Experts develop tools to talk to dolphins

Spielberg should be directing it, or perhaps Richard Curtis. Off the Bahamas a dolphinologist and an artificial intelligence specialist thrown together on board The Stenella are this summer developing a piece of hi-tech gadgetry that will, if it works, fulfil the 1960s' vision of talking to dolphins – and, if he shows up, ET as well. Although no photographs of it have been released yet, we know the Spielbergian-sounding Cetacean Hearing and Telemetry, or CHAT, interface is real, as are the hopes it carries. It is an iPhone-sized device with two hydrophones attached and a unique one-handed keyboard called a twiddler, which, when combined, is designed to be worn around a diver's neck while swimming with wild dolphins. Inside this box is a processor that contains a complex algorithm or pattern detector that, it is hoped, will learn to identify the fundamental units of dolphin acoustic communication to enable humans to decode dolphin and then reply. "CHAT is more a potential interface than a translator as it is supplying us humans with an acoustic bridge to allow exchanges between two acoustic species," says Dr Denise Herzing, of the Department of Biological and Psychological Sciences at Florida Atlantic University and founder of the Wild Dolphin Project. She began her long-term study of a pod of Atlantic spotted dolphins in 1985 and, with Dr Thad Starner from the Georgia Institute of Technology School of Interactive Sciences, is the brains behind CHAT. "As dolphins are likely to be the second smartest creature on the planet, with similar cognitive abilities and complex social structures to humans, this device will hopefully open the window for a great understanding and connection with other sentient beings. Similar interfaces created for chimps and parrots have already increased our understanding of the abilities of these species." Dr Herzing is already running workshops with SETI (Search for Extraterrestrial Intelligence) on how to identify non-human intelligences and how to then communicate with them. She is just the latest scientist to try to crack the dolphin's code since the idea that we could actually talk to dolphins emerged in the 1960s. Although the 3,500-year-old dolphin fresco on the walls of the Palace of Knossos in Crete is testament to humanity's long love affair with Flipper, it was only after psychoanalyst John C Lilly popularised the idea that dolphins could talk – in bestsellers such as Man and Dolphin (1961) – that researchers began to take the notion seriously. Lilly believed the size of the dolphin's brain relative to its body mass and its linguistic ability marked it out as being a kind of alien on Earth. He was one of 12 scientists, calling themselves the Order of the Dolphin, who founded SETI in 1961. Since then, researchers such as Dr Louis Herman, at the Dolphin Institute at the Kewalo Basin Marine Mammal Laboratory in Hawaii, have established that not only can dolphins understand a couple of hundred words using gestures and symbols, they understand syntax too, and even the difference between a statement and a question. This line of research, though, tapered off (the Dolphin Institute ended its research in this field in 2004) due in part to the expense and time it took to train dolphins and as research was focusing on understanding how the dolphin mind works and how they communicate with each other. There was also a feeling among some researchers that the species had been pushed as far as they could go. But Dr Herzing had a hunch that there was another reason. "Most scientists create a system of communication and expect the dolphins – especially those in captivity – to learn it by using fish as a reward, and they do it, but the dolphins are not empowered to use the system to request things from the humans." From 1997, Dr Herzing's first attempts to try and co-evolve a language with dolphins led to the creation of a large keyboard with symbols on – called a lexigram board – that were paired with a whistle, so when underwater, dolphins could either press a button with their nose or mimic the whistle to receive the toy they wanted. She now hopes CHAT is going to develop this work further. Initially divers will play one of eight words Dr Herzing has come up with for things that dolphins like, whether "seaweed" or "bow wave ride", and then the software will listen to see if the dolphins mimic them. Then after CHAT's algorithm has in effect learned dolphin by listening to it, it will try to identify the words and grammar of "dolphinese" in the hope that it will help scientists create their own dolphin-like signals; although in the end the interface still depends on humans to interpret what the dolphins actually mean. For Justin Gregg, from Dolphin Communication Project based in Old Mystic, Connecticut, whose own research eavesdrops on natural dolphin communication, CHAT is another way in which new technology is revolutionising how scientists study dolphins, whether it is the electronic tags dolphins wear or recording devices planted on the seabed. However for Gregg, while CHAT is "cool and innovative" and the questions it is designed to answer are "perfectly legitimate", the interface may struggle as the building blocks it is looking for may not exist in the first place. He believes the evidence for dolphins actually having a language "just doesn't add up". "They certainly have some of the attributes that a language requires, but we know now that many species do. In fact there is not more potential in dolphins; it's about the same as other animals." That doesn't stop people believing that they are special, Gregg says. "It's a hangover from the 1960s." While Dr Seth Shostak, SETI's chief astronomer, shares Gregg's concerns over just how special dolphins actually are, he is also worried about whether we will actually be able to understand anything the dolphins are saying if the interface does work. "As dolphins can't pick up a screw driver, they are never going to have the kind of technological civilisation that humans have, so even if we were able to pick out distinct dolphin words we would be unlikely to have any idea what they meant as their world view is going to be so different from ours." With the summer a "resounding success" so far for CHAT, Dr Herzing is confident that this is only the start of a journey in two-way communication. After all, if some day there was a close encounter it "would be nice if we'd had some practice with both etiquette and the ethics of interacting, as well as establishing some potential universal protocols". Glimpses of which, she believes, she has seen already in the mimicry, imitation and synchrony dolphins have initiated to engage with humans. So this time, if ET turns up, he won't have to phone home. Aping us: more animals that chat Dolphins aren't the only species that mankind has tried to do a Doolittle with. Since 1971 the Language Research Centre at Georgia State University has been using a lexigram board to learn and develop language skills in apes. The breakthrough for researcher Susan Savage-Rumbaugh came in 1982 when she realised that a baby bonobo called Kanzi had learned English just from watching his mother's language lessons on the lexigram board. After that she abandoned formal language training, seeing instead Kanzi and other apes as a member of a shared human-bonobo culture who were like a new-born human baby  just assumed to be able to learn language. Although other researchers question just how much Kanzi comprehends what he is saying, the bonobo now has a vocabulary of over 400 words of English that he can use to talk to his trainers and visitors through a three-panel, 384-key lexigram board. He has even held an online chat on AOL. On the other hand, Irene Pepperberg's work with Alex, an African grey parrot, blew away the assumption that only a large primate or dolphin  brain could handle language. Even though arguments still rage over the extent to which he was mimicking her, by Alex's death the Professor of Psychology at Brandeis University reported that Alex had a vocab of 150 words that he could use.