Judgments about animal suffering are of central importance in evaluating scientific protocols. To make such judgments we must confront a set of problematic practical questions: How are we to predict the effects on animals of particular scientific protocols? How are we to recognize and evaluate animal suffering when it does occur? How are we to alleviate that suffering? How do we address larger issues of harm that go beyond the individual animal? Here, we examine ways of answering such questions, concentrating upon critical anthropomorphism. Critical Anthropomorphism

Anthropomorphism is the attribution of human characteristics (including the projection of subjective states and feelings) to non-human entities. Such characteristics are usually applied to animals, but may also be applied to inanimate objects-such as when we talk about a "vicious" storm, an "angry" wind, an "aggressive" weed, an "insult" to the flag, and so on. Although these terms are obviously metaphors when applied to the inanimate, their use with animals can lead to uncritical thinking and invalid interpretations. (5,6,20) In its most extreme form, uncritical anthropomorphism can mean animals being "personified," or treated as if they actually were humans. In less extreme form, uncritical anthropomorphism may lead to misinterpreting an animal's behavior and hence to misunderstanding that animal's needs and emotional state (e.g., cattle kept outdoors in extreme cold may in fact be adapted to such conditions so that bringing them indoors can cause the animals heat-stroke; chimpanzees "grin" when aggressive or fearful and not, as uncritical anthropomorphism would claim, when amused or welcoming).

Criticisms of animal research often rely on an intuitive and uncritical anthropomorphism and a misplaced empathy that reflects more on human sensibilities than a real understanding of the state the animal may be experiencing. (7) In spite of these pitfalls, our intuitions about what is best for an animal, based on a knowledge of ourselves and other people, may often be more on than off the mark. (23) Indeed, if the similarities between animals and humans are deemed sufficient to extrapolate from animals to humans in medical, toxicity, neuroscience, behavior, and cognition research, by implication the possibility exists of extrapolating from humans to animals. Yet even in the human sphere, empathy based on self-knowledge alone cannot be the basis for scientific understanding. Similarly, our knowledge of people, no matter how sophisticated, will be insufficient for die task of understanding animals. Thus we need a critical anthropomorphism (CA), in which empathy is tempered by objective knowledge of the particular species' (or individual animal's) life history, behavior, and physiology.

While CA does not ignore empathy and the assault on our sensibilities of animal treatment, it requires a willingness to incorporate objective knowledge of the animal's natural history, nervous system, domestication, and prior experience, as well as an appreciation for the scientific literature on the treatment- It also incorporates the ethological view that animals are adapted to a range of environments by evolution and that removal of aspects of this natural environment (spatial, social, climatic, dietary, structural) should be done with care, preferably...