body language

As I said yesterday, technology might soon enable game characters to respond to players by reading our faces. There's a flip side to that. Players could better respond to characters if facial animations and body language were improved.

The most basic benefit of this is obvious: improved empathy. A character's dialog and actions have more emotional impact when those actions correspond well with the character's physical expressions.

Another benefit is to replace dialog entirely. Animating is not simple, but it's cheaper and easier to refine than hiring voice actors. It also saves valuable memory space. Silent expressions often have the greatest effect. But I expect refining such animations would be very difficult, considering that even good actors have difficulty replicating some emotions, like despair and terror.

The use of animations that brought this subject to mind is translation. Having taken many linguistics courses in college, I know that not all body language is universal across the globe, but the basics are.


On the one hand, you have games like The Saboteur. The game is set in Paris, so one might expect some of the dialog to be in French, even if the majority is English with a French accent. A bit of dialog local to the setting can add flavor, and good body language can make that dialog more meaningful.

On the other hand, there's the inherently national nature of any game. Games published worldwide either pay translators to adapt the dialog or apologetically ask foreign players to enjoy the game as it is. Something is always lost in translation, and body language helps to counteract those losses. And while great facial animations won't allow foreigners to play a dialog-centric game, it can be pivotal in action and puzzle games for making environmental dialog bearable.

Anyway, the usefulness of better facial animations and body language in games seems obvious. What's less obvious is whether or not each developer must tackle this issue on their own. Might it be possible for a company to develop a program for such animations to be used in a wide variety of games by different developers, similar to something like SpeedTree? Or must representation of the human face evolve only as a collective effort, like the modeling of skin or eyes?

localizing non-verbal language

Lots of companies localize their games these days. How good these localizations are I don't know.

Over the years, I've read about occasional quality problems in translations. Anyone who speaks multiple languages is aware that translating something literally is often a mistake, and that's the type of problem I see cited most often. Such mistakes are usually the result of appointing someone to the task of localization who is not a fluent speaker of both languages. Major publishers seem to have learned their lesson and avoid this mistake now.

So localization practices have improved, but I wonder about non-verbal language... meaning body language and similar things. How often is a lack of translation for that aspect of language problematic for gamers? Do designers pay any attention to it?

If it is a problem, can anything be done about it and be worth the expense? Afterall, we're not talking about text anymore.

For example, consider proxemics. Proxemics refers to the distance two or more people are expected to keep between each other when talking, as well as things like touching. How close you stand to someone and how you touch them communicates information, and what that information is interpreted to be depends on the culture of the observer. Two Iranian men are likely to stand much closer to one another as they casually chat than two American men would. It is also not uncommon in that part of the world for two men to walk down the street holding hands. A typical American is likely to interpret this to mean the men are gay, but it has no romantic connotation in the other culture.

Even within one country, this aspect of language can vary significantly. In rural areas of the United States, the act of shaking hands is often considered an essential greeting or seal to agreements, and a failure to do so can be taken as an offense or incite distrust. In some parts of America, women expect hugs even from strangers in casual settings.

Tone and volume are other points of misinterpretation. Many conversations between an American employer and employee would likely shock some foreigners because the employee is not being properly submissive in tone. Rich people often misinterpret a poor person's loudness as being obnoxious, whereas poor people often misinterpret a rich person's quiet speech as being snobbish.

Anyway, you get the general idea. Changing text is one thing, but changing non-verbal language involves extra voice-acting, extra animations, etc. Is this a problem worth dealing with?


Perhaps we should begin by simply asking how often it's a problem. Can you cite any examples in your own experience where a foreign game gave you the wrong impression?

For example, cutscenes and voice-acting in Asian games often seem melodramatic to me. Some of that's due to poor scripting and acting, but I believe some of it's due to cultural differences in loudness, tone, pitch, etc.