Language Log

« previous post | next post »

Rachel Premack, "Watch: This orangutan is uprooting what we previously knew about language", Washington Post 7/28/2016:

At first glance, this video of an orangutan imitating a trainer’s grunts may not seem incredibly significant. But primate researchers say Rocky, the 11-year-old orangutan in the video, could fundamentally alter how we think about spoken language.

“We don’t know exactly how human language emerged, but we do know that great apes have the cognitive ability to understand language,” said Robert Shumaker, executive vice president and zoo director at the Indianapolis Zoo. “We can no longer say that humans are the only species who can learn new vocalization and control [it] using the vocal folds or voice box.”

I can only hope that Robert Shumaker is being misquoted here, because (even leaving aside parrots and other talking birds, who use their syrinx rather than mammalian-style vocal folds), everyone knows that there are many animals who can "learn new vocalization and control [it]".

Back in 2003, I reported on a "University project to teach seals how to talk":

St Andrew's University has acquired its very own Dr Dolittle, with the arrival of a Harvard academic on a mission to teach seals to talk.

Tecumseh Fitch, a specialist in language evolution, plans to recruit undergraduates to "hang out" with young seals in the hope that the seals will pick up human speech patterns.  

The experiment may have echoes of Hugh Lofting's creation, but it is inspired by the bizarre but entirely genuine example of a talking seal called Hoover, who entertained visitors at an aquarium in Boston, Massachusetts, with entire phrases delivered in gravelly male tones.  

"He said things like 'hey Hoover, get out of there', and 'move over Hoover'," said Dr Fitch. "Not only did he do this, he said it with a Maine fisherman's accent."

In my comments on Tecumseh's efforts ("Talking seals and singing dogs", 11/28/2003), I noted that I once taught (or at least encouraged) a dog to sing.

And more significantly, in 1989 I watched Cynthia McLemore persuade a Yorkshire Terrier to perform a complex intonation contour in return for cheese. Unfortunately there are no videos of this (impressive) feat, but  YouTube is full of talking dog videos. And you can even find specific videos on how to teach your dog to sing, for instance this one (though it's more of a performance than a training manual):

You can even learn How to Teach Your Dog to Whisper.

And earlier this year, I posted on Cat Phonetics (3/13/2016), including a video in which a Lund University researcher asserts that "cats can vary their melody — their intonation — consciously".

Over the years, there have been several interesting articles about the ability of Asian elephants to imitate human speech — see e.g. "Elephant imitates Korean", 11/2/2012 — the quality of the imitation can be remarkable, as suggested by these audio clips:

annyong ("hello")
anja ("sit down")
nuo ("lie down")
choah ("good")

(The clips come from  a peer-reviewed article  — Stoeger et al., "An Asian Elephant Imitates Human Speech", Current Biology 2012).

The point of the reported orangutan research is apparently to demonstrate that a (limited) ability for conscious control of the vocal organs exists in non-human apes. Thus the scientific paper (Adriano R. Lameira, Madeleine E. Hardus, Alexander Mielke, Serge A. Wich & Robert W. Shumaker, "Vocal fold control beyond the species-specific repertoire in an orang-utan", Scientific Reports 2016) explains:

Vocal fold control was critical to the evolution of spoken language, much as it today allows us to learn vowel systems. It has, however, never been demonstrated directly in a non-human primate, leading to the suggestion that it evolved in the human lineage after divergence from great apes. Here, we provide the first evidence for real-time, dynamic and interactive vocal fold control in a great ape during an imitation “do-as-I-do” game with a human demonstrator. Notably, the orang-utan subject skilfully produced “wookies” – an idiosyncratic vocalization exhibiting a unique spectral profile among the orang-utan vocal repertoire. The subject instantaneously matched human-produced wookies as they were randomly modulated in pitch, adjusting his voice frequency up or down when the human demonstrator did so, readily generating distinct low vs. high frequency sub-variants.

In other words, Rocky can do what dogs do, though not yet what seals and elephants can.

It's true that some experts have claimed that non-human apes have no conscious control over their vocalizations. That notion has always seemed improbable to me, and I'm glad to see it definitively refuted. But it's too bad that the WaPo — and most of the many other mass-media articles on the subject — find it necessary to make claims about the discovery that are contradicted by  thousands of YouTube videos, and millions of people's life experiences.

 

July 30, 2016 @ 1:16 pm · Filed by Mark Liberman under Animal communication

Permalink