A speech scientist reveals why the viral 'laurel' versus 'yanny' argument has only one right answer
- The internet is aflame about whether a robotic-sounding recording says "yanny" or "laurel."
- Speech scientists say there's a simple reason for the audio trickery that has to do with the way our brains learn to quickly decipher vowels and words.
- The recording seems to be a slightly altered version of vocabulary.com's pronunciation aide for the word laurel. (Sorry, team yanny.)
Yanny or laurel?
The internet exploded in argument this week as people debated which of those words a robot was saying in this recording, posted by YouTube vlog-er Chloe Feldman:
Hardik Kothare, who works in the speech neuroscience lab at UCSF was quick to weigh in with his assessment of the sound. To his ear, this was definitely a recording of the word laurel.
If you're still on team yanny, take a listen to the original recording from Vocabulary.com, where it's the pronunciation key for just one word: laurel. (Apologies, yanny fans.)
The sound was recorded by a professional opera singer who was one of the original cast members in the Broadway musical Cats, according to Wired. The dictionary site hasn't revealed the man's identity, but said he was one of several trained singers enlisted to record hundreds of thousands of pronunciations, based on the rules of the international phonetic alphabet.
Kothare suggested that the recording was likely "cleverly synthesized" to trick our brain's powers of speech detection. He said there's a simple, logical reason why some folks who listen to the viral recording hear yanny while others pick up laurel.
"The human brain is trained to perceive and interpret speech on the fly in a remarkable way," Kothare said Tuesday on Twitter.Our ears learn at a young age to pick up clues about the vowels people are spitting out by focusing on frequencies at which certain sounds tend to resonate. The frequencies for each sound are a little different from person to person and language to language. And if you mess with the frequencies in a recording, you can change what people hear - it's similar to the way that our eyes can be tricked by an optical illusion.
The New York Times tried this out today and created its own yanny-versus-laurel audio switching tool.
It turns out that our brains can shift pretty easily between hearing yanni and laurel, just based on how low or high the frequency of the recording gets. Couple this with all the cultural and linguistic ways we've been trained to hear certain vowels, and you've got a perfect recipe for a little audio illusion.
"Speech perception and production depends heavily on an internal map of speech sounds," Kothare said. "You learn this map while learning to speak as a toddler, and also while hearing others speak on a day-to-day basis."
What do you hear?! Yanny or Laurel pic.twitter.com/jvHhCbMc8I- Cloe Feldman (@CloeCouture) May 15, 2018
Notice the white patch or jump after the /æ/ vowel? That may be perceived as the lack of a vowel sound or a consonant interspersed between the vowels. Hence the /n/ in Yanny. (14/n) pic.twitter.com/knbSRxYzLG- Hardik Kothare (@hardikkothare) May 16, 2018