Maps | General information (large file!) | Details and features | Specifications | Issue areas | Press release
Methodology | Background paper | The cartographic metaphor | Criteria | How the maps work (large file!)
For Instructors and Students | Importance of Turing debate | For instructors | For students | Protagonist index | FAQS
Commentary and Reviews | Commentary and reviews | Errata and corrections
Action Items | Buy the set of maps | How you can participate in this debate
Examples | View the maps. | Map 1 | Map 2 | Map 3 | Map 4 | Map 5 | Map 6 | Map 7 | (large files!)
MacroVU home page | Send us a message | Project Director's Home Page

We relied heavily on the metaphor of exploration and mapmaking during the research and production of the Can Computers Think? series. At the start, we thought a single map would suffice to chart the territory we were interested in. One map soon became two. Then four, then seven. The task we had set for ourselves grew more daunting as the boundaries of the territory grew. I kept reminding our team that, like the Lewis and Clark expedition, we could never know what was over the next hill. You cannot know the size of a continent until you reach the end of the landform. Our team experienced a similar sense of traveling through a "great unknown," which came from the existence of only a few good summaries of arguments in the field of artificial intelligence.

First expedition

These maps are the result of our first expedition into the jungles of a major academic argument. They bear all of the excitement and all of the imperfections of any such first expedition. The explorers have returned older and wiser from what is now an almost six-year adventure.

Like many explorers, we began our journey in familiar territory. Soon, we were into completely unknown (to us) areas of the debate. The difficulties of charting unknown territory was compounded by the fact that our team was making up the criteria and conventions for a new kind of cartography as we "traveled." Often, we had to stop and ask: How shall we present this? What kind of a map are we trying to make? At other times, we found ourselves wandering over land bridges into new continents-vast unmapped areas at the frontiers of the central debate about machine intelligence. Should we cross the boundary into the mind-body problem? Should we cross over into arguments about the nature of the mind? Should we develop a map of the argument about whether animals could think? A few further thoughts about the frontiers of the maps can be found in the "How to Read Argumentation Maps" section of this booklet.

Scale and cartographic distortions

Like all map-makers, we had to choose a scale to work with, and we had to stick to it. Cities on a roadmap are little rectangles or circles; arguments on an argumentation map are short summaries in claim boxes. We chose such a small scale because of the vastness of the territory, and because we wanted to capture the arguments at a broad, overview level---a level at which a lay reader can confidently navigate and understand the issues, seeing how they fit into broader conceptual contexts and territories.

Distortions are necessitated by any cartography. Geographic maps of the three-dimensional world are necessarily distorted when pressed into the confines of a two-dimensional surface. In addition to distortions of scale, these maps also distort chronological structures and sequences of debate as they are presented in certain general surveys, and the argumentation maps sometimes present issues differently than the authors themselves do. For example, although John Searle's Chinese Room and Chinese Gym arguments are part of the same cluster of published articles, and although they are variants of the same basic thought experiment, we separated them according to our conceptual division of the territory: the Chinese Room arguments appear in the context of the physical symbol systems debate (Map 4, which grows out of Map 3), and the Chinese Gym argument is located in the context of the connectionism debate (Map 5) .

Unmapped territory

A considerable number of arguments on the maps are imbedded in wider debates and discussions in philosophy, psychology, neuroscience, and other disciplines. Such imbedding illustrates how most philosophical thinking is broadly connected and interdisciplinary. In order to create a manageable and readable map, we left out extensive mapping of many such "peripheral" debates. An example of our treatment of this problem is the free will debates on Map 1, where we mapped only those arguments that specifically dwelt on Turing's question of machine intelligence, rather than on the degree to which human beings have free will, which is a longstanding philosophical question.

Index of Unmapped Territory

Map 2

Behaviorism/operationalism

Map 3

Other approaches to AI learning
Chomsky arguments
Language of thought arguments
Non-monotonic arguments
The frame problem
Gibson AI arguments
Chess and other game-playing

Map 5

Learning
Levels
Associationism
Connectionist representations
Systematicity arguments

Map 6

Consciousness
Functionalism
The mind-body problem
Epi-phenomenalism

Mapping jungles

We discovered some pretty thick jungles of prose during our exploration. Each of our team members ran into them, often got bogged down in the swamps, and frankly we sometimes had to detour around impassable areas. This is by way of saying that there exists a considerable amount of academic prose that is impenetrable, even to the most indomitable mappers. We may have missed some good arguments, and our summaries of others may be plainly wrong. We hope that these are not too many. But he or she who grows a jungle is likely to have it mapped as such.

Landmarks

If the reader is the traveler, and the links, or arrows, are the road, then the icons and illustrations that appear throughout the maps are landmarks and navigational aids. Their inclusion goes against the general tradition of refraining from using any kind of illustration or diagram in philosophical argument. As one early reader remarked, our project may have produced more visual accompaniment to philosophical discussion than existed in the entire previous history of philosophy. The benefits of including graphical landmarks are many. Illustrations serve as mnemonic devices for a reader who wants to return to a specific argument or region, for example. The graphics also sometimes aid quick comprehension of a complicated idea by providing a concrete clarification of an argument, in the form of an example. In addition, the illustrations add an element of levity to debates that can otherwise be quite obscure and abstract.