Processing of Polysemy in Information Retrieval: A Review of the Research


By Elijah Enos



This paper provides an overview of the challenges faced by information designers attempting to process instances of polysemy in natural language, as well as an overview of proposed solutions. Polysemy is the linguistic phenomenon in which one word will have more than one meaning. This creates a challenge for users of search engines and databases, who may have difficulty finding proper results, as well as for designers, who must create systems that can recognize nuances in natural language, even as this language changes and evolves. The issue of specialized disciplines with complex vocabularies is also addressed. Solutions proposed and tested include the addition of pseudo words to descriptor lists, and the utilizing of word webs – through Latent Semantic Analysis – to link similar terms together. Issues addressed for consideration in the future include designers’ ability to recognize polysemy within and across multiple languages and the role that voice recognition systems (i.e. Siri, Cortana) might play.


Full Paper (PDF File)

Reading Room - Portrait Eli

Elijah Enos

Elijah Enos lives in Sacramento, CA, where he works in the Collections Services department of the Sacramento Public Library. A book enthusiast since childhood, he holds a BA in English from California State University, Sacramento. Prior to finding a job at SPL, Elijah volunteered there as an undergrad, and credits this experience with fostering his desire to become a librarian. He intends to specialize in reference services and teen services during his time at San Jose State.