Editor’s note: Martin Edic is tech writer at Language Intelligence, a Rochester, New York-based translation and localization firm. This is an edited version of a post that originally appeared under the title, “‘It’s not free,’ don’t talk complicated, language extinction and AI SuperGLUE.”
As someone working within language, translations and cross-cultural learning, I’m always interested in reading articles related to language and culture around the world. Throughout August our team ran into the usual array of odd but interesting pieces. In this post I will share a few tidbits from four of these articles. On the most basic level, all of the articles are focused on human communication. That said, the articles cover a variety of topics: a look at cultural confusion that can mess up an intended message; how execs on earnings calls can make their message seem worse through the language they use; the terrible cultural loss when a language becomes extinct; and finally the complexity of explaining language to AI systems.
The secret to cross-cultural communication
Do you think knowing a language is enough to effectively communicate in other countries? You also need to understand the language of their cultural interactions:
“I have experienced shopkeepers in several places in the Middle East who insisted at checkout that I didn’t have to pay – it was ‘on the house.’ If you didn’t know about the local culture, you might be quite surprised, thank them and attempt to walk out of the store! Those familiar with the culture, however, would know that the token gesture was merely a sign of respect and that the appropriate response would be to insist on paying as usual.”
From the Association for Talent Development (ATD)
Research: Executives’ English skills affect the outcomes of earnings calls
An overly complex use of language can materially and negatively impact stock prices:
“When we analyzed the data using a regression model, we found that the use of non-plain English and erroneous expressions, which together we classified as linguistic opacity, had real consequences. Controlling for the actual earnings news, we found that executives who spoke opaquely suffered various capital market consequences, including lower trading volume, restricted price movement and inconsistent analyst forecasts.”
From Harvard Business Review
The year of indigenous languages: Making it your business
A look at 2,680 languages in danger of extinction – and why, when they go, the culture behind them is lost forever:
“The Endangered Languages Project refers to the unprecedented decline in languages around the globe as a form of ‘mass extinction.’ They estimate that around 40 percent of the world’s languages are endangered. Most of them are indigenous ones.
“This is particularly disturbing because indigenous languages are extremely rich in culture. They are complex knowledge systems, full of ancient wisdom and a unique understanding of the world. They are an integral part of our shared human heritage.”
From Nimdzi
AI researchers launch SuperGLUE, a rigorous benchmark for language understanding
Well, if you want an idea of how complex teaching AI to understand the complexities of meaning in language can be, including the ability to respond intelligently, try reading this article! Deep stuff.
“’SuperGLUE comprises new ways to test creative approaches on a range of difficult NLP [Natural Language Processing] tasks focused on innovations in a number of core areas of machine learning, including sample-efficient, transfer, multitask and self-supervised learning. To challenge researchers, we selected tasks that have varied formats, have more nuanced questions, have yet to be solved using state-of-the-art methods, and are easily solvable by people,’ Facebook AI researchers said in a blog post today.”