cognitive science exam 1
What is the Loebner Prize and how is it related to the test?
- make the test more formal - judges know some of the subjects are machines
Skinner claims that people gamble because they like feeling the excitement.
FALSE - Since Skinner is a behaviorist, he would not try to explain the behavior of gamblers using inner mental states like the feeling of excitement
Is imagery exclusive to vision?
NO "minds eye" - we can develop from all the other senses: remembering the fell or smell of something, mentally imagine sound of something
What project is Jackendoff undertaking? How does Jackendoff's terminology of "low, intermediate, and high" levels of visual processing correspond to Marr's terminology of the "primal, 2.5D, and 3D sketches"?
Where in the flow of information processing does consciousness arrive? Jackendoff thinks the intermediate level - correspond the same
According to Skinner, do human beings and animals learn in the same way?
here is little difference between the learning that takes place in humans and that in other animals. Therefore research (e.g., operant conditioning) can be carried out on animals (Rats / Pigeons) as well as on humans
What is the phenomenon of "inattentional blindness"?
the failure to notice something thats' fully obvious, when it's right in front of you
What was the full (or whole) report method? According to this method, what is the capacity of visual iconic memory? What is the problem with using this method? (What is the duration of visual iconic memory? What does the duration of this form of memory have to do with the previously mentioned problem?)
(briefly flashing some letters before his subjects and then, immediately after those (physical) letters disappeared, asking those subjects to report as many of the letter as they could from their visual iconic memory - capacity is around 3-5 letters - Many of his subjects reports that there were more than 3-5 letters in their visual iconic memory, but that these letters disappeared before they could report them (duration is about 1/2 - 2 seconds) - likely that the capacity of visual iconic memory was greater than Sperling's initial study suggested
What were the 3 groups of rats that Tolman and Honzik trained to run a maze? How were these 3 groups different from one another?
- 1st group received a reward each time they successfully ran the maze - 2nd group never received a reward - 3rd group was unrewarded first 10 days and then began being rewarded
What is the 2.5D sketch? What is represented at this stage of processing? (What are some of the tricks the visual system uses to extra information about DEPTH from a 2-dimensional image?
- 2nd stage of processing - starts with blobs in the primal sketch and creates lines on the basis of those blobs and puts in info about depth - hard to compute depth so the visual system does it in steps - representations of surfaces and their orientation s relative to the viewer (arrows point in different ways relative to the viewer given their current position) how can there be depth cues in a 2D image? TRICKS - occlusions - relative size - binocular disparity - shadows - elevation
What is Kosslyn's notion of "the visual buffer"? Where does this bit of visual processing occur? What role does it play in perception? What role does it play in imagery?
- A form of short-term memory storage between the visual senses and higher cognition - in cases of perception, it is primarily stimulated by the senses - but can also be stimulated by top-down processing, from higher cognition and long-term memory → result is imagery
In Segal's follow-up study, what did she learn about the Perky Effect? How did Segal further explore this phenomenon by projecting images that were different from what her subjects were asked to imagine? What conclusion about perception and imagery did Segal draw from her studies?
- Segal's initial attempts to replicate Perky findings failed → subjects could easily notice the projected patches - turns out necessary to induce a prior state of relaxation in her subjects - also tried projecting faint pictures that were quite different from the mental image she had asked her subjects to form - Some assimilated even this incongruous stimulus into their imagery, and still did not realize that a real visual stimulus was influencing their experience - some subjects were asked to imagine a New York skyline whilst a faint image of a tomato was projected on the screen. Several of them failed to notice the tomato, but reported imagining New York at sunset - CONCLUDED Perky effect does not show that mental images and faint percepts are inherently indistinguishable, the confusion between image and percept seems to occur because the processes involved in forming a mental image of the requested type interfere with the normal utilization of the mechanisms of perception
What are the important questions that Cognitive Scientists need to answer if they want to claim that imagery involves something like "pictures" in the head?
- are there compelling reasons for thinking that mental imagery involves picture-like representations BESIDES introspection? - Given that we don't think there are literal pictures in our heads/brains, should we say that mental imagery involves "picture like" representations in the head? But in what regards are these representations like pictures? - Given that we don't think there is an eye hidden somewhere in the brain that is looking at these pictures, should we say that our introspection of mental images involves something "eye-like"? But in what regards does this process involve something eye like? - What's the point of little picture-like entities in the head that are "seen" by eye-like mechanisms?
What is the "photographic fallacy" and how can it be used to respond to Dennett's arguments? What responses does Fodor give to Dennett's example of imagining a tiger? What response does Lyons give to the same case?
- argued that Dennett is operating with too narrow a conception of pictorial representation - consideration of photographs does not show that pictorial representation in general lacks this option - both stick figure drawings and photographs represent pictorially, they do so in very different ways - representation pictorially should not be forced into just snapshots - FODOR: It might be that there is some definite answer to the question, "how many stripes does my image-tiger have?" but that I cannot answer the question because images are labile (changing constantly, plastic); the problem is that we cannot hold onto our images long enough to count the stripes - Lyons says if you only getting a fleeting glance at an actual tiger you can't count the stripes either
What happens in Kosslyn's scanning studies? How do the results of this study suggest that imagery is a perception-like process?
- ask participants to imagine image like a boat, and scan from the end of the boat to the anchor at the front and compared scanning times - It took people longer to scan when the distance was longer - if imagery, like perception, is spatial, then it should take longer to mentally scan objects that are located further away from initial point - suggests imagery and perception are spacial - could be a distraction issue? critics say so: - now did it with a map and imagine flying from NC to FL and time it - then imagine NC to CAL and time it - took longer to scan all the way to CAL and no distractions in the middle - also did it with a small island map with spots on it
What happens in Kosslyn's study on distance and perceiving details? How do the results of this study suggest that imagery is a perception-like process?
- ask ppl to imagine elephant and rabbit - ask if the rabbit has whiskers and measure reaction for the answer - then asked to imagine rabbit next to a fly and ask same question → reaction time was faster than the rabbit next to elephant - Quicker to detect details on the larger image that fills the visual field
What is the difference between the beginning/end game, and the middle game, in chess
- begin from the same position so games take a while to differentiate - all positions including 6 or less pieces have already been solved - there's an opening book and an end book - middle is where it's most different and unique - you want to shrink the middle so that the opening and end connect - game begins when it gets out of book (novelty) and ends when it goes into book
EXTRA READING ABOUT RATS
- cognitive maps on rats? Spatial learning? - map had two groups! 1 based on the location E and W and one based on which way they turn (place learning and response learning) - 1st group did better to spacial information for animals is easier → evidence animals could form high-level representations of how environment laid out (COGNITIVE MAPS) - representations: stored info about the environment (fundamental of cognitive science) this is bad for behaviorism and good for cognitive
How does the evidence from brain damaged patients that Prinz discusses support Jackendoff's claim about consciousness arises at the intermediate level of visual processing?
- damage to the early visual areas should ordinarily eliminate activity in the intermediate visual areas, resulting in loss of visual experience (expectional cases with hallucinations) - damage to high level processing areas should leave experience intact because high level is postexperiential (so impaired visual recognition but vivid residual experiences) - if consciousness is located exclusively at the intermediate level, then extrastriate lesions and other areas that are intermediate in function should result in blindness (has been confirmed)
How does IBM's Watson work?
- extensive world knowledge and about to answer lots of questions - made to beat jeopardy (did it in 2011) - given access to lots of info (wiki, dictionaries, news, books, articles, etc.) - break down into keywords and phrases then search database to find answer - 1000s of algorithms - would do a secondary match to see if answer made sense - humans could come up with it faster when there's less contextual info
What are the challenges of speech recognition? (Why is it challenging to group phonemes into words?)
- first step in the process, whereby the acoustic speech signal is analyzed to determine the sequence of spoken words - need a speech spectrogram to extract phonemes from the segment of speech - phoneme to word is difficult because: - these concerns word boundaries (no pauses in spoken speech, hard to find start and end) - and phoneme variability (not all pronounced clearly and uniformly, variable in pitch and duration COARTICULATION - background noise - sound can be multiple words
How does Behaviorist methodology differ from that used by the Introspective Psychologist? What advantages are there to focusing on behavior versus introspective reports?
- focusing on behavior versus introspective reports - normally in a lab setting so it's repeatable phenomenon
What was Kohler's experiment? How did Sultan, the chimp, eventually solve the puzzle? Why did Kohler think that Sultan has arrived at this solution through insight and not through trial and error? What are some worries about Kohler's interpretation raised by Wynne and Udell?
- giving chimps a series of problems that required reasoning for the solution - block stacking task: banana hung from roof of enclosure of a reach w 3 big boxes - Sultan tipped box and climbed at horizontal distance and jumped to grab it - grabbed far away fruit by putting two sticks together but took an hour of playing - emphasis on sudden change btwn initial failure to success → higher mental faculties (insight) - but don't know much about animals previous training and 6 chimps at a time
What is the 3D sketch? What is represented at this stage of processing? (What are "geons"?)
- has info about depth but is more objective - not relative to a viewer perspective → neutral - GEONS: much simpler geometric shapes - uses trick to get complexity on the cheap
What is the algorithm/representation level? How is this level related to the phenomenon of ill-posed questions/problems? What occurs at this level of analysis?
- how does the system do what it does? - can we write the code to do this? - what assumptions, computations, representations? - how would we find out? Psychophysics! (asking people what they see or feel) can reveal the assumptions the human perceptual system uses to constrain ill-posed problems
What happens in Kosslyn's mental-walk task? How do the results of this study suggest that imagery is a perception-like process?
- imagine walking towards mental image of an animal - estimate how far away when the experience "overflow" (details were so close that they filled the visual field) - people had to move closer for smaller animals than larger → just like we would in real life - shows images are spatial, like perception
What is the partial report method? Using this method, what was the more accurate measurement of the capacity of visual iconic memory that Sperling arrived at?
- immediately after the stimulus disappeared, they would hear either a high tone, a middle tone, or a low tone. If they heard the high tone, they were to report letters from the top row of letters from within their visual iconic memory. If they heard the middle tone, they were to report letters from the middle row. Low tone, bottom row. - whatever tone was played, subjects were always able to report a decent number of the letters from the relevant row! - To determine the true capacity of visual iconic memory, Sperling took the number of letters from a given row that a subject could report when she heard the tone for that row, and multiplied that number BY 3 --- so the true capacity is 15 not 3-5
According to Christian, how was Deep Blue able to beat Kasparov in their last game? (And why should we think that this victory "doesn't count"?)
- kasparov bungled his 7th move and went into book → he could have defended better - he lost the game "in book" which isn't the real game
According to Miller's studies, what is the capacity of short term memory? (Is this capacity the same for both vision and audition (hearing)?)
- limited in absolute judgements (naming a color or identifying pitch) but not relative judgements - performance is good with 2 tones but falls off after 6 or more - similar with vision - our sensory systems are all information channels with roughly the same channel capacity (have built-in limits - 7 items at the same time (3 bits) - to increase capacity you can chunk information (relabel sequences of numbers with single numbers)
What is the cocktail party phenomenon? How does Broadbent's model explain this phenomenon? (What part of that diagram is meant to explain it?) What are the points of the arrows pointing backwards in Broadbent's diagram? (One of these arrow runs from "limited capacity channel" to "short-term store", the other runs from "store of conditional probabilities of past events" to "selective filter".)
- model of how sensory information is processed (psychophysics) - when at a cocktail party we can often hear many ongoing and unrelated convos but we can focus on just one we want to listen to.. how? auditory attention is selective (selective attention) - studied with dichotic listening experiments (different info in each ear) - information through senses to short term store and then sorted in selective filter for what we want to focus on for further processing (some people can hear the words of other convos but not extract meaning) - the arrow shows it's programmed by another system - information about the sorts of things that have led to that goal being satisfied in the past programs the selective filter - limited capacity info could be recycled back into short-term store to preserve if it is in danger of being lost
What evidence does Jackendoff cite for thinking our conscious experience arises at the level of the 2.5D sketch and not the primal sketch or 3D sketch (see p. 52)?
- of marr's 3 levels, only the 2.5d sketch corresponds to concious experience - neither too piecemeal nor too abstract
What is the the phenomenon of "latent learning" and how would a Cognitive Scientists describe what is happening in this phenomenon?
- rats were learning about the layout during the period without being rewarded → showed reinforcement was not necessary for learning and the rats were storing info when wandering even though no reward yet - what type of info is being stored? spacial layout or movement?
What are some of the strategies that we used by chatterbots in an effort to pass Turing's Test?
- relatively simple programs - These programs are typically pretty simple and come with a bunch of pre-programmed responses to certain questions - so, most people working in computer science and AI do NOT believe that chatterbots are actually intelligent or that they really "understand" what they are saying - one of the first to succeed was ELIZA (short script mimicking a psychologist) - PERRY imitating a paranoid schizophrenic (steer convo to preprogramed obsessions) - CATHERINE could carry on intelligent convos but only with bill Clinton - EUGENE persona was 13 yo Ukrainian boy (grammar and nonsequitors as language barriers) - CLEVERBOT statistically analyzing huge databases of convos to determine best responses (lack of personality and bad ability with new topics) - some store memories of previous convos to improve over time
How did each group perform in terms of its ability to learn how to navigate the maze? Which of these groups casts doubt upon the Behaviorist account of learning (and why)?
- rewarded rats quickly learnt to run the maze - unrewarded rats wardered aimlessly - BUT once 3rd group got rewarded they learnt the maze way faster than the 1st group - showed reinforcement was not necessary for learning
What is productive language? What is the babbling stage and when does it start? How do the sounds creating in the babbling stage relate to the infant's home language? What is the one-word stage and when does it start? What is the two-word stage and when does it start?
- starting to develop the ability to produce words - babbling: begins about 4 months, the stage of speech development in which the infant spontaneously utters various sounds at first unrelated to the household language - deaf babies will babble with their hands one word stage: from about age 1 to 3, speaking mostly in single words; now know sounds carry specific meanings → around 18 months capacity for learning new words jumps from one a week to one a day two word stage: age 2 speaking two word statements like telegraphic speech using mostly nouns and verbs and still doing it in correct order
What is pragmatic analysis? (How does this involve determining a "meaning" of a sentence that is beyond the meaning of words inside that sentence?) What are the kinds of statements described by Searle and how do these statements demand a response from the listener?
- the final stage of natural language understanding, produces a complete meaning for the sentence via the application of contextual info. this info includes data that have to do with the time and location of the utterance, who was saying it and to whom it was said - many sentences invite action on the listener's part → so these are the social rules that underlie language use Searle: - assertives: asserts his belief and suggests doing something about "it is hot in here" - directives: instructions dispatched from speaker to listener "turn down the radio" - commissives: commit the speak to a later action "I will take out the garbage later" that you need to verify/reward later - expressives: describe psychological state of the speak "I'm sorry for yelling at you" indicates sorrow and implies that the speaker trusts the listener - declaratives: utterance is an action "you are fired" now you must look for a new job
According to Kosslyn, what explains the picture-like quality of imagery? What explains our ability to scan or zoom in upon visual images?
- the info present in imagery is information in the VISUAL BUFFER - the VB is also responsible for the information present in our perceptual experiences - uses the same machinery as visual perception - so, if you think that perceptual experiences involve picture-like info, then so will imagery - doesn't fully answer the question, assumes that experience information is picture-like → but why is THAT picture like - In perception, the phenomenon of scanning, zooming, etc. is the result of various cognitive mechanisms operating upon info located in the Visual Buffer - visual imagery also operates using info from VB so the scanning present in imagery is the result of the SAME MECHANISM
What is syntactic analysis? How can syntactic analysis help an AI system determine the meaning of a word?
- the word sequence is analyzed via the use of knowledge of the language's grammar. this yields the sentence structure - perform the equivalent of using a phrase structure grammar to evaluate a sentence and break it down into its hierarchical constituents - understanding structure is necessary to get the meaning → if it's identified as a verb we know it represents action through *compositional semantics*
What kind of communication was PC Therapist III designed to use?
- topic of "whimsical conversation → not a topic but a manner - won Loebner prize - able to play off of a looseness in the contest rules, the natural behavior patterns of chatbots, and people's stereotypes about computers and humans - weird and funny in a normal sort of way - a jokester
According to Thorndike, what will happen when you put a cat in a puzzle box for the first time? How will the cat eventually figure out how to "solve" the puzzle?
- unable to see flashes of insight, apparent cleverness by chance and trial and error - well practiced cat could quickly escape and get its reward - if you give reward for action the action is stamped into the mind
What have brain scans revealed about perception and imagery?
- using brain imaging tech, both perception and imagery activate the *visual cortex* - saw an object, then object was taken away and they were told to imagine the object → same parts were active for both - Parietal, Temporal, Occipital, and Front are involved - problem of how images are stored and re remembered
What are the basics of how Deep Blue operates? To what degree does it use "search trees" to figure out what the best move is? What is a "heuristic" and how does it relate to Deep Blue's use of search trees?
1) a way to represent the board 2) a way to generate legal moves 3) a way to pick the best move - recursive minimax algorithm - could look at 300,000,000 positions per second - they need to have a point where you call off the search bc it goes so far - uses memoization: functions are stored and recalled - heuristic: kind of static guesstimate of how good that position seems, looking at things like who has more pieces, whose king is safer, etc.
What are the 3 levels of explanation Marr describes? Which level should we start with? What questions should we ask when working from that level? What is an "ill-posed question/problem"?
1. Computational Theory 2. Algorithm/representation 3. Hardware implementation comp theory - What info is being extracted and why? - What is the input? output? - can the output be straightforwardly inferred from the input? (nature of the computation) (usually no) - what extra knowledge/assumptions can make the problem tractable?
What are phonemes? Morphemes? Grammar?
phonemes: short, distictive sound units (a, t, ch) morphemes: the smallest units that carry meaning (speech is morpheme with s/puh/ee/ch phonemes) grammar: system of rules that enables us to communicate with and understand others
what is "introspection"
predates cognitive science - the method of turning one's attention inward to consider the current (conscious) contents of one's own mind - long been assumed that each of us has a kind of special privileged access to our own minds
observable behavior
public phenomenon that can be objectively measured and described
When it comes to translation, what's the difference between the "pure semantics" camp and the "pure empiricism" camp? Which camp does Google translation camp reside in?
pure semantics: tries to program linguistic understanding with the hope that the desired behavior will follow pure empiricism: tries to directly program linguistic behavior with the hope that understanding will either happen along the way or prove unnecessary - two approaches - Google uses statistical approach and leaves meaning out of it → regurgitates previous human translations (like cleverbot) - need a stable pov and consistency of style → need to be a specific person with a life
According to Marr, what is the purpose of vision? What is the input to vision? What is the output?
purpose: it represents the 3D layout of objects before the face in a way that allow us to recognize them input: light on the retina output: an internal 3d representation - it is an ill-posed question → how does the visual system start w retinal stimulation and end with 3d representation?
What is receptive language? When do infants start to do it?
receptive language: the ability to understand what's being said both to, and about us - as early as 4 months
What kind of communication was Eugene Goostman designed to use? (What are the advantages and disadvantages of this strategy relative to the strategy adopted by the designers of Cleverbot and Joan?)
runner up in 2008, 2005, and 2001 - need a single programmer to write responses so it has a personality
What is "situational awareness" and how is it related to an AI system's ability to understand what is being said?
social context is important and for AI to be natural and sophisticated they need this physically and socially
What is the primal sketch? What is represented at this stage of processing? (What is a zero-crossing?)
stage one of the algorithm/representation - starting with retinal simulation and then do operations on it that produce primal sketch zero-crossings: discontinuity in light intensity - little lines that show where light significantly changed - put together in little blobs in the primal sketch
What's the difference between surface structure and deep structure?
surface structure: organization of the sentence in the form that it is expressed in (how the sentence would be heard if it were spoken or read if it were written → variable and can be rearranged by transformational grammar deep structure: underlying meaning of a sentence and remains constant regardless of the specific form in which it is expressed
What is the basic set-up for Turing's Test? What does a computer have to do to pass this test?
test that he proposed in the 1950s for determining whether or not a computer could count as being genuinely intelligent - can a computer talk like a human? - a human judge has a text convo with unseen players, and evaluates their responses, to pass a computer must be able to replace one of the players without substantially changing the results (convo can't be distinguished from a human)
How did the psychophysics demonstration from the video show us something interesting about how color vision works?
the different color cars - all were Grayy - surfaces reflecting the same wavelengths of light → can appear different colors - What color is the illuminant - you must be solving by making an inference
What is the phenomenon of "change blindness"? Which of these phenomenon is relevant to demonstrations in the video?
the failure to notice a difference between what's there right now and what was there a moment ago video is change blindness
What is the fundamental idea of Cognitive Science?
the mind is an information processing machine (computer) - inner mental states = information processing states
How does this concern inspire the Behaviorist movement?
these concerns showed you could not build a science of the mind on introspection alone → not objective foundation → motivated Behaviorism
What, according to Turing, is the primary obstacle to a computer's passing the test? (What he right?)
thought memory storage was the issue BUT NO language is super complex - chatbots are baffled by pauses - need intuition to have a convo
How is Cognitive Science similar and different from Introspective Psychology and Behaviorism?
thread the needle btwn introspective and behaviorism - find a scientific (objective, reliable) way to explore inner mental states
2.5D sketch is not well suited to the project of recognizing objects.
true
How do we normally visually detect changes in our environment? What happens if the flicker or movement that is normally associated with an environmental change is hidden from the subject?
we visually notice change because our attention is drawn to the flicker or movement that associated with that change - The data these studies have generated suggests that we are actually kind of TERRIBLE at detecting change in situations like this
What is Simons' explanation for why so many people fail to notice the switch? How is his account a rejection of Marr's position?
when you look at another person you encode what's relevant (giving them directions) not what they're wearing - so the meaning hasn't really changed bc it's not something so imporatnt as gender - the goal of vision is not to build a complete picture and scene in your mind → it's to make sense of the meaning of the world around us
According to Turing (in 1950s), when will a computer be able to pass this test?
year 2000 with machines with 100 megabytes of memory
notion of universal grammar and How would the presence of a hard-wired (or innate) universal grammar help to explain how we learn language so efficiently?
yes languages have different rules but they also have many things in common → concept of universal grammar comprises the features that are instantiated in the grammars of all natural languages - each individual language at a fundamental level is not different from others but is a variation on a theme - considered as a collection of language rules hardwired into our brains from birth - modular aspect of mind and is a mental module: innate, genetically respecified, domain specific, and independent of other cognitive capacities - explain how we learn language so efficiently bc language acquisition requires the mastery of a large number of grammatical rules at different levels → the ease and rapidity humans learn this can be explained if there is at least some generic version of rules already present in the head at birth
How can the software/hardware distinction be applied to the mind? (And how does this help to explain how we can make progress understanding the mind even if we don't understand the brain?)
you can study the software of computer without knowing the hardware which parallels with understanding the mind without the brain
Why is it useful for our brains to have the ability to exert some top-down control over the contents of the visual buffer?
Information from the senses is often incomplete and degraded. This mechanism allows the brain to fill in some of the missing details - senses are pretty limited and don't work great all the time and this would ad info to the visual buffer and fill in missing detail
What non-human example do we have an organism learning a language through observation alone (and without any reinforcement)?
Kanzi the Bonobo - mom was learning english but he just tagged along → could pick things up like syntax faster than mom - language acquired spontaneously through observation without planned training and understood grammar, syntax
Has contemporary neuroscience found a single area of the brain corresponding to the primal sketch, another single area corresponding to the 2.5D sketch, and another single area corresponding to the 3D sketch?
Low: primary visual cortex (V1) along with sobcortical structures like visual nuclei of the thalamus and colliculus Intermediate: v2, v3, v4, v5 and also v3A and v8 (extrastriate areas just beyond v1 high: inferior temporal cortex such as TEO, TE, and portions of the STS, and LOC and some in VIP and PIP
What is Pylyshyn's account of imagery?
Mental imagery is an information processing phenomenon, is info is encoded in a word-like way, not a picture-like way - Our introspective experience is not reliable guide to how the info present is actually encoded
What is "reinforcement"? How does it related to "operant conditioning"?
Operant conditioning: method of learning that occurs through rewards and punishments for behavior - makes an association between a particular behavior and a consequence - behavior which is reinforced tends to be repeated; behavior which is not reinforced tends to die out-or be extinguished
What is an example of a mental process that can be described in information processing terms?
Perception (accessing info about 3d layout of the surrounding space) - memory (storing info to use later) - reasoning (drawing inferences to new info on basis of already known info) - action (using info to guide behavior)
What is "positive reinforcement"? What impact does it have upon a behavior?
Presenting the subject with something that it likes. e.g., Skinner rewarded his rats with food pellets - strengthens a behavior by providing a consequence an individual finds rewarding
What is "negative reinforcement"? What impact does it have upon a behavior?
Reward - in the sense of removing or avoiding some aversive (painful) stimulus. E.g., Skinner's rats learned to press the lever in order to switch off the electric current in the cage - strengthens behavior because it stops or removes an unpleasant experience hard to distinguish from punishment sometimes
What is a phrase structure grammar?
Sentences have distinct parts that are hierarchically related and the organization is called phrase structure - noun phrase and verb phrase → determiner, adjective, noun ; verb, noun - governs the use of phrase structures and imposes certain limitations on how a legitimate sentence can be put together - doesn't tell how we can rearrange a sentence to express new meanings
According to Skinner, how do young children learn what they need to know in order to progress through the above stages of productive language? How would Chomsky answer the same question? (What does Chomsky mean by "innate language" and ubiquitous grammatical categories/universal grammar"?)
Skinner: language is product of associative principles and operant conditioning → babies learn through reinforcement - if baby says mmm and mom brings milk then he associates them Chomsky: argued kids wouldn't reach full potention if learning was dependent on conditioning alone - innate learning and ubiquitous grammatical categories - says all languages share basic elements (Universal grammar): all contain nouns, verbs, and adjectives and humans are born with an ability to acquire language, and even a genetic predisposition to learn grammatical rules (we are HARDWIRED)
Kohler argued that Sultan solved the hanging-banana puzzle too quickly, and on the basis of too little "blind trial and error", in order for this to count as an instance of operant conditioning.
TRUE
What makes this form of short term memory "iconic"?
a form of short term memory that involves a brief, picture-like image - (The word "iconic" is meant to capture the fact that this stored information has a picture-like quality)
According to Behaviorists, how do rats, humans, and all other animals learn things?
all learning is the result of conditioning and conditioning depends upon processes of association and reinforcement - classical conditioning - strengthened association btwn a conditioned stimulus and an unconditioned stimulus - association eventually leads to organism to produce unconditioned response - ALL LEARNING TAKES PLACE THROUGH PROCESSES OF REINFORCEMENT AND CONDITIONING
What is "aphasia" and how can this phenomenon help us identify the parts of the brain responsible for language? What aspects of language are Broca's area and Wernicke's area responsible for?
aphasia: a neurological impairment of language, usually caused by left-hemisphere damage either to the Broca's area (impair speaking) or to Wernicke's area (impair understanding) - if you suffer a trauma to specific area that allows you to not speak but comprehend speech so we know that's where it's located - thinking and language are separate but also combined
What was the technique used in Perky's original study? (Did the subjects in that study report seeing the images that Perky projected on the wall? What evidence was there for thinking, despite what they said, subjects in this study were being influenced by these projected images?)
asked her subjects to fixate a point on a screen in front of them and to visualize various objects there, such as a tomato, a book, a leaf, a banana, an orange, or a lemon - subjects didn't know that a faint patch of color the right size and shape was projected onto the screen - none of the subjects ever realized they were experiencing real percepts, saw it as a product of imagination evidence: - some subjects expressed surprise at finding themselves imagining a banana "upright" rather than the horizontally oriented one they had been trying for - imagining an elm leaf after trying for a maple - also had actual imagined stuff such as vein of leaf and title of book - projected color patches in Perky's setup were clearly seen as such by witnesses who were not actively striving to form an image
What is a "Skinner box"? What is the basic idea of a "schedule of reinforcement"? What schedule of reinforcement is connected to gambling behavior?
box for the rats to measure responses - different patterns (or schedules) of reinforcement had different effects on the speed of learning and extinction - has effect on response rate and extinction rate - gambling: variable ratio reinforcement (hard to extinguish because of unpredictability)
What is a transformational grammar (and how does it relate to a phrase structure grammar)? What kind of conversions does transformational grammar allow us to do?
can change a sentence in three ways: active to passive, positive to negative, and question - need transformational grammar that allows us to transform one sentence into another bc phrase structure does not - set of rules for modifying a sentence into a closely related one - ex: the man read the book -> the man did not read the book - can use it to express two sentences that have different phrase structures but identical meanings
In the case of color vision, what's the function of color vision? Is it (color vision) an ill-posed question/problem? Is shape perception an ill-posed question/problem?
color vision - What is color vision for? to find fruit, to tell if it's ripe - where things are and the nature of those things - we want to determine the color of an object → but all we have is light coming to us from object → but light coming off the object is not only a function of the object but also of the light shining on the object - ill posed question: like saying AxB = 48, solve for a and b - we don't have enough information to come up with a solution implication: Inferring R from L requires other info or assumptions about L - many inferences in perception and cognition are ill posed like SHAPE PERCEPTION: each retinal image could have been cast by many diff objects, need other constraints to solve
What was Hofstadter's position on the ability to successfully play chess?
draws on central facets of the human condition, lie so close to the core of human nature itself - says computers brute-force will not be able to shortcut that fact - would need so much general intelligence that it wouldn't be considered a chess program at all
What is the primary methodology that the Behaviorist uses to study the mind?
early to mid 1900s - attempted to overcome some of the problems facing Introspective Psychology by shifting the emphasis away from introspective reports and towards observable behavior - meant to re-secure the foundations of psychology upon a more *secure, objective, and reliable* form of data - ignore or avoid any talk of the "inner causes of behavior" to get rid of subjective reports
What is the hardware/implementation level? What occurs at this level of analysis?
how is the system physically realized? (in neurons and brains, in computers, etc.) - need many levels of analysis to understand minds and brains and many methods to answer these questions - what are the nature of the neurons involved in solving this problem
What feature of your visual image of tigers is Dennett drawing our attention to? How does this feature cause trouble for the claim that imagery involves picture-like representations? How is this feature present in the examples of imagining a woman?
how many stripes does the mental tiger have? - question can't be answered though, our mental images don't contain that info - Mental images can be indeterminate with respect to visual properties (e.g., the number of stripes on a tiger). - Pictorial representations cannot be indeterminate with respect to visual properties. - So, mental images are not pictorial representations. - Even if you imagined a woman full-figure before you, your image likely did not go into sufficient detail to enable you to answer detailed questions - Your image can be inexplicitly noncommittal about whether she is wearing a watch, but a picture can only be explicitly noncommittal about it - a written description of a woman can be inexplicably noncommittal "our description might be very short, for example, which would make it impossible to tell whether the woman was wearing a watch or not. Dennett thus concludes that mental imagery has to be descriptional, and not pictorial"
In the case of someone asking for directions, approximately how many people failed to notice the switch?
1/2 the time 4/9 people didn't notice
what is "mental imagery"
internal representation of a sensory experience - see your dog playing and later you imagine it - tendency to focus on vision
What's the central concern facing introspective psychology movement? (What are some specific examples of this concern?)
lacks the objectivity and reliability necessary to support a "science" of the mind - ex: disagreements btwn subjects about when a stimuli can be consciously detected, the nature of some of our mental states (does thought involve imagery?) - hard to use on children and animals - what about mental illnesses?
What was Kasparov's track record against IBM's computer?
lost the first game - won next three and drawing the next two - 4 to 2 score - lost the next year -
What is the difference between the "form" and the "content" of a conversation? Which do we tend to focus upon when we are dealing with computers? With human beings? Which do we focus upon in a context, like speed dating, where we have limited time to communicate with someone?
machines authenticate on content, humans authenticate on form and verbal style - having a sense of a person and knowing about them are two different things - speed dating focuses on form to seem human and appealing → don't just answer questions like content
What kind of communication was Cleverbot (and Joan) designed to use? (What is "crowd sourcing" and how did this technique help Cleverbot come up with good replies to unusually specific questions?)
most human of 2005 and 2006 - very culturally attuned - borrows the intelligence of its users "conversational wikipedia" - thousands of prior convos to choose a response from (ghosts of real people) - doesn't have a unity of identity - crowdsourcing acquires an explosive growth in behaviors, but not consistent
What was the Behaviorist position on mental imagery?
originally philosophical thinkers assume imagery involves pictures in the head Behaviorism gave a hard turn: - DO NOT LIKE IT - find no hard evidence of imagery - "merely dramatizing" - NO introspection
what role does introspection play in the Introspective Psychology movement? (when did the movement take place)
A movement in psychology in the mid-late 1800s that attempted to construct a science of the mind on the basis of introspective reports of subjects
What is the phenomenon of learning though "insight"? How does this phenomenon cast doubt upon the Behaviorist claim that animals (and humans) only learn through operant conditioning?
According to Behaviorism, when an animal has to deal with a problem that it hasn't encountered before, it will basically try a range of random behavior and then, on the basis of positive or negative reinforcement (or even punishment), zero-in on the behaviors that seem to "work" - sometimes an animal seems to pick up a new skill suddenly and on the basis of hardly any trial and error. This phenomenon is know as "insight" - it seems to suggest that animals are NOT limited to learning things thought operant conditioning (trial and error)
