Mind as Machine: A History of Cognitive Science

Knowledge Graph: Mind as Machine: A History of Cognitive Science (Margaret A. Boden, 2006)

Editorial spotlight: ↑ computational functionalism — the paradigm shift

Concepts

• Boden's computational functionalism (importance 5): The central thesis: mind can be understood as information processing independent of its physical substrate. This is the unifying idea across all of cognitive science's history.. Source: (from training memory of book).
• Craik's mental models (1943) (importance 5): Kenneth Craik's The Nature of Explanation proposed that the mind constructs small-scale models of reality to predict events. First computational theory of mind.. Source: (from training memory of book).
• Wiener's cybernetics (importance 5): Norbert Wiener's framework treating control and communication in animals and machines as unified phenomena through feedback loops.. Source: (from training memory of book).
• Boden's cognitive revolution (1960s) (importance 5): The paradigm shift from behaviorism to information-processing models of mind. Enabled by computer metaphor and formal theories.. Source: (from training memory of book).
• GOFAI (Good Old-Fashioned AI) (importance 4): Boden's term for classical symbolic AI based on explicit rules and logical reasoning. Dominant paradigm 1956-1980s.. Source: (from training memory of book).
• Chomsky's transformational grammar (importance 4): Noam Chomsky's theory that language has deep/surface structure and innate universal grammar. Revolutionized linguistics and challenged behaviorism.. Source: (from training memory of book).
• PDP (parallel distributed processing) (importance 4): Rumelhart-McClelland's 1986 revival of connectionism with distributed representations and backpropagation. Challenged symbolic AI dominance.. Source: (from training memory of book).
• Marr's three levels of analysis (importance 4): David Marr's framework: computational (what/why), algorithmic (how), implementational (physical substrate). Influential methodology for cognitive science.. Source: (from training memory of book).
• embodied/situated cognition (importance 4): View that cognition is fundamentally shaped by bodily interaction with environment, not just abstract symbol manipulation. Challenge to classical AI.. Source: (from training memory of book).
• machine learning paradigm shift (importance 4): Boden's account of the shift from hand-coded knowledge to learning from data. Enabled by statistics and big data.. Source: (from training memory of book).
• Boden's computational creativity (importance 4): Margaret Boden's analysis of creativity as exploration/transformation of conceptual spaces. Can be computational without being routine.. Source: (from training memory of book).
• unified cognitive architectures (importance 4): Boden's theme: comprehensive models of mind (SOAR, ACT-R, etc.) attempt to unify symbolic/subsymbolic, declarative/procedural.. Source: (from training memory of book).
• Skinnerian behaviorism (importance 3): B.F. Skinner's view that behavior is shaped by environmental reinforcement without reference to internal mental states. Dominant in psychology pre-1960.. Source: (from training memory of book).
• Schank's scripts & Minsky's frames (importance 3): Structured knowledge representations for stereotypical situations and objects. Enabled natural language understanding systems.. Source: (from training memory of book).
• Dennett's intentional stance (importance 3): Daniel Dennett's view that attributing beliefs/desires is a useful predictive strategy, regardless of internal mechanism.. Source: (from training memory of book).
• Gibson's affordances (importance 3): J.J. Gibson's ecological psychology: environment offers action possibilities directly perceived without internal representations.. Source: (from training memory of book).
• Piaget's developmental stages (importance 3): Jean Piaget's theory of cognitive development through sensorimotor, preoperational, concrete operational, formal operational stages.. Source: (from training memory of book).
• evolutionary psychology (Tooby-Cosmides) (importance 3): View that mind contains specialized adaptations shaped by natural selection for ancestral problems. Extreme modularity position.. Source: (from training memory of book).
• Langton's artificial life (importance 3): Christopher Langton's study of life-as-it-could-be through simulation. Emphasized emergence and self-organization over top-down design.. Source: (from training memory of book).
• self-organization (Ashby) (importance 3): W. Ross Ashby's principle that complex order can emerge spontaneously from local interactions without central control.. Source: (from training memory of book).
• dynamical systems approach (van Gelder) (importance 3): Tim van Gelder's view that cognition is better modeled as continuous dynamical trajectories than discrete symbol manipulation.. Source: (from training memory of book).
• computational neuroscience (Sejnowski) (importance 3): Terrence Sejnowski's field bridging neuroscience and computation through mathematical models of neural circuits.. Source: (from training memory of book).
• statistical NLP revolution (importance 3): Shift from hand-coded linguistic rules to corpus-based statistical methods in 1990s. Enabled by large text corpora.. Source: (from training memory of book).
• McCarthy's common-sense problem (importance 3): John McCarthy's observation that AI systems lack everyday knowledge humans take for granted. Persistent challenge for AI.. Source: (from training memory of book).
• McCarthy's frame problem (importance 3): Problem of representing what doesn't change when actions occur. Fundamental issue in knowledge representation.. Source: (from training memory of book).
• Harnad's symbol grounding problem (importance 3): Stevan Harnad's question: how do symbols get their meaning if they only refer to other symbols? Requires sensorimotor connection.. Source: (from training memory of book).
• Hinton's distributed representations (importance 3): Geoffrey Hinton's idea that concepts are represented by patterns across many units, not localist grandmother cells.. Source: (from training memory of book).
• hybrid symbolic-connectionist architectures (importance 3): Boden's synthesis position: combine symbolic reasoning for some tasks with connectionist learning for others.. Source: (from training memory of book).
• Boden's conceptual spaces (importance 3): Structured possibility spaces defined by generative rules. Creativity involves exploring or transforming these spaces.. Source: (from training memory of book).
• Simon's bounded rationality (importance 3): Herbert Simon's concept that humans satisfice rather than optimize due to cognitive limitations. Realistic model of decision-making.. Source: (from training memory of book).
• Minsky's Society of Mind (importance 3): Marvin Minsky's 1985 theory that intelligence emerges from interactions of simple agents. No central controller.. Source: (from training memory of book).
• Gestalt principles (Wertheimer) (importance 2): Pre-computational theory that perception organizes wholes from parts via principles like proximity, similarity, continuity.. Source: (from training memory of book).
• Vygotsky's zone of proximal development (importance 2): Lev Vygotsky's concept of learning in social context through scaffolding within ZPD. Alternative to Piaget's individualist approach.. Source: (from training memory of book).
• swarm intelligence (Bonabeau) (importance 2): Collective behavior of decentralized systems like ant colonies. Intelligence emerges from simple local interactions.. Source: (from training memory of book).
• Maturana-Varela autopoiesis (importance 2): Theory that living systems are self-producing networks that maintain their organization. Biological alternative to functionalism.. Source: (from training memory of book).
• knowledge acquisition bottleneck (importance 2): Problem that extracting expert knowledge for systems is time-consuming and difficult. Limited expert systems' scalability.. Source: (from training memory of book).
• Edelman's neural Darwinism (importance 2): Gerald Edelman's theory that brain development involves selection among neural groups. Evolution-inspired neuroscience.. Source: (from training memory of book).
• Baars's global workspace theory (importance 2): Bernard Baars's model where consciousness is broadcast of information to multiple specialized processors. Computational theory of consciousness.. Source: (from training memory of book).
• qualification problem (McCarthy) (importance 2): Impossibility of specifying all preconditions for an action's success. Related to brittleness of symbolic AI.. Source: (from training memory of book).
• PDP microfeatures (importance 2): Subsymbolic features below the level of conscious concepts. Connectionist alternative to symbolic primitives.. Source: (from training memory of book).
• Smolensky's subsymbolic paradigm (importance 2): Paul Smolensky's attempt to reconcile symbolic and connectionist approaches through tensor product representations.. Source: (from training memory of book).
• H-creativity vs P-creativity (Boden) (importance 2): Historical creativity (new to humanity) vs. psychological creativity (new to individual). Most creativity is P-creative.. Source: (from training memory of book).
• emotion-cognition integration (Damasio) (importance 2): Antonio Damasio's somatic marker hypothesis: emotions are essential for rational decision-making, not opposed to reason.. Source: (from training memory of book).
• Picard's affective computing (importance 2): Rosalind Picard's field of computing systems that recognize, model, and express emotions.. Source: (from training memory of book).
• Suchman's situated action (importance 2): Lucy Suchman's critique that plans are post-hoc rationalizations; action emerges from situation. Challenge to classical AI planning.. Source: (from training memory of book).
• computational analogy (Gentner) (importance 2): Dedre Gentner's structure-mapping theory: analogy involves aligning relational structure. Computational model of reasoning.. Source: (from training memory of book).
• Kosslyn's mental imagery debate (importance 2): Stephen Kosslyn's pictorial theory vs. Pylyshyn's propositional theory of mental images. Computational models of internal representation.. Source: (from training memory of book).
• dual-process theories (Kahneman) (importance 2): System 1 (fast, automatic, intuitive) vs. System 2 (slow, deliberate, rational). Two modes of thinking.. Source: (from training memory of book).
• Schank's story understanding (importance 2): Roger Schank's work on scripts, goals, plans for narrative comprehension. AI needs causal reasoning about human actions.. Source: (from training memory of book).

Claims

• Newell-Simon physical symbol system hypothesis (importance 5): A physical symbol system has the necessary and sufficient means for general intelligent action. Core claim of classical cognitive science.. Source: (from training memory of book).
• Minsky-Papert XOR critique (1969) (importance 4): Perceptrons cannot solve XOR and other linearly non-separable problems. Triggered the first AI winter for connectionism.. Source: (from training memory of book).
• Chomsky's poverty of stimulus argument (importance 4): Children acquire language despite insufficient input, implying innate linguistic knowledge. Key argument against behaviorist learning theory.. Source: (from training memory of book).
• Searle's Chinese Room argument (importance 4): John Searle's thought experiment claiming that symbol manipulation cannot produce genuine understanding. Major challenge to strong AI.. Source: (from training memory of book).
• Fodor's modularity thesis (importance 3): Jerry Fodor argued for domain-specific encapsulated modules in perception and language. Influential in nativist cognitive architecture debates.. Source: (from training memory of book).
• Chalmers's hard problem of consciousness (importance 3): David Chalmers's argument that explaining subjective experience is fundamentally different from explaining cognitive functions.. Source: (from training memory of book).
• Rumelhart-McClelland past-tense debate (importance 3): 1986 paper showing neural network could learn English past tense without explicit rules. Sparked symbolic vs. connectionist wars.. Source: (from training memory of book).
• Pinker's words-and-rules theory (importance 2): Steven Pinker argued for dual system: associative memory for irregulars, symbolic rules for regulars. Defended hybrid architecture.. Source: (from training memory of book).
• Moravec's paradox (importance 2): Hans Moravec observed that hard problems for humans (logic) are easy for computers, while easy problems (perception) are hard.. Source: (from training memory of book).
• Churchland's eliminative materialism (importance 2): Paul Churchland's claim that folk psychology will be eliminated by neuroscience. Radical connectionist position.. Source: (from training memory of book).

Empirical results

• Miller's 7±2 (1956) (importance 3): George Miller showed working memory capacity is limited to ~7 chunks. Early empirical result supporting information-processing view.. Source: (from training memory of book).
• Tversky-Kahneman heuristics and biases (importance 3): Amos Tversky and Daniel Kahneman showed systematic deviations from rationality. Challenged ideal rationality assumptions.. Source: (from training memory of book).
• Weiskrantz's blindsight studies (importance 2): Larry Weiskrantz showed patients with V1 damage can respond to unseen stimuli. Evidence for unconscious visual processing.. Source: (from training memory of book).
• connectionist graceful degradation (importance 2): Neural networks degrade gradually with damage/noise rather than catastrophically. Biological realism advantage over symbolic AI.. Source: (from training memory of book).
• catastrophic interference problem (importance 2): Neural networks forget old knowledge when learning new tasks. Major limitation of standard connectionist learning.. Source: (from training memory of book).
• AI winters (1974, 1987) (importance 2): Periods of reduced funding and interest after overpromising. First after Lighthill Report, second after expert systems collapse.. Source: (from training memory of book).
• Shepard-Metzler mental rotation (1971) (importance 2): Roger Shepard showed rotation time increases linearly with angle. Evidence for analog mental representations.. Source: (from training memory of book).

Methods

• McCulloch-Pitts neural networks (1943) (importance 4): First formal model of artificial neurons as logical threshold units. Showed neural nets could compute any logical function.. Source: (from training memory of book).
• Hebbian learning rule (importance 4): Donald Hebb's 1949 principle: neurons that fire together wire together. Foundation of connectionist learning theory.. Source: (from training memory of book).
• backpropagation algorithm (importance 4): Learning algorithm for multi-layer networks by propagating errors backward. Solved the XOR problem and revived neural networks.. Source: (from training memory of book).
• Broadbent's filter theory (1958) (importance 3): Donald Broadbent's model of selective attention as early filtering of sensory input. First flow-diagram model of cognitive process.. Source: (from training memory of book).
• semantic networks (Quillian) (importance 3): Graph-based knowledge representation with concepts as nodes and relations as edges. Foundation of knowledge representation in AI.. Source: (from training memory of book).
• Newell's production systems (importance 3): If-then rule architectures for modeling cognition. Led to ACT-R and other cognitive architectures.. Source: (from training memory of book).
• Marr's computational vision theory (importance 3): Multi-stage theory of vision from primal sketch to 2.5D sketch to 3D model. Exemplar of computational approach to perception.. Source: (from training memory of book).
• Brooks's subsumption architecture (importance 3): Rodney Brooks's layered reactive control for robots without central representations. 'Intelligence without representation' slogan.. Source: (from training memory of book).
• Holland's genetic algorithms (importance 3): John Holland's evolutionary search method using selection, crossover, mutation. Applied evolution to problem-solving.. Source: (from training memory of book).
• expert systems (Feigenbaum) (importance 3): Edward Feigenbaum's knowledge-based systems encoding human expertise in if-then rules. Commercial AI success in 1980s.. Source: (from training memory of book).
• Sutton-Barto reinforcement learning (importance 3): Learning through trial-and-error interaction with environment using reward signals. Connects AI to behavioral psychology.. Source: (from training memory of book).
• Hopfield networks (1982) (importance 3): John Hopfield's recurrent networks with energy functions. Showed neural networks could store memories as attractors.. Source: (from training memory of book).
• Turing test (1950) (importance 3): Alan Turing's imitation game: can a machine fool a human interrogator? Operational definition of intelligence.. Source: (from training memory of book).
• von Neumann's cellular automata (importance 2): John von Neumann's self-replicating automata on discrete grids. Foundation for studying emergence and complexity.. Source: (from training memory of book).
• Holland's classifier systems (importance 2): Rule-based learning systems combining production systems with genetic algorithms. Hybrid symbolic-evolutionary approach.. Source: (from training memory of book).
• Quinlan's decision trees (ID3) (importance 2): Ross Quinlan's algorithm for learning classification rules from data using information gain. Practical ML success.. Source: (from training memory of book).
• support vector machines (Vapnik) (importance 2): Vladimir Vapnik's margin-maximizing classifiers using kernel trick. Dominant ML method in 1990s-2000s.. Source: (from training memory of book).
• temporal difference learning (importance 2): Richard Sutton's method for learning predictions by bootstrapping from successive estimates. Core RL algorithm.. Source: (from training memory of book).
• fMRI brain imaging (importance 2): Functional magnetic resonance imaging measures brain activity via blood flow. Enabled cognitive neuroscience's rise in 1990s.. Source: (from training memory of book).
• Hodgkin-Huxley neuron model (1952) (importance 2): Mathematical model of action potential generation through ion channels. Foundation of computational neuroscience.. Source: (from training memory of book).
• Boltzmann machines (Hinton) (importance 2): Geoffrey Hinton's stochastic networks using simulated annealing. Learned internal representations through energy minimization.. Source: (from training memory of book).
• constraint satisfaction networks (importance 2): Parallel relaxation networks that settle into solutions satisfying multiple constraints. Applied to perception and reasoning.. Source: (from training memory of book).
• McClelland-Rumelhart interactive activation (importance 2): Model of word recognition with bidirectional activation between letter and word layers. Showed top-down effects in perception.. Source: (from training memory of book).
• STRIPS planner (Fikes-Nilsson) (importance 2): Stanford Research Institute Problem Solver: classical AI planning using preconditions and effects. Foundation of AI planning.. Source: (from training memory of book).
• Selfridge's Pandemonium (1959) (importance 2): Oliver Selfridge's hierarchical pattern recognition with competing 'demons'. Early parallel processing model.. Source: (from training memory of book).
• blackboard architectures (Hearsay) (importance 2): Shared workspace where independent knowledge sources cooperate on problem-solving. Used in speech recognition.. Source: (from training memory of book).
• Schank's conceptual dependency (importance 2): Roger Schank's meaning representation using primitive actions and states. Canonical form for sentences.. Source: (from training memory of book).

Entities

• Turing's universal machine (1936) (importance 5): Alan Turing's abstract model of computation that processes discrete symbols according to rules. Foundation for treating mind as symbol manipulation.. Source: (from training memory of book).
• Dartmouth Summer Research Project (1956) (importance 4): McCarthy, Minsky, Rochester, Shannon convened the founding conference of AI. Coined 'artificial intelligence' and launched the field.. Source: (from training memory of book).
• Rosenblatt's Perceptron (1958) (importance 3): Frank Rosenblatt's learning machine that could recognize patterns. First implemented neural network with learning capability.. Source: (from training memory of book).
• General Problem Solver (Newell-Simon) (importance 3): First AI program to separate problem-solving method from domain knowledge using means-ends analysis. Landmark in cognitive modeling.. Source: (from training memory of book).
• Logic Theorist (1956) (importance 3): Newell, Shaw, Simon's program that proved theorems from Principia Mathematica. First reasoning program, presented at Dartmouth.. Source: (from training memory of book).
• McCarthy's LISP (1958) (importance 3): First language designed for AI, featuring symbolic list processing and recursion. Became dominant AI programming language for decades.. Source: (from training memory of book).
• Anderson's ACT-R (importance 3): Adaptive Control of Thought-Rational: comprehensive cognitive architecture combining symbolic and subsymbolic processing.. Source: (from training memory of book).
• Laird-Newell-Rosenbloom SOAR (importance 3): Unified cognitive architecture based on problem spaces and chunking. Aimed to model all aspects of intelligence.. Source: (from training memory of book).
• Conway's Game of Life (1970) (importance 2): Simple cellular automaton showing complex emergent patterns from simple rules. Popularized emergence and complexity science.. Source: (from training memory of book).
• MYCIN (Shortliffe 1974) (importance 2): Medical diagnosis expert system using ~600 rules and certainty factors. Pioneering expert system achieving specialist-level performance.. Source: (from training memory of book).
• Lenat's Cyc project (importance 2): Doug Lenat's massive common-sense knowledge base project started 1984. Attempted to hand-code all human knowledge.. Source: (from training memory of book).
• Lighthill Report (1973) (importance 2): James Lighthill's UK government report criticizing AI's lack of progress. Triggered first AI winter.. Source: (from training memory of book).
• Winograd's SHRDLU (1971) (importance 2): Terry Winograd's natural language system for manipulating blocks world. Showed language understanding requires world knowledge.. Source: (from training memory of book).
• Weizenbaum's ELIZA (1966) (importance 2): Joseph Weizenbaum's chatbot simulating Rogerian therapist using pattern matching. Raised questions about AI and understanding.. Source: (from training memory of book).
• Lenat's AM program (importance 2): Automated Mathematician discovered number theory concepts through heuristic search. Early computational creativity work.. Source: (from training memory of book).
• Hofstadter's Copycat (importance 2): Douglas Hofstadter's analogy-making program using parallel terraced scan. Models fluid concepts and perception.. Source: (from training memory of book).
• Loebner Prize (importance 1): Annual Turing test competition started 1991. No program has passed unrestricted version.. Source: (from training memory of book).
• BACON discovery system (importance 1): Pat Langley's program rediscovering scientific laws from data. Computational model of scientific discovery.. Source: (from training memory of book).

Relations

• Turing's universal machine (1936) enables Boden's computational functionalism
• Craik's mental models (1943) enables Boden's computational functionalism
• Wiener's cybernetics enables Boden's computational functionalism
• McCulloch-Pitts neural networks (1943) supports Boden's computational functionalism
• Boden's computational functionalism enables Boden's cognitive revolution (1960s)
• Dartmouth Summer Research Project (1956) enables GOFAI (Good Old-Fashioned AI)
• Logic Theorist (1956) exemplifies GOFAI (Good Old-Fashioned AI)
• General Problem Solver (Newell-Simon) exemplifies GOFAI (Good Old-Fashioned AI)
• Newell-Simon physical symbol system hypothesis supports GOFAI (Good Old-Fashioned AI)
• GOFAI (Good Old-Fashioned AI) exemplifies Boden's computational functionalism
• McCarthy's LISP (1958) enables GOFAI (Good Old-Fashioned AI)
• McCulloch-Pitts neural networks (1943) enables Rosenblatt's Perceptron (1958)
• Hebbian learning rule enables Rosenblatt's Perceptron (1958)
• Rosenblatt's Perceptron (1958) motivates Minsky-Papert XOR critique (1969)
• Minsky-Papert XOR critique (1969) enables AI winters (1974, 1987)
• backpropagation algorithm refutes Minsky-Papert XOR critique (1969)
• backpropagation algorithm enables PDP (parallel distributed processing)
• PDP (parallel distributed processing) supports Boden's computational functionalism
• Chomsky's transformational grammar enables Boden's cognitive revolution (1960s)
• Chomsky's poverty of stimulus argument refutes Skinnerian behaviorism
• Chomsky's poverty of stimulus argument supports Chomsky's transformational grammar
• Skinnerian behaviorism contradicts Boden's cognitive revolution (1960s)
• Miller's 7±2 (1956) evidences Boden's cognitive revolution (1960s)
• Broadbent's filter theory (1958) evidences Boden's cognitive revolution (1960s)
• Schank's scripts & Minsky's frames exemplifies GOFAI (Good Old-Fashioned AI)
• semantic networks (Quillian) enables Schank's scripts & Minsky's frames
• Newell's production systems exemplifies GOFAI (Good Old-Fashioned AI)
• Newell's production systems enables Anderson's ACT-R
• Newell's production systems enables Laird-Newell-Rosenbloom SOAR
• Anderson's ACT-R exemplifies unified cognitive architectures
• Laird-Newell-Rosenbloom SOAR exemplifies unified cognitive architectures
• Searle's Chinese Room argument contradicts Boden's computational functionalism
• Dennett's intentional stance contradicts Searle's Chinese Room argument
• Marr's three levels of analysis supports Boden's computational functionalism
• Marr's computational vision theory exemplifies Marr's three levels of analysis
• Gestalt principles (Wertheimer) precedes Marr's computational vision theory
• Gibson's affordances enables embodied/situated cognition
• embodied/situated cognition contradicts Boden's computational functionalism
• Brooks's subsumption architecture exemplifies embodied/situated cognition
• Brooks's subsumption architecture contradicts GOFAI (Good Old-Fashioned AI)
• Piaget's developmental stages supports Boden's cognitive revolution (1960s)
• Vygotsky's zone of proximal development contradicts Piaget's developmental stages
• Fodor's modularity thesis supports Boden's computational functionalism
• evolutionary psychology (Tooby-Cosmides) builds-on Fodor's modularity thesis
• Langton's artificial life contradicts GOFAI (Good Old-Fashioned AI)
• von Neumann's cellular automata enables Langton's artificial life
• Conway's Game of Life (1970) exemplifies Langton's artificial life
• Holland's genetic algorithms exemplifies Langton's artificial life
• Holland's classifier systems builds-on Holland's genetic algorithms
• swarm intelligence (Bonabeau) exemplifies Langton's artificial life
• self-organization (Ashby) supports Langton's artificial life
• Maturana-Varela autopoiesis contradicts Boden's computational functionalism
• dynamical systems approach (van Gelder) contradicts Boden's computational functionalism
• expert systems (Feigenbaum) exemplifies GOFAI (Good Old-Fashioned AI)
• MYCIN (Shortliffe 1974) exemplifies expert systems (Feigenbaum)
• Lenat's Cyc project exemplifies expert systems (Feigenbaum)
• knowledge acquisition bottleneck contradicts expert systems (Feigenbaum)
• knowledge acquisition bottleneck motivates machine learning paradigm shift
• machine learning paradigm shift contradicts GOFAI (Good Old-Fashioned AI)
• Quinlan's decision trees (ID3) exemplifies machine learning paradigm shift
• support vector machines (Vapnik) exemplifies machine learning paradigm shift
• Sutton-Barto reinforcement learning exemplifies machine learning paradigm shift
• temporal difference learning exemplifies Sutton-Barto reinforcement learning
• Edelman's neural Darwinism supports Langton's artificial life
• Chalmers's hard problem of consciousness contradicts Boden's computational functionalism
• Baars's global workspace theory supports Chalmers's hard problem of consciousness
• Weiskrantz's blindsight studies evidences Baars's global workspace theory
• fMRI brain imaging enables computational neuroscience (Sejnowski)
• computational neuroscience (Sejnowski) supports Boden's computational functionalism
• Hodgkin-Huxley neuron model (1952) enables computational neuroscience (Sejnowski)
• Hopfield networks (1982) enables PDP (parallel distributed processing)
• Boltzmann machines (Hinton) builds-on Hopfield networks (1982)
• constraint satisfaction networks exemplifies PDP (parallel distributed processing)
• McClelland-Rumelhart interactive activation exemplifies PDP (parallel distributed processing)
• Rumelhart-McClelland past-tense debate evidences PDP (parallel distributed processing)
• Pinker's words-and-rules theory contradicts Rumelhart-McClelland past-tense debate
• statistical NLP revolution exemplifies machine learning paradigm shift
• statistical NLP revolution contradicts GOFAI (Good Old-Fashioned AI)
• Moravec's paradox contradicts GOFAI (Good Old-Fashioned AI)
• McCarthy's common-sense problem contradicts GOFAI (Good Old-Fashioned AI)
• McCarthy's frame problem contradicts GOFAI (Good Old-Fashioned AI)
• qualification problem (McCarthy) builds-on McCarthy's frame problem
• Harnad's symbol grounding problem contradicts GOFAI (Good Old-Fashioned AI)
• Harnad's symbol grounding problem supports embodied/situated cognition
• Hinton's distributed representations supports PDP (parallel distributed processing)
• PDP microfeatures exemplifies Hinton's distributed representations
• connectionist graceful degradation evidences PDP (parallel distributed processing)
• catastrophic interference problem contradicts PDP (parallel distributed processing)
• Churchland's eliminative materialism supports PDP (parallel distributed processing)
• Smolensky's subsymbolic paradigm supports hybrid symbolic-connectionist architectures
• hybrid symbolic-connectionist architectures supports unified cognitive architectures
• Boden's computational creativity supports Boden's computational functionalism
• Boden's conceptual spaces enables Boden's computational creativity
• H-creativity vs P-creativity (Boden) exemplifies Boden's computational creativity
• emotion-cognition integration (Damasio) supports Boden's computational functionalism
• Picard's affective computing builds-on emotion-cognition integration (Damasio)
• Turing test (1950) supports Boden's computational functionalism
• Turing's universal machine (1936) precedes Turing test (1950)
• Lighthill Report (1973) enables AI winters (1974, 1987)
• Winograd's SHRDLU (1971) exemplifies GOFAI (Good Old-Fashioned AI)
• Weizenbaum's ELIZA (1966) exemplifies Turing test (1950)
• Loebner Prize exemplifies Turing test (1950)
• Suchman's situated action contradicts GOFAI (Good Old-Fashioned AI)
• Suchman's situated action supports embodied/situated cognition
• STRIPS planner (Fikes-Nilsson) exemplifies GOFAI (Good Old-Fashioned AI)
• Lenat's AM program exemplifies Boden's computational creativity
• BACON discovery system exemplifies Boden's computational creativity
• computational analogy (Gentner) exemplifies GOFAI (Good Old-Fashioned AI)
• Hofstadter's Copycat exemplifies computational analogy (Gentner)
• Kosslyn's mental imagery debate supports Boden's computational functionalism
• Shepard-Metzler mental rotation (1971) evidences Kosslyn's mental imagery debate
• dual-process theories (Kahneman) supports unified cognitive architectures
• Simon's bounded rationality supports Boden's cognitive revolution (1960s)
• Tversky-Kahneman heuristics and biases evidences Simon's bounded rationality
• Minsky's Society of Mind exemplifies unified cognitive architectures
• Selfridge's Pandemonium (1959) precedes Minsky's Society of Mind
• blackboard architectures (Hearsay) exemplifies unified cognitive architectures
• Schank's story understanding builds-on Schank's scripts & Minsky's frames
• Schank's conceptual dependency enables Schank's story understanding
• PDP (parallel distributed processing) supports Boden's cognitive revolution (1960s)
• machine learning paradigm shift supports Boden's cognitive revolution (1960s)
• computational neuroscience (Sejnowski) supports Boden's cognitive revolution (1960s)
• embodied/situated cognition contradicts Boden's cognitive revolution (1960s)
• unified cognitive architectures exemplifies Boden's computational functionalism
• Boden's computational creativity supports Boden's cognitive revolution (1960s)

© 2026 LOOMUS. Curation, importance scoring, edge typing, and visualization. Source paper © respective authors, quoted under fair use.