Clive Wynne misrepresented my published views in his Concepts essay “The perils of anthropomorphism” (Nature 428, 606; 2004). The concept of ‘critical anthropomorphism’ helps to establish ground rules for dealing with the inevitable anthropomorphic tendencies that we, as sentient human beings, confront in trying to understand the behaviour of other species.

Critical anthropomorphism involves not only careful replicable observation, but also knowledge of the natural history, ecology, and sensory and neural systems of animals as well. Such knowledge itself changes and has led me to reinterpret examples of animal behaviour involving, for example, perception and cooperation, in my own research. Problems can also arise when scientists fail to recognize inadvertent anthropomorphism in ‘objective’ studies; colleagues and I have documented erroneous conclusions in areas as diverse as communication, foraging, conservation planning and zoo design. As noted in your News Feature “True colours” (Nature 428, 596; 2004), “It's only when you begin to see the world as birds do — detecting light in the ultraviolet spectrum — that the full subtlety of their behaviour is revealed.”

Wynne asserts that I use critical anthropomorphism to argue that death-feigning behaviour in snakes is best understood by assuming that animals have conscious states. What we have shown is that hognose snakes monitor their visual environment and make adaptive decisions enabling escape behaviour while appearing completely unresponsive, and thus they are not unaware or unconscious of their surroundings when in this state, as previous authors had assumed. This problem with the word “conscious” is why I have argued for use of the term “private experience” (derived from behaviourist B. F. Skinner's “private events”) and not “consciousness” in the study of animal behaviour. The issues here are controversial enough to make careful attention to language essential.