How Apple Puts Indigenous Languages on the iPhone Keyboard
本文为原文前 6,000 字符的节选翻译,完整内容请查看原文。
How Apple Puts Indigenous Languages on the iPhone Keyboard
Chuck Hoskin Jr. fondly recalls watching a room full of Cherokee immersion students crowd around a screen to watch How to Train Your Dragon. Except Hiccup and Astrid weren’t speaking English—the protagonists were speaking Cherokee, dubbed by voices from the local community.
切罗基族酋长查克·霍斯金(Chuck Hoskin Jr.)深情地回忆起,他曾看到满屋子的切罗基语沉浸式教学学生围在屏幕前观看《驯龙高手》。只不过,片中的主角希卡普(Hiccup)和阿斯特丽德(Astrid)说的不是英语,而是由当地社区成员配音的切罗基语。
“They’re hearing it in our mother tongue—that didn’t happen a generation ago,” Hoskin Jr., principal chief of the Cherokee Nation, tells WIRED. “We’re a long way, thankfully, from rote classroom instruction and memorization to a world where kids can see the sort of creativity that sparks imagination and helps every kid become more articulate and literate.”
“他们正在用我们的母语聆听这一切——这在上一代人身上从未发生过,”切罗基族酋长霍斯金告诉《连线》杂志。“值得庆幸的是,我们已经走出了死记硬背的课堂教学模式,进入了一个新世界,孩子们在这里可以看到激发想象力的创造力,并帮助每个孩子变得更善于表达、更有文化素养。”
The Cherokee Nation has around 1,500 active speakers left, and most are over 70. Hoskin Jr. says the community is rapidly losing that generation, which is why it has spent time training up second-language learners. It even has a baby immersion program, hoping their first words are Cherokee.
切罗基族目前仅剩约 1500 名活跃的语言使用者,且大多数人已超过 70 岁。霍斯金表示,该社区正在迅速失去这一代人,这就是为什么他们花费时间培养第二语言学习者。他们甚至开展了婴儿沉浸式项目,希望孩子们学会的第一句话就是切罗基语。
“The foes we’ve had historically are the so-called Age of Discovery, colonization, the displacement of our people economically and geographically, boarding schools,” he says. “But it’s nothing compared to the mathematics we’re up against now. Which is why we’re filled with this urgency every day to do all we can to find partners that can help us accelerate what we want to do.”
“我们历史上所面临的敌人是所谓的‘大航海时代’、殖民主义、我们人民在经济和地理上的流离失所,以及寄宿学校,”他说。“但与我们现在面临的数学难题相比,这些都不算什么。这就是为什么我们每天都充满紧迫感,竭尽全力寻找能够帮助我们加速实现目标的合作伙伴。”
One of those accelerants is Apple. The Cherokee Nation has long collaborated with the iPhone maker, and Cherokee was the first Native American language included in iOS in 2010. Today, Apple says its iOS keyboard supports more than 200 languages and language variants, with a handful added last year in iOS 26, like Haitian Creole and Nez Perce. In iOS 27, it expanded support for Indigenous languages to include other Native American languages like Blackfoot, Comanche, and Northern East Cree.
苹果公司就是其中一个加速器。切罗基族长期以来一直与这家 iPhone 制造商合作,切罗基语是 2010 年首个被纳入 iOS 系统的美洲原住民语言。如今,苹果表示其 iOS 键盘支持超过 200 种语言和语言变体,去年在 iOS 26 中又增加了一些,如海地克里奥尔语和内兹珀斯语。在 iOS 27 中,它进一步扩大了对原住民语言的支持,纳入了黑脚语、科曼奇语和北东克里语等其他美洲原住民语言。
The goal is to meet young people where they already spend much of their waking hours: phone screens. Hoskin Jr. says that to revitalize and save a language, it has to become accessible. Apple says the effort isn’t as simple as translating words. It’s a multi-year endeavor that requires deep research, and the company leans on community members for their expertise.
其目标是在年轻人花费大部分清醒时间的地方——手机屏幕上——与他们相遇。霍斯金表示,要振兴和拯救一门语言,它必须变得触手可及。苹果公司表示,这项工作并不像翻译单词那么简单。这是一项需要深入研究的多年工程,公司非常依赖社区成员提供的专业知识。
But the process doesn’t start in Cupertino. It actually begins with Unicode. You may have heard of the name—it’s the organization that approves and standardizes the new emoji that appear on your smartphone every year. It also handles time zone localization, currency names, imperial versus metric, and so on.
但这个过程并非始于库比蒂诺,而是始于统一码联盟(Unicode)。你可能听说过这个名字——该组织负责批准和标准化每年出现在你智能手机上的新表情符号。它还处理时区本地化、货币名称、英制与公制转换等事务。
Before you can type or read a character on a smartphone screen, it needs to exist as a recognized character in the Unicode Standard. “Everything does start with Unicode in terms of a language, but it actually starts with characters and scripts,” says Unicode CEO Toral Cowieson. “We now have more than 157,000 characters that are included.”
在你能在智能手机屏幕上输入或阅读一个字符之前,它必须作为统一码标准中公认的字符存在。“就语言而言,一切确实始于统一码,但实际上它始于字符和书写系统,”统一码联盟首席执行官托拉尔·考维森(Toral Cowieson)表示。“我们现在已经收录了超过 15.7 万个字符。”
Cowieson calls Unicode a kind of digital Switzerland—a neutral consortium where rivals like Apple, Microsoft, and Google sit at the same table as linguists and volunteer contributors. (Cowieson told WIRED earlier this year that Unicode had only four full-time employees.) Proposing a language, verifying it, and then encoding the character set can take three to five years.
考维森将统一码联盟称为“数字瑞士”——一个中立的联盟,苹果、微软和谷歌等竞争对手与语言学家及志愿者贡献者坐在同一张桌子旁。(考维森今年早些时候告诉《连线》杂志,统一码联盟只有四名全职员工。)提议一种语言、进行验证,然后对字符集进行编码,可能需要三到五年的时间。
After a script is encoded, platforms like Apple and Google need to add font support. Without this step, a character could appear as an empty box—often called “tofu.” Next is orthography: a set of rules and systems for writing a language. Often, languages have a well-defined orthography, but when it’s not standardized, communities use ad hoc replacements because historically available keyboards don’t support it.
在书写系统编码完成后,苹果和谷歌等平台需要添加字体支持。如果没有这一步,字符可能会显示为一个空框,通常被称为“豆腐块”。接下来是正字法:即书写语言的一套规则和系统。通常,语言都有明确的正字法,但当它未标准化时,社区会使用临时替代方案,因为历史上现有的键盘并不支持这些字符。
Take the glottal stop—the speech sound made in the throat between the syllables when you say, “uh-oh.” Some languages have distinct characters for glottal stops, but as some keyboards don’t support them, many indigenous cultures use alternatives, like a question mark, to represent the sound. That’s when Apple connects with the community to figure out the best way to represent the sounds in digital text.
以喉塞音为例——即当你发出“uh-oh”时,音节之间在喉咙里产生的语音。有些语言有专门的喉塞音字符,但由于某些键盘不支持,许多原住民文化使用问号等替代符号来表示该音。这时,苹果会与社区联系,找出在数字文本中表示这些声音的最佳方式。
That brings us to the keyboard layout. When a community already has an established physical keyboard layout—like Osage—Apple adapts it directly to the software keyboard. Some cultures prefer the QWERTY layouts people are already familiar with.
这就引出了键盘布局的问题。当一个社区已经有了成熟的物理键盘布局(如奥塞奇语)时,苹果会将其直接适配到软件键盘上。有些文化则更倾向于人们已经熟悉的 QWERTY 布局。
Other languages need more custom approaches. For example, Nez Perce, or Nimipuutímt, does not use capital letters. Rather than having the Shift key toggle traditional uppercase letters, Apple repurposed it to reveal an accented secondary layer, so users don’t have to hold down individual keys to see those characters, which can slow typing. In the Cherokee language, each character mostly represents a consonant combined with a vowel sound, a form known as syllabary. Apple added adaptive vowel keys on the bottom row, so when a user types a consonant, these keys change to show the vowel combinations they can make.
其他语言则需要更定制化的方法。例如,内兹珀斯语(Nez Perce 或 Nimipuutímt)不使用大写字母。苹果没有让 Shift 键切换传统的大写字母,而是将其重新设计为显示带有重音的第二层字符,这样用户就不必按住单个键来查看这些字符,从而避免了打字速度变慢。在切罗基语中,每个字符大多代表一个辅音与元音的组合,这种形式被称为音节文字。苹果在底行添加了自适应元音键,因此当用户输入辅音时,这些键会发生变化,显示出可以组成的元音组合。
That’s a newer change to the Cherokee keyboard—one Apple approached the community about, meaning building a keyboard for a language isn’t a one-and-done project.
这是切罗基语键盘的一项较新改进,苹果为此专门与社区进行了沟通,这意味着为一种语言构建键盘并非一劳永逸的项目。
But digitizing an indigenous culture’s language hasn’t always been harmless. Genner Llanes-Ortiz, a Maya scholar and researcher at Bishop’s University, studies how Indigenous people use digital technologies to reclaim language heritage. He recalls how Māori speakers rejected proposals to sell their language to tech companies.
然而,将原住民文化语言数字化并不总是无害的。毕晓普大学的玛雅学者兼研究员詹纳·利亚内斯-奥尔蒂斯(Genner Llanes-Ortiz)研究原住民如何利用数字技术重拾语言遗产。他回忆起毛利语使用者是如何拒绝将他们的语言出售给科技公司的提议的。
“There’s a historical memory about the ways of dispossession that Indigenous peoples have experienced in the world,” Llanes-Ortiz says. “Māori working in their own revitalization efforts have been very firm in protecting access and storage of their linguistic materials because they have been exploited. They have experienced extraction of knowledge in the past.”
“原住民在世界上经历过各种被剥夺的遭遇,这已成为一种历史记忆,”利亚内斯-奥尔蒂斯说。“致力于语言复兴的毛利人非常坚定地保护其语言材料的访问权和存储权,因为他们曾被剥削。他们过去曾经历过知识被掠夺的痛苦。”
Language sovereignty and data ownership are top of mind, according to Chirasree Mandal, an Apple software engineer. “We don’t own anything, and we make it very clear that we are not trying to silo any data that we gain from this process,” she says. “It’s abo
据苹果软件工程师奇拉什里·曼达尔(Chirasree Mandal)称,语言主权和数据所有权是重中之重。“我们不拥有任何东西,我们非常明确地表示,我们不会试图将从这个过程中获得的任何数据据为己有,”她说。“这关乎……”