Google may use information about a person left on social networks to build suggestions for search queries and automatically complete them.

When you enter a search query into Google, Yahoo, or Bing, a drop-down suggestion list appears. On smartphones, autocompletion uses a dictionary from the device’s memory. This saves time and key-presses.

Google has approved a patent describing how the search engine could add words to a suggestion dictionary. The words would come from social networks where the user is registered. Effectively, the search engine collects information on which words the person uses for online communication and which words their friends use.

Why is Google looking for these words in social networks?

Theoretically, a user is more likely to type words they use in communication and that their friends and acquaintances use. Some users may use mostly slang in communication (e.g., teenagers) or professional terms not in general dictionaries aimed at a broad audience.

Text associated with a social network user may include the following content:

— profile pages from various social networks;

— pages where the user left data;

— forum pages and text-message history between social network users.

Various scoring methods are used to determine which words to include in the dictionary for a given user.

In addition to social network info, Google proposes using info from calendar entries, text documents, contacts, browser history. This could improve automatic query formation and create an individual user dictionary.

Overall, the patent is concentrated not on desktop search but on phone query autocompletion. That’s unsurprising given Google’s entry into the smartphone and smartphone-software market.

Judging from the patent, Google pays a lot of attention to the words people use when communicating online.