Σελίδες

Τετάρτη 29 Μαΐου 2013

Talking with Animals


Could there really be pets that understand everything we say? The dream of speaking to animals runs from Sigurd in Norse mythology right up to Dr Dolittle. Some researchers have claimed that they can actually do it, by teaching their subjects sign language or another mutually comprehensible means of communication. 

These included famous chimps like Nim and Washoe and the gorilla Koko. A few years ago, we were on the brink of a breakthrough in inter-species communications – so what happened?


Wittgenstein famously comm­ented that if a lion could talk, we would not be able to understand it.[1] He believed that different ‘forms of life’ would have such different viewpoints that there could not be any common ground. We might exchange phrases like “there are two zebras over there”, but cannot hope to understand lion ethics or æsthetics, or even know what ‘zebra’ really means to a lion. Think how hard it is for non-enthusiasts to ‘get’ a football fan’s enthusiasm and multiply by a million.

A few years later, the linguist Noam Chomsky suggested that language is an innate human skill and that we share an in-built ‘universal grammar’ common to all human languages. Animals, lacking this grammar, would not be able to communicate in the same way. By way of challenge, in 1967 Allen and Beatrix Gardner set about teaching a chimpanzee called Washoe to talk.

Washoe was given a highly stimulating environment, with toys, games, books and friends to play with. In many ways, she was treated like a human child, and had four main human companions who stayed with her for many years. Originally based in Washoe County, Nevada, she was later moved to the Central University of Washington.

Washoe’s carers communic­ated with each other in sign language, and she gradually picked it up herself, learning over 350 words. There are many anecdotal accounts of her language skill; she made up her own combination words, such as using the signs for “water” and “bird” when she saw a swan. Washoe lived until 2007, but research had moved on.

A second project in 1973 called “Nim Chimpsky” aimed at a more rigorous scientific approach. Like Washoe, Nim learned a significant vocabulary in sign language and could, for example, point to or ask for a banana or yoghurt. Meanwhile, in Project Koko, a lowland gorilla learned over a thousand words and is also credited with combining ‘water and ‘bird’ to describe a duck landing on a lake. These apes also proved capable of teaching others sign language and using it among themselves.

Critics objected that this was simple signing, and that neither Nim nor Washoe had any grasp of grammar in spite of the researchers’ claims. It looked more like conditioning, learning that a certain action brings a reward without any deeper understanding. And a closer look at film of the chimps producing what might be taken as intelligent communication showed something else: “The frame-by-frame analysis revealed trainers unconsciously prompting and modelling each word, and Nim imitating. Unaware of their own prompting, the trainers had credited Nim with producing sentences.”[2]

In 1980, Thomas Sebeok, a linguist at Indiana University and sceptic on ape language, organised a conference to review work in this field called “The Clever Hans Phenomenon: Communicat­ions with Horses, Whales, Apes and People”. 

Clever Hans, the 19th-century German horse and calculating prodigy, lost his calc­ulating powers when there was nobody around for him to pick up cues from. Sebeok suggested that similar effects were at work in the ape communications field, and that there was self-deception among researchers over just how well their subjects were communicating.

“Give orange me give eat orange me eat orange give me eat orange give me you,” was reportedly Nim’s longest utterance. This certainly communic­ates something, and might work as a lyric, but it does look like signalling rather than language.

While the believers kept on believing – as always – the mainstream view moved sharply away. There was too much suspicion that researchers were simply seeing what they wanted to see.

Ape intelligence studies continue, such as the one involving Kanzi, a chimp who can make and use tools and communicate via a keyboard.[3] He’s a remarkable chimp, but there are few who would claim that Kanzi can form grammatical sentences. “Get red ball,” “ball get red” and “red ball get” are all the same.

Meanwhile, there is also work with other species. Forgetting questionable efforts such as the Bowlingual translator (which ‘translates’ dog barks into human language), there is the Cetacean Hearing and Telemetry (CHAT) programme. Denise Herzing of the Wild Dolphin Project in Florida is working on a device for divers to talk to dolphins in something resembling dolphinese. The device will be able to detect and analyse dolphin squeaks and generate noises in reply. The idea is that a diver plays a given sound which the dolphin will associate with an action like “play with seaweed”, and that the dolphin will learn to use the sound itself.[4]

CHAT builds on previous dolphin efforts using underwater keyboards, and may one day be very useful for practical dolphin-human communications on the level of “get red ball”. But however smart dolphins may be, there’s little chance of a real conversation yet. As Hans and Nim and the others have shown, the risk is that when we think we’re talking to them, they’re just reflecting us and we’re really talking to ourselves.


NOTES
1 Ludwig Wittgenstein: Philo­sophical Investigations (1953).
2 The Animal Communication Project.
3 Great Ape Trust.
4 New Scientist, 9 May 2011.

Δεν υπάρχουν σχόλια:

Δημοσίευση σχολίου