While game developers are experts in making games functional, fun, and available to the masses, it is unlikely that their specific skill sets also include deep understanding of and real-life experience with SLA and CALL. Cognitive linguistics, psycholinguistics, sociocultural learning theory, systemic-functional linguistics, and complexity theory are all research-backed theoretical perspectives in SLA which would be useful for both CALL and game developers (Chapelle, 2009). In fact, the functional difference between CALL developers and game-based language learning (GBLL) developers is probably minimal. Why should the CALL apps, products, tools, and materials be boring and stuffy? And if they become ludified, what then differentiates them from a bonafide game?
An example from CALL research that would be useful for game developers wanting to know how their game can improve language learning is a study by Payne and Whitney (2002) in which they found that students who replaced half the oral conversation with written chat performed better than those who did not. This study suggests that a space for human computer communication can be exploited by CALL developers, and by extension, GBLL developers as well.
A frustrating problem arises when the game developers fail to notice and engage with the research community that has already been obsessively posing the same question for almost a century: How do humans best learn language? Game developers needn't remake the ever-present, ever-revolving wheel of theoretical research; it already exists! We just need to collaborate better!
Sadly, many attempts to involve linguistics experts in GBLL are either token (at best) or baseless (at worst). For example, the widely marketed, publicly known GBLL platform Duolingo does not use High Variability Phonetic Training (HVPT), which involves exposure to multiple native speaker voices repeating the target phrase, even though this method has been repeatedly shown to be the best way to learn L2 sounds. In fact, Applied Linguistics, one of the most respected journals in the field, will not currently accept new research articles on the topic of HVPT, considering it given knowledge. Duolingo instead uses a single native speaker to pronounce words and phrases, even though this is not research-based practice. Additionally, these are presented in a series of context-less sentences in an arbitrary sequence order that is not language-specific, despite the fact that order of acquisition of grammatical features is a complex, well-studied topic that has been shown to be both dependent on features of the language as well as features of the learners themselves.
The frameworks for learning language already exist, are well studied and researched, and are actually not always intuitive. Far too often the programmers, UX developers, and marketing experts wrongly assume that their own experience with language learning qualifies them as experts in SLA and CALL. Because of the great amount of humility it would take for developers to rely on linguists, perhaps the only real hope of revolution and progress in the field of GBLL will come when linguists themselves become developers. Learning Python seems like a more realistic goal for language experts than convincing a team of proud, intelligent tech experts to listen to us; after all, most of us are really good at talking about communication precisely because we so often find it elusive.
Chapelle, C. (2009). The relationship between second language acquisition theory and computer-assisted language learning. Modern Language Journal, 93(1), 741-753.
Kelly Hall, J. (2019). Essentials of SLA for L2 Teachers: A Transdisciplinary Framework. Routledge.
Payne, J. & Whitney, P. (2002). Developing L2 oral proficiency through synchronous CMC: output, working memory, and interlanguage development. CALICO Journal, 20 (1), 7-32.