Who plays video games?

Who plays video games? Many would describe those that play games as ‘gamers’ (Sheffield, 2013).
The term clearly describes the activity; however, this label is far from neutral in its actual use. To be a
gamer means to be part of a social construct with its own history and tensions (Shaw, 2012). Video
games and gamers have a long history of being associated with masculinity (Braithwaite, 2016; Chess
& Shaw, 2015; Cote, 2018). Recently, this association has been problematized in the community.
Considering Shaw’s (2013) suggestion of looking outside of established market logics to define people
who play video games, this research aims to leave space to discuss players of video games who exist
outside of the stereotypical young male.

What is a gamer? The term carries many connotations both within and outside of the gaming
community. Outside of the community, playing video games has long been pathologized. Video games
as a pastime have been demonized as juvenile wastes of one’s time and as violence-focused
simulations that drive young people to replicate that violence (Fromme & Unger, 2012; Jenkins, 1999).
Many video games also require heavy investments of time, which is often judged by those outside the
hobby as wasted time. Most importantly, they are associated with young white men and geek
stereotypes. The gamer identity is heavily gendered and only recently has encountered resistance to
this connotation.

Genres of Video Games

An important factor to consider when localizing video games is the genre of the game in question.
There are a lot of different video game genres and the genre of a game can impact the choice of
translation strategies for that particular game. As O’Hagan and Mangiron explain:

“As text types are significant in translation, game genres help identify similar characteristics of texts
and also often text volume (text-heavy games as opposed to action heavy), thus indicating the
particular translator competence required. Games belonging to a specialized domain such as military,
aviation, and various sports genres seek to achieve a great degree of authenticity and realism through
accurate visual and verbal representation for the given domain, including the precise use of
terminology” (O’Hagan and Mangiron, 2013: 70).

So how can video games be categorized into different genres? Video game site gamespot.com
categorizes video games into 68 different designations (Gamespot, 2017). Some of these are about
game mechanics (strategy, action, adventure, role playing etc.), some are about graphics (2D, 3D) and
some are about the business model of the game (free to play). A game can have a mixture of these
characteristics. So, as an example you can have a 3D action game, which is free to play. In addition to
the genre section games are also categorized into various themes. There are 30 different themes
(Gamespot, 2017). These include categorizations such as comedy, fantasy, historic, superhero etc. So,
the 3D, free to play, action game could be further categorized on whether it’s in a historic or fantasy
setting or whether it includes superhero elements or not.

Many other video game sites use similar genre categorizations. Simon Egenfeldt Nielsen et al. (2008:
40) argue that while these genres are useful for the purposes of the website, they aren’t derived from
any standard principle as categories such as ‘driving’ and ‘action’ imply different characteristics of
games. Instead, they propose a genre system based on the question ‘What does it take to succeed in
the game?’ They categorize games into four broad genres:

Action games
Action games video games, where criterion for success is motor skill and hand-eye coordination. Most
fighting games, racing games and shooters, as well as arcade games such as Pac-Man fall under this
category. These games test players’ reaction speeds more than their decision making (Nielsen et al.,
2008: 43).

Adventure games

Adventure games require deep thinking and great patience, and player must use skills of logic and
deduction in order to achieve the game’s goal. Games with puzzle solving elements and some role-
playing games fall under this category (Nielsen et al., 2008: 43).

Strategy games

Strategy games are games, where the player assumes the role of a general instead of being on the
battlefield (Nielsen et al. 2008: 43). Players “have to balance large numbers of interdependent
variables, paying careful attention to signals of other players’ choices and strategies” (Nielsen et al.
2008: 43). Popular games in this genre include WarCraft, StarCraft, Age of Empires, Civilization etc.

Process-oriented games

Process-oriented games provide players with a system to freely play with, instead of restricting them
to certain goals. Massively Multiplayer Online Games, where players assume the role of a character in
a vast, persistent universe; management games, where players manage a business like a zoo or a
theme park; simulation games trying to convey real world experiences like flying a plane fall under this
category (Nielsen et al. 2008: 44).

Text Types in Video Games

In addition to genre types, another factor is text type of the translatable assets within the game.
Determining the text type is an important process in translation since the function of the source text
will have an impact on which translation strategies to use to create the same look and feel in the target
language. Reiss categorizes texts into three types based on their function:

- Content-focused (informative) texts that are focused on content. News reports, scientific
articles, legal documents belong to this category (Reiss, 2000: 27-31).

- Form-focused (expressive) texts that are focused on form. Poem and novels belong to this
category (Reiss, 2000: 31-38).

- Appeal-focused (operative/persuasive) texts that are focused on appeal. Advertisements

Additionally, Reiss mentions a fourth group of texts: Audio-medial texts, “which are written to be
spoken (or sung) and hence are not read by their audiences but heard” (Reiss, 2000: 27). Texts
belonging to this category can also be categorized according to the above three types of texts but also
implement visual and audio elements that require special considerations (Reiss, 2000: 43-47).

As mentioned above, there are different assets to be translated when localizing a video game. These
include user interface, dialogues between game characters, game manuals, contents of the game
package etc. These different assets are of different text types and serve different functions. Below are
some examples of translatable assets in video games matched with corresponding text functions
according to O’Hagan and Mangiron’s table of taxonomy of narrative orient-ed game text:

User interface: informative function

Exposition/tutorial: informative function

Unvoiced dialog: informative/expressive function

Text in images (in-game): informative/expressive function

Text in images (not in-game): informative/some persuasive function

Voiceover: informative/expressive function

Songs performed by game characters/theme songs: expressive function.

Environmental sound: expressive/informative function

Manual: informative function

Box: persuasive/informative function (O’Hagan and Mangiron, 2013: 155¬158)

As it can be seen above, different parts of the texts translated in video game localization belong to
different text types. It is possible to encounter every text type when localizing a video game.
Translating the manual of the game and translating the text of a dialogue between in-game char-acters
are two very different endeavours. While translating a manual or the tutorial of the game, the
translator has to make the text clear so that the user can easily comprehend the instructions. On the
other hand, translating the narrative parts of the game requires aesthetic considerations. For this
reason, localizers of video games are usually given more freedom than localizers of other types of
software products. Mangiron explains the reason for the increased amount of freedom video game
localizers enjoy as follows:

“...when a localiser is assigned a project, s/he is practically given carte blanche to make any changes
deemed necessary, especially regarding the language, the cultural references and the humour, all of
which are key elements in games.” (Mangiron, 2006: 309).

Based on this, it can be concluded that parts of the game with an entertaining function (e.g. story,
dialogues) can be localized with more freedom due to the developer’s wish to preserve the
entertainment value of the product for every locale.

Therefore, it is not feasible to determine one strategy for a specific video game localization project as
a single video game may have texts of different types and those different text types will demand
different strategies and approaches to overcome the specific challenges belonging to that text type. It
is imperative that translators working on the project understand which part of the game they are
translating and what kind of strategies are suitable to keep the same look and feel.

LOCALIZATION PROCESS AND GAME-SPECIFIC CHALLENGES

Every genre of video games and every individual game have their unique challenges that the localizers
have to overcome in addition to the challenges of software and video game localization. Knowing how
a game works and the expectations of that specific game’s target audience are crucial to produce a
good localization of the game. As such, before analysing the Turkish localization of Football Manager
2015, it is important to take a look at the localization process and pinpoint the specific challenges
posed by the game.
These specific challenges call for a translator with specific qualities. The developer of the game, Sports
Interactive, lists the qualities a translator working on a Footballer Manager game should have as
follows:

*An aptitude for translation (preferably with relevant experience or qualifications)

*Mother tongue speaker for target language and in particular football terms and colloquialisms

*Very good knowledge of English and in particular football terms and colloquialisms

*Lots of experience in playing FM games

It can be seen that, in addition to usual qualifications expected from the translators (i.e. good
knowledge of both languages), there are other skills required from the localizers of the game.
Knowledge of source and target culture football terminology as well as experience playing previous
games in the series demonstrate that localizing specific video games require unique skills depending
on the genre and theme of the game. These skills are needed to cope with the specific challenges a
game such as Football Manager 2015. The following sections aim to underline these challenges.

Text Strings

The main point to keep in mind when analysing the localization of a Football Manager 2015 game is
that most of the text in the game is created through the use of text strings that have variables in them.
As highlighted by O'Hagan and Mangiron, translators working on localization projects translate these
types of short fragments of decontextualized texts (O’Hagan and Mangiron, 2013: 102). This is
explained in the official forums of the game as follows:

The main thing to remember is that what you see on the screen is rarely a direct translation as most
sentences (strings as we call them) contain variables within them. So, if you are thinking “How could
the translator make such a basic error?” it may not be as simple an error as it appears to be.

If we look at the simple string:

Code:

The English Premier Division match between Arsenal and Man Utd originally scheduled for 21st
October has been moved to 26th October due to international call ups.

This is how it appears in our code.

Code:

The <%fixture_name#1-long> match between <%team#1-short-notype> and <%team#2-short-

As can be seen, the localizers can only translate the non-variable parts of the text. The rest is filled in
depending on what happens in each game. This is fundamentally different from translating the text of
a novel or translating subtitles for a movie, where the text is fixed. It is easy to see how the translation
of these text strings might cause problems for localizers due to decontextualization. As mentioned by
Esselink, localizers have to have "in-depth programming, product or content knowledge” (Esselink,
2000: 63) to overcome the challenges posed by this type of translation environment.
Another challenging aspect of the use of variables is the structural differences of different languages
causing issues for the localizer. Bernal-Merino illustrates a potential challenging scenario of variable
in-game text as follows:

“Variables can be used in many complex ways to enhance players’ immersion by addressing them, their
chosen profile, and their performance directly. For example, strategy games allow the player to choose
among different nations to conquer the world. When someone attacks you, the message received says
“/n name of nation /n is attacking you!”. Names of nations vary widely from one language to another.
They may take an article or not; they can be singular or plural, as well as masculine or feminine. Not
only do translators there-fore have to be careful with the syntax of the sentence and the possible
relocation of the variable, but they also need to be aware of potential changes due to the morphology
of in-dividual linguistic items. The formula in question could generate a sentence such as “Rome is
attacking you!”, but it could also produce “The Vikings is attacking you!”. Whenever possible,
programmers and designers opt to rephrase sentences to avoid these grammatical hurdles of natural
languages. The above message, for example, could be rewritten in the passive voice, as “You are being
attacked by /n nameofnation /n!”, so the formula would allow for both “the Vikings” and “Rome”.
However, this might not be an option for all target languages” (Bernal-Merino, 2007: 6).

As suggested by Bernal-Merino, different languages require different solutions for overcoming this
particular challenge.

For example, let’s have a look at the term in the “Football Manager 2015” game, ‘manager’. 'Managers'
in a football context are people who are in charge of football clubs. Their responsibilities can range
from selecting the squad, determining tactics, managing training sessions to making decisions on which
footballers to transfer and offer contracts to. However, there are no ‘managers’ in Turkish football
culture. The closest term is ‘teknik direktor’

Modding, ROM hacking, and fan translations

Computer games, while having developed separately from console games, share many similarities with
them (Newman, 2017). One of these similarities is the ability for users to modify, or ‘mod,’ the game’s
code to create new levels, characters, and even entirely new game experience. Game modding
highlights labor practices that have developed in new media environments that blur the line between
official and unofficial as well as fan and producer.

Modifying computer games is a practice that grew in popularity with the rise of PC gaming in the 1980s
and 1990s (Sotamaa, 2010). Kücklich (2005) specifically traces the development of the practice back
to the PC game Doom in the mid-1990s. While mods of this game were not the first, they did establish
a framework that continues even today. Specifically, Doom’s publisher id Software created tools that
allowed users to modify elements of the game. This structure eventually led to certain computer
games’ engines being used to develop other games. Game engines are frameworks licensed and owned
by publishers. Mods, on the otherhand, are often created using tools provided by publishers. Even so,
modding tools can be used to essentially create entirely new games on top of an existing one. This was
the case with what Kücklich describes as one of the most popular mods in history, the game Counter-
Strike. Based off of the Valve game Half-Life, Counter-Strike was a highly successful mod due to its
creators selling the game to Valve and joining the publisher. Despite starting out as fans, these creators
joined the industry.

Fans have taken advantage of new technology to bring themselves closer to publishers. Sotamaa
(2010) and Kücklich (2005) emphasize the importance of the relationship between the modder and the
game publisher. Publishers of computer games have responded by allowing fans to create mods within
terms that they dictate. One reason they do so is the opportunity to profit from the unpaid labor
influenced by dedication toward the game being modded. Mods can also lead to heightened player
engagement. Kücklich points to how a game’s modding community can extend the life of a game, which
ultimately helps sales for the publisher.

In contrast to computer game modders, who were often motivated by adding content to games they
already had access to, fan translations of video games are motivated by a desire to make the game
accessible to a larger community of players. Consalvo (2016) refers to this as “unintended travel” (p.
42). While a game producer might not intend for a game created for the Japanese market to leave the
country, fan communities can come together and bypass official channels in order to release games to
larger audiences. These communities’ work is arguably like that of fansubbers; however, the process
is far more complicated and reveals how much labor fans are willing to invest in accessing games they
enjoy.

Two main elements of video game fan translations are the following: the translation itself and hacking
the game’s code. Translators and programmers work in tandem in order toaccomplish the main tasks
of translating the game’s text, including its narrative script, and inserting that text into the game’s
programming code. Muñoz Sánchez (2009) describes the programmers involved in these projects as
ROM hackers due to the fact that they hack the original game file (read-only memory). He also
identifies some tools used by ROM hackers to access and modify game code. Unlike PC game modding,
however, these are not official tools provided by publishers. ROM hackers must work without any
documentation to guide them in the process of understanding how a game’s code operates.

Fan translations thus combine elements of fansubbing and game modding. The work done by ROM
hackers in translation projects is similar to fansubbing in that it bypasses language barriers and
computer game modding. Unlike fansubs, the text contained within a ROM’s programming code can
be arranged in a variety of ways Muñoz Sánchez (2009). Once the ROM hacker locates this text and
figures out how it is incorporated into the game’s overall code, he or she can use tools to remove that
text and present it in a format for the translator to work on. The translator relies on the work of the
ROM hacker to work with the game’s programming logic and extract the text into a form that is more
amenable to being read by humans instead of a computer.

Fansubbing and computer game modding in comparison to fan translations and ROM hacking present
examples of how different fan communities can develop similar practices in different contexts. Fan
communities in the context of the game industry also provide yet another avenue for understanding
new practices that have been developed by fans within the new media environment (Jenkins 2006,
2013). Even with fans’ increased ability to modify games and bypass localization decisions by
publishers, they do so without the permission of game publishers. Console game publishers keep
details of their development practices deeply guarded and mostly communicate with fans through
their own channels and game press.

The task of the localizer

Translation is only one part of localization; however, translation theory also illuminates some of the
philosophical conundrums inherent in localization. Changing a game to keep it the same echoes the
difficulty of translating the form and meaning of a text. The task of the translator, according to Walter
Benjamin, maywell be futile before he or she even begins the task. Languages do not map onto one
another neatly. They can differ in grammar, sentence syntax, writing systems, and the way in which
they depict the world. The translator cannot recapture the original work exactly in his or her native
language.
Schleiermacher (2012) discusses two theoretical approaches to translation: paraphrasing and
imitation. Benjamin (1923/2012) refers to these concepts as fidelity and freedom, respectively.
Paraphrasing and fidelity are word-for-word translation. Schleiermacher associates it with the sciences
and defines paraphrasing as mechanical, even math-like. He claims that the “paraphrast treats the
elements of the two languages as though they were mathematical signs that can be reduced to the
same value by means of addition and subtraction” (p. 48). The imitator, on the other hand, works to
“contrive a copy, an entire work comprised of parts that differ noticeably from the parts of the original,
yet which in its effect comes so close to the original as the differences in the material permit” (p. 49).
Schleiermacher associates imitation with the arts. The twentieth century was marked by a general
preference in translation theory for fidelity and accuracy because of the rise in prominence of scientific
and commercial documentation. Efficiency and ease of communication were priorities rather than
artistic license (Bastin, 2019).

At first, a word-for-word translation seems like a simple enough task. However, translating the words
that comprise a text is different from translating the meaning. Benjamin argues that a word-for-word
translation ironically obscures meaning when he states, “with regard to syntax, word-for-word
translation completely rejects the reproduction of meaning and threatens to lead directly to
incomprehensibility” (pp. 80-81). The meaning that comes out of a text is more than the syntax and
grammar rules that govern language. Meaning communicated by language is more than the sum of the
components of a spoken or written sentence. Schleiermacher (2012) and Benjamin (1923/2012) both
argue for going beyond this dichotomy, though they do not provide a clear solution to the problem.

These theoretical concerns raised by Benjamin (1923/2012) and Schleiermacher (2012) also apply in
the case of video game localization. However, the task of the localizer, in most cases, is not to prioritize
linguistic fidelity, but rather fidelity to the gameplay experience. The gameplay experience is
subjective, and thus localizers must make their own decisions as to how to best Whereas the deeper
meaning communicated by language is the challenge for the translator, the deeper meaning of playing
a game is the challenge for localizers. Understanding how localization professionals make changes in
games to maintain the gameplay experience requires dissecting the components of the gameplay
experience.

Game localization is more than translation, it is also creative writing. Localization professionals fill in
what is possibly “lost in translation.” Language and meaning are precarious during this process, yet the
lack of structure also provides opportunities for creativity. Localization professionals take advantage
of the instability of this space and blurred boundaries between languages to creatively transform
meaning. O’Hagan and Mangiron (2013) apply the concept of “transcreation” (p. 196) to the
localization process. Transcreation refers to how localizers tactically maneuver within the limits
imposed on them from producers. Originally, the concept of transcreation referred to post-colonial
translation practices in Brazil and India that actively pushed back against Western colonial power
(Munday, 2016; Vieira, 1999). These transcreative approaches question the primacy of the original text
and reconceptualize translation as two-way instead of simply from source to target locale. In the
context of game localization, Mangiron and O’Hagan (2006) define transcreation as the “freedom
granted to the translator, albeit within severe space limitations” (p. 11). The translator subjectively
makes decisions as tohow to apply their creativity to preserve the original gameplay experience. Game
localization is unique among other kinds of translation because of the creative license that is given to
localization teams. Translators are allotted a significant amount of agency in many cases while also
complying with rules and guidelines imposed from above by game producers (Šiaučiūnė & Liubinienė
2011).
Overall, the game localization process is a liminal process located between source and target locales.
It operates in the spaces between languages, which in this case, are English and Japanese. In this in-
between space, the grammatical and semiotic structures of language break down (Turner, 1979; van
Gennep, 1981). This concept was first applied to rituals and later to physical spaces such as border
areas between nations (Thomassen, 2014). Localization is the border region between languages and
cultures, where the rules and conventions of Japanese language flow into those of English. Localization
professionals must navigate this space as they work to transfer meaning from one culture to another.

Localizers also operate in a liminal space between industry and players as well as between producers
and game text. Video game producers set the limits from above, and expectations of player
preferences guide the creative decisions that many localizers make (Chandler & Deming, 2011). Before
the establishment of the Entertainment Software Rating Board, Nintendo had strict content policies
for any game published on its platforms. Such policies severely limited religious references, alcohol,
and sexual content. As a result, localizations of games for Nintendo platforms had to modify or remove
any elements that violated their policy (Grealy, n.d.; Mandelin 2018a). In this space, localizers do more
than simply transfer linguistic signs from one language and culture to another. As game localization
has developed since the 1980s, the process increasingly demands more intricate changes to
gamecontent. Many localization teams consist of translators and writers who specialize in the target
language (O’Hagan & Mangiron, 2013).

Finally, a key part of game localization is the importance of invisible localization. A major goal of
changing game content to maintain the perceived original, source language, experience of playing the
game is to make the finished product as seamless as possible. Localizers consider a good localization
to be one where the player cannot tell that the game was modified at all. Two concepts from
translation theory, domestication and foreignization, explain the different ways that localizers
approach the process. Domestication brings the foreign author closer to the reader and prioritizes
making the translation sound fluent in the target language. Foreignization, on the other hand, carries
the reader closer to the foreign author by keeping traces of the source culture and deliberately
challenging the reader by introducing them to new perspectives (Gambier & van Doorslaer, 2013;
Schleiermacher, 2012).

Invisibility in game localization, which is a domesticating strategy, refers to two strategies. First,
localizers aim to remove Japanese cultural references that may confuse or offend the target locale.
Secondly, modifications made to games should be seamless. Even games that embrace Japanese
cultural references as part of the setting incorporate changes like the insertion of English text to
replace Japanese text as seamlessly as possible.

Translating the gameplay experience

Foreignization and domestication have so far been discussed like binaries. In the actual localization
process, these two strategies are means of achieving what is considered to be a “quality localization.”
Ultimately, maintaining the immersive nature of the diegetic game world is the goal of localization.
Cultural markers are not the only elements of games that affect the play experience. Localizers
interviewed prioritized translations that made the player feel immersed in a game’s world. Some ways
that they achieved this was by keeping track of glossaries of specialized terms for long-running game
series and avoiding direct references to a game’s control scheme (Chojnowski, 2016). Much of this
work required creativity from the localization team. As mentioned earlier, localization is also creative
writing. This is the work necessary to “change it to keep it the same.” The less creative route is
attempting to do an exact translation from Japanese to English. Creativity, according to localization
professionals interviewed, made the difference between a good and bad localization. A bad
localization does not draw on the creativity of the localization team to make changes while keeping
the game the same. A good localization ideally navigates this balance to create what localizers consider
to be necessary changes.

A ‘bad’ localization disrupts the seamlessness of the gameplay experience and immersion by exposing
the fact that the game is from another country. The most obvious type of bad localization is a
translation error, especially when it inhibits gameplay in some way. Interfering in the player’s gameplay
experience opposes the industry’s goal of replication the player experience of the original, since the
original game likely did not have such errors (Schules, 2012).

The most basic localization errors can occur in the game’s interface. The text box, a common element
of a game’s interface, primarily contains text to communicate with the player. Narrative text and
character dialogue, for example, are often displayed in dialogue boxes on the screen. If the game’s
programming properly accounts for text box length, then text will always be formatted to stay within
the box. If the localization team does not have access to the development team, however, then doing
this is more difficult. Localizers interviewed considered text box flowing outside of boxes as
embarrassing and a sign of sloppy localization.

Figure 4. Text improperly formatted in a dialogue box. From French localization of the game I Am
Setsuna.

Another localization error that operates on a deeper layer of text is affects how textual instructions
orient the player in the game’s world. When dialogue or text that informs the player is translated
incorrectly, the player is prevented from progressing in the game. In the worst case such errors can
prevent game progression entirely. Dietz (2006) termed these kinds of errors as “linguistic plot
stoppers.” Many of these kinds of mistranslations are in older games or games whose localizations had
a limited budget (Mandelin & Kuchar, 2017). This kind of error is especially detrimental to immersion.
The player will either be prevented from progressing or will have to stop playing to look for information
outside of the game.

At an even deeper level, text informs players of details important to establishing the game’s world.
Schules’ (2012) analysis of the Nintendo DS game Lux-Pain is a notable example of a localization error
at this level. The game’s localization not only has English grammar mistakes and typos, but it also
creates confusion as to the game’s setting due to mistakes made in text for important story segments.
Confusion as to whether the game is set in the U.S. or Japan impedes the player’s immersion in the
diegetic game world. Schules argues that language itself is a component of gameplay. His argument
points to the complexity of the gameplay experience. Localization affects not just text or the narrative,
its influence touches all elements that contribute to the emergent experience of gameplay.

Aside from these technical localization issues, the greatest challenge in localization is transferring
meaning between languages. In this case, ‘bad’ localization as defined by localization professionals was
a translation that tried to be too faithful to the text itself instead of the meaning it communicates. A
word-for-word localization was considered a failure. One interviewee stated:

It's not like literally translating things word for word. We're also not about "oh, just forget about what
the Japanese said. Let's go do our own thing" either. We want to be faithful to the Japanese, but being
faithful in localization doesn't mean doing word to word translation. (E. F., personal communication,
March, 2019)

Here, he alludes to a balancing act that localization professionals manage as part of their work. A word-
for-word translation may seem like a faithful translation; however, this kind of translation is not the
goal of game localization.

A veteran translation summed up his philosophy on translation in the game localization context:

I think the number one reason for a poor translation—and I’m not talking about an inaccurate
translation. It can be very accurate, but it’ll just read flat and not be interesting.

The number one reason is not that the translator did not have the skill to do it; it’s that they didn’t
take agency. They didn’t take that extra step to make the changes that needed to be made, which is—
it’s an egotistical thing to say, and it’s exactly antithetical to the—we want the original with no gussying
up and no censorship and that sort of thingthat people make a lotta noise about. Yet, my argument in
defense of changing things is that if you don’t recreate the experiences…

…if you directly translate, you’re sabotaging the scene. You’re actually not doing justice to the original.
It’s better to mix things up and change things around. (Veteran Localizer, professional communication,
June, 2016)

A bad localization, according to the interviewee, is one where the translator does not exercise their
agency to go beyond a word-for-word or literal translation. Not every game localization project gives
the localization team the time or budget to completely express their agency (Schules, 2012). The
quoted interviewee had started on smaller projects with less freedom at first and was gradually
promoted based on their good work. Even with the limitations that they faced early on, they took
advantage of the freedom they had on projects to distinguish the localized text and their work.

A good localization endeavors to change the original Japanese game to maintain the gameplay
experience of the original. The Final Fantasy X example shows how this strategy was used to change
dialogue in an important story cutscene. The translator (and other members of the localization team)
must make the choice to exercise his or her creative agency. In doing so, they endeavor to move
beyond a word-for-word translation and engage with the deeper meaning communicated by the game
text. This requires taking risks because localizers draw on their own experiences and interpretations of
the game text to make creative decisions.

Good localizations are the result localizers having the freedom to creatively edit the game text.

I think you have to find out that word is making the Japanese dialogue work in its own terms and then
find a way to make the English dialogue work on its terms, not rely on the storytelling methods of the
Japanese texts, which, often, really don’t fly in English. It’s not that the text is bad, although, sometimes
it is, but in a perfect world, even, with a really well written original text, you can’t use the same tricks
that they use because it doesn’t compute when you bring it over to English. (Veteran Localizer,
professional communication, June, 2016)

However, using one’s agency to make changes in service of the game text is risky. This interviewee
further pointed to the fact that players would have their own interpretations of game localizations.
Overstepping and making players feel as though the localization strayed too far from the original could
also happen. When the effort is successful, a particular game will often be remembered for having a
localization that went above and beyond expectations.

An example of an excellent localization that received praise from players and critics alike is the game
Vagrant Story (2000). The localization team took advantage of the fact that the game’s setting was
based on medieval Europe. In this case, the English language had an association with this setting that
Japanese lacked. Their approach to the game’s English script was to “make it Biblical,” which meant
adopting a Shakespearean style for the game’s dialogue (Learned, 2017). They described their work on
the game as maintaining the meaning of the original Japanese script while enhancing the English script:

You can call that "improving" on the original script, but the original was fine in Japanese—great, even.
You could also call it giving it our own spin, or even needlessly embellishing, but everything in the
English is directly inspired by the Japanese text andcharacters. At the end of the day, we just took what
was there and did our best to write the way we would have expected to see in that story had it been
originally written in English. (Learned, 2017)

The Japanese script was not badly written. They embellished the English dialogue based on a seemingly
artistic desire to make the game more interesting to themselves and English-speaking players. Despite
Vagrant Story having a low print run and not becoming a best-seller, the game is still remembered for
its English localization among game players and press (Robinson, 2017).

Figure 5. English dialogue from Vagrant Story. The localized used stylized English to embellish the narrative.
The successful localization of Vagrant Story points to a philosophy expressed by an interviewee:
“change it to leave it the same” (Veteran Localizer, Personal communication, June, 2016). Since
localization professionals are trying to preserve the gameplay experience and not the text itself, they
justify changing text and other assets of games to serve this purpose. This approach carries some risk
and requires that the localization team have the time, budget, and freedom to make such changes.

Game Localisation Processes

Game localisation can happen concurrently with development or afterwards, either in-house or
outsourced to a localisation agency. It involves a similar localisation engineering step to software
localisation, where files or assets are prepared for translators with instructions, reference materials,
code, and style guides. Before translation, translators are often expected to familiarise themselves
with the game by playing and reading about it. The IGDA guidelines suggest a length of at least three
days for this period of familiarisation, and up to a month for massively multiplayer role-playing games.
Translators are then expected to work with developers on a target-language style guide and glossary
of common terms. The term ‘translatable assets’ is used for the various elements of a game to be
localised. This includes in-game text, art, video, and audio assets within the game, and printed
materials in case of a physical release. Some dedicated tools have been developed to manage the
localisation process for this diverse set of assets, such as XLOC and Gridly, and game development
engines such as Unity and Unreal Engine incorporate some localisation-specific features and guidelines
(Rivas Ginel, 2022). In-game text may be in the UI in menus, system or help messages, and tutorials, as
well as descriptive texts and written dialogues. In the past, these have usually been translated using
spreadsheets, but there is increasing use of CAT tools in the game industry, particularly memoQ, and
even some use of MT. Consistency of terminology for the game or platform is important, but narrative
and dialogue text needs to be creative, engaging, and not repetitive. Art assets such as maps or signs
embedded within the game need to be translated, although this is often managed by developers who
will just include the text string in files sent to translators. Where cost is an issue, these assets may not
be translated at all.

Video and audio assets are translated using subtitles and/or dubbing, although the workflow is unlike
standard audiovisual translation. Subtitles are translated with other translatable assets as above. The
IGDA guidelines simply suggest the use of brightly-coloured text with a dark outline. For audio, dubbing
scripts are prepared and studios and voice actors hired. For a voiceover section, the dubbed audio may
not need to be synchronised with video, but for much of the game, these will need to be either time-
constrained, synchronous with other audio, or lip-synchronous. In addition, there may be ‘stitches’:
short audio clips that can be triggered at appropriate points within the game, possibly with variables
that can be slotted together (O’Hagan & Mangiron, 2013). Players will go through a series of game
situations within a game, with different levels of interactivity and different implications for dubbing
synchrony: the player may receive a task in audio or written instruction, they may be in interactive
game action, there may be a restricted interaction for dialogue, or a cinematic clip with no interaction
possible at all. A fully localised and customised AAA game may be dubbed into eight or nine languages,
with over 100 actors per language, which represents a huge cost and effort. At the time of writing,
games localisation companies are looking at using artificial voices, in order to save costs for some
secondary characters.

First of all, testers need to be gamers with strong target-language skills. They are charged with finding
bugs and might also query some translation decisions, then with testing that bugs have been fixed
without causing new issues. Once a game has been published, the process will usually repeat with
regular updates, patches, ports to other platforms, and possible sequels, with the sometimes with the
added complication of user input, as some games allow modification and co-creation by gamers under
the limits of a EULA (End User Licensing Agreement).