[NOTE: This is the full text of a paper we submitted to FDG'11. It was rejected for being too subjective and not citing enough other work. Fair enough. But poorly referenced subjective rants are just what blogs are for, right? So here it is.]
Ethical Choices in Videogames: Lessons from Moral Psychology
By Dan Staines and Malcolm Ryan
How do we create engaging ethical scenarios in games? This question has been taken up with seriousness by many designers, wanting to see their work grow beyond pure action and address deeper aspects of our lives. We are making progress, but existing designs are still too simplistic. They neither engage us as strategic gameplay nor as meaningful stories. To answer the question we must look deeper into moral reasoning itself to learn the skills it involves and how to engage them. In this paper we investigate the Four Component model of moral psychology to see what light it can shine on the problem. The result is a pattern for a holistic system of ethical gameplay, incorporating ethical identity, investigation, choices and challenges.
In recent years there has been a surge of interest in ethical decision making in computer games. A few stand-out titles (Mass Effect, Fallout 3, BioShock) have shown us the potential narrative power of placing the player in a morally challenging situation requiring them to make a difficult choice. These kinds of scenarios have been the stock and trade of storytelling for millennia, but we are only beginning to understand how to make them work in an interactive context. There is potential for great narrative power to be had in such interaction but sadly so far we have typically squandered this power on choices that are obvious, over-wrought and ultimately insignificant.
Part of the solution lies in gaining a better understanding of the processes of moral decision making itself. In this paper we look to the cognitive-developmental psychology of the ”Neo-Kohlbergian” school of researchers to gain insight into the variety of skills that comprise ethical reasoning. The Four Component Model of James Rest and colleagues will help us to think more broadly about the entire process of recognising an ethical problem, solving it and enacting the solution. We then consider how this understanding can help us to improve the way we represent ethics in our games, n terms of both strategy and narrative.
2. Moral Psychology: The Four Component Model
The Four Component Model is a cognitive-developmental approach to moral psychology, developed by James Rest and colleagues and expanding on the work of Jean Piaget and Lawrence Kohlberg, but diverging from its predecessors in terms of their emphasis on rationality. Where Kohlberg saw the capacity to make rational moral judgements as the basis of moral maturity, Neo-Kohlbergians such as Rest contend that it is merely part of a much larger and more complex cognitive apparatus. According to their model, there are in fact four key psychological components that comprise the complete moral agent:
- moral motivation – the desire to act on moral decisions and focus on them to the exclusion of other concerns,
- moral sensitivity – the ability to recognise and respond to moral phenomena,
- moral judgement – the capacity to engage in moral reasoning and make moral choices,
- moral action – the ability to act on moral decisions and see them through.
Taken together, these form the Four Component Model of moral functioning.
When we think about designing ethical scenarios in our games, it is normal to focus on moral judgement – deciding the right thing to do in a presented situation – but if we do so we are only engaging a fraction of our player’s moral faculties. So lets look in more detail at each of the four components before investigating how they might influence our design choices. In what follows we draw on the elaborations of Narvaez and Lapsley to the Four Component model.
2.1 Moral Motivation
Moral motivation is the impetus to act on our judgement, the desire to do the right thing. It is about cultivating an ethical identity which values courtesy, honour, generosity and fairness. Without these qualities, judgement lacks motivational force and will seldom result in action.
2.2 Moral Sensitivity
Moral sensitivity is the ability to discern that an issue is moral in the first place and recognise how it might be interpreted by different participants. It involves a number of skills including the ability to connect to others and understand their perspectives. Communication (both speaking and listening) is a key skill of moral sensitivity, as is the ability to recognise and control one’s own social bias.
2.3 Moral Judgement
Moral judgement is the act of determining a proper course of action in a moral situation. This is the component that we are most familiar with. It requires skilled reasoning and the ability to identify and employ different judgement criteria. It also requires the ability to predict and evaluate the possible outcomes of our actions to plan our behaviour.
2.4 Moral Action
Moral action is employing the skills to carry out the solution. Skills of moral action include the ability to negotiate, to solve interpersonal problems, to take initiative for action, and to show courage to persevere in the face of obstacles.
3. Game Design: Mechanics and Metaphor
Every game is the interplay of two systems: the mechanical system of the rules and goals (the ‘mechanics’) and the narrative meaning attached to those mechanics (the ‘metaphor’). Thus for example, chess’ mechanics describe the movement of various pieces and the victory condition, but its metaphor is one of battle between ‘pawns’ and ‘knights’ concluded by ‘capturing the king’. In chess the metaphor is very loose, in other more abstract games (such as tic-tac-toe) it is completely absent, but in most computer games it is a very important part of the game.
The mechanics and the metaphor both give meaning to events and objects in the game. The mechanical meaning of an event is its strategic importance. An event that advances progress towards the goal is positive, an event that obstructs process is negative. The narrative importance of an event, on the other hand, depends on the significance given to it by the metaphor. To return to chess. the strategic significance of check is that it forces the player to defend the king, the narrative significance is that the monarch is in danger and one of his servants must urgently act to defend him. In designing a game with a significant metaphor, care must be taken to make the ‘intrinsic narrative’ of the mechanics and the ‘extrinsic narrative’ of the metaphor coincide, else the player may experience dissonance.
Ethical decisions are intrinsically decisions about other people and how they are affected by our actions. An abstract game is amoral, there is nothing good nor evil about capturing an opponent’s knight or threatening his bishop – it is ‘just a game’. It is only through metaphor that we can bring the player to regard these actions ethically. Thus maintaining narrative immersion is important. Dissonance between the mechanics and the metaphor is a threat to this immersion. It takes the player out of the story and reminds them they are playing a game.
We contend that ethical systems in most existing games are dull, both strategically and narratively. They are over-simplified and engage the player’s ethical skill set in a very shallow fashion. They are also regular sources of dissonance, as the narrative meaning of a moral event has little connection to its strategic importance. Let us consider the Karma system of Fallout 3 as a particular example.
3.1 Karma in Fallout 3
The Karma game mechanic in Fallout 3 has been praised as a victory of Utilitarian ethics but we see little reason to praise it as good game design. It consists of a single axis variable, ranging from +1000 to -1000, representing how ‘good’ or ‘evil’ the player character (PC) is. It starts out at zero representing an ambivalent ‘neutral’ character and is affected by certain player choices throughout the game. The list of actions which affect your Karma shows a certain perverseness and inconsistency. For example, you lose 100 Karma for killing a ‘non-evil’ towns-person but nothing for devastating a raider base (as raiders are automatically deemed ’evil’). Furthermore, you can earn back that Karma by giving away two bottles of water to beggars.
The consequences of Karma are limited. Some companions will require a particular level of Karma (good, bad or neutral) before they agree to accompany you. If your Karma is very high or very low you will begin to randomly encounter powerful enemies sent to kill you. So for early, weak characters there is a strategic incentive not to act too moral or immoral to avoid danger.
The mechanics of Karma are not strategically interesting. It is easy to adjust your Karma to whatever setting you desire, with relatively little sacrifice. We suspect that this is deliberate, that the designers wanted to treat Karma as a largely cosmetic attribute that had no major strategic importance giving the player the freedom to play whatever morality they wanted without endangering the outcome of the game. However, in doing so they have actually weakened its narrative significance, as the dissonance between mechanic and metaphor destroys the player’s suspension of disbelief.
The Karma system aside, Fallout 3 is known for presenting the player with a number of moral problems. These range in sophistication from the farcical (the Power of the Atom quest which invites the player to detonate a nuclear bomb in the middle of an innocent town – in return for a handful of currency) to the clever (the Oasis quest which asks the player to assist in the euthanasia of a person in a paralysed state). It is interesting to note that this latter decision, which is by far the most ethically sophisticated in the game, has no Karmic consequence, unless the player chooses the outright evil option (deliberately causing the victim to die in extreme pain).
This disconnect is informative: a single Karma scale can only represent characters who are cartoonishly heroic or villainous; real moral character is sophisticated. This realisation is useful; it means that there is room to create an ethical system that is both strategically rich and narratively subtle. We will next look at each of the four components and consider how they can help us build such a system.
4. Designing ethical games
Designing a meaningful ethical game is more than just creating ‘moral choices’; it is about creating an entire system that engages the skills of moral sensitivity, judgement, motivation and action. As a strategic system it should reward exploration, offer interesting choices and challenging obstacles. As a narrative system is should present a world populated with engaging characters, a dramatic plot and room for self-expression through role-play. Let us now look at each of the Four Component in turn and consider how it might inform this design.
4.1 Moral motivation
The key game dynamic of moral motivation is identity construction, that is giving the player the ability to decide who they are. From a narrative point of view, this requires provision for a diversity of moral character beyond just ‘good’ and ‘evil’. Ultima IV took a step in this direction, representing eight different virtues: honesty, compassion, valour, justice, honour, sacrifice, spirituality and humility (although by requiring the player to optimise all eight, it ultimately homogenised the character and suppressed self-expression). There are also a variety of ways that a character can be evil; consider the seven deadly sins for example. An interesting character is not ‘good’ or ‘evil’; she has a mixture of virtues and sins, even existing at times in opposition. If we are want moral content in our games, we need to start by allowing complexity in their character.
Narrative also requires a href=”http://www.amazon.com/gp/product/1592003532/ref=as_li_tf_tl?ie=UTF8&tag=woonpl-20&linkCode=as2&camp=217145&creative=399353&creativeASIN=1592003532″>character growth. The character’s moral identity should change over time as she interacts with the world. This growth should drive towards establishing the uniqueness of that character, in the same way that the sun and the wind decide the particular shape of a tree as it grows. Looking at the character one should see the scars of the history that shaped her: the humiliating moment that made her humble or perhaps spiteful, the praise that fueled her honour and her pride.
Strategically this means a constructing multivariate representation of ethical identity. The system should be an interesting construction toy, allowing multiple affordances through different combinations of identity choices, but with constraints that create a dynamic feedback system rather than allowing unfettered growth in any direction.
A richly interactive toy such as this allows the player to express himself by setting his own goals for his character. One possibility might be for the player to select a ‘self-image’ at the outset of the game and then measure their progress against that image. The cost of an action that violated this image would be a loss of self-approval, hindering the player’s progress until it was remedied.
4.2 Moral sensitivity
The key game dynamic of moral motivation is investigation. A typical ethical conundrum in a game lays out the facts and then presents us with a set of choices; in the real world things are not presented so neatly. Moral sensitivity is the skill of being able to discern the importance of a situation in the first place, and one of its essential components is the ability to identify and express emotions. The extent to which one can practice this skill in a game depends largely on the emotional complexity of the characters the player encounters. A character possessed of complex emotions invites scrutiny and empathy, encouraging the player to recognise the role emotional states play in motivating their behaviour. A one dimensional caricature, as many of our NPCs are, does not invite emotional scrutiny before it doesn’t have any emotions worthy of scrutiny.
Moral sensitivity also requires recognising diverse moral stand-points. As with the PC, we tend to give our NPCs simplistic labels as ‘good’ and ‘evil’ rather than give them depth of moral character. The Megaton quest in Fallout 3 is again a rather egregious example. Sheriff Simms is an affable, fatherly stereotype of ‘goodness’; Mr Burke is a scheming villain. These kinds of caricatures deny ethical engagement. More interesting characters need more complex moral stand-points and internal complexity. Simms could be scrupulously upright and hostile to strangers, but with a genuine love for his son and a blind spot when it comes to pleasing him. Burke could be an outcast with a grudge, a proud man who has faced discrimination (warranted or not) on the basis of his connections. Such characters would provide much more material to engage the player’s moral sensitivity.
To establish this depth, we must give the the player time to explore, to get to know the characters, build relationships with them and uncover their problems. Far too often we encounter scenes in RPGs which begin with “Welcome stranger! Please help us answer our ethical problem…”. Such an abrupt introduction prevents exploration. The ‘show; don’t tell’ mantra applies here. Rather than front-loading all the characters’ motivations in introductory dialogue, the designer should reveal them slowly through ongoing interaction with the player. This also gives the player time to form attachments that can later be tested when conflict arises.
4.3 Moral judgement
The key game dynamic of moral judgement is interesting choice. A binary choice between the ‘good’ or ‘evil’ outcome is rarely interesting. Once the player has determined their desired moral direction such choices are usually obvious. A diversity of moral criteria helps here: choices can become genuine moral dilemmas, tradeoffs between different virtues and vices, informed through investigation and dramatic in terms of their effect on characters.
Strategically, these choices need to balance risks and rewards, offering what Schell terms ‘triangularity’. They should be informed through investigation, but not obvious, and they should be consequential, having long term impact on the game and feeding back on future choices. It is through these choices that the player constructs their character and they need to provide a complex strategic space.
Narratively, choice should be be dramatic. That is, it should address conflict between characters or conflict within characters (such as a character fighting a drug problem). To avoid being over-wrought, conflict should be personal. That is, it should deal with characters the player knows, rather than the anonymous masses. A good story is less about saving the world and more about saving just one person or saving yourself. Again, the ‘show; don’t tell’ mantra applies. The consequences of a choice should be reflected in ongoing changes to the player’s relationships. A one-off judgement affecting an anonymous character who will never be encountered again is not drama. We must give narrative context and consequences to our choices.
4.4 Moral action
The key game dynamic of moral action is challenge. Making a moral judgement may be easy, but carrying it out should be difficult, requiring courage and perseverance. This implies an ongoing and possibly costly commitment. Mechanically, ethical action should be balanced against risk of failure and some kind of material cost. Choices should not be resolved instantaneously, but should require action over time with potentially mounting costs and opportunities to abandon the choice. This provides long-term vs short-term tradeoffs and enriches the strategic space. In terms of narrative, ongoing action provides drama by allowing us to place obstacles in the player’s path, threatening the goal.
The most challenging problem for game design is to allow moral action with subtlety. In purely physical systems like combat we have a continuous space of action with emergent gameplay that permits the player to act with a high degree of subtlety once they master the controls. Our ability to implement social interaction (usually dialogue) falls far short of this mark. Real moral action often requires careful employment of interpersonal skills. This kind of care does not translate well to selecting from discrete dialogue choices. While we could simulate social interaction with continuous combat-like mechanics, representing that simulation with a convincing metaphor would be very difficult. We simply lack the emotional and linguistic AI to do so.
Making our games ethically engaging calls for more than just cribbing moral dilemmas from an ethics textbook; it requires the construction of a narrative world with meaningful characters and a diversity of moral perspectives. It requires space for the player to form a detailed ethical identity and to explore and grow that identity through meaningful choices and challenging gameplay. In short, it requires morality to be the heart of the game rather than a cosmetic feature.
Our next plan is to put our money where our mouth is and attempt to implement such a game. We intend to reimplement the Megaton scenario from Fallout 3 in an interactive fiction engine such as Inform and then modify it along the lines suggested in this paper. We do not expect this to be easy. Designing such a game will be challenging but we hope we have shown how the insights of moral psychology can help guide the way.
- T. Hartmanna, E. Toza, and M. Brandona. Just a game? unjustified virtual violence produces guilt in empathetic players. Media Psychology, 13(4):339–363, 2010.
- L. Kohlberg. Essays on Moral Development – Vol. 1. Harper and Row Publishers, San Francisco, 1981.
- D. Narvaez and D. K. Lapsley. The psychological foundations of everyday morality and moral expertise. In D. Lapsley and C. Power, editors, Character Psychology and Character Education, pages 140–165. University of Notre Dame Press, 2005.
- J. Piaget. The Moral Judgement of the Child. Routledge & Kegan Paul Ltd, London, 1932.
- J. R. Rest, D. Narvaez, M. J. Bebeau, and S. J. Thoma. Postconventional moral thinking: A Neo-Kohlbergian Approach. Psychology Press, 1999.
- J. Schell. The Art of Game Design: A book of lenses. Morgan Kaufmann, 2008.
- Schreiber, L. Hughes, C. Seifert, B. Cash, C. Pineda, T. Robertson, J. Preston, and L. Law. Choosing between right and right: Creating meaningful ethical dilemmas in games [online]. 2009.
- M. Schulzke. Moral decision making in fallout. Game Studies, 9(2), 2009. Available from: http://gamestudies.org/0902/articles/schulzke.
- L. Sheldon. Character Development and Storytelling for Games. Course Technology PTR, 2004.