Why Does Repeating a Word Make It Lose Its Meaning?

Language— a complicated process that involves syntax, semantics, and auditory processing—allows people to communicate needs, wants, and ideas.¹ Thus, multiple networks across the brain enable the ability to process and produce language.

According to cognitive psychologist Michael Vitevitch at the University of Kansas, language errors provide useful insights into the underlying patterns of how the brain deals with language. His group studies these language mistakes and related auditory illusions, like when a speech phrase becomes a song in one’s head, to learn more about language mechanisms and hearing disorders. “By looking at these errors, we can kind of get a better understanding of what steps your mind goes through to assemble a sentence and try to convey thoughts and ideas to people,” he said.

This work also provides insights into another common language phenomenon: semantic satiation, or the experience of a repeated word, like language, losing its meaning to an individual.

Communication requires individuals to both perceive and produce linguistic information. One model that synthesizes both of these concepts is the Node Structure Theory, which hypothesizes that speech perception occurs in a distinct network from speech production.² Neurons responsible for encoding the meaning of a given word form what’s called a lexicon node and, separately, clusters of neurons that store the syllables of a language comprise a phonological node. These two systems are interconnected so that they can cooperate to provide an individual with the semantic and composite information to understand and create words.

Vitevitch said that the model can also serve as a basis to understand how speech errors and auditory illusions, like semantic satiation, occur. When a person initially says a word, neurons in the semantic node fire, providing a speaker with the learned meaning of the word.

As the person repeats this word, though, these neurons fatigue; the constant stimulation from the repetition prevents the cells from going through their normal refractory period, so they stop signaling. “If the semantic information here stops firing…the meaning is gone. And so, all you're doing is just going through the phonological information and the motor movements to make those sounds over and over and over again,” Vitevitch said.

Vitevitch explained that the reason why these semantic-encoding neurons fatigue while syllable-encoding neurons continue to function is akin to the “fast twitch” and “slow twitch” types of muscle cells. During conversation, the faster fatiguing, slower recovering semantic neurons allow people to switch to new topics more easily. Meanwhile, speakers and listeners must be able to produce and perceive the same syllable in many different contexts, so they fatigue slower and recover quicker.

“I was always interested in illusions like this because systems don't break randomly. They break where there's a weak point, and so where that weak point is could help us understand how the system overall is built and better understand how to potentially address different disorders and that sort of thing,” Vitevitch said.

In the lab, Vitevitch said that his group and others often use self-report surveys in which participants either listen to or speak a word or phrase and indicate when their perception of the information changes.³ He added that researchers are also exploring the use of electrophysiological methods to explore where these linguistic phenomena occur in the brain, while other scientists are using deep learning models to understand how they arise.^4,5