The player can manipulate the switch, but not the ship model.
The main goal of this analytical part is to show how the previous theoretical work can be used to say something more detailed about specific games. At the same time, the text performs a simple movement from games exhibiting the classical problems of interactive fiction, to games that can be seen as a re-interpretation of the utopia of interactive fiction.
Firstly, I'll be studying Myst to identify where the combination of game and narrative fails or succeeds. Then I'll examine the action games Doom and Unreal, as well as what happens when these games are used as multi player games. Finally I will re-evaluate the utopia of interactive fiction and examine Witness and Last Express. I think that these two final games point not to an interactive narrative but towards a computer game that is thematically closer to the movie or the novel.
To read a computer game is not the same as reading a poem, a movie, or a novel. In the traditional media, the object read is (still) assumed to contain a meaning that can be interpreted from the object. (No matter what claims have been made to the contrary.) Conversely, there is no tradition for seeing computer games as works of substance or meaning; there are no publications about the interpretation of computer games, no faculties, no tradition for this.
According to my model of non-linear texts, a computer game has a program level and a level of material. The material is traditionally interpretable elements such as text, graphics, and sound. Computer science does have (many) procedures for analysing programs, but this is not oriented towards interpretation. What is possible is to interpret the combination of program and material. In the simulation game Sim City (Wright 1989), the player is the mayor in a city and has to control such parameters as tax, buildings etc.. Sim City claims to be (and is often taken as) a realistic model of a city. But any simulation is obviously based on assumptions, and in Sim City crime comes from nowhere and can only be countered with police stations. And this is a clear ideological statement. The formula for crime in Sim City is not directly available to the player. (Even if explicitly created by the programmer.) The output of this formula could alternatively be calculated as a product of industrialisation and presented as pollution. (An environmentalist version.) The same material (the term "crime" and the graphical representation of the city) could easily be controlled by the reverse formula, where crime develops as a product of the number of police stations. (A combination that sees humans as fundamentally good, but corruptable.)
As a user it is hard to describe the program independently of its representation on screen - you do not have access to it. But there is a program, typically developed in a programming language like C++. As in the example above, it is hardly possible to perform sensible interpretations of an independent formula, independent of its context - in casu the material. The interesting part is the combination of program and material.
So far I have tried to describe the computer game as phenomenon and to sketch interesting ways of looking at his. I have presented a theoretical description of the game, but the final goal can't be just to confirm this theory. I will rather be using the theory to examine a number of things:
I will be making a distinction between game passages and narrative passages. The latter means 1) that the player cannot do anything 2) that time is compressed and that the time narrated is shorter than the time of reading. (In an earlier Christian Metz quote, the narrative was characterised by creating one kind of time using another kind of time. (p.*)) Game passages are interactive and happen now. The narrative is linear and has a non-correspondence between time of reading and time of the narrated.
When the computer game is counted as empty of content, it is usually because it is being measured by the same standards as used for traditional media. Myst (Cyan 1993) is undoubtedly the recent computer game that has been most accepted in cultural circles. For example, The Danish newspaper Weekendavisen chose to review Myst in the book section, concluding that:
Rigmor Kappel Schmidt and Peter Frost-Olsen praise the beautiful graphics and ambient sound. The game is not very well described. If we were to look at it more analytically, we might say that Myst is a very traditional game - the classical hunt for things to interact with, puzzles to solve. Graphically and thematically it was possible one of the first games to utilise the large amount of storage offered by a CD-ROM. At the same time Myst quite handily navigates the most obvious pitfalls of the genre.
In Myst you have to click around a small island world. According to the manual, "you" have been reading a book which you have suddenly been sucked into. You then gradually discover that this world has been created by Atrus, but that he and his two sons have been trapped in separate books. You hardly ever meet these characters during your playing; you get to see them in a book, but apart from a brief meeting with Atrus you do not interact with them directly.
In Myst, the player is never described. You see from your own eyes and are not represented on screen. According to the manual, you are "yourself". Thus you do not have to identify with another person, but act as you would. "[...] lose yourself in this fantastic virtual exploration and act and react as if you were really there." So the manual asks the player to create models for how he/she would react if he/she were actually in the fictive world. Obviously you are not "yourself", but play as a projection of yourself in the game world. That you do not have to identify makes some things easier: the game doesn't have to characterise the main character, and the player's possible reluctance to accept another personality is avoided.
The interface is the most problematic part of Myst. The frame narrative claims that you are inside a book, present in another world. But the interface does not represent any bodily presence in this world; no possibility of movement using the cursor keys. The mouse is the only option:
The graphics give the impression that all visible objects can be manipulated and that you can move in all directions. And many objects can be manipulated: meaning that they are present in both program and material. Other objects are only in the material. This is the same problem as in text-based interactive fiction, where the ambient description often contains objects that can not be manipulated. This problem is obviously not present in narratives, where the narrator decides what needs to be examined more closely, what is foreground and what is background. If a minor character only has two lines relevant for the plot, we only hear two lines. In a game, there is a risk that the player will want to ask more detailed questions, and that the minor character ends up repeating the same lines. And this of course breaks the illusion of a free game world that you can navigate.
As an example the player early on gets to a switch next to a fountain, You can turn the switch on and off, but you cannot pick up the ship model in the fountain:
The player can manipulate the switch, but not the ship model.
So while the introduction claims that you really are there, the interface posits very strict limits to what you can actually do. It would seem logical according to the game world (the material) that you could pick up the ship, but you can't (the program).
This means that the focus ends up on the two-dimensional picture. The effort needed to move the mouse pointer is not related to the distance between these two objects in the game world; the interface only relates to the two-dimensional representation.
In Myst, time stands still when the player doesn't act. There is never any urgency, no demands on the reaction time of the player. This is in some ways similar to the temporal freedom granted to the reader of a novel.
I have pointed out that you basically can't have interaction and narration at the same time. If a computer game is to have both interaction and narration, it has to alternate between the two. The top selling Myst confirms this, albeit in an intriguing way. On a concrete level, the player's actions in Myst are quite uninteresting. The player moves around collecting hints on how this switch must be flipped, how this wheel must be turned etc... All of this to collect the missing pieces of the books, Atrus and his two sons Sirrus and Achenar are captured in. Hints are found primarily through artefacts in the game world, like the two following.
Two artefacts with a story.
The paper has a double function of helping the player get further in the game, and of indicating a sender and receiver of the message (a man and a woman). Knowledge is presented and a part of story is told. We do not need the story of Atrus and Catherine to complete the game, but it adds some atmosphere. The book on the right is pure narration about the prehistory of Myst, but does directly give anything needed for the playing of the game.
Playing Myst, you meet many things that tell of Myst's prehistory. This narration is done through artefacts in the game world, but is always about what happened before the player started playing. The player can always leave a text and do something else. So a direct confrontation between the temporal logics of game and narration is avoided: It is not the time of the player that is compressed, but the time that is told.
There is one obvious contradiction in the time of Myst, since Myst contains a time machine. It ought to be possible to get to a point in time where the inhabitants of Myst were still around, but this is not possible. Myst is always empty. This abandonment and this displacement between the prehistory and the time of the game are always active in Myst.
The challenge given to the player by Myst is primarily in discovering how different mechanical devices in the game must be manipulated to gain further access to the game world. The program contains some logical conditions where a given "object" (represented in some material) must be manipulated in a certain way to give access to more material. Logical puzzles work well in a game context because they are purely causal. There is no doubt that the reason that you arrive at the end sequence is because you've entered the correct code. So it is possible for the player to discover the structure of the game, and this means that the player has reason to play. Puzzles work because they connect the program and material. One obvious criticism of this kind of game is that they are always about turning handles and finding keys - as they have been since Adventure. The real strength of Myst is the extent that this solving of puzzles also uncovers a prehistory.
There are both temporal and technological reasons for the game world being abandoned: It is very hard to create satisfying and convincing interactions between a player and computer-controlled characters. The most obvious reason being that computers do not understand natural language. The narrative frame does seem slightly forced and determined by technological and temporal possibilities.
To reconstruct a story during the playing of the game gives a feeling of connection between the story and the actions of the player. At the same time it would be quite easy to change the material of the game without changing the program. Myst could be made into a game where you - for example - live on a planet threatened by an asteroid and have to connect machines to defend the planet. Or the book theme could be replaced by a video cassette theme (which would be more logical since the books contain video anyway), but then Myst would probably not have been reviewed in the book section... But perhaps that is not the point. Instead I will say that Myst works well since the fairly inflexible program (Myst could more or less be made to work on two slide show projectors) is matched by a narrative frame that explains why the game world is abandoned. This can unfortunately seem slightly forced. As a game, Myst is a traditional puzzle-based interactive fiction, but one with high production values, and a narrative frame that concerns writing and fiction. And this makes Myst highly valued by traditional cultural criteria.
A large part of narratology has focused on the placement of the narrator in the universe of the text. That is, what grammatical traits are attached to the enunciation, what does the narrator know about the people and the world that is being told? In interactive phenomena it may be useful to focus on the placement of the player. The reader/player of an interactive work is necessarily placed; the player has certain possibilities of influence, and thus an interactive work sets the player as a part of the text world. In the same way that a narrator can be placed unclearly in a novel, so can the player in a game.
The Danish author Svend Åge Madsen's novel Af Sporet er du kommet ("You have come from traces/tracks") (Madsen 1984) contains a short story called Spor ("Traces/Tracks"). Spor is claimed to be a short story written by the computer Datam 18 to create a life story for the fictious character Rikkard Duebo Vem. Spor consists of 24 pieces of text, labelled A-X. The 23 of these refer to two or more of the remaining pieces, and the reader has to make choices as to what should be read the next time. The protagonist is an undefined "he". It is possible to discuss what kind of text, Spor is. There is a game element not only since it is clearly influenced by forking stories such as the Choose your own adventure series. It is also explicitly marked as such in the novel when the characters read Spor:
They all have been aware of the existence of an "exit" - meaning a good ending. And it has been possible to cheat. So Spor is at least described as being a game. According to what we might term the "contract" of games, the player has a fixed degree of influence in the text/game universe. Most pieces in Spor do live up to this, for example D:
So the reader makes a choice that corresponds to the choice of the protagonist. But in other pieces, the reader's choice is about what should happen as a consequence of the protagonist's actions:
And finally, the choice "Choose to act on his own" in piece E leads to the protagonist raping a woman against his will. So in E you do not even control the protagonist's actions, where in D you controlled the world. So the reader is placed in three different positions: As the protagonist, as an omnipotent God/writer, and as something controlled by the animal lusts of the main character.
The interesting part is that Svend Åge Madsens two other forking texts, Den slette fortæller ("The evil narrator", 1970) and the novel Dage med Diam ("Days with Diam", 1972) also claim to fork according to the choices of the protagonist, but that their forks also place the reader in several different position. I have to admit that I find it difficult to explain why this is so. But all three texts are perhaps not as much games as they borrow game structures. We might say that the narrativity wins over the game elements. But the primary point must be that we should not take the self-description of the text at face value, and that the placement of the reader is worth studying.
Historically, most action games for home use followed the arcade game principle where the player is never allowed to take a break. I am not going to claim that Doom introduced pauses to the action game, but it has been massively influential. In Doom, the player is not constantly followed by a threat. Rather, Doom begins with the player being safe. When you enter a room, all monsters suddenly begin moving and attacking you. When these monsters have been killed, you can once again take a break. Doom is still real-time, but the game world has been created so it is possible to take a break in that real time. It is an open question whether it is realistic that the monsters do not try to find the player when he/she stands still. Shouldn't they be running from room to room? On the other hand we don't know much about these monsters - and that is one of the advantages of placing a game in a not very well known setting.
Doom was received ecstatically by game players, but in other circles it has been declared extremely dangerous. On a nominal level, this is understandable since Doom is a very violent game, and has spawned a whole genre of equally violent games. There is something uncomfortable about the psychology of the genre, where the player kills thousands of monsters to get further in the game. It may remind us of the egoism of a small child before the child begins to see itself as placed in a social context. And it is a bit like the lack of understanding other people found in autism. Doom does seem to have been built exclusively for the player.
The different elements of Doom fit together extremely well. The interface is very simple: You can move forwards and backwards, turn, strafe, change weapons, open doors, and shoot. There is no possibility of using signs or at all communicating with the enemies. And that is the point: The opponents you meet in Doom are clearly marked as purely evil, unthinking beings that evidently have to be exterminated. This means that the simplicity of the interface is not a problem, because you do not wish to try to communicate. The minimal (and basically ignored) narrative frame of Doom is no literary masterpiece, but it serves the purpose of presenting the monsters as evil:
These opponents are not very intelligent. They are mainly occupied with attacking the player, but occasionally attack each other as well. There is also an element of randomness in their behaviour. This means that they, despite of their following well-defined rules, can act differently every time you meet them. This is a bottom-up construction, meaning that rather than describe an overall structure and choreographing every element to match that structure, you construct a number of small elements that interact.
The newer game Unreal (Epic Megagames 1998) is a 3d-shooter like Doom, and is also real-time. But Unreal adds some new tricks, as in the following pseudo-narrative passage:
1. Entering a hallway. 2. A body flies through the air.
3. The attacking monster steps forward
In the first image we are entering a hallway. To get this far may have taken a long or a short time. But whenever you reach the hallway, you always hear a scream and see this body flying through the air (and the monster coming towards you). It is not possible to construct any causal reason why this should be so. But the player's presence starts the scene, and as a player you are guaranteed to witness this scene, no matter when you arrive.
I call this pseudo-narrative because there is no variation in speed and no distance between the time of the narrated and the time of the reading: It is the temporality of the game; that time is not modified and that the perspective is fixed. But here some predefined sequences are triggered by the player's arrival. This gives the action some of the economy found in narratives: The player will always experience this interesting event.
From a non-technical point of view there is no reasonable explanation why the monster should wait attacking till the player arrives. The problem with this technique is that it makes the game world less credible. By playing many times, you get to understand that this coincidence is a construction.
So Unreal has a double temporality. Most of the game events are real time, other are on hold in a kind of pseudo real time, to get suddenly activated (and happen in real time) when the player arrives.
I have mentioned, but not yet discussed, that Doom and Doom II have a playing mode where the game is not specifically for one player, but is constructed for the simultaneous playing of several people:
Doom II: Two players watching each other (on different computers).
Multiplayer Doom requires several computers in a network. It does not presuppose physical proximity, even if playing over large distances can be unsatisfying due to delay between the computers (lag). All 3d-shooters that I am aware of have a multiplayer mode. A multiplayer game changes three basic things in relation to the games previously discussed:
These three factors make the computer game a complication, theoretically and practically. The evaluation of the player becomes more of a social function. The goal of the player is not just to acquire a structural understanding and performance in relation to the game, but also to understand and master the interaction between the game and the other players. The individual player is still at a specific screen and keyboard, but the situation now contains other sentient beings, that also create mental models of the game world. This makes lying, cheating, and faking relevant, and recursive situations arise where A considers B's mental model of A's mental model etc..
The most popular type of multiplayer game at the time of writing (1998) is the 3d shooter, where the players are free to play indefinitely on a single level (according to how the game it set up). The historical precursor of the modern multiplayer games is the original MUD (Bartle & Trubshaw) from 1980: MUD, often taken for "Multi user dungeon" is a multiplayer version of Adventure: A game world built (textually) from a number of rooms that you can move between using textual commands. MUD continues to be a large underground phenomenon on the Internet with thousands of MUDs with different themes and structures. The revolutionary part in this context is that most MUDs are continuously active. This means that the game goes on 24 hours a day, with no interruptions. MUDs usually contain characters controlled by the program (non-playing characters or NPCs), which means that a MUD can be active even when nobody is watching. (The programs internal representation of the game world is updated without it being represented visually). The persistence of such a game world leads to an absolute abolishment of temporal variations.
The multiplayer game posits steep demands on the temporal relations and equally sharp steep demands on the program construction. I am not aware of any multiplayer games built on narrative passages - this is since the multiplayer game works much against temporal variation, and also because games constructed according to an fixed sequence are much to inflexible to accommodate several simultaneous players.
The structure of Doom allows for variation and flexibility: In a way, the whole point of the action game and of most games as such is that they consist of a number of small components that can be combined in large number of ways. On top of this we typically find an ending principle influenced by these small elements. In Doom the energy of the player runs out by getting hit by opponents - accordingly each monster is connected to an amount of tension. In a board game like Monopoly it is the financial status of each player that risks going to zero. This kind of bottom-up design makes the game world infinitely more flexible than if designed according to a plot or a fixed sequence.
The temporal aspects are not the only things that distinguishes a computer game from a novel, there are other basic differences: The narrative basically works with fixed sequences, human relations, but computer games are almost always about physical movements in space. In Space Invaders you move a small spaceship, in Myst you move "yourself" around an island. With a sweeping generalisation we might say that narratives are about time, but that computer games are about space, or at least about space on a level of detail that is not relevant in narratives.
When the computer game is not capable of telling stories, but is yet often seen as to dry, abstract, or emotionless, these elements have to be introduced by other methods than that of narration. We have to acknowledge that the computer game has a variety of devices that work very well, and that there is no specific reason to work against these. So we have to make a distinction between narratives and the themes found in narratives. Most narratives are about things like human relations, ambition, and desire. Since these themes can not be introduced as simple additional symbolisation (narrative or not) of a simple game, the game structure has to be extended to contain models of the human relations presented. The themes have to be implemented in the program, be more than postulates on the packaging.
We need to get to a situation where the player can interact with a world that develops no matter whether the player acts or not. And the player has to be able to act without the game loosing its progression - the player still needs something to play for. I think some games point towards this utopia, but to describe this I will start by going back 15 years:
Witness (Infocom 1983) is from the golden age of the text adventure. Witness is text-based; the game world is represented textually and the player acts textually. Witness is the oldest game to get analysed here, but it also contains some fairly modern traits, traits I believe could be part of a future for interactive fiction. The main things is the relative autonomy of the game world, i.e. not everything revolves around the player.
Witness works as a dialogue between "the game" and the player. "The game" describes the scenario, and the player can react. (">" Marks the writings of the player.) The illusion is occasionally broken since there is no direct connection between the vocabulary used in the description and the vocabulary that the player can use:
>wait
The rain is falling heavily now.
>drink rain
(Sorry, but the program doesn't recognize the word "rain".)
Witness suffers from the recurring problem of interactive fiction, that not all the information in the description (textual or graphical as in Myst) can be interacted with.
Witness is a detective story with the player as the detective that has to examine the circumstances surrounding the death of a Mrs. Linder. Using the textual interface you have to examine things, collect evidence, interrogate suspects (to the degree that this is possible). You can then arrest a suspect when you feel certain, and the verdict of the court depends on having collected sufficient evidence. (No matter if you have arrested the right person or not.)
Espen Aarseth has made an analysis of the predecessor of Witness, Deadline (Aarseth 1997, p.115-128), where he describes Deadline as an autistic game (lack of relation to pain etc.), and concludes that the voice of the game is hard to place since it both describes the scenario, comments its own abilities (as in the above quote), and ask overall questions about the game flow ("This case has ended. Would you like to start your investigation from scratch? (Y/N)").
The question is then how to explain this unreliable not-really-narrator that the player interacts with. I think the inspiration is from the computer itself. The original Adventure was created at a time when almost all operating systems were based on command lines. If a command is typed correctly, it is executed and the results are displayed. Mistyping results in an error message. This dialogue between user and computer works on many different levels; failure to understand the user's command; failure to execute the user's command; the result of the user's command. Another possible inspiration is Joseph Weizenbaum's classical program Eliza (1966), where a program works as a therapist, answering answer or question with a new question. (This connection has been suggested by Niesz & Holland 1984).
Where the time of Adventure is rather abstract, with a "move" as the unit, time in Witness is measured in minutes. The game starts at 8:00. Every action performed takes a minute in the game time. It is possible to wait for a person. ("wait for Monica") or to wait to a certain point ("wait to 8:59"). If something happens in the meantime, you are asked if you want to go on waiting. This allows the player to skip dull periods.
The really interesting part about Witness is the way in which the player's actions interact with some fixed events that take place with or without the player being present. To show this, we can look at a passage early in the game. You have arrived at the Linder family, have greeted everyone and followed Mr. Linder into his office:
[...]
This is obviously the office of Mr. Linder's company, Pacific Trade Associates. At the west end of the office, a massive desk of teak and mahogany faces toward the window. It has no drawers, but the top is covered with piles of letters, some newspapers, a telephone, and various souvenirs from the Far East.
Behind it is a large ornately-carved chair, like a cruiser escorting a battle ship. A simple wooden chair, polished smooth by visitors, flanks the desk on the other side. On the north wall is a lounge, upholstered in green velvet and a bit lumpy, with a framed wood-block picture hanging over it. On the outside wall, next to a door and window, stands a grandfather clock, ticking relentlessly. A file cabinet stands in the corner.
The door to the interior hallway is open.
A cat is sleeping in the corner.
Linder sits down in the carved chair.
"If you'll just take a chair, I'll explain what this is all about."
The rain outside is falling heavily now.
>sit on wooden chair
You are now sitting on the wooden chair.
Linder begins his story. "My late wife, may she rest in peace, got involved with a young man named Stiles. Naturally I tried to stop this affair, but without much success. Since my wife passed away, this Stiles fellow has gone off the deep end, I'm afraid, and blamed me for her death. I tried my best to ignore him, but he seems to have lost his senses. This morning I received this note and decided to ask your help." He hands the note to you.
[...]
>wait
Time passes...
The clock chimes 9 times to mark the hour.
You hear the door bell ring.
Linder looks toward the window and says, "I don't think Phong has answered the door bell yet." He reaches toward the butler's button and at the same instant shouts "Stiles!" You turn around and dimly see a figure outside. Suddenly there is a flash of light and an explosion, and the window falls into dozens of shiny shards. The cat bolts and disappears somewhere. The figure outside turns and runs before you can see the face. When you turn back around, you see Linder slumping down in his chair, with a bloody stain spreading across his silk shirt. He teeters on the edge of the seat, then falls onto the floor, quite dead.
It's now 9:04 p.m.
If you don't accept the chair, the following happens:
>wait
Time passes...
You hear the door bell ring.
Linder looks toward the window and says, "I don't think Phong has answered the door bell yet." He reaches toward the butler's button and at the same instant shouts "Stiles!" You turn around and dimly see a figure outside. Suddenly there is a flash of light and an explosion, and mortal pain radiates from your heart. As blood fills your lungs and a scream fills your brain, you feel sure of only one thing: you should have taken a chair when Linder asked you to.
There are other possibilities: You can choose not to keep your appointment and rather go to the back yard to see Stiles sneak towards the window.
Stiles always sneaks into the backyard just before 9:00. Correspondingly, Linder's daughter Monica always goes to the movies just after the game has begun, and she always returns at a certain time. Where Myst or Unreal openly feature video or events that only happen when the player arrives, the events in Witness have an autonomy that makes the game world more credible and lifelike.
In Reading for the Plot (1984) Peter Brooks makes the claim (using Todorov) that the detective novel is a meta-novel in the sense that it is about a person that using clues (discourse) reconstructs a story - a description of the crime. (Brooks p.29). This means that the detective and the reader perform the same movement, and that the detective novel gets much strength from this parallel between reader and detective. A computer game based on a criminal case works well since the player also uncovers a story that is earlier in time than the game events. The detective novel is obviously not just the uncovering of a story, but a story in itself. In Witness, the uncovering is a game, which is not entirely the same thing. But it is a detective game, and it carries a minimal ideal story - the detective finds and arrests the culprit. You then play to discover how to actualise this ideal story. The uncovering of the structure of the game also leads to an uncovering of the prehistory. Myst is not a detective story, but follows the same strategy - through your movements in the game you uncover a prehistory. In both cases the uncovering of the past has some relevance for the game: It may for example tell us who would have motive to kill Mr. Linder (or the player).
The world of Witness does not - unlike that of Doom or Unreal - seem created just for the benefit of the player. Witness gives the impression of working independently of the player. Furthermore it succeeds in building dramatical tension based on human relations. And this happens without locking the player in non-interactive sequences. The weak point of Witness is primarily the mode of interaction: You try to communicate using normal sentences, but the program turns out to be incapable of understanding anything but the most basic constructs. Correspondingly it turns out to be quite difficult to interrogate the game characters, since they understand very little and can only answer with a few canned phrases. Again this is a conflict between the promises of the material - the characters are textually presented as characters - and what the program can deliver - a few predefined responses.
Witness is an early example of a game about human relations, but one that works without locking the player in an inflexible duality between narration and interaction.
Last Express: Miss Wolff, the protagonist Robert Cath, and a conductor.
The Last Express (Brøderbund 1997) is a newer and more advanced attempt at creating an interactive fiction that does not trap the player within a few pre-defined choices. It is not a perfect game, neither does it match the utopia I have described. But it is a good example of how the themes of narratives can be added in a context that is more game and less of an inflexible tree structure. In some respects, Last Express is a lot like Witness - it is partly a detective story, and you uncover a prehistory while playing.
Graphics, narrative frame and dialogue are of unusually high quality. But the most interesting element is the use of meaningful events that take place in the real-time of the game (rather than being told in the same time). You can witness discussions and events that are meaningful in the overall game, but without being locked in time or without interactivity. This solves some problems, since the games I have previously discussed (with Witness as an exception) only introduce interesting material by adding artefacts or having straight narrative passages, where the player can't do anything. Last Express including interesting events in the real time of the game.
The interesting element of Last Express is that it is a game world built around a number of fairly autonomous characters that interact with each other and with the player. The player can move around six train carriages, talk to the characters and listen to conversations.
During this thesis I have described the computer game as a structure with an arbitrary narrative frame, but in this case I have to acknowledge that the Last Express has an excellent frame - conflicts on the last Orient Express before the breakout of World War I. The game contains many different nationalities - British, French, German, Russian, Serbs. They all speak their own language (the game is subtitled). There are various discussions of the political situation, the young Russian anarchist's declaring his belief in the class-less society etc.
According to the manual, Last Express has a context-dependent interface. This basically means that it works like in Myst: You can only activate certain things and perform certain actions, and this is indicated through the shape of the mouse pointer. Dialogue works by simply clicking on a person, which makes the protagonist say something that the player cannot actually control. A single click may result in a long or short conversation. On some occasions, the protagonist becomes part of a longer and clearly narrative passage.
Some things are not entirely logical: When you first enter your compartment and find the body of Cath's friend Tyler, your clothes get bloodstains all over. Tyler's jacket hangs on the wall, but you can only put it on once you have hidden the body.
At first you can't pick up the jacket, later you can.
Within the game world there is no reason for this - it is of course a limitation within the program, and a limitation according to the flow of the game. It is not an open world to interact with.
If you make a mistake or wish to redo some action, it works in a pretty unique way. Using a clock you can rewind time to the point you wish to start from:
Using a clock you can turn back time.
In some cases you do not have the option of being present at a specific time since the game is locked in a narrative passage, and the clock will only let you jump to a point in time where the player can interact.
You assume the role of the young American Robert Cath, boarding the Orient Express only to find that his friend Tyler has been murdered. Cath then decides to pretend being Tyler. This is a sensible construction: As described, there are problems in getting the player to identify with another person - to act in a situation is very different from simply understanding why another person would act in a certain way. There is a symbolic difference between identifying with a character and actually performing that character's actions - even if it is just a game. Last Express does not blindly seek this problem - the protagonist is a blonde hero - but at the same time, an extra effort is made to evade the identification problem. Firstly by letting the protagonist be an American on a European train, leading to a slightly distanced view on the game world. Secondly by letting the protagonist assume another person's identity, thus minimising the potential conflict between the identities of the player and the protagonist.
But Last Express is also about uncovering what happened before the game began. This goes for the other character's relation to each other, and it is also relevant to the question of what the protagonist has done previously. On the first night you are introduced to Miss Wolff - still under a false name - and our protagonist tells her "I believe we've met before." But this is never uncovered. The "I" of the protagonist possesses knowledge that is not communicated to the "I" of the player.
The game uses a variety of filmic devices: You basically move about watching the world in first-person perspective. When you are in a dialogue with the game characters, you change to an outside view. If killed by an opponent, the game stops. If arrested (you are a fugitive from the law), the game is also stopped, and a female voice reads from a diary describing the events:
A diary explains what has happened to the protagonist.
The diary belongs to a fellow passenger. In this game, death is clearly also connecting to increased knowledge for the player.
The placement of the player in Last Express is ambiguous. On one hand most of the game is about moving around the train and watching with the eyes of the main character. Accordingly the player is staged as the protagonist. When interacting with other people, you simply click on them, which leads to a narrative passage. In such a narrative passage, Cath says (and does) a number of things not specified by the player, so the player is rather set as a counsellor that can push Cath in different directions.
Of the games discussed here, Last Express is the game to most successfully combine narrative material and interactivity. Structurally, it is due both to the (more or less) bottom-up construction with interacting characters, but as a game it is in a way closer to the action game than Myst is: You often die in Last Express. As in Witness, the action game, and most interactive fictions, the narrative frame has one positive ending and a plethora of negative endings: Getting caught by the police, killed by opponents, blown up by a bomb. In every case, the game ends and you restart before you committed the last error. So basically it is a simple structure with an ideal line to follow. The point is that within this you can explore the game world, discover the same thing in different ways, listen to a number of conversations.
Placing the game in a setting with strong spatial and temporal limits - a train - makes it easier to create a flexible game world. At the same time, time is (so to speak) meaningful, since the train moves through different countries in a Europe on the brink of war.
Finally, there is good reason for creating the graphics in the style of the 1910's, Art Nouveau - since Art Nouveau has large monochrome areas, the images are easily compressible and take up very little space on the CD-ROM. So this style has a technological advantage over photography.
Last Express does unfortunately contain much of the non-repeatability of the narrative. Having completed the game, there is not that much reason to play again, even if there is some attraction in trying to complete the game in a different way. I consider Last Express the most successful attempt in creating a game that uses the themes of the novel or the movie, but at the same time evades the worst conflicts between narrativity and interactivity.