Actually, terms such as "people" or "people" are supposed to include a broad mass of people, regardless of gender. But this is not the case: An evaluation of 630 billion words has shown that the English term for people, »people«, primarily refers to men. The team led by April Bailey from New York University in the USA reports on this in the journal Science Advances.
Bailey and her colleagues compared how gender-neutral terms and those describing women and men were used on nearly three billion mainly English-language websites. For this purpose, an intelligent computer program calculated the "word embeddings", in German word embeddings, of terms such as "person", "person", "man" and "woman". Word embeddings show how often and in which context individual words appear within a text. Word embeddings of terms used interchangeably such as "researcher" and "scientist" are very similar; those of very different words, such as "researcher" and "instead", on the other hand, have little in common. The procedure is based on the assumption that words that are used in a similar context are also similar in content.
As the analysis showed, the word codes of actually neutral terms such as "person" were equal to the embeds of "man" or "he" stronger than those of "woman" or "you". "Men" and "people" are used more often. In another calculation, Bailey and her team found that neutral terms and stereotypical male verbs and adjectives also appear in a similar context. So in general, an article was about people, they were described as "arrogant" or "directly" and less as "gullible" or "giggling".
The fact that terms such as »men« are linguistically almost equated with »people« allows conclusions to be drawn about what idea of a person prevails within a society. If they see themselves as rather male, this has an effect on various areas of coexistence such as health care or safety aspects, the researchers explain. Non-male persons could be disadvantaged as a result. At the same time, the understanding of the male norm is passed on to children through language and is also consolidated in younger generations.
In addition, biased language has a far-reaching influence on digital life: artificial intelligence (AI), such as that used in smartphones and computers, creates word embeddings to learn a language. The embeddings are based on texts written by people. With this, the AI takes over the linguistic imbalance between the sexes. This was shown, for example, by a work from 2019: intelligent machines automatically translated "the doctor" as "the doctor" – although the phrase "The doctor asked the nurse to help her with the procedure" (in German: "The doctor asked the nurse to help her with the procedure") even showed that a doctor was meant. In order for the discrimination of women not to continue, the researchers conclude, the programs would have to learn to calculate word embeddings without discrimination of a gender.
Quality Velocity Subtitles
How neutral the German language is in comparison can only be assessed to a limited extent on the basis of the English-language study. Presumably, however, masculine terms are also preferred in German, because the German language has a grammatical gender and gives each noun one of three specific articles. One speaks of »the table« or »the girlfriend«, instead of using the same article for »the table« and »the friend«. The more such gender-specific information is in a language, the less gender-neutral expressions it contains, the researchers explained. This, in turn, the scientists speculate, could lead to the generic masculine being used more frequently: Then there is talk of a doctor or a teacher, although the gender of the person is unknown.
