Talking with Alex*

Humans are immersed in an ocean of words — we understand them, think with them, repeat them, play with their sounds and meanings, and create novel combinations that are understandable to other humans adrift in our part of the ocean. The skill with which a human uses words in speaking has long been a key element in the perception of his or her intelligence. We have a marked tendency to treat people who are articulate as more intelligent than those who are speak poorly. For example, I am startled when a well-spoken ‘customer service representative’ on the telephone is unable to spell psychology or to retain a seven-digit phone number. I am surprised in the opposite direction when someone who speaks English poorly, or with a large proportion of uninterpretable slang, suddenly conveys a subtle and discerning point or quickly fixes a complex problem with my computer software. This language test also has traditionally been used to separate the intellectual ability of humans from that of other animals. Put simply, if other animals can’t understand and generate words or signs in sentencelike combinations, we judge them as less intelligent than humans. The language-based divide that we have constructed between humans and other animals is well defended. During the last 30 years, work suggesting language-related communication in species ranging from chimpanzees to sea lions has attracted a firestorm of criticism that questioned its accuracy, efficacy, and importance. Objections have included claims of outright fraud, inadvertent cuing, and dismissal of results as ‘mere trial-and-error learning’. Perhaps the most common academic complaint is that the language-like productions of nonhumans lack the grammatical constraints and generativity characteristic of human language. Why are humans, certainly academic humans, so intent on distinguishing human language reception and production from the abilities of nonhuman animals? One possibility is that language is the key to a number