Colin Allen

As arti® cial intelligence moves ever closer to the goal of producing fully autonomous agents, the question of how to design and implement an arti® cial moral agent (AMA) becomes increasingly pressing. Robots possessing autonomous capacities to do things that are useful to humans will also have the capacity to do things that are harmful to humans and other sentient beings. Theoretical challenges to developing arti® cial moral agents result both from controversies among ethicists about moral theory itself, and from computational limits to the implementation of such theories. In this paper the ethical disputes are surveyed, the possibility of a `moral Turing Test ’ is considered and the computational di culties accompanying the diŒerent types of approach are assessed. Human-like performance, which is prone to include immoral actions, may not be acceptable in machines, but moral perfection may be computationally unattainable. The risks posed by autonomous machines ignorantly or deliberately harming people and other sentient beings are great. The development of machines with enough intelligence to assess the eŒects of their actions on sentient beings and act accordingly may ultimately be the most important task faced by the designers of arti® cially intelligent automata.

Cognitive ethology is the comparative study of animal cognition from an evolutionary perspective. As a sub-discipline of biology it shares interest in questions concerning the immediate causes and development of behavior. As a part of ethology it is also concerned with questions about the function and evolution of behavior. I examine some recent work in cognitive ethology, and I argue that the notions of mental content and representation are important to enable researchers to answer questions and state generalizations about the function and volution of behavior.

It is a truism among ethologists that one must not forget that animals perceive and represent the world differently from humans. Sometimes this caution is phrased in terms of von Uexküll’s Umwelt concept. Yet it seems possible (perhaps even unavoidable) to adopt a common ontological framework when comparing different species of mind. For some purposes it seems sufficient to ­anchor comparative cognition in common-sense categories; bats echolocate insects (or a subset of them) after all. But for other purposes it seems necessary to find out more about how organisms organize their perceptions into biologically significant and perhaps cognitively meaningful states. Complex animals have high bandwidth sensory channels that feed into large nerve networks with very complex dynamics. Even for relatively simple animals belonging to species believed to have a small, fixed number of neurons, the odds are very much against any two animals of the same species, let alone different species, having exactly the same couplings to the environment, the same dimensionality in their nervous systems, or the same dynamics. Given such diversity (which von Uexküll himself recognized), how should we think about shared representation, shared meaning, and cognitive similarity between individuals and species?

Recently something close to a consensus about the best way to naturalize the notion of biological function appears to be emerging. Nonetheless, teleological notions in biology remain controversial. In this paper we provide a naturalistic analysis for the notion of natural design. Many authors assume that natural design should be assimilated directly to function. Others find the notion problematic because it suggests that evolution is a directed process. We argue that both of these views are mistaken. Our naturalistic account does not simply equate design with function. We argue that the distinction between function and design is important for understanding the evolution of the physical and behavioral traits of organisms.

Many psychologists and philosophers believe that the close correlation between human language and human concepts makes the attribution of concepts to nonhuman animals highly questionable. I argue for a three-part approach to attributing concepts to animals. The approach goes beyond the usual discrimination tests by seeking evidence for self-monitoring of discrimination errors. Such evidence can be collected without relying on language and, I argue, the capacity for error-detection can only be explained by attributing a kind of internal representation that is reasonably identified as a concept. Thus I hope to have shown that worries about the empirical intractability of concepts in languageless animals are misplaced.

Primatologists generally agree that monkeys lack higher-order intentional capacities related to theory of mind. Yet the discovery of the so-called “mirror neurons” in monkeys suggests to many neuroscientists that they have the rudiments of intentional understanding. Given a standard philosophical view about intentional understanding, which requires higher-order intentionality, a paradox arises. Different ways of resolving the paradox are assessed, using evidence from neural, cognitive, and behavioral studies of humans and monkeys. A decisive resolution to the paradox requires substantial additional empirical work and perhaps a rejection of the standard philosophical view.

Cognitive ethologists are in need of a good theoretical framework for attributing intentional states. Heyes and Dickinson (1990) present criteria that they claim are necessary for an intentional explanation of behavior to be justified. They suggest that questions of intentionality can only be investigated under controlled laboratory conditions and they apply their criteria to laboratory experiments to argue that the common behavior of approaching food is not intentional in most animals. We dispute the details of their argument and interpretation of the laboratory experiments. While criteria such as those suggested have a role to play in comparative studies of cognition, both laboratory and field studies are important for assessing the applicability of intentional explanations across different taxa.