Most researchers regard language as unique to humans, something that makes our species special. But they fiercely debate how the ability to speak and listen evolved. Did speech require our species to evolve novel capabilities, or did we simply combine and enhance various abilities that other animals have, too? A new study with a language-trained chimp suggests that when it comes to understanding speech, the basic equipment might already have been present in our apelike ancestors.
The notion that language evolved only in the human lineage and has no parallels in other animals has long been attributed to the linguist Noam Chomsky, who argued beginning in the 1960s that humans had a special "language organ" unique to us. But more recent studies have shown that other species are surprisingly good at communication, and many researchers have abandoned this idea—even Chomsky himself no longer holds to it strictly.
However, some scientists continue to argue that humans have evolved unique ways to perceive and understand speech that allows us to use words as symbols for complex meanings. These contentions are based in part on a notable human talent: We can recognize words and understand entire sentences, even if the sounds of the words have been dramatically altered until they are a pale shadow of their linguistically meaningful selves.
So a team of researchers turned to Panzee, a 25-year-old chimpanzee, to test the assumption that only humans have this talent. Humans raised Panzee from the age of 8 days, and her caregivers exposed her to a rich diet of English language conversation about food, people, objects, and events. Panzee can't talk, so she communicates with those around her using a lexigram board of symbols corresponding to English words (see photo). She can point to 128 different lexigrams when she hears the corresponding spoken word.
A team led by Lisa Heimbauer, a cognitive psychologist at Georgia State University in Atlanta, set out to see how well Panzee could duplicate the human talent of understanding English word sounds when they are so badly distorted that they are difficult to recognize. The team used two electronic methods to distort the words: noise-vocoded (NV) synthesis, which makes words sound very raspy and breathy; the other, known as sine-wave (SW) synthesis, which reduces words to just three tones, is something like converting a rich color photograph into a stripped-down black and white version. (The words included chimp-friendly terms such as banana, potato, tickle, and balloon.)
Panzee performed well above chance when she heard distorted versions of 48 words that she knew and had to choose among four lexigrams, the team reports this week in Current Biology. Thus, while a chance result would have been one out of four correct choices, or 25%, Panzee scored 55% with NV words and about 40% with SW words, which are particularly difficult to understand even for humans. This was almost as good as the performance of 32 human subjects using the same 48 words, who chose the correct NV word 70% of the time but, like Panzee, the correct SW word only 40% of the time.
Heimbauer and her colleagues say that Panzee's strong performance argues against the idea that humans evolved highly attuned speech-recognition abilities only after they split from the chimp line some 5 million to 7 million years ago. The finding that Panzee passed a challenging test for speech recognition implies, the team writes, that "even quite sophisticated human speech perception phenomena may be within reach for some nonhumans." Still, the team says that its experiments don't rule out that humans have evolved additional speech-perception abilities that our ancestors and chimps lacked.
The authors have come up with a "nice result," says biologist Johan Bolhuis of Utrecht University in the Netherlands, but it shouldn't come as "a big surprise." For example, zebra finches have been shown to be able to distinguish very small sound differences in words spoken by humans, including ones that differ by only one vowel. That's a talent Bolhuis considers "even more remarkable" than Panzee's because it so closely parallels the way humans perceive speech.
J.D. Trout, a psychologist and philosopher at Loyola University Chicago in Illinois, thinks that the authors are far from proving their case. "These experiments don't bear on the question of whether speech is a special adaptation of humans," Trout insists, noting that the human subjects had to pull matching words out of their vocabularies of about 30,000 words, whereas Panzee had a much smaller vocabulary to search through. But Heimbauer points out that unlike the human subjects, Panzee had never been exposed to distorted speech before the experiment, making her performance all the more impressive.