At Language Insight, proofreading is an integral part of the services we offer. From transcription to translation, a thorough proofread is always carried out as part of the process to ensure the resulting document is accurate. That includes every news story and blog post we write too.

However, it would seem we’re not alone. A new study has revealed the world is made up of proof-readers. In fact, we’re all at it!

A research team at the Department of Brain and Cognitive Sciences (BCS) at Massachusetts Institute of Technology has published its new report in the Proceedings of the National Academy of Sciences. In the paper, they reveal that when spoken information is lost between two people as a result of external noise and distraction, the listener edits it in order to make sense of the information they have received.

We’re all editors

According to the research, which was supported by the National Science Foundation, humans use a mental editing process to make sense of language. When the information they receive is confusing, there are specific strategies that are used in order to make it comprehensible.

For instance, if someone says: “I have given the sausage a dog”, this could sound like a riddle or simply leave the listener confused. However, it’s likely they will mentally edit what they have heard to: “I have given the sausage [TO] a dog.” This makes a lot more sense and is something the listener will do first rather than ask for clarification.

Edward Gibson, a professor at BCS, explains that hearing speech over what he calls a “noisy channel” can cause information to get lost. His team’s paper, ‘Rational integration of noisy evidence and prior semantic expectations in sentence interpretation’, explains that because of this, “proof-hearing” is employed.

Leon Bergen, co-author of the study and a PHD student at BSC, tells MIT News: “What we’re getting is quantitative evidence about how exactly people are doing this proofreading. It’s a well-calibrated process.”

Interestingly, the researchers discovered that the more nonsensical a sentence is, the more likely the recipient is to think it is correct. In the example “on to the cat jumped a table”, the words ‘on to’ need to be deleted and re-inserted between ‘jumped’ and ‘a table’. However, the vast majority of test subjects who heard the sentence regarded it as correct and thought it was deliberately surreal. Meanwhile, in the sausage dog example, half of the subjects recognised there was an error. The researchers concluded that this is because only one amendment was needed in the sausage dog example, as opposed to the two required in the cat sentence. In addition, subjects were more likely to think something was missing from the sentences they heard than that they contained too many words.

Researchers identified other editing processes employed by listeners as well. One was that the more incorrect information a person is presented with, the more they are likely to judge it to be correct. Should they hear several incorrect sentences, they will search for meaning within these and assume they are right, whereas if only the odd sentence is nonsensical they will spot that it is obviously wrong.

Mr Bergen explains: “People seem to be taking into account statistical information about the input that they’re receiving to figure out what kinds of mistakes are most likely in different environments.”

The research has already caused excitement in language circles. Roger Levy, a University of California professor of linguistics, tells MIT News: “People are trying to reverse-engineer what the message is, to make sense of what they’ve heard or read.” However, this process occurs without people even being aware of it.

Predicting the future

Linguistics is certainly a psychological subject and the language a person speaks can even influence the way they perceive the world. Economist Keith Chen went so far as to theorise that someone’s mother tongue can affect their country’s saving rates.

In his paper, he explained that some languages are futureless while in others the speaker differentiates between present and future. German is a futureless language, so a person might say: “It snows tomorrow,” while English differentiates so the sentence would be: “It will snow tomorrow.” Because in German, the present and future are not separated, the future does not seem as far away. This means people who speak German are, on average, more likely to save for their retirement than they are in English-speaking countries, where the future psychologically is perceived as further away. As a result, the savings rates are affected.