A recently published article concerning a talking beluga whale is in the news. I answered some questions for Michael Marshall at New Scientist about this topic (article here), and also emailed with Ed Yong over at Not Exactly Rocket Science. For extra in-depth insight (i.e., the mad ramblings of someone who had not yet had his morning coffee), here is a full copy of the email I sent to Ed in reply to questions he asked about this news story:
I think this is pretty cool indeed. Although the recording sounds like a drunken beluga playing a kazoo. Here are some thoughts/answers:
This is interesting in that there are anecdotal accounts of belugas imitating human speech, and even an article from 1979 about it (which is the one I’ve seen cited before in the scientific lit), but this is the first case where it was recorded and analyzed to verify what was happening. So it’s cool to have this added to the body of literature. Although listening to the recording, it does not sound exactly like human speech (i.e., I have no idea what the whale is “saying”), but they’ve certainly made the case that the whale was trying very hard to produce human language sounds. As you probably know, many dolphin species have been observed imitating human speech patterns or artificial signals, so this is not a novel finding in that sense. But this is one of only a handful of cases where it happened spontaneously (i.e., without the animal being trained), so that’s cool. This study is also further evidence of the mechanism odontocetes use to make their vocalizations (i.e., vibrating their phonic lips like human vocal cords as opposed to “whistling” by passing air over a surface).
The question is, why would the beluga be doing this? Only a handful of animals have the ability to mimic sounds or alter parameters of their vocalizations to match an external sound they hear, including songbirds, parrots, hummingbirds, cetaceans (including humpback whales and all toothed cetaceans like belugas and dolphins), seals, elephants, and bats. This ability, which is retained into adulthood for these species, has different functions. For songbirds, adding new calls to their repertoire allows them to advertise their fitness and/or engage in territorial defense. For dolphins (and likely belugas) it has a social function. Learning new vocalizations allows one to broadcast identity information (and possibly to address another individual by, for example, using that individual’s signature whistle), and to adopt new vocal patterns when one changes partners or group affiliation. Bottlenose dolphins that are long-term partners (like in male alliances in Shark Bay) sometimes co-create new whistles, converging on a whistle (or set of whistles) that is unique to that partnership. It’s likely similar for belugas, since they are very social. There are a couple other explanations as to why toothed whales are so good at changing their vocal parameters and thus imitation. One is that these animals are constantly needing to modulate their echolocation signals for a variety of reasons when using their echolocation to obtain object and environmental information. So they have an amazing ability to alter parameters of their vocal tract already, which should make it all the more easy to replicate human speech sounds via the same mechanism. Also, because the quality of their vocalizations (e.g., whistles) will change depending on how deep they are in the water (i.e., water depth will distort the quality of their vocalizations), it’s likely the case that toothed cetaceans are very sensitive to tweaking parameters of their vocalizations on a constant basis to retain the desired whistle (or other) structure as they move through the water column. Again, this makes them very flexible with the ability to change their vocalizations and acutely aware of shifts in frequency patterns, which should predispose them to being more capable of imitating sounds than other species.
In any event, this mimicking ability it fairly rare in the animal kingdom. Although, you probably saw that article on mice imitating vocalizations/songs, which is one of the first times a terrestrial mammal species shows this ability as well, and suggests that perhaps many more species have this ability than we currently know about.
On a side note, you should check out this new article: Preliminary Study of Object Labeling Using Sound Production in a Beluga http://www.comparativepsychology.org/ijcp-vol25-3-2012/02.Murayama_PDF.pdf
It’s a brand new study from Japan where they’ve trained a captive beluga to produce and understand four different vocalizations that were matched with four different objects. This technique has been used many times before in animal language experiments with dolphins and usually starts out as the basis of two-way communication experiments that never go anywhere. But it is an example of belugas’ ability to learn to imitate sounds.
Ridgway S, Carder D, Jeffries M, & Todd M (2012). Spontaneous human speech mimicry by a cetacean. Current biology : CB, 22 (20) PMID: 23098588