When Characters Speak Off the Page


“Oh, John, you don’t believe that Martin took that money from the commissary till! You don’t believe it, do you, John?”

He did not answer.

“Say you don’t believe it!” she urged. “Tell me it isn’t true!” She had come close to him. As he bent toward her, the wind blew a tendril of her hair against his cheek.

“And if it were?” he slowly asked. “You’d love him just the same?”

“Oh yes,” she whispered shivering. “Oh, yes, no matter what he’d done.”


Perhaps Martin took that money from the till, or perhaps he did not; learning the truth would require reading a bit more of Helen Ellsworth Wright’s story “A Covenant Redeemed” than I have any interest in doing. Far more interesting to me is how John questions Molly. He does so slowly, and oddly enough, this may have affected the speed with which we read his words.

Bo Yao and Christoph Sheepers, researchers at the University of Glasgow, asked participants to silently read a number of passages containing quoted speech. Sometimes, the character spoke his words slowly and other times quickly. After tracking participants’ eye movements, the researchers discovered that participants read the slow speech more slowly than the fast speech. Readers, the authors concluded, were “hearing” the character’s voices inside their heads as they made their way through the texts.

So what to make of this? An immediate thought is that readers may not have been listening to voices at all. Rather, the idea of anything being done slowly could be enough to make the concept of slowness more salient, slowing readers down much the way the smell of popcorn encourages us to eat or seeing an aquarium reminds us to DVR Shark Week on the Discovery Channel. But the researchers also looked at reading times for indirect speech (e.g., John slowly asked if she’d love him just the same) and found no effect whatsoever. This at once casts doubt on the more general (and less interesting) “salience” explanation and supports the authors’ initial interpretation: because indirect speech is a summary of what is said, not a faithful rendering, we shouldn’t expect readers to “hear” the characters in quite the same way. (It’s an open question how readers respond to speech following quotative like, the subject of last week’s post, but I’d bet my postdoc salary on our treating it like direct, rather than indirect, speech.)

Of course, there must be a limit to our embodiment of characters’ speech. Were John to speak loudly, would we read his words differently than if he’d spoken quietly? Perhaps, but any relationship between the speech and our embodiment of it would be much messier, as we can’t silently read more or less silently. It’s also not entirely clear whether readers are responding to actual versus relative speed: halfway through a novel full of characters who speak as though they were in an Aaron Sorkin film, we may find ourselves fully habituated to a “new normal” rate of speaking. Were we sensitive to every drawl, how could we ever get through a novel like Gone With the Wind?

Still, the research has the comfortable, if fallible, backing of human intuition. According to an exceedingly informal poll inflicted on my friends and family, many of us do hear voices when we read. That’s why when a beloved book is turned into a movie (or even an audiobook) we so often feel flummoxed that someone else’s take on a character’s voice could be so different from our own.

Permission required for reprinting, reproducing, or other uses.

Jessica Love holds a doctorate in cognitive psychology and edits Kellogg Insight at Northwestern University’s Kellogg School of Management.


Please enter a valid email address
That address is already in use
The security code entered was incorrect
Thanks for signing up