If you were to ask Siri, Apple’s intelligent personal assistant, to help you grasp calculus, she would most probably not understand you. This is because Siri has a hard time understanding metaphors. But new research could change that, as it suggests that digital assistants could one day learn the ancient algorithms that humans have used for centuries to create and understand metaphors.
A team of researchers at UC Berkeley and Lehigh University in Pennsylvania has mapped out 1,100 years of the metaphoric (English) language to find patterns in how speakers have added metaphors to their vocabulary. Their results demonstrate how humans have used language that originally described palpable experiences to describe intangible concepts (for example, “grasping an object” has become “grasping and idea”).
Interestingly, the team found that words associated with digestive organs, solidity, wetness, plants, and textiles were more likely to provide sources for metaphorical expression, while emotional and mental states like excitement, fear, and pride were more likely to be the targets of metaphorical extension.
These results provide first large-scale evidence that the creation of new metaphorical word meanings is systematic, and could, therefore, serve as a base for designing natural languages processing systems such as Siri and other digital assistants, to help them understand and use creativity in human language.
Source:
University of California – Berkeley via ScienceDaily (https://www.sciencedaily.com/releases/2017/06/170627152606.htm)
The Wonderful World of Algorithms 