INTRODUCTION

What are thoughts made of? Do we think in pictures? In words? In symbols? What is the currency of human cognition and how do the representations that make up thinking come to be in our minds? In this chapter, we explore the rich sources of input that humans receive from perception and language and how combining information from these two input streams can be used to create the amazing complexity and sophistication of the human knowledge system.

Cognitive science is often seen as emerging from the confluence of two research programs: Chomsky's nativist critique of behaviorist learning theories and the rise of artificial intelligence. Together these two tides lead to a seemingly inevitable pair of conclusions: we think in language-like symbols, and the primitive symbols used in thought are innate. If we think in innate language-like symbols, then obviously we do not think in English, or Russian, or Kuuk Thaayorre. Instead, we think in the universal language of thought – Mentalese” (Fodor, 1975). This conclusion has been explicitly defended by some in cognitive science, but more often it is an unarticulated background assumption. For example, in the literature on concepts and categorization, conceptual representations are often described using structured lists of linguistically labeled features, and researchers rarely suggest that the words used in their theories correspond to mental representations that are radically unlike words.

For some time now, Paul Churchland has been defending a connectionist theory of mental content, alliteratively labeled state-space semantics. The theory is at once holistic, prototype-based, and neurally reductionist. If you mix those ideas into a single theory, one thing is guaranteed: you will hear from Jerry Fodor and Earnest Lepore. The exchange between Churchland and the Rutgers duo has spanned several papers, and several voices have been chiming in from the sidelines. Both sides have claimed victory, but these declarations are premature. I think Churchland has been successful in addressing some of the objections levied by Fodor and Lepore, but others remain. Fodor and Lepore recognize that there may be a way out for Churchland. He could overcome some of their more pressing concerns if he embraced an extreme form of concept empiricism. In pointing this out, they intend to highlight the fatality of their objections. Empiricist theories of concepts are unacceptable on other grounds. No one, not even Paul Churchland, wants to resurrect Hume. If empiricism is the only way to navigate state space, it's time to give up on that program and set course for another semantic theory. This is a rare point of agreement between Churchland and his critics.

I will argue that empiricism isn't as unworkable as it appears. If I am right, then state-space semantics may be able to withstand the objections of Fodor and Lepore. But that does not mean Churchland wins the debate. Empiricism raises further concerns about state-space semantics.

Introduction: The Problems of Consciousness

Reading the philosophical literature on consciousness, one might get the idea that there is just one problem in consciousness studies, the hard problem. That would be a mistake. There are other problems; some are more tractable, but none is easy, and all interesting. The literature on the hard problem gives the impression that we have made little progress. Consciousness is just an excuse to work and rework familiar positions on the mind–body problem. But progress is being made elsewhere. Researchers are moving toward increasingly specific accounts of the neural basis of conscious experience. These efforts will leave some questions unanswered, but they are no less significant for that.

To move beyond the hard problem, I would like to consider some real problems facing consciousness researchers. First, there is a What Problem. This is the problem of figuring out what we are conscious of. What are the contents of conscious experience? For those looking at the brain, it is closely tied to a Where Problem. Where in the brain does consciousness arise? Locating consciousness may not be enough. We need to address a How Problem. How do certain states come to be conscious? This can be construed as a version of the hard problem, by asking it with right intonation: How could certain physical states possibly be experienced? But there is another reading that is also worth investigating.

In the 18th century, David Hume said that every idea is built up from copies of prior impressions. In modern terminology: all concepts are built up from stored records of perceptual states. This empiricist credo is enjoying a resurgence these days. Researchers in numerous fields are seriously investigating the hypothesis that thought has a perceptual basis. They are questioning the rationalist assumptions that have dominated cognitive science since its inception in the 1950s. Empiricism is still a fringe movement, however. It is often dismissed as gratuitously radical and utterly indefensible. That attitude has become something of a dogma, but it derives from two serious worries. One of them has to do with innate ideas. Empiricists have tended to be antinativists, and the current orthodoxy in cognitive science is to postulate a considerable amount of innate knowledge. I think the orthodoxy is mistaken, but I will have little to say about that here. The second major reason for rejecting empiricism has to do with abstract ideas. We often think about things that are far removed from sensory experience. There is no way to paint a mental picture of truth, justice, democracy, or necessity. If empiricism is to have any hope of success, it must be able to explain how we come to think about things at this level of abstraction.

I think the objection from abstract ideas must be dismantled piecemeal. There are different kinds of abstract ideas, and these must be accommodated in different ways.

Two Theoretical Approaches to Emotion

There seem to be two kinds of emotion theorists in the world. Some work very hard to show that emotions are essentially cognitive states. Others resist this suggestion and insist that emotions are noncognitive. The debate has appeared in many forms in philosophy and psychology. It never seems to go away. The reason for this is simple. Emotions have properties that push in both directions, properties that make them seem quite smart and properties that make them seem quite dumb. They exemplify the base impulses of our animal nature while simultaneously branching out into the most human and humane reaches of our mental repertoires. Depending on where one looks, emotions can emerge as our simplest instincts or our subtlest achievements. This double nature makes emotions captivating, but also confounding. Researchers find themselves picking one side at the expense of the other, or packaging seemingly disparate components into unstable unions. I will defend a more integrative approach. For a more thorough treatment, see Prinz (forthcoming).

Noncognitive Theories

As I will use the terms, a cognitive theory of the emotions is a theory that maintains that all true emotions involve cognitions essentially. Noncognitive theories maintain that emotions do not necessarily involve cognitions. It is no easy matter to say what cognitions are. A failure to define this key term can easily lead to unproductive cross-talk. Despite that caveat, I will proceed without a definition. One can capture the difference between cognitive and noncognitive theories by considering some examples.