1/43
Looks like no tags are added yet.
austro-german classical music (classical paradigm)
musical tradition centred on written scores and formal rules, used as a metaphor for classical cognitive science where mind is rule-governed system executing central instucitions
ex. orchestra strictly following written score under direction of a conductor
reprents the classical sese-think-act framework, where behaviour is guided by internal metnal representations (scores) and cnetralized control
msucial romanticism (connectionist paradigm)
musical movement focusing on emotion and subtle regularities that resist rigid symbol rules, used as a metaphor for connectionist neural networks that learn statisticl patterns from experience
ex. a musician improvising or playing with expressive timing that cannot be captured by rigid mathematical sheet music
illustrates connectionisms strength in modelling biological, pattern0based processing htat goes beyond hand-coded symbolic rules
modern/minimalist music (embodied paradigm)
a msucial movement built on simple repetive loops shifting out of phase without central planning, used as a metahor for embodied congition where complex behaviour emeges without central cotnrol
ex.steve reich’s its gonna rain, where complex musical textures emege from simple acoustic interactions
demosntrates how organized, complexpatterns emerge naturally from physical interactions rather than a master control program
conduit metaphor
Micahael Reddys’ concept that communication works like a pipeline: sender packs internal thoughts into a container (words or notes) and transmits it to a passive reciver who unpacks it
ex. viewing a musical score or speech purely as a package of symbols sent frrom composers head to a listeners brain
serves as primary taget of critique for embodied theorists who reject the idea that cognition is just passive decoding of internal symbols
corporal (embodied) music cognition
view that musical understanding is not a disembodied mental decoding of symbols, but a physical engagement involving body movement, acoustic resonance, and action
ex. a listener understanding a rhythym by tapping their foot, dancing, or feeling physical acoustic vibrations in a live concert hall
replaces passive, disembodied representational theories with direct perception-action coupling and physical engagement with environment
generative grammar of music (GTTM)
rule-based model applying Chomsky’s lingusitic principles to explain an expereinced listener’s unconsious musical intutiion through formal rules
ex. listener automatically grouping individual notes into muscial phrases and identifying metric beats based on hierarchial rules
proves that classical, rule-based mental representations apply effectvley to non-verbal perceptual domains like music
tonal hierarchy (probe-tone method)
Krumhansl’s finding that human liteners organize musical tone into a mental spatial hierarchy of stability relative to a key signature→ listeners rate how well a probe note fits a musical context, revealing internal mental space where pithes are organized hierarchially by stability relative to a key
in C major, listeners judge the tonic note (C) as most stable, followed by E, G, sclae nots, and lastly non-scale chromatic notes
demosntrates how the mind constructs internal representational spaces to organize sensory experience
musical connectionism
the application of artificial neural networks to learn, categorize, and generate musical structures directly form audio exposure without hand-coded rules, includes fillin in mising sensory inputs (pattern compeltion)
ex. multillayer perceptron discovering hidden representations of chord relationships, just by training on raw pitch vectors
shows that neural networks can discover subtle, ineffable statistical regularities (like style or mood) that elude formal symbolic AI. ANNS naturally exploit context and cooperative unit interactions to handle noisy, incomplete real-world inputs without symbolic logic
musical stigmergy
decentralized musical cooridnation where a musciians action alters the physical/acoustic environment, and that environmental change triggers thenext action by other performers or computers
ex. perfrms in terry riley’s in C deciding when to move to the next musical motif based on what they hear surrounding them in the room
demonstrates how complex social and cognitive coordiantion can occur though environmental cues withotu central conductor or master bleuprint
digital msucial instruments (DMIs)/ behavioral objects
Interactive hardware/software interfaces that decouple physical control from sound generation, acting as active tools that scaffold and extend the musician's cognitive abilities.
ex. The reacTable (a tangible tabletop synthesizer) or Swarm Music software that interacts in real time with a performer's physical gestures.
provides concerete empircial eviedence for cognitive scaffolding and the extended mind hypothesis, showing cognition extending beyond physical brain
hybrid cognitive architectures
Systems that combine classical symbolic rules, connectionist pattern learning, and embodied environmental actions into a single integrated framework.
ex. NetNeg, a music composition system using neural networks for fluid melody generation alongside rule-based agents enforcing strict counterpoint rules.
proves that the 3 schools of cognitive science are not mutually exclusie paradigms, but complementary levels of analysis that can be unified
hybrid theories of cognition (synthesis)
integrated cognitive architectures that combine classical symbolic rules, connectionist pattern learning, and embodied environmental actions into a unified system
ex. NetNeg, a music composition system using neural networks to propose fluid melodies alongside rule-based agents enforcing counterpoint rules.
proves that the 3 schools of cogntive science are not mutually exclusive or irreconciliable paradigms, but complementary perpectives that can be synthesized to provide complete explanations of mind
crisis of control
emerged when external contorl methods like punched tape were too slow to take advantgae of the speed of new electornic computers
early solutions, like plug boards, took advantage of speed, but were time consuming to chnage for another task. crisis led to the invention of the stored program computer
important because it revealed the classical notion of internal, central control ist necessary, but merely arose to solve a practical computer problem
distal stimulus
the actual 3D physical object, event, or scene existing in the external physical world
ex. real wooden cube sitting on desk in fornt of u
represents real world physical object that a visual system must accurately perceive and interpret
proximal stimulus
the 2D pattern of energy (such as light projeected onto the reina) that directy hits an organisms sensory receptors
the flat 2D image formed by light rays striking the retin at the back of ur eye
serves as the raw sensory input for cision, demonstrtaing that the brain doesnt have direct access to 3D objects, but only to 2D projections
perceptual underdetermination (poverty of the stimulus)
the mathematical reality that a single 2D proximal stimulus on the retina can be produced by an infiinte number of different 3D distal scenes
a flat line drawn on a retina could be a small near stick, a large far pole, or a tilted wire- 2D signal alone cannot tell u which one it is
proves that vision is an inverse problem- the visual system cannot passively record the world, but most actively use computational rules to construct a single 3D scene
unconsious inference
hermann von helmholtz concept that the visual ssytem automatically and unconisously calculates the most likely 3D scenbased on prior expereience and sensor evidence
ex. seeing a square shadow on a wall and instantly inferring a box is nearby, without conioiusly calcualting the angles
established early in that perception is an active computational process of problem-solving rather than passive photographic recording
t
the new look
jerome briner’s classical framework proposing that perceptionis driven form top down by an individuals conisous beliefs, expectations, goals, and conceptual categories
ex. mistaking a coiled garden hose for a dangerous snake in the dark cause u were recently warned about snaes in the area
represents the extreme classical ‘seeing is thinking’ view, emphasizing how cognitive context shapes visual interpretation
top-down (theory-drive) vision
visual procesisng where high-level mental expectations and semantic knowledge directly guide and alter how low-level sensory information is interpreted
ex. reading mispelled words in a sentence effortlessly because ur knowledge of grammar and context fills in missing letters
highlights the roles of stored knowledge and internal representations in constructing perceptual experience
bottom up (data-drien) vision
visual processing that flows strictly from raw sensory signals up to higher level interpretation without relying on beliefs or cosnious expectations
ex. automatically detecting a sudden flash of light or sharp edge in ur peripheral vision regardless of what u were thinking about
demonstrates how specialized visual hardware can process environmental inputs rapidly and automatically
natural computation
david marr’s approach to vision, modeling how early visual systems solve inverse problems using physical mathematics rather than high-level cognitive reasning
ex.mathematical algorithims that derive 3D surface orientation directly from subtle gradients in light intenisty (shape form shading)
showed how complex pereptual tasks can be solved by specialized, low level computational modules operating prior to central thought
natural constriants
universal physical properties and enduring laws of the real world that are ‘wired’ into visual procesisng to resolve perceptual underdetermination
ex. the rigidity constraint→ assuming that moving visual points that maintain fixed relative distnaces belong to a single solid, rigid object
explains how early visual hardware can accurately interpret the 3D world bottom-up without needing top-down beliefs or expectations
the compromise view (tri-stage model of vision)
a hybrid framework that rsolves top-down vs. bottom-up debate by divind perception into 3 sequential funcitonal stages: visual detection, visual cognition, and siotropic conceptual reasoning
ex. detecting raw color edges (detection), spotlighting features into an object file (cognition), and recognizing it as ‘my car’ (isotropic reasoning)
reconclies historical clash between pure data-driven vision (Marr/gibson) and pure thoery-driven (helmholtz/bruner)
feature integration theory (FIT)
Treisman’s model proposing that basic visual features (color, orientation, shape) are extracted automatically in parallel, but require serial attention to bind together
ex. spotting a single red line among green lines instantly (parallel search), vs finding a red horizontal line among green horiozntal lines and red vertical lines (serial conjunction search)
provides a ofundational classical model of visual attention explaining search times and how feature-binding errors occur
attnetional spotlight
mehanism of feature integration theory that serially focuses visual attentionon a specific spatial location to bind co-occuring features into signal object representation
ex. scanning a crowded room person-by-person to locate a friend wearing a specific blue coat and red hat
explains how limited cognitive resources select and synthesize sensory details into coherent metnal representations
visual routines
ullman’s concept of serial, software-like attentional procedures built from basic operations (indexing, boundary tracing, area spreading) toc ompute spatial relationships
ex. tracing a line through a complex maze to determine if a dot is inside or outside a boundary
bridges low level feature extraction and high-level reasoning by demonstrating how visual system computes abstract spatial relations
FINST theory (visual indexing)
plyshyn’s mechanism where early vision preattentively attahces a small set of non-conceptual, referential tags directly to physical objects in the world
ex. holding 4 mental ‘tethers’ onto 4 specific moving targets in a crowd so u can keep track of them without describing their appearances
solves corresondence problem by directly grounding internal mental symbols to distal physical objects in real time without cosntnat memory updates
multiple obejct tracking (MOT)
an experimental task where subject must ivisually track multiple movig tagrgets among identical moving distractors over time
ex. wathcing four specific white circles move and bounce around unpredicatably among eight identical white circles on a screen
serves as key empirical evidence for FINST theory, proving htat human visual capacity is hard-limited to tracking approximately 4 items simultaneously
subitizing
the rapid, accurate, and effortless recognition of small qunatities (1-4 items) without counting them one by one
ex. instantly seeing 3 apples on a table and knowing there are 3 without goung ‘one, two, three”
demonstrates an architectural limit of visual indexing → subsitize up to 4 items because we only have 4 FINST tags available
motor anchors
preattnetive motor-command indices that tag physical locations in spatial or body coordinates to guide physical movements
ex.reaching ur arm directly towards a glass on ur desk while keeoping ur eyes fixed ton computer monitor
extends visual indexing into action, showing how perception and moto r execution are linked in embodied cognitive architectures
visuomotor coordination
the real-time mapping between visual spatial indices (FINSTs) and motor pointers (ANCHORs) to coordinate physical actions with visual targets
ex. a basketball player catching a pass and instantly hoothing wihtout stopping to visually measure distance
illusstrates the embodied connection between visual sensing and motor action, bypassing slow internal planning buffers
index projection hypothesis
pylshyn theory that mental imagery does not use a literal internal scree, but instead projects visual indices onto real physical structures in the surrounding environment
ex. imagining a tic tac toe game in ur head while looking at a blank wall and anchoring ur imagined moves to specific points on the wall
reconciles symbolic theories of imagery with embodied cognition by showing how mental representations use extenal spatial scaffolding
cognitive impenetrability of early vision
principle that early visual processing mechanisms operate automatically in fixed architecture and cannot be altered by consious beliefs, goals, or knowledge
ex. continuing to see muller-lyer arrow lines as unequal in length even after measuring them with a ruler and knowing they are identical
defines the hard boundary between modular visual hardware (cognitive impenetrable) and central thought processes (cognitvely penetrable)
pylshyn’s hybrid theory of visual cognition
an integrated framework that combines classical symbolic object representations, connectionist neural tracking networks, and embodied physical indexing into a single model of vision
neural networks track motion preattentively (connecitonist), FINSTs tag physical items (embodied), and object files assign concepts like ‘dog’ (clasiscal)
demonstrates that the 3 schools of cognitive science are not opposing paradigms, but complementary leveles of an integrated cognitive system
cognitive dialectic
philosphical framework proposing that theoretical progress in cognitiv science evolves dynamically through 3 sewuential stages: thesis, antithesis, and synthesis
ex. classical symbol manipulation (thesis) was challenged by neural networks and robotics (antithesis), driving the field toward unified hybrid architectures (synthesis)
explains theoretical conflcits between schools of thought are not signs of a permanently broken field, but necessary steps toward a complete, integrated science of mind
thesis (classical cognitive science)
the initial foundational paradigm of cognitive science proposing tha tcognition is rule-governed symbol manipulation occuring in an internal, centralized ‘pure cognitive system’
ex. turing machnes, production systems, or computer programs that manipulate discrete symbols using formula rules
established core baseline assumption that cognition is a form of information processing, serving as the historical starting point for all subsequent cognitive theories
antithesis (connectionist and embodied cognitive science)
the counter-movements that arose in direct reaction against classical cognitive science, rejecting disembodied, serial symbol processing in favor of alternativve processing mechanisms
ex. connectionism introdusicng parallel neural network pattern recognition, and embodied cognition intoduicing sense-act loops and behaviour based robots
exposed limitations of rigid symbolic AI by highlighting the importance of biological plausibility, learning from experience, and real time environmental interactions
synthesis (unified cognitive science)
the ultimate goal of the cognitive dialectic, merging the valid mechanisms and insights of the classical, connectionist, and embodied schools into a single, cohesive framework
ex. pylshyn’s hybrid theory of visual cognition (combining connectionist preattetnvie tracking, embodied spatial indexing, and classical object files)
prevnets the field from fracturing into isolated sub disciplines, proving that all 3 approaches investigate necessary, complementary parts of a complete cognitive system
pre-paradigmatic science
kuhns term for an early stage in a scientific discipline characterized by deep debates over methods, core definitions, and basic architectures without a single dominant consensus
ex. cognitive scientists debating wether the minds basic architecture is a serial CPU, a parallel neural network, or a physical sense- act loop
shows that current clashes between cognitive schools reflect the young age of the discipline rather than an irreconcilable scientific failure
fable of the six blind men and the elephant
metaphor for how cognitive science schools misinterpret their partial findings as the entire truth
ex. one blind man feeling the tusk calls the elephant a spear (classical focusing on internal planning); another touching the knee calls it a tree (embodied focusing on physical action)
demonstrates that each school explores a different, real aspect of mind, and that true progress requires integrating their partial observations into a single complete model
pure cognitive system
normans concept describing how early classical cognitive science isolated internal reasoning processes (‘pure cognition’) from perceptual input and motor output
ex. early production systems that modeled abstract problem-solving (like playing chess) while compleltely ignoring how eyes see the board or how hands move pieces
illustrates how classical AI became a ‘victim of its own success’ by over-focusing on central planning, making it necessary for embodied cognition to reconnect mind to body and world
natural computation
marr’s pragmatic approach to vision that derives mathematical solutions to perceptual problems by combining physical environmental laws with computational and biological constraints
ex. combing physical smoothness constraints (embodied world laws), neural network connectivity (connectionist hardware), and formal feature maps (classical representations) to compute motion
serves as a working model of cognitive synthesis, proving that vision research succeeded precisely because it abandoned idealogical loyalty to a single school
reverse engineering (analytical methodology)
top down research strategy of observing an intact, fully funcitoning cognitive agents behaviour and decomposing its internal software/hardware mechanisms step by step
ex.recording human verbal protocols during a puzzle task to deduce the internal production rules stored in human memory
the primary methodology of classical cognitive science, essential for validating wether artificail computational models aree strongly equivalent to human minds
forward engineering (synthetic psychology)
bottom up methodology where simple physical componetns are conneted into an artificial agent, placed in an environment, and observed to see what complex behaviors emerge
ex.braitenberg’s simple sensorimotor vehicle or Walter’s robotic tortoises exhibiting complex fear or aggression through simple wiring interacting with physical light
pushes individual architectures to their limits, demonstrating how much complex behavior emegres ‘for free’ from physical environment interactions without complex internal programs