - tuning a radio while talking, for example. While not necessary for completion or indeed success, the affordances for players who want to multitask in this way opens up dialogue selection to be approached as a challenging activity requiring skill.

Oxenfree, by affording its dialogue system the same design and interactive affordances as its spatial environment, paves the way for far more involving 2D narrative game experiences.

Interface - We Should Talk

We Should Talk (2018) is a short narrative game designed by Carol Mertz that features a unique, ‘intention’-based dialogue mechanic. The player spends an evening at a bar, flitting between conversations with other patrons of the bar, and texting with their partner (who is at home). The player and the patron NPCs are presented as animated 3D models, while the player’s interactions with their partner happens through a text message interface on their phone. Dialogue is presented in a cascade down the middle of the scene - modelled after contemporary messaging UI - either filling the gap between the player character and their animated scene partner or else filling the screen of the player character’s phone. This interface elegantly bridges the gap between fictional and ludic, ecological and emphatic; the ‘voiced’ lines are ecological, but the intentional modifiers - which make up the main interactive mechanic -emphatic (and where the real ludic interaction lies). Instead of selecting what their character will say, the player is tasked with committing to an intention - what they mean by what they say. On an even more microscopic level, the player must determine the emotional intention of each clause, rather than the line as a whole; an interaction that would usually consist of a single button press is now a series of considered and cascading linguistic and psychological decisions. This visual separation of the dialogue’s semantic elements, in contrast to Oxenfree’s more instinctual abstract response options (and its seductively feedback-heavy interruption mechanic), encourages a level of emotional ‘close reading’ normally not seen in games, largely due to the analytical ‘slowdown’ it can engender in gameplay. Resources like The Game Narrative Toolbox (Heussner et al, 2015), for example, guide writers towards clarity in dialogue over almost all other considerations, in order that the majority of players parse the emotional intention of lines more easily, and to guard against players feeling a lack of control over their character. But by framing intention as the ludic element of the game, We Should Talk turns the parsing of dialogue choices into a challenging activity requiring skill. It also means that when the player’s emotional intention is ‘misread,’ it feels like the result of the narrative rather than a ‘break’ in ludic-narrative cohesion. This avoids the loss of control sometimes associated with more emotionally involved games.

Of course, by designing an interface that encourages the scrutiny of each line in this way, Mertz naturally increases concentration on the task at hand, but by framing that interface in an integrated and fictional way - the hyperfamiliar cell phone - she hooks into both the direct feedback loop and the time-altering properties of that device’s design. The increased character animation, environmental animation, and background audio of the ‘in-person’ dialogue sequences, when compared to the more statically presented phone interface, draw the player’s attention to this altered sense of time, and recontextualizes the dialogue cascade as superimposed and emphatic - a neat metaphorical frame for the emptiness of these one-sided conversations. Indeed, this feeling of being wilfully misunderstood in person, but heard clearly on text, creates a clear series of narrative and emotional goals leading the player back to their phone. The end result is a surprising granularity in emotional player activity. This, coupled with the visual and tactile familiarity of the game interface, increases the player’s feeling of being in-role and likely encourages a loss of self-consciousness.

We Should Talk iterates on the standard dialogue interface to create an emotionally and narratively involving experience. Though short in length, it demonstrates that a granular approach to text selection can facilitate flow, and points the way for further developments in narrative interface design.

Where Next? - Design Recommendations

By applying the lenses of time, space and interface to the above three case studies, it is evident that dialogue systems which deviate from the ‘standard dialogue meta’ can successfully facilitate flow in players. However, mimicking the designs of these games will only result in their absorption into a new standard dialogue meta. What follows is a list of new avenues for dialogue design, inspired by the work undertaken in this paper.


Frequency (singular) - despite the prevalence of duologues (plays between two people) in theatre, games writers have seemed reluctant to explore a singular dialogue system that might represent two people talking for an entire game. Perhaps this is a function of games designers leaning into the format’s strengths, multiple characters being cheaper to add than live performers are to hire, but a game of significant length staging a conversation between only two characters feels oddly radical.

Frequency (repetitive) - innovation can often be found in pushing traditional concepts to their limits, and a game which locks a player into a repeating dialogue sequence that they must learn to escape might have quite the diachronous effect.

Speed (stretch) - Signs of the Sojourner has shown how powerful slowing down a player’s experience of dialogue can be; using stretch to increase a player’s reaction time even further, maybe encouraging the reading of body language in animated characters or the close analysis of, as in We Should Talk, clauses or even individual words, might be a productive path.


Layout (parallel) - despite the existence of split-screen gaming, one layout barely touched upon in dialogue design is parallel. An interactive fiction version could be easily prototyped using Ink, in which, say, two conversations progress down either side of the screen and it is up to the players how long they remain ‘in control’ of each one. Alternatively, choices in one could ‘rewrite’ the other, as when films flash back to show a different perspective on events.

Spatial opposition - as raised in this paper, a system for the procedural varying of line length and dialogue shape could easily be developed. A ‘stamina’ system could fuel the length or intensity of a player’s lines (more stamina required to deliver longer lines), and the dialogue system could respond by matching intensity during ‘rising’ action and opposing it during conflict.

Spatial segmentation - though identified in this paper as a site of flow disruption, a possible solution to this might be to maximise the feeling of segmentation rather than minimise it. The first act of the game might not include a dialogue system at all - maybe it’s an exploration of a dream-like 3D environment; the second might be only a dialogue system, reflecting on the player’s experience in the first; the third could unify the two, or return to the first but inflected with information from the second.


Integrated, fictional, ecological - dialogue systems that evince these interface frames are few and far between, but perhaps, given recent advances in shader technology, it might be possible for a game’s dialogue to be delivered as a texture on a 3D environment. Gone Home (The Fullbright Company, 2013), but with the found text written on every object in the house; a player might interact by ‘writing’ their character’s inner monologue onto the surface, marking the house with their own thoughts.

Integrated, ludic, emphatic - a dialogue system that stages the process of editing, making each word a game piece that might be interacted with by the player/editor in some way - red strikethrough for a suggested cut, green highlighter for an enthusiastic remark. The next draft returns with certain changes made, others not, and the palimpsestic effect of these sequences builds up a relationship between the player/editor and the writer over time, although they never directly speak.


Though game academics have long engaged (however contentiously) with ideas of narrative, little attention has been paid specifically to the design of dialogue systems and sequences which, though far from the only deliverers of game narrative, are certainly some of the more audience-familiar. By identifying and appropriating multiple established structures for considering game narrative and general design - Wei, Calvert and Bizzocchi’s work on game narrative time and space; Jorgenson’s gameworld interface theory - this paper ‘translates’ them via a wide range of examples onto the specific field of game dialogue design. While some categorisations fit only loosely (and might signal the need for a wider sample, or further research), others provide compelling opportunities for close analysis of dialogue systems and sequences. A vocabulary for analysing dialogue systems through time, space and interface is successfully developed, and by extending these structures into a broad spectrum analysis of the ‘standard dialogue meta’, the sites most commonly disruptive to flow are easily identified. Three case studies, each focusing on a different structure, demonstrate that the innovation of these dialogue designs lies in these structures themselves: if time, space and interface are indispensable tools for analysing dialogue systems, they may also be crucial tools for designing them.

Finally, using observations noted during research, writing and the performance of case studies, this paper extrapolates a number of design recommendations and possibilities for innovative dialogue systems. At least one of these ideas will form the basis of the writer’s final project, in fulfilment of their Masters degree, but their hope is that the rest might provide inspiration to other designers looking to further investigate the disruptive space between narrative and play.

Key Sources

Csikszentmihalyi, M. (1990) Flow: the Psychology of Optimal Experience. NY: Harper Perennial.

Echodog Games (2020) Signs of the Sojourner [Video game]. Echodog Games.

Jorgenson, K. (2013) Gameworld Interfaces. Cambridge MA: MIT Press.

Mertz, C. (2018) We Should Talk [Video game]. Mertz, C.

Night School Studio (2016). Oxenfree [Video game]. Night School Studio.

Wei, H., Bizzocchi, J. & Calvert, T. (2010) “Time and Space in Digital Game Storytelling” in International Journal of Computer Games Technology, Vol 2010.


Aarseth, E. (2001) “Computer game studies, year one” in Game Studies. 1(1), pp1–15.

Aarseth, E. (2007) “Allegories of space: the question of spatiality in computer games,” in Space Time Play: Computer Games, Architecture and Urbanism: The Next Level, von Borries, F., Walz, S. P. and Bottger, M. Eds., pp. 44–47, Berlin: Birkha ̈user.

Adams, E. (2010) Fundamentals of Game Design. 2nd Ed. Berkeley, CA: New Riders.

Austin, J. L. (1962) How To Do Things With Words. Oxford: Clarendon Press.

Bal, M. (2009) Narratology: Introduction to the Theory of Narrative. 3rd edition, Toronto: University of Toronto Press.

Bateman, C. (2007) Game writing : narrative skills for videogames. Boston, MA : Charles River Media.

Bauer, R., Suter, B. & Kocher, M. (Eds) (2021) Narrative Mechanics. Bielefeld, transcript Verlag.

Bauer, R. & Suter, B. (2021) ‘Narrative Mechanics: Strategies and Meanings in Games and Real Life’ in Narrative Mechanics, Bauer, R., Suter, B. & Kocher, M (Eds). Bielefeld, transcript Verlag.

Boluk, S. & LeMieux, P. (2017) Metagaming: Playing, Competing, Spectating, Cheating, Trading, Making, and Breaking Videogames. Minneapolis: University of Minnesota Press.

Bridgeman, T. (2007) “Time and space,” in The Cambridge Companion to Narrative, Herman, D. Ed., pp. 52–65, Cambridge, MA: Cambridge University Press.

Brusk, J. & Bjork, S. (2009) Gameplay Design Patterns for Game Dialogues. Proceedings of DiGRA 2009.

Calleja, G. (2009) ‘Experiential Narrative in Game Environments.’ DiGRA 2009 Conference Proceedings: Breaking New Ground: Innovation in Games, Play, Practice and Theory.

Calleja, G. (2011) ‘Emotional involvement in digital games.’ International Journal of Arts and Technology. 4. pp 19 - 32.

Chatman, S. (1995) Story and Discourse: Narrative Structure in Fiction and Film. Ithaca: Cornell University Press.

Chen, J. (2007) “Flow in Games” <https://www.jenovachen.com/flowingames/thesis.htm> (Accessed on: July 16 2021).

Collins (2021) Dialogue. <https://www.collinsdictionary.com/dictionary/english/dialogue> (Accessed on: June 25th 2021)

Costikyan, G. (2013) Uncertainty in Games. Cambridge, Massachusett: MIT Press.

DiGRA (2012) Digital Library <http://www.digra.org/digital-library/> (Accessed on: 25 June 2021).

Echodog Games (2019) What’s a Narrative Card Game Anyway.<https://www.echodoggames.com/blog/2019/04/19/what-s-a-narrative-card-game-anyway> (Accessed 19 June 2021).

Echodog Games (2019) Prototyping New Features in Signs of the Sojourner.<https://www.echodoggames.com/blog/2019/10/14/prototyping-new-features-in-signs-of-the-sojourner> (Accessed 19 June 2021).

Edgar, D. (2021) How Plays Work. London: Nick Hern Books.

Ellison, B. (2008) Defining Dialogue Systems. 8 July 2008. <https://www.gamasutra.com/view/feature/3719/defining_dialogue_systems.php> (Accessed 19 July 2021).

Eskelinen, M. (2004) “Towards computer game studies,” in First Person: New Media as Story, Performance and Game, Wardrip-Fruin, N. and Harrigan, P., Eds., pp. 36–44, Cambridge, MA: MIT Press.

Fernandez-Manjon, B., Fernandez-Vara, C., & Manero, B. (2013) ‘Stanislavky’s System as a Game Design Method: A Case Study.’ DiGRA 2013 Conference Proceedings: DeFragging Games Studios.

Fernández-Vara, C. (2014). Introduction to game analysis. New York: Routledge.

Frasca, G. (2004). ‘Videogames of the Oppressed: Critical Thinking, Education Tolerance, and Other Trivial Issues.’ In N. Wardrip-Fruin & P. Harrigan (Eds.), First Person: New Media as Story, Performance, and Game. Cambridge, MA: MIT Press.

Herman, D. (2002) Story Logic: Problems and Possibilities of Narrative. Lincoln: University of Nebraska Press.

Heussner, T., Finley, T., Hepler, J. & Lemay, A. (2015) The Game Narrative Toolbox. London: Routledge.

Hitchens, M.(2006) “Time and computer games or “no, that’s not what happened,” in Proceedings of the 3rd Australasian Conference on Interactive Entertainment, pp. 44–51.

Ingold, J. (2018) Sparkling Dialogue: A Masterclass. 1 December 2018. <https://www.youtube.com/watch?v=_vRfNtvFVRo> (Accessed June 21st 2021).

Ingold, J. (2019) Designing Investigate Conversations. 23 October 2019. <https://www.gamasutra.com/blogs/JonIngold/20191023/352685/Designing_Investigate_Conversations.php> (Accessed June 21st 2021).

Ingold, J. (2020) Dreaming Spires: Dynamic Narrative, Layer by Layer. 22 May 2020 <https://www.gamasutra.com/blogs/JonIngold/20200522/363472/Dreaming_Spires_Dynamic_Narrative_Layer_by_Layer.php> (Accessed June 21st 2021).

Iser, W. (1991) The act of reading: A theory of aesthetic response. Baltimore: Johns Hopkins Univ. Press.

Jayemanne, D. (2020) “Chronotypology: a Comparative Method For Analysing Game Time” in Games and Culture Vol 15, Issue 7.

Jenkins, H. (2004) “Game design as narrative architecture,” in First Person: New Media as Story, Performance, and Game, Wardrip-Fruin, N. and Harrigan, P., Eds., pp. 118–130, Cambridge, MA: MIT Press.

Juul, J. (2001) “Games telling stories”, in Game Studies. 1(1), pp1–12.

Juul, J. (2004) “Introduction to game time,” in First Person: New Media as Story, Performance, and Game, Wardrip-Fruin, N. and Harrigan, P., Eds., pp. 131–142, Cambridge, MA: MIT Press.

Juul, J. (2005) Half-Real: Video Games between Real Rules and Fictional Worlds. Cambridge MA: MIT Press.

Juul, J. (2007). ‘A certain level of abstraction. Situated play.’ DiGRA 2007 conference proceedings (pp. 510–515). Citeseer.

Koenitz, H. (2018) “Narrative in Video Games” in Encyclopedia of Computer Graphics and Games, Lee, N. (Ed).

Lindley, C. (2004) Narrative, Game Play, and Alternative Time Structures for Virtual Environments. Springer.

Mateas, M. (2004). A Preliminary Poetics for Interactive Drama and Games. In N. Wardrip-Fruin & P. Harrigan (Eds.), First Person: New Media As Story, Performance, and Game (Vol. 1, pp. 19-33). Cambridge, Massachusetts; London, England: The MIT Press.

Murray, J. (1997) Hamlet on the Holodeck: the Future of Narrative in Cyberspace. New York: Free Press.

Nitsche, M. (2007) “Mapping time in video games,” in Situated Play, Proceedings of DiGRA 2007 Conference, pp. 145–151.

Perron, B. & Schroter, F. (2016) Video Games and the Mind: Essays on Cognition, Affect and Emotion. North Carolina: McFarland.

Prince, G. (2003) Dictionary of Narratology. 2nd edition, Lincoln: University of Nebraska Press.

Ryan, M. (2001) Narrative as virtual reality: immersion and interactivity in literature and electronic media. Baltimore: Johns Hopkins University.

Sali, S., Wardrip-Fruin, N., Dow, S., Mateas, M., Kurniawan, S., Reed, A. & Liu, R. (2010) ‘Playing with words: from intuition to evaluation of game dialogue interfaces’ in Conference: Proceedings of the Fifth International Conference on the Foundations of Digital Games.

Schechner, R. (2006) Performance Studies. London: Routledge.

Searle, J. Speech Acts: An Essay in the Philosophy of Language. London: Cambridge University Press, 1969.

Short, E. (2016) Beyond Branching: Quality-Based, Salience-Based, and Waypoint Narrative Structures.<https://emshort.blog/2016/04/12/beyond-branching-quality-based-and-salience-based-narrative-structures/> (Accessed: August 11th 2021)

Short, E. (2019) Storylets: you want them.<https://emshort.blog/2019/11/29/storylets-you-want-them/> (Accessed: August 11th 2021).

Short, E. (2021) Design of Orwell’s Animal Farm – Emily Short speaking to the London IF Meetup,January 2021. 4 April 2021. <https://www.youtube.com/watch?v=B0Ccatc8Pa8> (Accessed: August 11th 2021)

Tanenbaum, T. J. & Tanenbaum, K. (2008) ‘Improvisation and Performance as Models for Interacting with Stories’, Proceedings of Interactive Storytelling, First Joint International Conference on Interactive Digital Storytelling.

Tanenbaum, J. (2011). ‘Being in the Story: Readerly Pleasure, Acting Theory, and Performing a Role.’ In Si, M., Thue, D., André, E., Lester, J., Tanenbaum, J. & Zammitto, V. (Eds.), Interactive Storytelling (Vol. 7069, pp. 55-66): Springer Berlin / Heidelberg.

Tychsen, A., Hitchens, M. & Drachen, A. (2009) “Game time: modeling and analyzing time in multiplayer and massively multiplayer games,” Games and Culture, vol. 4, no. 2, pp. 170– 201.

Wolf, M. J. P. (2001) “Space in the video game,” in The Medium of Video Game, Wolf, J. P. M., Ed., pp. 51–75, Austin, TX: University of Texas Press.

Yang, R. (2018) Notes on “Sparkling Dialogue”, a great narrative design / game writing talk by Jon Ingold at AdventureX 2018.<https://www.blog.radiator.debacle.us/2018/11/notes-on-sparking-dialogue-great.html?m=1> (Accessed June 29th 2021).

Zagal, J. P. & Mateas, M. (2007) “Temporal frames: a unifying framework for the analysis of game temporality,” in Situated Play, Proceedings of DiGRA 2007 Conference, pp. 516–522.

Zoran, G. (1984) “Towards a theory of space in narrative” in Poetics Today, vol. 5, pp. 309–335.


A Sharp (1999) King of Dragon Pass[Video game]. A Sharp.

Amplitude Studios (2021) Humankind [Video game]. Sega.

Bethesda Game Studios (2008) Fallout 3 [Video game]. Bethesda Softworks.

Bethesda Game Studios (2011)The Elder Scrolls V: Skyrim [Video game]. Bethesda Softworks.

Bethesda Game Studios (2015) Fallout 4 [Video game]. Bethesda Softworks.

Bioware (1998) Baldur’s Gate [Video game]. Interplay.

Bioware (2000) Baldur’s Gate II: Shadows of Amn [Video game]. Interplay.

Bioware (2003) Star Wars: Knights of the Old Republic[Video game]. LucasArts.

Bioware (2007) Mass Effect [Video game]. EA.

Black Isle Entertainment (1999) Planescape: Torment [Video game]. Interplay.

Blizzard Entertainment (2004) World of Warcraft [Video game]. Blizzard Entertainment.

Campo Santo (2016) Firewatch [Video game]. Campo Santo.

Case, N. (2014) Coming Out Simulator [Video game]. Case, N.

CD Projekt Red (2015) The Witcher III [Video game]. CD Projekt.

Crowther, W. & Woods, D. (1977) Colossal Cave Adventure [Video game]. Crowther, W. & Woods, D.

EA Sports (2020) Fifa 21[Video game]. EA Sports.

Exxos (1988) Captain Blood [Video game]. Infogrames.

Failbetter Games (2015) Sunless Sea [Video game]. Failbetter Games.

Infocom (1980) Zork I [Video game]. Personal Software.

Inkle (2020) Pendragon [Video game]. Inkle.

Interplay Productions (1997) Fallout [Video game]. Interplay Productions.

Introversion Software (2001) Uplink [Video game]. Introversion Software.

Ion Storm (2001) Deus Ex[Video game]. Eidos Interactive.

Larian Studios (2014) Divinity: Original Sin [Video game]. Larian Studios.

Larian Studios (2020 - early access) Baldur’s Gate III [Video game]. Larian Studios.

Lucasfilm Games (1990) The Secret of Monkey Island [Video game]. Lucasfilm Games.

Naughty Dog (2013) The Last of Us [Video game]. Sony Computer Entertainment.

Naughty Dog (2020) The Last of Us 2 [Video game]. Sony Computer Entertainment.

Nerial (2020) Animal Farm[Video game]. The Dairymen.

Nintendo EPD (2017) Legend of Zelda: Breath of the Wild [Video game]. Nintendo.

Obsidian Entertainment. (2010) Alpha Protocol [Video game]. Sega.

Obsidian Entertainment (2015) Pillars of Eternity [Video game]. Paradox Interactive.

Obsidian Entertainment (2016) Tyranny [Video game]. Paradox Interactive.

Paradox Studios (2012) Crusader Kings II [Video game]. Paradox Interactive.

Perfect Entertainment (1999) Discworld Noir[Video game]. GT Interactive.

Playabl Studios (2005) Facade [Video game]. Playabl Inc.

Playground Games (2021) Forza Horizon 5 [Video game]. Xbox Game Studios.

Quantic Dream (2018) Detroit: Become Human [Video game]. Quantic Dream.

Remedy Entertainment (2001) Max Payne [Video game]. Gathering of Developers.

Rockstar North (2013) Grand Theft Auto V [Video game]. Rockstar Games.

Rockstar Studios (2018) Red Dead Redemption 2 [Video game]. Rockstar Games.

Supergiant Games (2020) Hades [Video game]. Supergiant Games.

The Fullbright Company (2013) _ Gone Home_ [Video game]. The Fullbright Company.

Telltale Games (2012) The Walking Dead [Video game]. Telltale Games.

Troika Games (2004) Vampire: The Masquerade: Bloodlines [Video game]. Activision.

Valve (2004) Half-Life 2 [Video game]. Valve.

Valve (2020) Half-Life: Alyx [Video game]. Valve.

Valve (2011) Portal 2 [Video game]. Valve.

Volition (2013) Saints Row IV [Video game]. Deep Silver.

ZA/UM (2019) Disco Elysium [Video game]. ZA/UM.

11bit Studios (2018) Frostpunk [Video game]. 11bit Studios.