ALCStudies Journal

Riqui’s en Quito

I’m in touch with new friends in Ecuador through an application on my phone called WhatsApp. Their English is often far better than my Spanish. I’ve been using Google Translate to help me correspond. When I arrived in Ecuador for the first time in December last year, I knew barely a lick of the lingua franca. My mental phrase book was limited to Hola, Como esta? and Como se llama? I’ve learned a few more words and phrases since then, including some curse words and slang for communicating inspiration, frustration and desire.¹

Technologies exist now that purport to remove the necessity for our ever having to learn a new language. A person from a local university showed me an app on her phone that allows her to hold it up to a person in a foreign country, ask them to speak into it, and receive an approximation of the corresponding phrases in her own idiom. While this eliminates the effort (and the fun) of learning a language, I’m not sure shoving your phone in someone’s face to hold a conversation will endear yourself to local residents.²

Douglas Hofstadter rips Google Translate a new one in the Atlantic. Hofstadter’s beef is that GT’s algorithms don’t understand the meaning behind what you are trying to say. Therefore, it can’t really replace a real human translator (and take his job). I find Google Translate useful. Of course GT doesn’t give a rat’s ass about what you are trying to say. I’m guessing it correlates words and phrases based on proximity and context against a history of similar words and phrases in a different language stored in it’s vast Google-ian deep-thought-like³ repository. The actual translation is likely based on probability and pattern recognition. That’s a guess. Anyone with more knowledge in this field feel free to correct me.

Artificial intelligence, natural language processing and machine learning have been around as concepts in business, academia, science fiction and prognostications of the future for years.⁴ The technology has gone from theory to practice and is finding its way into call centers, banking, medical transcription, and most usefully, assistive devices for the vision- and hearing-impaired and cognitively and physically disabled. It now allows us to look up phrases and get almost instantaneous pretty good okay-ish translations on the web that that give the gist and allows us to understand and be understood.

As Hofstadter points out, speaking and writing across cultures contains idiosyncratic and unique phrases and meaning.⁵ Even human translators find balancing nuances of meaning against accuracy and clarity challenging. Grammar checkers and thesauri included in word processing software reject work by writers such as Hemingway and (especially) James Joyce as ungrammatical and wrong. Machine-only translation often produces borderline gibberish.⁶

As a bit of assistive technology, GT has its place for helping those who sometimes want a quick and dirty⁷ way to communicate in a different language. Replacing real translation with this sort of tool makes us complacent and robs us of the worthwhile work and pleasure of finding and appreciating the beauties and subtleties of another language.

In the utilitarian bottom-line world we live in, too often we believe we have no other choice. The scary thing is when people start taking these pieces of technology as gospel and assuming they are the only game in town. – CDL