Preface
1 AI: what is it? 
definitions: what would it look like if I saw one? 
True AI Story: 1.1-ELIZA meets PARRY:
the syntax is willing but the semantics is weak 
a history of scaling down 
categorizations of AI work 
the goals of AI research 
the heuristic programming approach 
the Samuel phenomenon 

2 AI and the Science of Computer Usage: The Forging
of a Methodology 
how to use the essential tool? 
first specify, then verify 
the nature of AI problems 
a methodology of incremental exploration 
rapid prototypes to the rescue? 
supportive environments 
forging a new methodology 
is AI so different? 

3 The Major Paradigms 
symbolic search spaces 
planning intelligent solutions 
SSSP infrastructure 
the pivotal role of searching strategies 
heuristic pruning 
connectionism: a possible alternative? 
connectionism: the second coming 
on not losing their inhibitions 
the need for decay 
subsymbolic connectionism: the good news 
when is an AI system like a piece of fine china? 
subsymbolic connectionism: the real news 
reasoning with amorphous complexity 
the myth of empirical guidance 
what's the stopping rule? 
single-minded models 
philosophical objections 
potential solutions to the dilemma 
formal analysis 
software support systems 
approximate translation-the truth about mendacity 
the SSSP and the CP: integration, bifurcation,
or annilation? 
simulated evolution: guess and try it out 
'bad' paradigms 

4 The Babel of AI Languages 
it's all done by manipulating symbols 
LISP 
flexibility 
the magic of recursion 
code-data equivalence 
the special assignment 
lists of properties 
PROLOG 
the independence of declaration
loss of control: better or worse? 
extralogical pollutants 
negation as failure 
verify or compute 
bidirectionality 
pattern matching 
the promises of PROLOG
parallelism 
a specification language 
heuristic controls 
object-oriented programming 
programming environments 
LISP environments
LOOPS 
POPLOG 
True AI -Story: 4.1. DIMWIT (Do I Mean What I Tell):
A PA (Programmer's Assailant) System 

5 Current Expert Systems Technology (CEST)
experts with tunnel vision

6
the basic assumptions and the criticisms
what can be CESTed?
explanations and context sensitivity
updating knowledge bases and machine learning
let's dig deeper
logical decision making
human and computer decision making
classes of human decision making
connectionism: a possible answer?
knowledge elicitation
knowledge engineers and the third degree
automatic learning from examples
empirical techniques
CEST: where is it and where is it going?
Knowledge Representation: A Problem of Both
Structure and Function
why networks?
why neurons?

pandering to.evolution: beware of classical reconditioning
neural architectures: in the beginning
knowledge representation: structure and function
the SSSP and the CP: representational issues
knowledge representation in the CP
functionally distributed representations
symbolic connectionist representations
winner-takes-all subnets
hybrid connectionism
totally distributed representations
path-like architectures in the CP
bath-like architectures in the CP
knowledge representation in the SSSP
logic-based representations
procedural representations of knowledge
semantic networks
elements of structured knowledge: frames,
scripts, and schemata

7 Vision: Seeing is Perceiving
bottoming in: operators canny, uncanny,
and cannyless
pixel processing 
edges and lines 
vertices or junctions 
texture: a truly superficial feature 
illumination, reflectance, and other sources
of nuisance 
the intrinsic image 
model-based vision systems 
True AI Story: 7.1 
beer cans, broomsticks, etc.
seeing as perceiving 
oversight and hallucination 
the modularity of human vision 
eyeballs and nervous optics 
biological feature detectors
humaa perceptual behavior 
breaking up context
structuring top-down information 
a cognitive model of word recognition 
the eye of the robot 
general theories of visual perception 
the vision of connectionists

8 Language Processing: What You Hear is What You Are
natural language
what mode of natural language?
the goals of AI-NLP
natural language: the essential ingred'ients
phonetics and phonology
the lexical level and above 
generation and analysis
natural language generation (NLG)
text generation systems
empirical guidance for NLG
natural language understanding (NLU)
syntax, grammars, and parsing
grammars
furious transformational grammarians
sleep curiously 
transition networks: augmented and otherwise
unification and the new grammatism
semantic definite clause grammars (SDCG)
NLP and a formal complaint
semantics 
the meaning of semantics 
the atomic struture of meaning
the case of the missing-blocks world 
True AI Story: 8.1 SHRDLU and a "SORRY" story 
revolting computational linguists 
scripted NLU and its dependencies 
True AI Story: 8.2 Try it again SAM 
the conceptual dependency notation 
a Swale of a tale 
True AI Story: 8.3 Another SWALE of a tale 
giving semantics preferential treatment
bidirectional NLP
pragmatics? 
machine translation (MT)
natural language interfaces (NLI)
networks for NLP 

9 Learning To Do it Right 
can we have intelligence wit�1out learning?
can we have AI without learning?
learning paradigms in AI
learning as the accretion of symbolic structures
learning as the adjustment of link weights
external tutoring: learning by being told
learning on the path
learning in a bath - taking the plunge
climbing hills because the 're there
rote learning: if it might be useful, store it
learning generalities
induction
overgeneralization and refinement 
a first guess and generalization 
True Al story: 9.1. Underneath the arches:
an everyday story of concept learning
competitive learning 
learning particularities: removal of unwanted
generalization 
EBG, or is it EBL? 
the EBL viewpoint 
mechanized creativity
learning by introspection
rediscovering things
learning by analogy 
learning at the knowledge level 
soaring through search spaces 
the more you know the slower you go
on finding needles in haystacks
when to learn and what to learn
giving credit where it is due
unlearning

10 Foundations of AI: Can we find any?
foundations: why dig for them?
formal foundations
a disinterested user's guide to the FOPC
the curse of nonmonotonicity
logical odds and ends 
True AI Story: 10 .1 It is not a closed world after all
the epilogic
methodological foundations 
the roles of programs in AI 
programs as theories
programs as experiments
rational reconstructions in AI 
sorting out AI methodologies
philosophical foundations
there's nothing special about you, or me
building the foundations on the CP
undermining the foundations of the CP
total disbelief: let's not be Searle-ish

11 Prognostications, or W(h)ither AI? 
abstract AI and concrete AI 
is the mind an appropriate object for scientific study? 
True AI Story: 11.1. Sand in the works
AI as a magnifying glass 
AI: can it be practically useful?
AI: just wait till we get into parallel hardware
last words
References 
Author Index 
Subject Index