The Strong Story Hypothesis and the Directed Perception Hypothesis

I ask why humans are smarter than other primates, and I hypothesize that an important part of the answer lies in what I call the Strong Story Hypothesis, which holds that story telling and understanding have a central role in human intelligence. Next, I introduce another hypothesis, the Directed Perception Hypothesis, which holds that we derive much of our commonsense, including the commonsense required in story understanding, by deploying our perceptual apparatus on real and imagined events. Then, after discussing methodology, I describe the representations and methods embodied in the Genesis system, a story-understanding system that analyzes stories ranging from précis of Shakespeare’s plots to descriptions of conflicts in cyberspace. The Genesis system works with short story summaries, provided in English, together with low-level commonsense rules and higher-level reflection patterns, likewise expressed in English. Using only a small collection of commonsense rules and reflection patterns, Genesis demonstrates several storyunderstanding capabilities, such as determining that both Macbeth and the 2007 Russia-Estonia Cyberwar involve revenge, even though neither the word revenge nor any of its synonyms are mentioned. Finally, I describe Rao’s Visio-Spatial Reasoning System, a system that recognizes activities such as approaching, jumping, and giving, and answers commonsense questions posed by Genesis. The Right Way Just about everyone agrees that much has been accomplished since Turing published his seminal paper, Computer Machinery and Intelligence (Turing 1950). On the other hand, most would also agree that less has been accomplished than expected. Although applications of Artificial Intelligence are everywhere, we still do not have a computational theory of human intelligence. A team of dedicated first-class engineers can build systems that defeat skilled adults at chess and Jeopardy, but no one can build a system that exhibits the commonsense of a child. What has been tried? Turing argued that human intelligence is a matter of complex symbolic reasoning. Minsky argues for a multiplicity of ways of thinking coupled into a reasoning hierarchy with instinctive reactions on the lowest level and self-conscious reflection on the highest level (Minsky 2006). Brooks argues that whatever human intelligence is, studying it directly is beyond the state of the art, and we must instead organize systems in layers of competence, starting with the objective of understanding low-level layers that produce insect-level intelligence (Brooks 1991). Still others, in many seminal papers, have suggested that the right way is, for example, through architectural design (Laird, Newell, & Rosenbloom 1987), neural mimicry (McClelland & Rumelhart 1989), or statistical methods (Pearl 1988). Each of these approaches has made important contributions, especially from an engineering perspective, but none has shown us light at the end of the tunnel, not yet at least. What is missing, I think, is an approach centered on asking what exactly makes humans different from other primates and from early versions of ourselves. For guidance, I ask about the early history of our species, and I find provocative suggestions in the speculations of paleoanthropologists, especially those of Tattersall (Tattersall 1998). Basically, Tattersall believes we are symbolic and other primates were not and are not. He says we were not symbolic either, until about 50,000–70,000 years ago. Before that, we were structurally rather modern for perhaps 100,000 years, but during that earlier 100,000 years, like the Neanderthals, all we could do was make simple stone tools and work with fire. Then, we started making art, and eventually produced the drilled seashell jewelry found in the Blombos Cave, the cave paintings at Lascaux, and the figurines at Brassempouy. Such art, Tattersall believes, requires symbolic thinking and its appearance is evidence of becoming symbolic. Tattersall argues that we became symbolic rather suddenly, probably in southern Africa, possibly in a population reduced to a few thousand or a few hundred individuals. It was not a matter of slowly growing ability proportional to slowly growing brain size. More likely, it was a evolutionary accident, with nonlinear effects, that unleashed the power of other faculties previously evolved by selection for benefits other than producing human-level intelligence. Of course, saying we are symbolic does not take us very far toward a computational theory. Chomsky, who fre-