Homophone Phobic? Hear’s Help

-1Technology giveth, and technology taketh away. What it gives in instantly accessible information, digital readouts, and spell checkers, it takes away in research and memory skills, the ability to tell clockwise from counterclockwise, and spelling proficiency. And once we have started down the path of any of these innovations, it is hard, if not impossible, to turn back. Thus from a generation reliant on computerized spell checking has emerged a new problem: the ubiquitous homophone. Every professor I know has encountered the problem:

  • He retired at the peek of his career.
  • The Pilgrims learned to cook maze.
  • She soon came to her census.
  • What a waist of time!
  • For all intense and purposes …

A few of these are genuine spelling errors; most are typos that a spell checker either approved (peek) or offered to correct with the correct spelling of a homophone (offering waist, for example, for the misstyped wast). Having lost their proofreading skills, students don’t catch these homophones. Occasionally the result is a refreshingly hilarious moment for their professors. Since I teach fiction, my two recent favorites were a detective who’d “known about it from the gecko” and an onanistic prison cellmate “who stank of sweat and seamen.”

More often, we end up groaning, swearing, cursing spell checkers, and lecturing glassy-eyed students on the importance of proofreading. Now we can send them to Homophone Check, the brainchild of the composition instructor Jason Braun and the computer programmer Dan McKenzie, where they can enter their text in the window and see the words they might have misused highlighted in yellow. They won’t get the right words handed to them, but at least they’ll stop to check what might have slipped through before. Right?

Well, not so fast. First, right now Braun and McKenzie have loaded their site with 40 of “the most commonly confused homophones,” with sample sentences rather than dictionary definitions to gently guide the writer toward the correct word. The first problem is that English contains about 10,000 homophones. So for the passage “He told hymn he wood take a brake in an our. Even if our sails sore, the prophets wee turn inn will be phew,” Homophone Check highlighted only the two uses of our. I spoke with Jason Braun about the relative ineffectiveness of a checking tool with such a small lexicon. He said:

I wanted to get the homophone checker out there and see if people would use it. It’s not perfect. I make no claims about that. … The great thing about software or apps or Web apps is that people are OK with beta versions. If someone creates a new app that’s useful but not perfect, people will use it, and help improve the app.

In other words, let’s see if we like this idea, then build the glossary. Sounds reasonable. And Braun, a dyslexic who spent much of his academic career “as the knucklehead who thought he was Baudelaire,” is ambitious to build a version of Homophone Check geared toward dyslexics as well as one for ESL students.

The second problem is more complicated. If the homophone is close in meaning and used the same way syntactically—think affect/effect—the highlighting and sample sentences may not help much. Let’s say you’ve written, “Your possessiveness effects how I feel about you.” First, you’ll find the homophones listed only in sets under the first alphabetical member, so you would need to guess that the typo in effects has to do with the first letter and that the homophone begins with a. You’d then find sample sentences like:

  • She hopes to effect change once she’s promoted.
  • Her poise will affect her chance of promotion.

I don’t know about your students. Mine will doubtless move on to the next highlight.

But where there’s Wi-Fi, there’s a way. “Given the time and resources,” Braun says, “we could bring in someone to help us come up with an algorithm that would predict, to a greater or lesser degree, when a particular homophone was used incorrectly.” Or, if users were willing to let their texts remain in the program’s memory, its developers could enact what Braun calls “machine learning” in which “we might be able to use this data to let Homophone Check learn from itself.”

I know: sigh, groan. But the spell-checker gremlin is already at large in the world. Eventually, we’ll let loose some sort of homophone checker to try to catch him. If Braun’s wizardry is to fare better than the spell cast by the sorcerer’s apprentice, he needs our help and not our derision. You can reach him at

Return to Top