The Use and Reception of Varieties of Spanish in Videogames

With more than 585 million speakers, Spanish promises to be a profitable target language for videogame sales. However, as Spanish has such a wide geographical distribution, the question of variety deserves some attention. Companies and websites alike have recently started to differentiate between European Spanish and Latin American Spanish. This paper presents a three-phase study to delve into the matter. Based on a catalogue of 3003 videogames available on Steam in Spanish and published between 2006 and 2016, a survey among 172 developers is carried out to confirm the companies’ preferred varieties. Then, the analysis of a small corpus extracted from the catalogue shows whether in-game text actually represents the chosen variety of Spanish. Finally, a survey among 569 gamers provides some insight into their tolerance for foreign varieties. This paper aims to describe the current situation and provide a starting point for further reflection on some of the questions posed by the results of this study.


Introduction: The Videogame Industry
The videogame industry has experienced stable growth in recent years and the forecast for the near future is optimistic (Newzoo, 2020a;Palandrani, 2021;Wijman, 2020;Witkowski, 2021). This success can only be reached with the aid of localisation, as players expect entertainment content to be available in their local language (Bartelt-Krantz, 2011). Sales of localised versions of videogames represent an important share of the total revenue, accounting for up to 50% (Chandler & Byte Level, 2006) or 70% of the sales of a given game (Giné, 2009, cited in Bernal-Merino, 2015Klimov, 2017a).
This preliminary case study focuses on the use and reception of varieties of Spanish as a target language in videogames. Spanish has traditionally been one of the most popular target languages, together with French, Italian, and German, collectively known as FIGS. More recently, the market has been gradually shifting from the traditional FIGS (or EFIGS, if English is not the original language, but rather one of the target languages) to a more flexible choice of languages due to several factors.
Firstly, the number of indie studios is increasing thanks to digital distribution systems (Toftedahl et al., 2018). Indie, which comes from the word independent, is the familiar name given to studios that do not enjoy the financial support and are not constrained by the creative restrictions of a large game publisher. The availability of free game engines (such as Unity or Unreal Engine) and the possibility to publish games on computer platforms with reduced fees (such as those offered by Itch.io, Steam or Epic Games) allow solo developers and small studios from all over the world to publish their creations. When Japan and the United States were the main videogame developers, it was only logical that Japanese and English were the main original source languages (O'Hagan & Mangiron, 2013). Now, however, the existence of indie developers from different countries may encourage the publication of videogames in their native languages, including Spanish.
Secondly, China, Japan, and the Republic of Korea have positioned themselves as three of the four top countries in terms of game revenue, a position they share with the United States (Budapesto, 2021). This has made the Asian market a desirable prospect, though one with its own difficulties due to the required cultural adaptation to appeal to the Asian public (Dong & Mangiron, 2018). France, Italy, Germany and Spain remain among the top 10 countries by game revenue (Budapesto, 2021;Newzoo, 2020b), followed closely by Russia, Mexico and Brazil (Budapesto, 2021). Furthermore, not all genres are equally popular in different countries; the genre of the game will influence its popularity regardless of localisation. Charlie Oscar, the Lithuanian indie studio behind the strategy board game Gremlins, Inc., recommended Russian above FIGS because they had significantly more gamers from Russia than from the FIGS countries combined (Klimov, 2017a). In hindsight, one of the developers reflected that this may also be related to the fact that they were able to communicate constantly with players in Lithuanian, Russian and English, the three languages spoken by the studio (Klimov, 2017b). Conversely, Jumb-O-Fun, the Canadian developer of Wagers of War, found English and French speakers to be the most profitable markets (Yoccoz, 2017). For Mega Dwarf, the Canadian developers of God of Word, German was by far the most profitable language after English (Yoccoz, 2017). For developers who can only afford to localise their game into a few languages, it remains to be seen if the traditional FIGS will continue to be the languages selected by default.
In general, videogames produced and distributed by major publishers (commonly known as AAA games) are being published in more languages than the traditional FIGS. For example, Red Dead Redemption 2, a text-heavy game and one of the most popular games of 2018 (Polygon, 2018), was subtitled in at least 11 languages: Chinese, English, French, German, Italian, Japanese, Korean, Polish, Portuguese, Russian and Spanish 1 ; Horizon Zero Dawn, a major action role-playing game (RPG) of 2017, is available in Steam, a popular videogame digital distribution service, with subtitles in 17 languages: Arabic, Chinese, Danish, Dutch, English, Finnish, French, German, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Russian, Spanish and Swedish 2 . Developers have also started to differentiate between varieties of the same language. In the two games mentioned above, subtitles are available in both European Spanish and Latin American Spanish.
After this introduction to the gaming industry, which aims to show how the market is evolving with regard to languages into which games are translated, including Spanish and its varieties, the following exploratory research applies an observational approach. In the first place, the definition of variety and the most common varieties of Spanish in the videogame industry will be briefly discussed. Next, the methodology will be addressed to explain the scope of the analysed population and the limitations of the results. To the best of our knowledge, no previous research has combined information from developers and gamers to analyse varieties of Spanish in videogames. The goal of this paper is to describe developer approaches and the reactions of gamers. Special attention is paid to how many games are developed in Spanish and which varieties are chosen for localisation. Subsequently, we look into whether the chosen variety is actually the one used in the game, and, finally, gamers have been asked to evaluate how important it is to have games translated into their native variety of Spanish.

Varieties of Spanish in the Videogame Industry
When a language is used across a vast territory, inhabitants of different parts of that territory speak differently and, given enough distance, the differences in usage may become so great as to be unintelligible to speakers of the other variety (Penny, 2004). The term dialect is commonly used to refer to a set of linguistic features common to a geographical region and distinct from the corresponding set of linguistic features used in other regions where the same language is spoken. In reality, it is difficult to delimit a given dialect because each individual feature that is typical of a given region (be it a term, a particular pronunciation or a grammar element) extends across an area that does not necessarily coincide entirely with the area in which another linguistic feature of the dialect is common (Penny, 2004). Still, following Penny's contribution to the topic (Penny, 2004), we use dialect to describe geographically delimited sets of linguistic items and variety to describe a set of linguistic items delimited both geographically and socially. Examples of varieties could be Castilian Spanish and Andalusian Spanish, but also the Spanish of the older generation, or the Spanish of Galician fishermen.
According to the data collected for this paper, which is confirmed by my experience as a project manager in the audiovisual translation (AVT) industry, the videogame industry distinguishes three main varieties of Spanish that are both based on geographical distinctions and socially delimited (or even constructed in the case of neutral Spanish). Although this paper will not focus on the linguistic characteristics of each of them, they are briefly described below for purposes of clarity: • European Spanish: also known as Spanish from Spain, Castilian Spanish (Penny, 2004) or Peninsular Spanish (Pountain, 2016). It represents the Spanish spoken in Spain and until recently it has been the main (and usually only) variety of Spanish available in videogames. Among its many unique characteristics, one of the most salient ones may be the use of the personal pronoun tú/vosotros as opposed to the use of usted/ustedes in most countries of Latin America (Mazzitelli & Garrido Domené, 2019;Moreno de Alba, 2011;Morgan & Schwenter, 2019). This particular aspect will be revisited in the corpus analysis (see section 4.2.).
• Latin American Spanish: this covers the varieties of Spanish spoken in all Spanish-speaking American countries and it purposely avoids the characteristics of European Spanish. This term is mostly used in the AVT industry outside the Spanish-speaking community (López González, 2019). From a linguistic point of view, there is no such thing as a common dialect for all Spanish-speaking countries in America, nor a dialect for each country. They are rather divided according to patterns of settlement, contact with other languages, and the proportion of urban and rural communities (Lipski, 2012).
• Neutral Spanish: the objective of this artificial variety, which is not native to any Spanish speaker, is to allow the commercialisation of a given product in all Spanish-speaking countries. Its origin may be linked to the international agreement to use a neutral variety for early film dubs (Fuentes-Luque, 2019). It should avoid country specific characteristics and it should ideally be accepted in any Spanish-speaking country (Gómez Font, 2013). In practice, it is rather difficult to achieve neutrality.
Bearing in mind that there are more than 585 million Spanish speakers (Instituto Cervantes, 2020), the complexity and consequences of choosing one variety over another are well worth studying. Spain is one of the countries with the highest game revenue: it was the 9 th country worldwide in 2018 with a turnover of 2,202 million dollars (Budapesto, 2021) and 10 th in 2020 with 2,656 million dollars turnover (Newzoo, 2020b). Mexico does not fall far behind: it generated 1,577 million dollars in 2018 (Budapesto, 2021) and has about three times more players than Spain (Newzoo, 2018;Brugat, 2019), which makes it a very promising market.
Thanks to digital distribution systems, developers are able to collect more data from players and learn about their playing habits. Players are also able to give feedback to developers and voice their dissatisfaction when a game is not available in their native variety (Roszak, 2018). This may be part of the reason why the number of languages available in videogames is increasing in comparison to the time when FIGS were the norm. Proof of this changing trend is the fact that Steam has distinguished between Spanish from Spain and Spanish from Latin America since around October 2018 3 .
However, the popularity of each variety of Spanish is difficult to appreciate by studying a few independent cases and little is known about the reception of any given variety among the speakers of a different linguistic variety. Likewise, the original language of a videogame is normally not specified in the description of the game nor in the stores or on websites where games are purchased. It is therefore difficult to estimate how many videogames are developed in Spanish and how many are localised afterwards. By combining information from developers, videogames, and gamers themselves, this paper hopes to contribute to a general understanding of the use of varieties of Spanish in videogames.

Methodology
To delve into the developers' localisation strategies and the gamers' reaction to localisation choices, an observational study was carried out between 2017 and 2020. The methodology includes three consecutive phases, each dependent on the previous one: a probabilistic sampling of videogame developers, a corpus-based product analysis of said developers' games, and a non-probabilistic sampling of videogame players.

Process: Probabilistic Sampling of Videogame Developers
At the beginning of 2017, a catalogue was compiled listing the 3003 videogames available in Spanish and published during the previous decade (2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016) in Steam. We chose Steam from among other options because, at the time, it was the most popular videogame website according to the daily visits it received 4 . Nowadays, Steam distinguishes between Spanish from Spain and Spanish from Latin America, but in 2017, when this research was being conducted, all varieties of Spanish were presented as a common language: Spanish. As the website does not offer the option to download data automatically, we manually copied the title, the publication date, the name of the developer, and the name of the publisher into an Excel spreadsheet. Next, the estimated number of players for each game was copied from SteamSpy, a website that provides information about Steam statistics. Then, we tried to find a contact email or form for each developer. This was not always possible, because some of the companies had already closed down (at least 53, according to different websites and blogs), their websites were no longer active, or they did not provide any contact details. When no email or form was found, we resorted to social media. In some cases (less than 10%) we did not find any way to contact the developers.
The key feature of probabilistic samplings is the possibility to make assumptions for the whole population based on the analysis of a relatively small sample. The factors that measure the veracity of the assumptions are the margin of error and the confidence level. The former indicates the range in which an assumption can be wrong. The latter indicates how many times the said assumption will be within the estimated range. If the level of confidence is 95% and the margin of error is 5% (figures that usually apply in scientific research) and the results show that half of the sample speaks Spanish, between 45% and 55% of the population will speak Spanish in 95% of cases. The bigger the sample, the higher the confidence level and the lower the margin of error.
The sample must be randomly selected to ensure that all candidates have an equal chance of selection. We extracted a random sample of the population and invited the developers to take part in the survey. If no answer was received within a few weeks, a follow-up email was sent to the same address and any other form of contact available (e.g., a contact form and a social media account in addition to the contact email). As the answers were not enough to meet the margin of error we wanted to work with, we needed to extract four more samples, but did not go past half of the population to allow for randomness in the choice of participants. We finally elicited a total of 172 answers from the developers, which allows conclusions to be drawn with a 95% confidence level and a 7.5% margin of error. Two main reasons justify the relatively small scope of the first sample. On the one hand, many videogame companies enforce strict non-disclosure agreements. One of the developers we contacted replied that it was explicitly forbidden to take part in surveys or provide any information related to the localisation process. On the other hand, as stated by a few respondents, some developers are involved in the creative process but are not familiar with the decisions made during the localisation process, either because it was the publisher's responsibility or because they work exclusively in creative/developing aspects. Unfortunately, it was not always possible to find a contact email for the person responsible for localisation.

Product: Corpus-Based Analysis
The games forming part of the corpus needed to be games whose developers had participated in the survey to guarantee that we had received reliable information from the developers about the localisation process, such as the original language of the game, the varieties into which the game was localised, and the profile of the translator.
The popularity of the game was also important. We used YouTube as a source for in-game text because it contains freely available tutorials and gameplay videos uploaded by players. It also allows videos to be paused and revisited when needed. Walkthroughs with the most views were given preference. Some video walkthroughs include the whole game and others are divided into several parts. If all the parts of the most popular video walkthrough were not enough to extract the desired number of words, the second video walkthrough with the most views was added. To ensure that we could find enough resources to analyse the game, we discarded games with fewer than 30,000 players. At the time of gathering information for the catalogue, the author of SteamSpy (see previous section) warned users about the inaccuracy of data for games with fewer than 30,000 players (Galyonkin, 2017), which provided us with a limit to be used in our corpus-based analysis.
The 82 eligible games were divided into four categories according to the information provided by the developers: games translated by a professional translator or agency (exhaustively revised), games translated by a professional translator or agency (not exhaustively revised), games translated by volunteers, and games translated using an automatic translation software. Additionally, a fifth category was added for games where the developers were highly familiar with the concept of transcreation and the translator was granted freedom to deviate from the literal meaning of the source text (a game could theoretically fall under this category and any of the previous ones, but this did not occur in our corpus). The two most text-heavy games for each category were chosen for inclusion in the corpus. Games were not selected based on the variety of Spanish used. Incidentally, only one game was localised into Latin American Spanish among the ten games in the corpus. All other nine games were localised into Spanish from Spain. Up to 2,000 source and target words were transcribed for each game and organised in bilingual segment pairs, understood as in computer-assisted translation tools, i.e., from a full stop, quotation mark, question mark, or colon until the next of any of these symbols. The limit of 2,000 words was based on the characteristics of the games: four of the games included no more than 1,300 words and one included only about 2,200. It was decided to establish the limit at 2,000 words to avoid great differences between games with fewer words and those with more.
Based on the work of numerous authors (Martí Ferriol, 2010;Merino-Álvarez, 1986;Molina & Hurtado Albir, 2002;Pajares, 2010;Ros Abaurrea, 2015;Vázquez Rodríguez, 2014), an analysis model was created with a total of 19 categories. This paper will focus on the category we have called "geographical variation". A segment is classified into this category when a word does not match the variety of Spanish requested by the developers.
The main references used in this section are the Royal Spanish Academy and its dictionaries or affiliated organisations. The Royal Spanish Academy is an official institution with the mission to ensure the stability of the Spanish language. It is based in Spain and works together with national language academies from 22 Spanish-speaking countries. FundéuRAE is a non-profit organisation created in collaboration with the Royal Spanish Academy. One of its main tasks is to answer questions related to the use of the Spanish language. The use of the words mentioned in this section can also be verified in the Reference Corpus of Contemporary Spanish (known as CREA), which belongs to the Royal Spanish Academy and includes texts from the last 25 years, classified by year, author, country, topic, title, and publication.

Reception: Non-Probabilistic Sampling of Videogame Players
In order to extrapolate the results of the study to the whole population, we would have liked to carry out a probabilistic sampling of Spanish-speaking players. However, there are more than 17 million videogame players in Spain alone (Brugat, 2019) and this research project lacked the resources to contact them. Faced with the impossibility of extracting a randomly selected sample of the Spanish-speaking videogame players, there was no choice but to proceed with a convenience snowball sampling (Rojo, 2013, p. 95), in which the participants helped to recruit more research subjects. The main difference between a probabilistic sampling and a convenience sampling is that the results of a probabilistic sampling can be generalised to the whole population, whereas the results of a convenience sampling cannot be generalised because the sample has not been randomly chosen and is not representative of the general population. This last type of sampling is acceptable only if critically controlled by the researcher's knowledge of the field (Gile, 1998) and it can be useful as an exploratory research.
The survey was distributed in digital format via email and social media to reach as many players as possible. Participation was restricted to individuals older than 15 years of age who understood Spanish and played videogames for at least one hour per month. Although the threshold does not seem high, it was deemed more helpful for the research to have more opinions rather than to limit the answers to individuals with a higher frequency of playing. No statistical differences were found between high frequency players and lower frequency players when Chi-squared tests were applied in relation to the questions described in the analysis. Out of the 672 answers received in 2020, 569 participants claimed to meet the three criteria. To lessen the risk of fake answers, the survey could only be answered once on the same device and the participation criteria were not shared with the subjects.

Analysis and Results
The results of the analysis are divided into three sections. The first, focusing on the localisation process, analyses the developers' answers to the survey. The second describes the results of the corpus analysis of ten videogames. The third focuses on the reception of the analysed videogames and summarises the players' answers.

Localisation Process Analysis
According to the survey respondents, 82.6% of the videogames were written in English in the first place and only 1.8% were originally written in Spanish. As shown in Figure 1 below, there were nine different original languages among the 172 games analysed in the survey and only English stood out among the other languages. In one specific case, the developers confirmed that the game was simultaneously developed in English and Russian. Figure 1.

Original Language of the Videogames in the Process Survey According to Their Developers
Source: author's own study.
In 25% of the cases, the reason to choose English as the original language was due to it being the native language of the studio or the screenwriter, but the main reason to write games in English was to reach a wider audience (59.6%). In 54.4% of the games written in English, none of the original screenwriters were native speakers of the language they were working with.
As for the target language, European Spanish was the most popular variety of Spanish according to the survey. Figure 2 shows that 59.6% of the games were localised into European Spanish, followed by 30.1% of the games localised into neutral Spanish. Latin American Spanish (not country-specific) was used only in 13.8% of the cases. Localisation into a Latin American country-specific variety is scarce and represents only 1.8% of the total (it was not confirmed what country-specific varieties were used). The sum of the previous percentages surpasses the limit of 100% because they include the games that were localised only into one variety of Spanish and the games localised into two different varieties, which amounts to 7.8% of the total. The exact percentages can be seen in the figure below.

Variety of Spanish in Which the Videogame is Available
Source: author's own study.

Product Analysis
According to the information provided by the developers, only one of the ten games included in the corpus was localised into Latin American Spanish. For all the other games, European Spanish was the preferred variety. In general, this matches the results of the textual analysis, as nine of the ten games include pronouns and verbal forms corresponding to the mostly European Spanish-specific tú/vosotros ( Although the personal pronouns match the variety of Spanish requested by the developers, the analysis shows that two out of nine games include Latin American expressions which do not correspond to the requested European variety of Spanish. According to the survey respondents, those were the only two games in the corpus commissioned to volunteers (in contrast to paid professional translators or agencies) and the only games that contained geographical variations. The names of the games will not be mentioned for confidentiality reasons, as agreed with the developers, In the first game, a total of 13 cases of geographical variation were found among the 2000 words analysed. These range from nouns to verbs to idioms. In Table 1, a reflexive verb is highlighted in bold. This verb is used as a transitive verb in European Spanish and is only used pronominally in Latin America as an intransitive verb (FundéuRAE, 2018). Table 1.

Reflexive Verb From a Latin American Variety of Spanish in a Game Translated Into European Spanish
English Spanish He refused to believe it.
Source: author's own study.
The beginning of the segment shown in Table 2 contains an adverbial locution commonly used in Argentina and Mexico. According to the Royal Spanish Academy's Spanish Dictionary, it is correct, but not common in Spain, and the Dictionary recommends replacing it with después de que (FundéuRAE, 2009). Table 2.

Adverbial Locution From a Latin American Variety of Spanish in a Game Translated Into European Spanish
English Spanish After they […].
Source: author's own study.
The second game translated by volunteers contains five cases of geographical variation, found among the 2000 words analysed for this study. Table 3 depicts one of these cases in the form of the expression hacer reversa. According to the Royal Spanish Academy, this term is used in Colombia, El Salvador, Honduras, Mexico, Panama, Puerto Rico, and the Dominican Republic. Its European Spanish equivalent could be dar marcha atrás. Table 3.

Noun Used in Several Latin American Countries but not in Spain
English Spanish Damn, the street is blocked, I got to reverse the truck.
Source: author's own study.
The second example from this game, shown in Table 4, includes the adverb nomás. This adverb could be back translated as "just," which is not present in the source text but could be implied from the meaning: "we can't [just] leave those guys like this!". According to the Royal Spanish Academy's Spanish Dictionary, nomás is commonly used in Argentina, Bolivia, Chile, Colombia, Costa Rica, Ecuador, Honduras, Mexico, Nicaragua, Paraguay, Peru, and Venezuela. Although European Spanish speakers would understand it, it is uncommon and could have been replaced by an alternative such as sin más. Table 4.

Adverb Used in Several Latin American Countries but not in Spain
English Spanish I'll make another lap of the park, we can't leave those guys like this! Daré otra vuelta. No podemos dejar a esas personas así nomás.
Source: author's own study.

Reception Analysis
As was to be expected from a convenience sampling, the characteristics of the sample who took part in the reception survey are highly influenced by the social environment of the researcher. In general, this means that the proportion of university graduates is greater (69.1%) than that of the general population (40-47%) (Bernardo, 2017;Europa Press, 2018;Strambotic, 2019). The average age of the participants (M = 28.84, SD = 6.33) matches the statistics that report 20-to 35-year-olds to be the age group that plays most (El Publicista, 2019;EpData, 2020;Orús, 2020). The number of younger participants (aged 20 or younger) is a little lower than expected in comparison with other statistics (El Publicista, 2019;EpData, 2020;Orús, 2020). Spanish native speakers comprise 88.3% of the 569 participants and the remaining 11.7% claim to speak Spanish well enough to play; 88.9% of the sample live in Spain and 4.6% in Latin America (Argentina, El Salvador, Guatemala, Mexico, and Peru). The remaining participants are distributed across non-Spanish-speaking countries.
Of the native Spanish speakers, 47.9% play mainly with Spanish audio and another 47.9% play with English audio. The next most popular language for videogame audio is Japanese (2.2%). Regarding subtitles and on-screen text, 69.0% of the native Spanish respondents choose Spanish as the main language and only 30.2% choose English instead.
Given the choice, 44.6% of the players prefer to have games with a lot of text completely localised into their native language, i.e., with translated texts and dubbed audio; 30.2% prefer partial localisations (translated texts and the original audio) and another 10.7% prefer to play in the original language of the game (without any localisation) if they understand the original language. 5.8% do not mind which language they are playing in.
When it comes to playing in a different variety of their own language, 25.7% of the native Spanish speakers feel greatly bothered that they cannot play in their native variety of Spanish, 36.2% feel slightly bothered and 32.9% are not bothered. As shown in Figure 3, the rest of the answers depends on the genre of the game or on other factors.

Level of Annoyance at a Game Localised Only Into Another Variety of Spanish
Source: author's own study.
The significance of this reality is highlighted in comparison with the number of players in the same sample who feel bothered by a game that is not localised into Spanish at all (see Figure 4). Only 15.5% of the sample are greatly bothered when a game is in a different language, and 20.1% are slightly bothered. Another 21.9% say their answers depend on the languages other than Spanish that are available.

Level of Annoyance at a Game not Localised Into Spanish
Source: author's own study.
The level of annoyance is also notable for 23.9% of the participants who would not consider buying a game localised into a variety of Spanish different from their own. A further 42.9% would decide whether to buy a game depending on its genre. Thanks to Fisher's exact test, which analyses contingency tables, we can verify if the answers from the participants who live in Spain statistically differ from the answers of the participants who live in Latin America. Based on the results of the test, it appears that the reaction is shared by all Spanish speakers, without significant statistical differences between residents of different countries (p = .075 for the level of annoyance and p = .260 for the willingness to buy games in a different variety).
Figure 5 compares how many native Spanish speakers from the sample are willing to buy a game that is not localised into Spanish (represented in red) and how many would buy a game localised only into a variety of Spanish that is different from their native variety (in orange  Figure 5.

Willingness to Buy a Game Which is not Localised Into Spanish (red) or a Game Localised Into Another Variety of Spanish (orange)
Source: author's own study.

Conclusions
First and foremost, the limited scope of the three analyses needs to be highlighted. The population of the study is limited to computer games available in Steam. If we had analysed console games or even computer games from a different website or year range, the results may have been different. There is a 7.5% margin of error in the analysis of the localisation process (2.5% above the usual margin of error for scientific research), which may make results somewhat inaccurate. Furthermore, the reception analysis is based on a convenience sampling, and it is important to stress that its results cannot be generalised. However, there is room for some conclusions and considerations.
The limited number of answers obtained from the developers may be due to the strict confidentiality of the industry (Muñoz Sánchez, 2011) or to a lack of interest in research about localisation. Although the margin of error is not as low as we had hoped for, the results provide an approximate view of the reality of the industry. According to the sample, the vast majority of the games was originally developed in English (82.9%), but more than half of those games were not written by native English screenwriters. Instead, almost 60% of the developers decided to write in English to reach a wider audience. Further study about writing in a foreign language or about self-translation in videogames could help to assess the consequences of this practice. ). There are about three times more players in Mexico alone than in Spain (Brugat, 2019;Newzoo, 2018), but Latin American players are proportionally underrepresented, perhaps for economic reasons, since, despite its smaller population, Spain still has a higher game revenue (see Introduction). As we have seen, 7.8% of the developers already localise their games into two different varieties of Spanish and time will tell if this strategy becomes more popular in the future. It is yet to be confirmed whether neutral Spanish, preferred by developers over Latin American Spanish but not native to any country, will be accepted in the gaming context by the wider Spanish-speaking community. In the case of early films, neutral Spanish proved not to be a popular option on either side of the Atlantic (Fuentes-Luque, 2019; Mazzitelli & Garrido Domené, 2019), but more research needs to be conducted in the context of recent multimedia interactive entertainment software.
The corpus, though small, already shows some inconsistencies related to the language variety. Despite using the pronouns tú/vosotros, as befits European Spanish, two out of nine localisations into European Spanish include terminology that is only used in Latin American countries. They are the only two games in the corpus for which the translation had been commissioned to volunteers. Unfortunately, the scope of the corpus is not large enough to confirm whether this was a random coincidence or whether translations commissioned to volunteers result in more inconsistencies among different varieties of Spanish than translations commissioned to professional translators or agencies.
The sample who took part in the reception survey are not representative of the population of the study (videogame players) nor of Spanish-speaking countries (only 4.6% of the participants live in Latin America). However, the number of participants is considerable (569 of whom 499 are native Spanish speakers) and it hints at some phenomena worth studying in the future. Maybe not surprisingly, as Spain has traditionally been a dubbing country (Chaume, 2007;Hayes, 2021;Matamala et al., 2017), 44.6% of the Spanish-speaking respondents prefer fully localised games and 30.2% appreciate partial localisation above other options. As the reception study is not comprehensive, conclusions cannot be generalised, but it seems that Spanish-speaking players would rather play in English than in a different variety of Spanish, as only 6.3% of the respondents refuse to buy a game that is not available in Spanish while 23.9% of the respondents refuse to buy a game in a different variety of Spanish. In this regard, it is even more important to avoid inconsistencies between different varieties of Spanish, such as those found in the games commissioned to volunteers in the corpus. Further research on the impact of each transfer mode would help to establish whether a change in commissions toward more Latin American Spanish localisation would be beneficial for developers.
The industry is ever-changing, and this paper has provided only a preliminary overview of some linguistic choices made during the decade starting in 2006. Replication of similar research (e.g., based on another time frame or another platform) or a deeper analysis of some of the questions posed by the results of this study (e.g., the reception of neutral Spanish and different varieties of Spanish across Spanish-speaking countries or the link between translations by volunteers and the level of inconsistencies among different varieties of Spanish) may benefit the industry, the academic community, and the design of specialised training.