Automatic translation is now mainstream. It’s simple enough to have text on a web site or in a text field translated between languages. The Google translate app on a smartphone combines (AR) “augmented reality” and translation algorithms to create a screen image that substitutes what’s in front of the camera with something like the same text in your own language. That’s handy for reading street signs, museum captions and menus.
It’s also handy when a non-Chinese speaker wants to see the nutritional content of food products.
Here is a photograph of the back of the pack along with its automatic translation appearing in real time as AR on the smartphone screen.
The translations are not always accurate, but they don’t need to be. Here’s something I discovered in Ukrainian, along with the app’s flickering real time translations. Click on the picture to see the detail.
This is automated translation, but it also bears a resemblance to the products of surrealist writing, literary collage, and the play of metaphor.
Guess it’d be gratifying to know whether your Ukrainian text sample recounts ingredients, prayers, or both at once. Also, would you suppose that the ingredient: “dried mosquito fish wire” sprang from a frustrated concrete-poet-turned-product-label-writer (versus automated deformity), for that may be its destiny(?) … in all, the proportion of signal-to-noise is praiseworthy. Google Translate is perhaps more akin to shimmering locks betwixt undulating (language) seas … transmitting more-or-less bent interpretations that wriggle through shuttered gaps. If it fails to satisfy the pursuit of artefact-free translation, then it may live on as a surreal poetry engine. Which programming group deserves praise for the logical wheat, which for the poetic chaff–or is this two sides of the same semiological coin? BTW: always wonder whether you hail to “coin,” or “koi-né.” As an avid reader of the MIT book series, always feel sheepish I’m garbling its author’s name.
Looks like koi-né means “common language” in some tongues. So that will do. These translation tools seem to work by finding best statistical match between phrases in the the source and target language from a vast database of texts. So there’s already something of a mash-up going on between different texts. Thanks for the insights Daniel.