Google’s efforts to search out an integrative language in search

Google’s John Mueller and Martin Splitt had two special guests on their podcast show this week to talk about how Google is tackling inclusive language in search – a topic we’ll briefly cover only covered over here once. The short version of this 30-minute podcast is that Google search automatically processes including language, as it does with any new type of term or slang.

The two guests were Zineb Ait Bahajji and Bruno Cartoni, who work in this area at Google.

Inclusive language aims at avoiding offense and fulfilling the ideals of egalitarianism by avoiding terms that express or imply sexist, racist, or otherwise biased, biased, or derogatory ideas towards a particular group of people (and sometimes animals). The use of inclusive language could be seen as a form of political correctness; the term “political correctness” is often used to refer to this practice, either as a neutral description by proponents or commentators in general, or with negative connotations by their opponents.

Zineb defined inclusive language as “language that is actually free of words or phrases that are stereotypical or discriminatory. As a result, the masculine form has been the standard term in any context for a very long time and in many languages ​​such as French. For example, people would say, “During the board meeting, the chairpersons have the discussion.” But that could be a chairperson.

About 8 minutes into the podcast, the Google rep describe some of the efforts they are making on Google search to make it more inclusive.

In French and German there is an increasing tendency to abbreviate the feminine and masculine forms together. For example, instead of saying “étudiant” or student in the masculine form and then adding “étudiante”, which is the feminine form, we would just pull everything together with a special character. So it would be “étudiant” followed by a special character, such as a hyphen or a slash, and then the final “e”, the feminine form. And recently, in French in particular, we have seen an increase in the use of a very special character, the middle point, which is called the point median in French. You can now find it on Google Gboard on the mobile phone by long pressing the period key.

John Mueller explained that Google indexes in three ways for this purpose.

(1) Indexing – Finding the words in documents and then expanding those words to include “some of those words in the appropriate versions”. John said, “Maybe if that’s not a natural word boundary, but a sign that you’re referring to your different versions, you might leave out the punctuation.”

(2) Order of precedence – John said that the order of precedence that John said is “all about the serving side of things”. John explained, “We saw from previous episodes of the podcast that we automatically expand the query that is displayed based on known synonyms, abbreviations, different versions of different words, and in practice these systems tend to run automatically.” So Google is able to handle inclusive language and rank it automatically.

(3) Entities – John said the point is to understand entities, both in content and in queries and the individual attributes assigned. “For example, we know that the Eiffel Tower is a building and that it is a certain height. You can ask Google how tall the Eiffel Tower is. For other entities, gender can also play a role,” explained John.

John first said this using the knowledge graph. The knowledge graph “is usually built automatically based on the content we find online. However, that doesn’t mean it’s always automatically correct. For example, if you ask who the second lady of the United States is, well, the die Role is still the same, but now it’s different because the US Vice President is a woman and she’s married to a man, so the title is different now, and that kind of bias needs to be improved both in our languages ​​and of course in Google systems. “

Here is the podcast, it’s an interesting podcast to listen to:

When you write, you have to think about how everyone feels involved when they read what you write, said Zineb and Bruno. They also said that the search engines have to adjust to the way people write, not tell them how to write. But there is a trend for people to write in a more inclusive way. Also, each language has its own integrative language challenges, including masculine or feminine and other domains – some are easier to work with and others more difficult – everything is language dependent.

Google is still in the early stages to work on this topic in the various Google services and products.

Here are the key points that John sums up at the end:

(1) Ideally, search engines would automatically display a variety of content across all different aspects.

(2) Search engines should automatically be able to use the various popular forms of inclusive writing when they are found on a website. So the goal should be that these pages are processed automatically and are accessible for all related searches.

(3) Ideally, website owners should only use inclusive writing techniques whenever appropriate.

(4) People like you or listeners should keep the search engines moving to make them work better and better because without that gentle nudge down the line, I think it’s very easy for search engines to say, “Well, it seems to be working well.” Maybe we don’t have to change anything. ” But maybe we need to change some things. Again, it’s an interesting topic and I recommend you listen to it. I don’t think this is a purely SEO thing, it’s a reminder that we should all remember to write like this, and it’s not easy. Trust me, I mess this up all the time. “

(5) Zineb added, “Write for your users. And your users are diverse. They don’t just have an audience of male or female readers. So make sure everyone feels included by reading everything you do write online. “

Forum discussion at Twitter.

Leave a Reply

Your email address will not be published. Required fields are marked *