02_Week 2 Notes (History).pdf

September 13, 2024 Mental Representations Nodes or symbols that are translated or computed by cognitive processes to process the information “cinq”, “5”, “V” = all represent the same content/concept Words, symbols HOW CAN...

September 13, 2024 Mental Representations Nodes or symbols that are translated or computed by cognitive processes to process the information “cinq”, “5”, “V” = all represent the same content/concept Words, symbols HOW CAN THE MIND “REPRESENT” SOMETHING? what’s the nature of a mental representation, the internal code we use to represent knowledge How neurons code for a concept Electrophysiological neuron recordings representations as images Resemble what they are about (except abstract) To know something is to have it i.e. representation of a red ball because we have the concepts “red” and “ball” beforehand Historical Foundations Plato & Socrates Plato’s problem: how can we know given our fragmented limited experience? - how do we know more than we are explicitly taught or experienced? ❖ Socrates: inherit knowledge from earlier souls (genetic endowment, innate knowledge) that just needs “awakening” - Encoded knowledge (implicit): by assumption certain properties of the world become part of the natural kinds to which the organism responds - exaptations: exploitation in a novel function of a characteristic that was originally an adaptation for a different function Descartes, Hobbes, Hume 1. Mental representations are ideas that can take different forms (codes, images) These ideas enter into cognitive processes the mind’s eyes = imagining (not seeing) Imagining can result in anything you want Imagination is not like perception Responses are based on knowledge 2. Mental representations are calculations (computations) that are language-like, abstract, symbolic they are the elements of thoughts, cognitive processes images need to be interpreted when perceived If we stored raw images, then we would need to reinterpret during recall  September 13, 2024 Still need to rely on some code to recall them Needs some form of tag, metadata, or description Descriptions of the scene/event Descriptions are represented as symbolic codes (propositions) action (verb) with predicates (adjectives, subject, adverbs, objects) Encoding the event 3. Perceptual and cognitive systems (some) work as a function of the laws/ rules they follow rules work regardless of the content of parts of the equation (except “or” or “not”) The reasoning process is carried out simply for its form (syntax or structure) Computations over symbolic representations are operations driven by rules ** Behind most perceptual or cognitive processes are sets of rules governing processes DESCARTES - knowledge is acquired from senses - Knowledge is put in the mind by God (innate) - Ideas, representations, are the certainty we have of our existence ❖ A lot of what we do is mechanical ❖ Much is part of our creative endowment - Personal alterations - Distinction between perception and judgement (cognition) ❖ i.e.: Difference between looking at a chair and understanding it is a chair **first cognitive revolution Early In uences Introspectionism (Wundt, Titchener) - physics: inspection, observation ❖ Observation = laws - Psychology: introspection ❖ “Observation” of our mind ❖ “Inspection” of one’s own thoughts, processes - Subjects would report on the contents of their “thoughts” as they perceived or processed a stimulus - Problems: 1) many unconscious inferences ❖ as we are in contact with stimulus, we are not consciously processing it - i.e. listening to a sentence, our brain does not go “this is a verb”, etc. 2) Reports are faulty: memory & attention bottlenecks 3) Implicit knowledge: how can we know what kind of knowledge enter into a given process? fl  September 13, 2024 Behaviourism (Watson, Skinner) - psychology as the science of behaviour ❖ No introspection ❖ No “consciousness” - Behaviours are S-R pairings - Learning by reinforcements and punishments ❖ Ex: Language acquisition: words are learned by reinforcements ❖ Language is bootstrapped (ex: all past tense verbs end in -ed, “goed”) BUT: Physics can only tell you what the properties of the object are, not how it is represented nor why it is represented that way Chomsky (reaction to Behaviourism) - Said that “Skinner was postulating “internal representations”, not ‘observable behaviours’” - - What is possibly innate is the initial state of grammatical knowledge that yields the grammar of any language given exposure to any linguistic community at an appropriate maturational stage The Generative Capacity of language ❖ Ex: recursion: - Mary thinks Paul loves Susan - John believes Mary thinks Paul loves Susan ❖ Etc. - The “poverty of stimulus” argument (how can you have knowledge not taught) ❖ Behaviourists account for cognitive phenomena was insufficient ❖ i.e. toddlers creating sentences that they never heard before with words they’ve encoded using the rules they know 2nd Cognitive Revolution - Human factors after WW2 ❖ skill acquisition, performance in complex tasks Theories of Communication & Attention - information and information processing - Studies of Attention (Cherry) ❖ Dichotic listening paradigm ❖ Allocation of attention to different tasks ❖ Selection of which “stimulus” you attend to Computers & Computer Science - physical systems that store and use/transmit information - Rich metaphor for human information (Turing) - “Intelligent” program  September 13, 2024 TURING MACHINE Anatomy: head: scanner and printer of symbols Infinite memory Tape: squares start blank and end up with 0,1 from head Limited representations and limited se of rules but compute an infinite number of computations i.e. language Limited number of words Limited number of categories (verb, noun, etc) Limited combinations (syntactic and grammatical structures) Unlimited of sentences Information processing system (physical) can behave just in virtue of its rules and representation can explain cognitive behaviour by specifying the nature of the rules and representations a system encode Psychological Processes as information processes (Miller) - influence of Theory of communication - plans: internal representations (a set of instructions) one must have to guide one's actions - Quantifying information based on units transmitted - STM: 7+/- units Chomsky - Linguistics ❖ Competence/performance distinction ❖ Language as a computational system Neisser: Cognitive Psychology Book - new methods of investigation - Ration time as a measure of processing load or complexity Flow Diagrams - used to model stages of components of cognitive processes - The mind as an information processor ❖ i.e. cognitive mode of lexical decision - “Is the item divisible into prefix and stem?” Yes/no, etc.  September 13, 2024 Levels of explanation Functional Level Behaviourism: S-R laws non-observable mental states do not come to play Perform actions (R) that might differ from what stimuli (S) might invoke Radical Dualism: no connection brain and mind are 2 different entities How does the mind cause behaviours How does the non-physical mind cause physical behaviours/event? Functionalism: explaining the machine’s behaviour Identifying the actions (inputs) and the outputs (outcomes) No claims about internal mechanisms behaviour of the machine can be expanded as a function of Machine’s internal state Machines causal structure **Can account on behaviour of any machine by simply describing the rules Behaviouralistic explanation input and outputs only Cognivistic explanation internal states, non-observable 2+ possible states Outputs depend on Input Machine’s current state given Cognitive theorizing the brain = the organ of the mind Representationalism: emphasis on mental representations Functionalism: emergent properties CAN MACHINES THINK? if cognitive is information processing, then non-brain machines can think Learn patterns using algorithm to interpret data  September 13, 2024 Rise of Cognitive Psychology (Timeline) Tolman: concept of a mental map - understanding of where to go and how to get there (navigation) Cherry: Attention experiments Turing: Turing machine Digital computers: model cognitive processes on the available computers Skinner: verbal behaviour Broadbent: first to divide flow diagrams of cognitive processes Chomsky: beginnings of linguistics, verbal behaviour (review of Skinner) Miller: plans and the structure of behaviour Neisser: first cognitive psychology book ***Cognitive science is a mix of different disciplines