Introduction to the Philosophy of Cognitive Sciences | Coursera

Brief Information

Week 1

This week, we will explore scientific interpretations of how our minds evolved, and some of the methodologies used in forming these interpretations. We will relate evolutionary debates to a core issue in the philosophy of mind, namely, whether all knowledge comes from experience, or whether we have ‘inborn’ knowledge about certain aspects of our world.

1-1) Stone-Age Minds Part I

  • Topic of this lecture
    • How human brains and human cognitive structures evolved.
  • Old debate in philosophy
    • Whether we can have knowledge which isn’t learned
      • = Whether the mind is a blank slate
    • Whether all our knowledge comes from our experience of the world
  • Evolutionary psychology
  • Social learning and culture
  • Lots of human behaviors can be explained as a product of culture and cultural evolution.
  • Natural selection can be seen as a process of environmental filtering.
    • The environment filters out the organisms that are least suited to living in it.
  • Adaptations are physical features or behavioral traits that an organism has as a result of natural selection.
  • Necessities for a trait to be a product of natural selection
    1. The trait is heritable.
    2. Some organisms have the trait but the others do not.
    3. The trait is advantageous to the organism’s reproductive or survival success.
  • Example 1: Beavers
    • Beavers like wet areas.
    • Behavior adaptations
      • Beavers build dams to block the rivers and make wet areas.
      • Beavers cut down trees to make dams.
    • Physical adaptations
      • Beavers have big rugged teeth to cut down tree.
      • Beavers have a damming instinct stimulated by the sound of water running over obstacles.
    • Selection pressure
      • The pressure for beavers to adapt to their environment
    • Beavers take the strategy to change their environment so that the selection pressures are reduced.
  • Can we use perspectives of natural selection and adaptation to understand the evolution of our cognitive traits?
  • Natural selection and adaptation are the same event. But they have different perspectives.
    • Natural selection: The nature selects the traits of organisms.
    • Adaptation: Organisms adapt to the nature for their survival or reproduction.

1-2) Stone-Age Minds Part II

  • Ongoing debates in cognitive science
    • Debate 1. How much of our brain is a product of natural selection.
    • Debate 2. How much of our brain is a product of social learning.
  • Evolutionary psychology
    • = a subfield of psychology that maintains that most of our cognitive capacities are results of natural selection
  • The 4 claims of evolutionary psychology
    1. The human brain is a product of natural selection.
    2. The human brain adapted to solve particular problem which our ancestors faced.
    3. The cognitive capacities are inheritable.
    4. The brains we have now are the brains our ancestors evolved long  years ago.
  • How we should understand the role of natural selection in shaping the human brain.
  • Our modern skulls house Stone Age minds.
  • The vast majority of our species’ history wad spent living in small hunter gatherer groups in the African savanna.
    • The environment where we have evolved the modern brains is called the Environment of Evolutionary Adaptation.
  • Mal-adaptations
    • The adaptations that were beneficial to us when we were hunter gatherers, but that were not beneficial to us in our modern urban environment.
    • Example 1: Craving and salt and sugar
    • Example 2: a visceral reaction to snakes and spider
  •  What were the kind of problems that our hominid ancestors faced?
    • Who should I cooperate with to do my hunting and gathering?
    • Who in my environment is free riding? – the cheater detection problem
  • We have a cognitive adaptation which evolved to allow us to detect cheaters in our environment.
  • We are good at detecting when a social norm has been violated.
  • The modular view of the mind. the modularity of the human mind. 
    • The mind is a series of mini-computers.
    • Each mini-computer is specialized  to do a particular cognitive task. For example, detecting friends from foes, detecting who is a cheater, detecting who is free riding, mathematical calculation, spacial cognition.
  • The incredibly complex organ, the brain is likely to evolve over millions of years, changing incrementally in tiny steps.
  • Example. The human eyes.
    • light sensors → retinas → cones and rods → fine color vision
    • The separate parts of the eyes seem evolved separately because even if a part of eyes do not work properly, it is not the case that we cannot totally see anything.
  • Natural selection can act on one module of the brain.

1-3) Stone-Age Minds Part III

  • Ecological inheritance
    • = environment inheritance
    • = non-genetic inheritance
    • Inheritance through social learning, not through genes
    • Organisms inherit their environment.
    • The ability to learn or modify your behavior in response to your environment is very widespread in the nature.
    • Ecological inheritance is rife in humans.
    • Examples: house, knowledge, behaviors
  • Social learning is learning from the behavior of others.
  • Understanding social learning is crucial to understand human evolution.
  • Culture is the collective term of for large repertoires of behaviors which are transmitted by social learning.
  • Two types of differences: cultural differences, genetic differences
  • Animal cultures are simple. If a single individual invents a new behavior, then it spreads through the population as the animal is copied by others.
  • The human culture is are complex. A single individual inherits a behavior. Then, you modify the behavior and improve its design. Then, you pass on the modified behavior. This kind of culture is called cumulative culture.
  • Living in large complex social groups are not instinctive behaviors but socially learned.
  • This kind of modifications we make to our environment massively change the selective pressures acting on us.
    • We insulate ourselves from hostile environments with clothing and house.
  • Example: lactase persistence
    • General rule: After weaning, it doesn’t need to produce lactase.
    • In a third of the world’s human population, adults keeps producing lactase, which is called lactase persistence.
    • They benefit from drinking animal milk, which has rich protein and fat.
    • Lactase persistence is the result of genes.
    • The populations with lactase persistence are with a long history of dairying.
  • Gene culture coevolution
    • It is reliable that dairying tradition makes their genes adapt to the tradition. “Tradition changes gene.”
    • The gene to produce lactase made the dairy tradition. “Gene changes tradition.”
    • Genes changes their environment to adapt. The environment changes genes to adapt.
  • The human language is a uniquely powerful, flexible communication system.
    • As long as you know the meanings of the words and the rules of the language, then you can understand any sentence that follows those rules.
    • No other species has a communication system that works like language.
    • Attempts to teach human language to non-human animals have met with only limited success.
    • Language is a product of cumulative culture.
  • Pinker, S. & Bloom, P. (1990) Natural language and natural selection. Behavioral and Brain Sciences 13 (4): 707‐784
    • This paper argued that language is a complex biological trait that appears to be designed for communication.
    • ★ The only way to explain such traits is to appeal to natural selection.
  • Our historical record for languages doesn’t actually stretch back very far.
    • Writing was only invented 5000 years ago.
    • Languages must have been around a lot longer than 5000 years.
    • In this respective, language is largely influenced by culture than by genes.
    • Among cumulative cultures, language is has a very long period of time to evolve.
  • In order to survive, it has to be easy to learn language.
    • As a result, language evolves to have rules, patterns, and regularities.
  • Why language is a cumulative culture
    • [modification] People modify their  language to meet their communicative needs.
    • [culture] People highly rely on language in most of their life. Interaction using language is very frequent. Thus, people share many traits of language.
    • [inheritance] People learn language from people. They don’t create language as a whole.
  • How to research on the human language (subordinate reasearch topics)
    • How language is learned in the present
    • How language changes over historical time
    • Simulate these two processes of learning and changing either in a computer.
    • Examine the processes with real people.
  • It is not proved that language is a product of culture than a biological adaptation.

1-4) Handout

1-5) Additional Reading

  1. Cosmides, L. and Tooby, J. (1997). Evolutionary psychology: a primer. Available at: http://www.cep.ucsb.edu/primer.html
    • an introductory reading on evolutionary psychology
  2. Evolutionary Psychology | Stanford Encyclopedia of Philosophy
    • an overview of philosophical issues in evolutionary psychology
  3. Evolutionary Psychology | Internet Encyclopedia of Philosophy
    • an overview of philosophical issues in evolutionary psychology
  4. The Language Evolution and Computation unit, the University of Edinburgh
    • more on the evolution of language
  5. Niche Construction, The neglected process in evolution
    • Niche = environment. Niche construction is the process that organisms’ modify their environment to lower natural selection pressures.
  6. Why chimpanzees might be less ‘cultured’ than humans | Live Science
    • Chimps just aren’t as motivated to learn from one another as humans are.
    • SUMMARY

Week 2

One of the hardest problems in science is the nature of consciousness. We know that we have consciousness. We do not just blindly process information, make discriminations, take actions. It also feels a certain way to do so from the inside. Why do creatures with brains like ours have consciousness? What makes certain bits of our mental life conscious and others not? These questions form the heart of consciousness science, an exciting field to which psychologists, neuroscientists and philosophers contribute. This session will explore these questions, and introduce some recent progress that has been made towards answering them.

2-1) What is Consciousness Part I

I. Introduction
II. What Do We Mean By Consciousness
  • Consciousness is a folk concept that arises out of our everyday interests.
  • A scientific understanding of consciousness should approach our folk talk about consciousness.
  • Folk concepts of consciousness
    • 1. sentience
    • 2. wakefulness
    • 3. access consciousness
    • 4. phenomenal consciousness
  • Concept 1: Sentience
    • = the state that
      • a creature acts in an intelligent way and
      • it is responsive to its environment.
    • e.g.) ‘The spider is aware that we’re here and has sensibly taken evasive action.’
  • Concept 2: Wakefulness
    • = the state that a creature is wake, not asleep.
  • Concept 3: Access consciousness
    • = the state that a creature can report ‘what it is thinking now’
    • The majority of our mental life is not access conscious.
  • Concept 4: Phenomenal consciousness
    • = qualia
    • = the state with subjective feelings that accompany many episodes in our mental life
III. The Hard Problem
  •  The hard problem of consciousness
    • to explain how creatures {have/produce} phenomenal consciousness
    • Two perspective to consciousness
      • 1. From the subjective point of view: phenomenology
      • 2. From the objective point of view: neuroscience
  • Phenomenology
    • Reflecting on our conscious life a subjective point of view, via introspection
  • Neuroscience
    • Consider your brain as a physical object.
    • However, we have no idea how brain activity produces phenomenal consciousness.
  • Why the hard problems of consciousness so hard?
    • It is difficult to link phenomenology and neuroscience together.
  • A number of philosophers, including Frank Jackson argued that
    • science will never reductively explain phenomenal consciousness in terms of brain activity.
IV. Mary the Color Scientist

  • Frank Jackson’s thought experiment
    • Mary is born and grows up inside a black and white room.
    • Mary learns all about how the human brain detects and processes color information.
    • One day, Mary’s released from the room.
    • Mary spots a red rose. Mary experiences color for the first time. Mary learns about the subjective feelings that accompany seeing color.
    • Summary: Mary knows all about how the brain works, yet the facts about phenomenal consciousness still elude her.
    • Conclusion: Even if we complete neuroscience, we would still be stuck with the hard problem of consciousness.

2-2) What is Consciousness Part II

I. Scientific Perspectives: Introduction
  • Lecture 2-1: philosophical approaches to consciousness
  • Lecture 2-2: scientific approaches to consciousness
    • = empirical approaches
II. States of Consciousness

  • It is useful to consider states of consciousness as a combination of 2 factors: wakefulness and awareness.
  • Wakefulness
    • the degree to which we are awake
  • Awareness
    • the degree of our capacity to think, feel, and perceive ourselves and our environments
  • States of consciousness
    • conscious wakefulness, drowsiness, light sleep, deep sleep
    • lucid dreaming, REM sleep, deep sleep
    • general anesthesia, coma
    • minimally conscious state, vegetative state
    • lock-in syndrome
  • There is no single brain area whose activity is solely responsible for either awareness or wakefulness.
  • Wakefulness is highly dependent on activity in the subcortical structures.
  • Awareness is highly dependent on activity in the cortex.
  • Awareness can be divided into two elements.
  • External awareness
    • the awareness we have whenever we navigate through and interact with the external world
    • depends on activity in the frontal and parietal lobes of the cortex.
  • Internal awareness
    • the awareness we have whenever we navigate through and interact with the internal world; daydreaming, retrieving memories, planning for the future
    • depends on activity in a network of regions that are on the media side of the brain.
  • It is rare to be doing both things at the same time.
    • The activity of the two networks involved in awareness, is negatively correlated.
  • Changes in wakefulness affect activity in the cortex.
    • While we are awake areas of the cortex busily communicate with one another.
    • As we fall asleep, these areas’ communication is reduced.
    • For example, areas of the frontal and posterior cortex.
III. Perceptual Awareness: How Much of the World Around Us Do We Experience?
  • We are aware of a surprisingly small subset of the information enter our brain through our senses. We are aware of the information we are attending.
  • Inattentional blindness
    • We become blind to the information that we are not attending.
    • This represent the intimate link between awareness and attention.
  • The capacity limits of our visual working memory
    • = the store of visual information that is available for our immediate use
    • We start to forget elements when we perceive more than four elements.
IV. Perceptual Awareness: What Process Shape Our Experience?
  • Bistable images
    • e.g. the Necker Cube, the Rubin face vase
    • Bistable images help us discover dissociation between perception and awareness.
    • While staring at a bistable image, the external stimulus does not change, but our perception does change. The change is called a perceptual switch.
    • The areas of the brain active at the same time as perceptual switches
      • occipital lobe(visual areas)
      • frontal and parietal regions
    • The activities of these areas just mean the correlation with perceptual switches, not the causation

  • TMS(Transranial Magnetic Stimulation)
    • TMS is used to figure about a certain brain region has a causal influence.
    • TMS works by applying a brief powerful magnetic pulse to the surface head.
    • TMS interferes temporarily with the activity of the area of cortex right underneath it.
    • Experiment 1
      • Apply TMS to the Broca’s area where produces speech.
      • Speak continuously.
      • Operate TMS.
      • We can see TMS interferes with the speech production.
    • Not every part of the brain that we apply TMS to will have immediate observable consequences.
    • Experiment 2
      • Apply TMS to certain part of parietal cortex. ⇒ The switching becomes slower.
      • Apply TMS to slightly different parts of parietal cortex. ⇒ The switching becomes faster.
      • Conclusion: The parietal cortex is causally involved in bistable perception.
V. Perceptual Awareness: Can We Perceive Things Unconsciously?
  • How to investigate unconscious perception?
    • Unconscious perception: People are shown to things, but they actively suppress the things from awareness.
  • Backward visual masking technique
    • Experiment 1
      • One image is shown very briefly.
      • Another image is shown for longer at the same location.
      • People cannot report the first of the two images. Often, they deny the first one was there.
    • Experiment 2 (masked priming technique)
      • A word A is shown.
      • A is masked by a meaningless pattern.
      • A string B of letter is shown.
      • People are asked to decide B is a real word or not.
      • Suppose B is a real word.
        • If B is semantically related to A, then they detect faster than if B is not.
      • Example
        • If the masked word is ‘infant’, people are faster to recognize the word ‘child’ than they to recognize the word ‘orange’.
      • Conclusion: The masked word activates a semantic network in the brain.
    • Another research result
      • Masked words activate visual area of the brain more than meaningless strings of letter do, even when people remain unaware of masked words.
  • Now, there is no theory that offers a full unified account of consciousness.
  • The current theories just offer agendas for the future research.

2-3) Handout From Mark’s Lectures

  • What do folk mean by ‘consciousness’?
    1. Sentience — A creature is receptive to its surroundings and it can act in an intelligent way
    2. Wakefulness — Not asleep or otherwise incapacitated
    3. Access consciousness — A thought that is widely broadcast in a creature’s brain and guides many of its actions
    4. Phenomenal consciousness — Subjective feelings that accompany many episodes in our mental life
  • Our focus is phenomenal consciousness.
  • Two main ways of accessing facts about our mental life
    1. Introspective reflection on our own experience (phenomenology 현상학)
    2. Examination of our brains and behavior from outside (psychology, neuroscience)
  • The hard problem of consciousness
    • Explaining how subjective feelings (accessed by phenomenology) are brought about by our brains and behavior (studied by the natural sciences) is a particularly difficult problem.
  • Frank Jackson’s argument that the hard problem will never be solved is called the Knowledge Argument.
    • Even if we are in the lucky position of Mary — having a completed neuroscience — we could still not be able to predict or explain the phenomenal feelings that accompany seeing color.

2-4) Slides From David’s Lectures

  • 2-3 contains the content of these slides.

2-5) Additional Readings

  1. [Conscious Pictures]. (2013, Apr. 8). David Chalmers on the “hard problem” of consciousness – Chronicles 1. [Video File]. Retrieved from https://youtu.be/kdbs-HUAxC8.
    • David Chalmers explained the hard problem of consciousness.
  2. Pinker, S. (2007) ‘The brain: the mystery of consciousness’, Time
  3. Van Gulick, Robert, “Consciousness”, The Stanford Encyclopedia of Philosophy (Winter 2016 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/win2016/entries/consciousness/>.
    • Great overview of philosophical problems concerning consciousness with many suggestions for further readings
  4. Simons, D. (2010) ‘The monkey business illusion’
    • About limitations of perceptual awareness
  5. Simons, D. and Chabris, C. F. (uploaded by Simons, D. 2010) ‘Selective attention test’
    • About limitations of perceptual awareness

Week 3

Recent years have seen a revolution in the kinds of tasks computers can do. Underlying these advances is the burgeoning field of machine learning and computational neuroscience. The same methods that allow us to make clever machines also appear to hold the key to understanding ourselves: to explaining how our brain and mind work. We explore this exciting new field and some of the philosophical questions that it raises.

3-1) Intelligent Machines Part I

  • Machine learning
    • = the subfield of computer science that gives machines the ability to learn from their experience without being explicitly programmed.
    • Machine learning helps us to understand how our own brains work.
  • Computation
    • = A way of solving a problem by following a set of instructions called an algorithm
  • Algorithm
    • = A series of small steps that solves a certain problem
  • Alan Turing
    • an English mathematician
    • Turing was obsessed with the project of trying to create an intelligent machine.
    • His famous paper
      • Turing, A. M (1936) On Computable Numbers With an Application to the Entscheidungsproblem. Proceedings of the London Mathematical Society
    • In this paper, Turing introduced a universal computing machine called a universal Turing machine.
  • Universal Turing machine
    • = a machine that can replace any of our computer if given the right instructions
    • Key components of a universal Turing machine
      • 1. a long paper tape that is able to store all data
      • 2. a head that is able to scan along the tape, read data from and write data on the tape
  • The phrase ‘if given the right instructions‘ matters to create intelligent behavior.
  • Finding the right algorithm is crucial to create intelligent behavior.
  • What kind of algorithm produces intelligent behavior.
    • 1. Algorithms involving language-like symbols and rules
    • 2. Algorithms involving networks inspired by the human brain, so-called connectionist networks
    • 3. Probabilistic algorithms that represent a range of different outcomes with their probability and learn from experience using Bayes’ rule
  • ★ Performing the right algorithms to make a machine intelligent might be what makes humans intelligent. The algorithms are not just engineering tools but also the key to explain how the human brain works.
  • A cognitive scientist, David Marr, said that computation could help us to answer three questions about the brain.
    • Q1. Which task does the brain solve?
    • Q2. How does the brain solve that task?
    • Q3. Why is that task important for the brain to solve?
  • Three levels of computational description
    • 1. Computational level description
      • = a description of which mathematical function, addition, subtraction, multiplication, the device compute.
      • Which function does the device compute? Why does the device compute?
      • input —[computation]→ output
    • 2. Algorithmic level description
      • = a description of how the device solves its tasks
      • How does the device compute that function?
      • Each algorithm spends different time and uses different memory capacity.
    • 3. Implementation level description
      • = a description of how the physical parts inside the device map onto steps in that algorithm
      • How do the physical components of the device map onto steps in its algorithm?
      • How to find the implementation level description
        • Keep the device open. Watch what changes inside when computation proceeds.
        • Intervene(rewire, move components) on the device. See how the intervene affects the device.
  • Cognitive scientists want to know
    • which computation the brain performs,
    • which algorithm it uses for performing that computation, and
    • which physical bits of the brain are relevant for implementing that computation.

3-2) Intelligent Machines Part II

  • The brain might work like a probabilistic machine.
    • Our brain guesses what is the best action in the future based on what is present in the world.
    • Our brain computes the probability of each action to be best. Then, our brain chooses the action with the highest probability.
  • Hermann Von Helmholtz proposed that our visual perception was the result of unconscious inference.
    • Unconscious inference is the unconscious process that the brain completes missing information using past knowledge and construct a hypothesis about our environment.
  • The idea that the brain is a probabilistic machine proposes that
    • the brain works by constantly forming hypotheses about environment and the actions to take.
  • This idea has been formalized from machine learning and statistics using conditional probabilities and the Bayes’ rule.
  • The probabilistic inference using the Bayes’ rule is called Bayesian inference.

  • H: hypothesis, D: current environment
  • Bayesian inference predicts that the optimal way to find the best action is to combine information from multiple modalities but weighting the information according to its reliability.
  • McGurk Effect
    • ‘Bah’ sound with ‘bah’ lips is perceived as ‘bah’.
    • ‘Bah’ sound with ‘dah’ lips is perceived as ‘dah’.
    • Conclusion
      • Our brain unconsciously combines the visual and the auditory information in our perception of speech.
      • This combination creates a new mixture that might be different from the initial sources of information.
  • In everyday life, the brain combines information from different sensory modalities.
  • Prior distribution(?) serves as a summary of all previous experience.
  • ★ In the current situation of strong uncertainty, we rely maximally on our previous experience.
  • Our brain makes assumptions all the time.
  • Visual illusions
    • Concave faces look convex. the hollow mask illusion.
    • Objects look symmetrical, smooth.
    • Orientations look horizontal or vertical.
  • In the cases of visual illusion, the interpretation chosen by the brain is automatic and unconscious, and cannot be modulated voluntarily. These assumptions make sense because most objects in the world conform to those expectations.
  • While illusions, we tend to perceive reality as being more similar to what we expect than it really is.
  • situation of strong uncertainty = ‘The objects do not conform to the average statistics.’
  • Bayesian inference can be applied to all domains of cognition.
  • Bayesian models are useful for psychiatry.
  • Schizophrenic patients are more likely to make a decision after a very small number of observations. They conclude in insufficient observations. They are biased. Their prior belief is too strong or too weak. This could explain why the patients experienced the world differently.
    • Bayesian modeling might help diagnosis comutationally.
  • Bayesian models are very useful for describing perception and behavior at the computational level.
  • The focus of the current research in neuroscience
    • Which areas of the brain would be involved
    • How Bayes’ rule of the terms of these equations would be implemented

3-3) Additional Readings

  1. Rescorla, Michael, “The Computational Theory of Mind”, The Stanford Encyclopedia of Philosophy (Spring 2017 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/spr2017/entries/computational-mind/>
    • Recommendable as an introductory reading
  2. Garson, James, “Connectionism”, The Stanford Encyclopedia of Philosophy (Winter 2016 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/win2016/entries/connectionism/>
    • Recommendable as an introductory reading
  3. Clark, A. (2012) Do Thrifty Brains Make Better Minds?. The New York Times
  4. Clark, A. (2012) Prediction and the Brain: A Response. The New York Times
  5. [BBC]. (2010, Nov. 10). Try The McGurk Effect! – Horizon: Is Seeing Believing? – BBC Two. [Video File]. Retrieved from https://youtu.be/G-lN8vWm3m0
    • A YouTube video illustrating the McGurk effect, from the BBC
  6. [Scientific American]. (2012, Oct. 11). What Is the Ames Illusion? – Instant Egghead #23. [Video File]. Retrieved from https://youtu.be/gJhyu6nlGt8
    • A YouTube video illustrating the Ames illusion, from Scientific American
  7. Motion perception | Wikipedia
  8. [eChalk]. (2012, July 20). The rotating mask illusion. [Video File]. Retrieved from https://youtu.be/sKa0eaKsdA0
    • A YouTube video illustrating the rotating mask illusion, by eChalk Scientific

Week 4

  • Instructor: Andy Clark, Barbara Webb

Cognitive Science has recently taken a strongly ’embodied turn’, recognizing that biologically evolved intelligence makes the most of the opportunities provided by bodily form, action, and the material and social environment. This session explores the way this impacts our vision of minds, brains, and intelligent agents, and asks whether there can be a fundamental science of the embodied mind.

4-1) Embodied Cognition Part I

  • Cognitive science is the science that studies the mind.
  • Within cognitive science, there is a field study called ’embodied cognition’.
  • Embodied cognition is all about the huge difference that having an active body and being situated in a structured environment make to the kind of tasks that the brain has to perform in order to support adaptive success.
  • Embodied cognition focuses on a complex interplay between what the brain is doing and what the body is doing.
  • Robotics is one of the key places where brain, body, and world come together.
  • Example 1: The passive dynamic walking machine built at Cornell University
    • This robot has no motors or controllers
    • Its motion is begun by falling forward on a slight slope, pendulum motion.
    • Taking the physical into account is important.
    • There is something going on inside us when we walk.
    • Understanding and actually building the physical system tells us more about how we walk than trying to introspectively think about how we move our legs, swing the foot forward.
  • Example 2: The behavior of desert ants
    • A desert ant simply alters the size of it zig-zag proportionally to how unfamiliar its current heading appears.
      • If things looks familiar, it is heading straightly.
      • If things looks unfamiliar, it swings around until it sees some view that it saw before.
      • The ant does not distinguish individual objects as objects.
      • The ant just processes the scene of objects into low resolution vision as a whole.
    • The ant senses polarized light to deduce the position of the sun and use the position as a compass cue.

  • Biology provides the physical structures designed to solve the problem robotics faces.

4-2) Embodied Cognition Part II

  • Brains and their bodies are coevolved.
    • Brains are evolved based on their bodies.
    • Bodies are evolved based on their brains.
  • Example 1: the blue-fin tuna
    • The tuna makes the most of its watery environment.
    • The tuna steps into water currents to go faster.
    • The tuna flaps its tail to create small vortices to blast off efficiently or turn more quickly.
    • The tuna makes short-term changes to its immediate environment to improve its swimming.
    • Many creatures make long-term changes to their environment. These changes can structure the world to make their interaction with the world more successful.
    • Conclusion: Creatures change their environment to interact with the world successfully.
  • Example 2: How do ants build a building across the entrance to the nest to guard themselves?
    • Rule 1: move around → meet a particle → pick up the particle → cross the nest entrance → smell the odor from the eggs → drop the particle
    • Rule 2: move around → meet a particle → pick up the particle → move around → meet another particle → drop the particle
    • Rule 1 makes the wall near the entrance.
    • Rule 2 makes the wall get bigger.
    • Two simple rules make the wall easier.
    • It is not introduced when the ants stop building the wall.
    • Conclusion
      • Interactions with the world structures the world.
      • Interactions with the world make complex tasks easier.
  • Question: How does interaction with the world scale up to higher cognition: thinking, reasoning, and planning?
  • Example 3: Babybot
    • Babybot uses its own actions to help it learn about the world.
      • Babybot sweeps its arm in its field of vision.
      • If Babybot encounters an object, it learns about the object by using its arms and vision.
  • Example 4: Multi-modal self-stimulation
    • A robot pokes, pulls, and shoves objects.
    • It moves around objects.
    • It makes objects noises.
    • It brings an object to its face and smell the object.
    • These interactions build higher cognition.
  • Example 5: How to make a robot recognize a chair
    • We understand a chair by the concept of sitability.
    • To understand a chair, a robot should interact with (or sit on) chairs, not just see their shape.
  • Example 6: the urban challenge robots
    • The urban challenge robots are robots embedded in cars who had to drive around a real urban environment, i.e., other cars, roads, road signs, etc..
    • The robot has multi-layer control systems.
      • Low level control system
        • keeping the wheels on the road
        • detecting if a collision is about to occur
        • tracking moving objects
        • making appropriate evasive actions
      • High level control system
        • Planning how to get from A to B
  • The naked brain fallacy
    • = the fallacy of assuming that all the interesting cognitive action is always going on in the brain
  • We should not think of minds as disembodied computers.
  • We should think minds are integrated with our physical capacities and our interactions with the world.
  • To create artificial minds, we will have to build the mind systems that are embodied.

4-3) Embodied Cognition Slides

  • Body, action, and world!
  • Sensory, motor, and neural systems
  • Perception, action, and neural systems
  • A unified science of the mind encompasses ecological context, action, timing, bio-mechanics, dynamics, computation and representation.

4-4) Additional Readings

  1. Wilson, Robert A. and Foglia, Lucia, “Embodied Cognition”, The Stanford Encyclopedia of Philosophy (Spring 2017 Edition), Edward N. Zalta (ed.), URL = <https://plato.stanford.edu/archives/spr2017/entries/embodied-cognition/>
  2. Tutorial on Embodiment | the European Network for the Advancement of Artificial Cognitive Systems, Interaction and Robotics
  3. Wilson, A. D. (2012) A Tale of Two Robots. Psychology Today
    • Comparing the different approaches to walking taken by traditional and embodied robotics: Asimo and Big Dog
    • SUMMARY
  4. [AMBER-Lab] (2010, Oct. 18) McGeer and Passive Dynamic Bipedal Walking [Video File] Retrieved from https://youtu.be/WOPED7I5Lac
    • A video about passive-dynamic walkers
  5. The current crop of robots from the Cornell University Biorobotics and Locomotion Lab
  6. Robot Ethology | Vassar College

 

Leave a Reply

Your email address will not be published. Required fields are marked *