To bolster our claims about the experience or mind of a dog, we will learn how to ask the dog if we’re right. The trouble, of course, with asking a dog if he is happy or depressed is not that the question makes no sense. It’s that we are very poor at understanding his response. We’re made terribly lazy by language. I might guess at the reasons behind my friend’s recalcitrant, standoffish behavior for weeks, forming elaborate, psychologically complex descriptions of what her actions indicate about what she thinks I meant on some fraught occasion. But my best strategy by leaps is to simply ask her. She’ll tell me. Dogs, on the other hand, never answer in the way we’d hope: by replying in sentences, well punctuated and with italicized emphases. Still, if we look, they have plainly answered. For instance, is a dog who watches you with a sigh as you prepare to leave for work depressed? Are dogs left at home all day pessimistic? Bored? Or just exhaling idly, preparing for a nap? Looking at behavior to learn about an animal’s mental experience is precisely the idea behind some cleverly designed recent experiments. The researchers used not dogs, but that shopworn research subject, the laboratory rat. The behavior of rats in cages may be the single largest contributor to the corpus of psychological knowledge. In most cases, the rat itself is not of interest: the research isn’t about rats per se. Surprisingly, it’s about humans. The notion is that rats learn and remember by using some of the same mechanisms that humans use—but rats are easier to keep in tiny boxes and subject to restricted stimuli in the hopes of getting a response. And the millions of responses by millions of laboratory rats, Rattus norvegicus, have greatly informed our understanding of human psychology. But the rats themselves are intrinsically interesting as well. People who work with rats in laboratories sometimes describe their animals’ “depression” or their exuberant natures. Some rats seem lazy, some are cheery; some pessimistic, some optimistic. The researchers took two of these characterizations— pessimism and optimism—and gave them operational definitions: definitions in terms of behavior that allow us to determine whether real differences in the rats can be seen. Instead of simply extrapolating from how humans look when pessimistic, we can ask how a pessimistic rat might be distinguished by its behavior from an optimistic one. Thus, the rats’ behavior was examined not as a mirror to our own but as indicating something about . . . rats: about rat preference and rat emotions. Their subjects were placed in tightly restricted environments: some were “unpredictable” environments, where the bedding, cage mates, and the light and dark schedule were always changing; others were stable, predictable environments. The experimental design took advantage of the fact that, hanging out in their cages with little to do, rats quickly learn to associate new events with simultaneously occurring phenomena. In this case, a particular pitch was played over speakers into the cages of the rats. It was a prompt to press a lever: the lever triggered the arrival of a pellet of food. When a different pitch was played and the rats pressed the lever, they were greeted with an unpleasant sound and no food. These rats, reliably like lab rats before them, quickly learned the association. They raced over to the food-dispensing lever only when the good-harbinger sound appeared, like young children rallied by the jingle of an ice-cream truck. All of the rats learned this easily. But when the rats were played a new sound, one between the two learned pitches, what the researchers found was that the rats’ environment mattered. Those who had been housed in a predictable environment interpreted the new sound to mean food; those in unstable environments did not. These rats had learned optimism or pessimism about the world. To watch the rats in the predictable environments jump with alacrity at every new sound is to see optimism in action. Small changes in the environment were enough to prompt a large change in outlook. Rat lab workers’ intuitions about the mood of their charges may be spot-on. We can subject our intuitions about dogs to the same kind of analysis. For any anthropomorphism we use to describe our dogs, we can ask two questions: One, is there a natural behavior this action might have evolved from? And two, what would that anthropomorphic claim amount to if we deconstructed it?