During our team’s weekly terminology meeting, we examined some poorly translated specimens, such as extraordinarias amortizaciones de capital translated as “extraordinary amortizations of principal.” One of my colleagues, Rafael, half smiled and half snickered when he said “Oh, sounds like Google Translate.”
We translators are detail oriented in the best of times. In the worst of times, we are fault finders. Identifying a collage of symptoms is not diagnosis, however, and we collectively tend to misidentify as Google Translate output a translation that is more likely the product of poor writing skills.
We might think that the more experienced we are as translators, the better we are at spotting errors and the nature of those errors. But human psychology points to habituation, where our eyes increasingly get used to seeing an erroneous or nonsensical expression (a phrase like “scientifically formulated”) as normal. Even the expression “the new normal” is suspect if subject to careful scrutiny.
How can we, trained language specialists*, be prone to misread an agrammatical or erroneous statement and consider it normal? One answer could be that it’s one side of language evolution. Language users push the boundaries of what’s conventional until a critical mass of users is reached, users who agree that a newly formulated expression is normal. By force of habit, no less. To the trained eye, a sentence, phrase or question that sounds, walks and reads too far apart from convention is considered incorrect. To the bristled consciences out there, the binomial correct/incorrect is a requisite judgment function. Yes, you can talk and write any way you want. But if you want your writing to be meaningful to others beside you and convey a message to others, you have to play by the rules. Rules set up by the majority of users. Grammar rules, syntax rules, vocabulary-forming rules.
Instead of blaming our machines for our own mistakes, we need to look in the mirror and see who is actually the writing instrument at play. Is it possible to write so poorly that we can mistake the product for the GIGO trash spat out by computer? Certainly. But AI, MT and their offspring such as Google Translate have enough blemishes and misshapen innards and brains, we do not need to torture them any further by misplacing on them the responsibility—and the guilt— that is
so distinctly and humanly ours to face.
*The expression is broadly and generously applied here.