Exam 2 Flashcards

Question

what is more favoured by modern science? linguistic determinism or relativity?

Answer 1

relativity

Answer 2

- space - number - substance - colour - exact vs approximate math

Answer 3

- langauges vary in their spatial terms - relative (left/right/in front of) depend on the viewer's position relative to the object - absolute (north, south, up/down the mountain) do not depend on the viewer's position - most languages have both but some (e.g., Tzeltal) have only absolute | not “the spoon is to the right of the cup” but “downslope of the cup”

Answer 4

- they confuse mirror images but have excellent ability to identify absolute directions (even indoors) - e.g., "does the hot water come out of the upslope or downslope tap?"

Answer 5

Tzeltal or Dutch speakers asked to memorize the array and “make it the same” - Tzeltal --> Absolute match - Dutch --> Relative match

Answer 6

1. reasoning, not interpretation: What does “make it the same” mean language just seems to nudge participants one way or the other in an ambiguous situation. 2. determinism, not habit: unlikely – english speakers are worse at telling directions, not hopeless. experiments find that both English speakers and Tzeltal speakers are capable of using both absolute and relative coordinate systems. 3. causation: Tzeltal speakers live on the slope of a mountain, where absolute directions are very useful – so couldn’t it just be that the language reflects the coordinate system that is most useful for them? Another Mayan people (the Tzoltil) speak a language with both types of terms – and they behave just like the Tzeltal on the arrangement task.

Answer 7

- vary in the number and precision of their number words - many languages used by hunter-gathering people (e.g., Piraha) have only 3 number words – “one”, “two”, “many.

Answer 8

- they use these words in an imprecise way (“hoí” doesn’t mean “exactly 2”) - they don’t use them recursively (e.g., “hoí-hoì”) - they are very bad at tasks that require counting

Answer 9

1. reasoning, not interpretation: yes – this is a genuine failure in an objective task. 2. determinism, not habit: yes – Piraha don’t ever seem to be capable of basic counting or number-related skills. 3. causation: no hunter–gatherer language has an elaborate counting system, and no language that is widespread in urbanized societies lacks one. It’s much more likely that reverse causation is true: hunter–gatherer people have no need for counting with large numbers, and languages evolve to introduce number words when the need arises. Other hunter–gatherer languages with moderately complex number systems (e.g., Mundurukú) also use them approximately rather than precisely. The key seems to be having a counting algorithm (pairing objects with number words) rather than the words themselves

Answer 10

- count: "apple" "pebble" typically refer to bounded objects - mass: "applesauce" "sand" typically refer to unbounded stuff they follow different grammatical rules

Answer 11

- pluralization: two pebbles vs /two gravels - quantifiers: a pebble vs /a gravel; many pebbles vs /many gravel - bare singular: gravel is expensive vs /pebble is expensive

Answer 12

2-year-olds were presented with unfamiliar objects and substances - naming phase: “this is my tulver.” - test phase: “point to the tulver.” - object trials → shape match - substance trials → material match syntax didn’t help because they had not learned the count/mass distinction. this means construals came before the linguistic distinction

Answer 13

languages require us to pay attention to different things in order to form sentences

Answer 14

- tenses (e.g., english) --> when something happened - evidential inflections (e.g., Turkish) --> how something is known (hearsay vs direct observation) - english spatial terms (e.g., ‘in’ vs ‘on’) --> support vs containment - Korean spatial verbs --> whether the fit is tight or loose

Answer 15

speaking a language may ingrain mental habits that we follow even when we aren’t speaking

Answer 16

- languages carve up the colour spectrum differently - unlike english, russian has words for light blue (goluboy) and dark blue (siniy) but no word that includes both

Answer 17

- english and russian speakers were given a colour discrimination task - trials could be easy (e.g., crossing 8 gradations) or hard (crossing 3 gradations) - trials could cross the goluboy/siniy boundary or not (e.g., swatch 2 vs 5 or swatch 9 vs 12)

Answer 18

on hard trials (not easy), russian speakers were better for trials that crossed boundaries

Answer 19

- solving complex problems requires a memory buffer - e.g., trying to add 3+5 vs multiplying 17 * 7 - in the more complex case, you are breaking the problem into a sequence of steps and storing intermediate results

Answer 20

- visuospatial sketchpad: visual working memory (storing visual information about shapes and colours; maybe how you solved 17 * 7). - phonological loop: uuditory working memory (storing ~2 seconds of auditory information; usually how we memorize lists of digits).

Answer 21

we might exploit the sounds of language as a memory hack for the phonological loop

Answer 22

participants were trained on a set of math problems (e.g., 54 + 79) in one of the two languages they were tested on both exact problems and estimation problems: - old problems in the original language - old problems in the other language - new problems in the original language

Answer 23

switching languages had as big of a cost as switching problems – but only for exact problems

Answer 24

- many studies purport to demonstrate linguistic determinism – the idea that language constrains thought because we think in words – but the evidence is remarkably weak - weaker versions of linguistic relativity could have some truth – cultivating habits of attention, circumventing bottlenecks in working memory - language appears to be primarily for communication, not for thought

Answer 25

- binding problem - frame problem - grounding problem - qualia problem

Answer 26

how does the mind integrate different features of an object into a unified representation?

Answer 27

- when you see an apple, you separately process its color, shape, and motion, yet perceive a single whole (intramodal binding) - when you hear someone speak, you integrate visual information about their lip movements and auditory information about the sounds (cross-modal binding)

Answer 28

- these different types of information are processed separately (by different “modules” in distinct parts of the brain), which has important computational and evolutionary advantages - but how does the information get put together? - there isn’t a little person (“homunculus”) who lives in your head to observe and integrate these streams

Answer 29

1. object files 2. predictive processing

Answer 30

- early stages of visual processing are automatic and “preattentive,” in which features are independently processed - later stages require attention to select part of a “master map” of locations at which features were detected - when we attend to a location of the master map, features at that location are bound and stored in an “object file”

Answer 31

- the brain generates predictions (as unified, coherent representations) about potential causes of sensory input - features are compared to those predictions, and the brain attempts to minimize prediction error by selecting the most likely interpretation

Answer 32

how do we update the web of information in our minds without having to check every possible implication?

Answer 33

- a robot is trying to save its spare power supply, which is on a cart that has a bomb on it - it pulls the cart out of the room, but the bomb is still on it - okay, so the robot considers all the implications of its action - the bomb explodes while it’s thinking about whether the wallpaper will change colour

Answer 34

the robot must distinguish between relevant and irrelevant implications, but doing so still requires it to consider an infinite number of possible implications

Answer 35

- it’s critical to update our beliefs in light of new actions or knowledge, yet only a small number need to be updated - how do we know which ones to update? - it’s not feasible to check every possible one, nor do we seem to

Answer 36

how are symbols in the mind connected to things in the world? that is, how do symbols refer to real things?

Answer 37

the chinese room - a monolingual English speaker uses a rule book to match Chinese characters that are input into the room to outputs he writes down - he is manipulating symbols according to the rulebook, but he does not understand Chinese

Answer 38

- a purely syntactic system – using only symbol manipulation – might be able to get the right answers by consistently applying the rules of logic - but how could those answers be “about” anything? - it’s like trying to use a dictionary where the entries are all in a foreign language

Answer 39

how (and why) do we have phenomenal or subjective experiences that accompany certain forms of information processing?

Answer 40

- when our eyes and brains process a certain wavelength of light, we have an experience of redness - when our skin and brains process certain tactile information, we experience it as pain - when a bat relies on echolocation to determine the locations of objects, what (if anything) “is that like”?

Answer 41

- it does not seem like these experiences play any necessary causal role in guiding action, so why do we have them? - we have absolutely no idea about how qualia could arise from the brain, or even how they are consistent with the laws of physics

Answer 42

- we’ve made progress on some (the binding problem) - we need better theories for others (the frame problem, grounding problem) - some might remain mysteries (qualia)

Answer 43

- when we look at a non-living natural object like a stone, it could easily have arisen through random processes or even existed forever. - conversely, when we look at a man-made object like a watch, its design is obvious - its parts are perfectly arranged to carry out a function - it is exceedingly improbable such a thing would come about without being intentionally designed – there are no watches on the moon - but the design of lifeforms is far more sophisticated – more like watches than like rocks

Answer 44

- they believed in an intelligent designer - there was no other explanation of life's design

Answer 45

humans had domesticated dogs from wolves - perhaps god domesticated us?

Answer 46

entities that can make copies of themselves

Answer 47

1. variation in that trait 2. differential consequences for different variants 3. transmission of variants from one generation to the next

Answer 48

3d molecular machines that carrry out all sorts of biological activities

Answer 49

sequences of just 22 amino acids

Answer 50

dna is a code that tells our cells what proteins to manufacture

Answer 51

millions of nucleotides (a, t, c, g) in a specific order

Answer 52

- sets of 3 nucleotides that code for a particular amino acid - example: cta --> leucine; cca --> proline

Answer 53

a sequence of codons that codes for a protein

Answer 54

different versions of a gene (swapping a nucleotide or inserting/deleting a sequence)

Answer 55

- random copying errors - example: if cta were miscopied as cca, a leucine would be replaced with a proline, which could change the protein it codes for

Answer 56

- the sequence of nucleotides in the genetic code - this is where variation comes from

Answer 57

- the observable consequences for the organism - this is why variation matters

Answer 58

- physical characteristics (height, eye color, facial symmetry) - biological characteristics (genetic diseases, innate immune system) - cognitive characteristics (risk-taking, perceptual abilities)

Answer 59

- the squares have a genetic code much like ours - once in a while, a copying error is made that changes a square’s colour - aaa = blue; aat = red; aca = green

Answer 60

some phenotypes change the odds that an individual with that phenotype survives or reproduces

Answer 61

the environment

Answer 62

factors that affect survival and reproduction

Answer 63

- predator avoidance – speed, camouflage (depends on environment, e.g., tundra vs jungle) - viability of offspring – incest avoidance, resource investment - resource gathering – size (for predation), specialized body parts (e.g., pelican’s pouch)

Answer 64

suppose squareworld is inhabited by triangles (predators) with poor vision: - a blue square has a 50% chance of being eaten - a red square has an 80% chance of being eaten - a green square has a 5% chance of being eaten

Answer 65

differential consequences (survival/reproduction) for a variant of a replicator will change the frequency of that replicator

Answer 66

the average number of copies a replicator leaves in the subsequent generation

Answer 67

over subsequent generations, an organism becomes more adapted because its genes code for phenotypes that increase fitness

Answer 68

the proportion of green squares increases over generations: - generation 1 --> 10% green - generation 2 --> 16% green - generation 3 --> 33% green - generation 4 --> 50% green

Answer 69

- replicators are entities (like genes) that make copies of themselves. in genetic evolution, different genotypes result in different phenotypes - genes encode instructions for building proteins, resulting in different phenotypes which can be transmitted across generations - if a phenotype leads to greater fitness, its genotype will become more common in the next generation – natural selection

Answer 70

ground finches use their beaks to crack seeds.

Answer 71

it depends on many genes, so it can continuously vary and is transmitted across generations

Answer 72

- finches with deeper beaks can crack bigger seeds as adults - finches with deeper beaks are less likely to survive to adulthood

Answer 73

- there were fewer small seeds. during the drought, only finches with deeper beaks could crack the remaining (large) seeds - after the drought, finches with shallower beaks could crack seeds and were more likely to survive to adulthood

Answer 74

- we have our ingredients – variation, consequences, and transmission - polygenic traits can vary continuously even though each allele is binary - multiple selection pressures can push in opposite directions - changes in the environment change the balance of selection pressures - we ordinarily think of evolution as slow, but strong selection pressures can produce rapid changes

Answer 75

- sight is useful in all sorts of ways, but the eye seems far too complex to evolve all at once - it has way too many parts to arise simultaneously, and “part of an eye” isn’t useful — like an iris without a retina

Answer 76

through small, continuous steps, each providing a fitness advantage. darwin wrote: “... if numerous gradations from a simple and imperfect eye to one complex and perfect can be shown to exist, each grade being useful to its possessor, as is certainly the case; if further, the eye ever varies and the variations be inherited, as is likewise certainly the case and if such variations should be useful to any animal under changing conditions of life, then the difficulty of believing that a perfect and complex eye could be formed by natural selection, though insuperable by our imagination, should not be considered as subversive of the theory.”

Answer 77

- light-sensitive spot (many invertebrates): gives information about light intensity, e.g., presence of a predator - depression – eye pit (limpet): gives information about the direction of the light - deeper depression – eye cup (beyrich’s slit shell): gives more precise information about direction. pinhole eye (california abalone): forms images on light-sensitive tissue - closed eye with transparent cover (turban shell): protects from damage and parasites - lens eye (atlantic dog whelk): focuses light to make more precise images.

Answer 78

- each transitional form provides a fitness advantage over the previous one - even though each step is unlikely, it’s not astronomically unlikely because each step is small - eyes have evolved dozens of times independently – convergent evolution

Answer 79

after copulation, males literally hook themselves to the female for hours

Answer 80

- benefits: prevents other males from copulating, reducing sperm competition - costs: time-consuming, prevents the male from mating with others or doing anything else

Answer 81

the sex ratio in the population — more females makes it less advantageous; more males makes it more advantageous

Answer 82

adjusting behaviour in response to the sex ratio could optimize fitness, but it has costs — time, energy, and risk of error

Answer 83

- florida --> consistent 1:1 sex ratio - oklahoma --> variable sex ratio (can vary from 2:1 to 1:2 across regions)

Answer 84

- oklahoma bugs adjust their behaviour – it’s plastic - florida bugs don’t – their behaviour is canalized

Answer 85

plasticity itself can vary between individuals and be heritable

Answer 86

behaviours, not just physical traits, can evolve via natural selection if variable, consequential, and transmissible there are two ways behaviour can adapt: - canalization: behaviour is pre-programmed and evolves across generations - plasticity: behaviour is flexible and changes within an individual - plasticity isn’t always better — depends on benefits (higher when environment is unstable) and costs (time, energy, errors, nervous system complexity)

Answer 87

in many species (e.g., baboons), males are bigger than females

Answer 88

in terms of gamete size — females produce few large gametes (eggs); males produce many small gametes (sperm)

Answer 89

- females are more limited in how many offspring they can produce per time period - males are less limited - females invest more in lactation, nurturing, and protection (in many species) - females are choosier about mates - males have higher variance in offspring numbers and more competition for mates

Answer 90

because fighting ability applies more to males in species with greater intrasex competition for mates

Answer 91

- larger body size --> better at fighting, less vulnerable to predators. - smaller body size --> faster maturity, less food needed.

Answer 92

because it’s costly signaling: since a male can pay this cost and survive, it signals high genetic quality to females

Answer 93

energy to grow and maintain, and increased difficulty hiding or fleeing from predators

Answer 94

because females are choosier; to males it’s worth the cost to compete for females, while for females it usually isn’t

Answer 95

because of sexual selection — competition and display are stronger for the less choosy sex, usually males

Answer 96

- selection pressures can come from one’s own species (sexual selection). two kinds of sexual selection: intrasex competition (e.g., fighting among males) and mate choice (e.g., costly signaling) - sexual selection pressures usually stronger for the less choosy sex - for costly signals to work, they must be visible and reliably indicate hard-to-observe traits (like gene quality). - selection acts at the individual level, not species level — even if the species as a whole would be better off if males stopped competing, each individual male benefits from winning.

Answer 97

from an evolutionary perspective, we should be selfish — only the fittest survive

Answer 98

if selfish individuals survive best, why do some species cooperate at a cost to themselves?

Answer 99

finding ways that seemingly selfless behaviour is actually selfish — aligning individual and social incentives

Answer 100

family (genetic relationship), friends (repeated interactions), and strangers (within a larger group)

Answer 101

- the replicators are genes, not individuals - your genes can benefit from sacrificing you if it helps produce more copies of themselves

Answer 102

- kin selection - reciprocal altruism - reputation signaling

Answer 103

if (i) b is sufficiently genetically related to a, and (ii) the help raises b’s fitness enough relative to the decrease in a’s fitness

Answer 104

“would i lay down my life to save my brother? no, but i would to save two brothers or eight cousins.” – j.b.s. haldane

Answer 105

- a helps b with the expectation that b will help a later - it’s useful because individuals access different resources at different times, but risky due to possible exploitation

Answer 106

when individuals interact repeatedly with the same partners

Answer 107

- a helps b because c might observe the helping behaviour - this communicates to c that a is a reliable social partner, increasing chances of future cooperation with c (or others c gossips with)

Answer 108

- the environment imposes multiple selection pressures that can push traits in opposite directions - the balance of pressures can vary across geography, time, or sexes - what is good for a gene’s success (replicator fitness) is not always good for the species.

Answer 109

a phenotype that arises not directly via selection, but as an indirect consequence of an adaptation

Answer 110

heat from lightbulbs; arches and spandrels

Answer 111

- belly buttons (umbilical cord is the adaptation) - mozzi’s howling at fire alarms (communication with his “pack” is the adaptation) - art (preference for symmetry and landscapes may be the adaptations)

Answer 112

a trait becomes more or less prevalent in the population just through random chance

Answer 113

for ‘neutral’ variants that have no effect on fitness

Answer 114

smaller populations (e.g., bottlenecks caused by random catastrophes)

Answer 115

non-coding regions of dna

Answer 116

a visualization of all possible combinations of a species’ structural traits, such as size, coiling, and elongation in seashells

Answer 117

most of it is empty

Answer 118

there are no animals with wheels

Answer 119

- lack of possible mutations (no set of genes could produce that trait) - not adapted to any environment (e.g., wheels aren’t useful for natural terrain) - violate laws of physics or chemistry (e.g., wheels not physically feasible – though we know this last one is false)

Answer 120

- modern organisms are descended from ancestors whose phenotypes caused their genes to spread more in past generations - we don’t need to be “trying” — a trait can have adaptive value without awareness or intentional deployment

Answer 121

- every animal has adapted to its environment - differences among species reflect different environments and survival strategies, not “progress.”

Answer 122

most genes affect many traits (pleiotropy), and most traits are influenced by many genes (polygenic)

Answer 123

- not all evolution is due to natural selection, but evolution that produces “design” generally is - evolution is limited by genetic material, possible environments, and the laws of physics - evolution is a blind process, even to the organisms that are evolving.

Answer 124

the complexity of life has emerged from natural selection, which requires variation, fitness consequences, and transmission

Answer 125

yes — drift and spandrels — but natural selection is the only unintentional process that explains design

Answer 126

pressures imposed by the environment, including other members of the species: - competition for scarce resources - predation by other species - intrasexual competition for mates - sexual choice by mates

Answer 127

the ability to attain goals in the face of obstacles by means of decisions based on rational (truth-obeying) rules

Answer 128

bacteria can orient toward certain chemicals and away from others

Answer 129

movement of an organism in response to a stimulus

Answer 130

when an attractant or repellant chemical is sensed, runs (straight-line propulsion) become more common and tumbles (random turns) less common

Answer 131

to reproduce, replenish proteins, and repair dna

Answer 132

- photosynthesis: obtaining and storing energy fro sunlight (plants and some bacteria) - respiration: obtaining and storing energy from sugar (other life) (animals, fungi, and some bacteria)

Answer 133

an explosion of lifeforms: eukaryotic cells, multicellular life, and eventually enimals including those with brains

Answer 134

waiting vs killing

Answer 135

aiting + external digestion: releasing spores that float around, growing filaments when they land on a carcass, and releasing enzymes that digest tissue externally

Answer 136

killing + internal digestion: using an inner cavity for digestion and releasing enzymes internally

Answer 137

it required rapid and complex decision-making, including movement

Answer 138

1. internal digestion 2. neurons 3. muscles

Answer 139

- all-or-none firing: a neuron can’t “half-fire”; once threshold is reached, it fires. - rate coding: firing rate encodes information about continuous quantities - adaptation: threshold adjusts to higher or lower levels of stimulation (e.g., light receptor cells) - chemical synapses: neurons communicate via neurotransmitters - excitatory and inhibitory connections: firing can make connected neurons more or less likely to fire

Answer 140

features are conserves across species, and neurons appear to have evolved only once

Answer 141

coral polyps and jellyfish

Answer 142

a net of independent neural circuits, each with a reflex

Answer 143

simple goal-directed, environmentally sensitive behaviours

Answer 144

swallowing – requires one set of muscles to be active and another to be inactive; a simple circuit of excitatory and inhibitory neurons can implement this

Answer 145

nematodes - tiny worms

Answer 146

1. bilateral body plan 2. valence-responsive neurons 3. input integration 4. context-sensitivity

Answer 147

- radially symmetric animals (like coral or jellyfish) would need to detect and move in any direction to move toward food - bilaterally symmetric animals (bilaterians; ~99% of animal species) can “steer” toward food using a simple algorithm: if food smell increases, keep moving forward or if food smell decreases, turn - n ematodes can’t see, but they still find food rapidly by circling in on it

Answer 148

- valence: goodness/badness of a stimulus --> steer toward good things, away from bad things - in simple organisms like nematodes, valence is hard-coded - when a positive (vs negative) stimulus is detected by a sensory neuron, the organism is wired to steer toward (vs away from) that stimulus

Answer 149

- some stimuli aren’t always good or always bad (e.g., temperature or food smell) and whether to approach/avoid should depend on the current state - CO2 is released by both food and predators - nematodes will approach CO2 when hungry, but not when full

Answer 150

1. neurons with different threshold: approach higher temperature if too cold; colder temperature if too hot. have a “too hot” neuron and a “too cold” neuron 2. chemicals that track internal states: approach food if insufficient energy; do not approach if sufficient. animals release hunger signals (e.g., ghrelin) versus full signals (e.g., leptin). neurons that detect food smell become more responsive when hunger signals are present and less responsive when full signals are present

Answer 151

- mere approach/avoidance responses fail when stimuli are not continually present - “moods” help organisms to maintain a behaviour pattern triggered by a stimulus even when that stimulus is transient

Answer 152

- arousal (high v low): amount of movement - valence (positive v negative): type of movement

Answer 153

- chemicals that have widespread and long-lasting effects on other neurons - different from signaling of excitatory vs inhibitory neurons, which only affect the specific neurons they connect to - released by specific neurons (e.g., dopamine neurons) - neurons have different numbers and types of neuromodulator receptors - have various effects (e.g., inhibition, activation, changing sensitivity) on other neurons that have the matching neuromodulator receptor

Answer 154

- triggered by nearby reward - activates arousal, pursuit (exploitation)

Answer 155

- triggered by consumption of reward - inhibits reward pursuit (satiation)

Answer 156

- triggered by negative stimuli - activates fight-or-flight response (escape) - downregulates energy-consuming activities like digestion, cell growth, immune activity (acute stress response)

Answer 157

- many stimuli aren’t good or bad in themselves, but because they are correlated with something else - a chemical that is released by a predator - the saltiness of water can be associated with what food grows there - in organisms without brains, such associations must be hard-coded

Answer 158

- the environment changes faster than biological evolution can keep up - leads to a selection pressure for an organism to learn from experience within its lifetime

Answer 159

- reflexes that are conditional and depend on the animal's experience with that stimulus - evolutionarily ancient – present in bilaterally symmetric animals - involuntary - the first form of memory

Answer 160

it changes

Answer 161

strengthening of response with more exposures to an association

Answer 162

weakening of response if the correlation weakens

Answer 163

- after a delay, repsonse re-emerges - a form of ltm - broken association could be transient

Answer 164

- an old association is easier to (re)learn than a new one - an old association is less likely to be a fluke than a new one

Answer 165

the world contains many stimuli at any given time - how do we know which association to learn?

Answer 166

1. eligibility traces: a stimulus only is “eligible” to form an association for a short period of time (e.g., 2 seconds) 2. overshadowing: more intense cues (e.g., a strong smell) ”overshadow” weak cues (e.g., a dim light) 3. latent inhibition: cues that have been commonly experienced in the past are less likely to form associations (new ones are preferred) 4. blocking: once a cue is established, it “blocks” other cues from forming associations

Answer 167

during the cambrain explosion - a period of huge diversificaiton of life

Answer 168

they share a similar brain structure and developmental path

Answer 169

outcomes with positive/negative valence reinforce/punish recent actions, making them more/less likely to occur

Answer 170

evolved in vertebrates (mammals, reptiles, fish) but not in simple bilaterians (nematodes)

Answer 171

- which action in a (possibly long) sequence of actions should be reinforced? - the bilaterians’ tricks for associative learning won’t work because the actions occur over an extended period of time

Answer 172

- this is not just a problem for animals, but a problem for AI that was identified in the 1960s and not solved until the 1990s - example: Getting a computer to play a game like checkers. which moves should be reinforced?

Answer 173

No, because early actions might have been more important

Answer 174

- with enough experience, maybe the “right” moves will get reinforced - no, because there are too many possible combinations

Answer 175

use predicted rewards rather than actual rewards to reinforce behaviours - critic: predicts the likelihood of success after each action. e.g., evaluating the board position for your chance of winning - actor: chooses which action to take and is rewarded when the critic thinks the action increases the likelihood of success - the signal for learning is the temporal difference in predicted reward from moment to moment. e.g., the moment when your maneuver puts you in a position to win the game

Answer 176

- implement the reinforcement signal - triggered by the prediction of the reward not the reward itself (once association is learned) - more sophisticated than the role of dopamine in early bilaterians: repurposed from a signal of nearby reward to a signal of predicted reward

Answer 177

1. time: dopamine release tracks the cue–reward delay and shorter delays produce larger spikes 2. probability: cues with a stronger association with reward produce larger dopamine spike 3. omissions omitted rewards are punishing (disappointment) and omitted punishments are rewarding (relief)

Answer 178

- basal ganglia - input from cortex (monitors environment and action) - output to motor system (inhibits them by default) - both input and output to/from dopamine neurons in the hypothalamus

Answer 179

learning to repeat actions to maximize dopamine

Answer 180

- the hypothalamus responds to actual reward, not predicted reward - at first, actual reward is reinforced - later, the basal ganglia learns to predict the hypothalamus’s reward signal - the inhibitory output to the hypothalamus ”cancels” the actual reward signal

Answer 181

- specific neurons detecting specific things: can recognize few things and new things require evolutionary change - patterns of neurons encode things: can recognize more things and learn to recognize new things patterns, but vertebrates can | early bilaterians couldn’t recognize patterns but vertebrates can

Answer 182

neuron action patterns of different stimuli can overlap

Answer 183

- dimensionality expansion – smaller number of olfactory neurons connect to larger number of cortical neurons - sparseness: olfactory neurons only connect to a few of those neurons - the result is pattern separation – the activation pattern of cortical neurons is much less overlapping than for olfactory neurons

Answer 184

similar stimuli (which should be grouped together) don't have identical activation patterns

Answer 185

- cortical neurons synapse back to other cortical neurons - so a partial pattern will reactivate the other neurons and complete the pattern (auto-association) - this allows the cortex to recognize a pattern that is similar but not identical

Answer 186

- requires more energy - can hunt when it is cold (e.g., at night)

Answer 187

- early warm-blooded animals – therapsids – became dominant - went nearly extinct when food became scarce (Permian-Triassic mass extinction) - the surviving cynodonts (small plant-eating therapsids) evolved into the first mammal

Answer 188

- first-mover advantage - this led to a selection pressure for simulating actions before they occurred – not learning by doing, but learning by imagining - these computations are more energy-intensive and time-consuming than anything we’ve seen so far and could probably have evolved only in warm-blooded animals

Answer 189

the new brain structure --> 6-layer structure that surrounds most of the brain

Answer 190

- vision, hearing, smell - touch, pain - movement - language and even music (in humans)

Answer 191

- in vertical columns - neurons within a column all respond to similar stimuli - many connections within a column; few connections across columns - different columns are wired the saem way and are interchangeable - the inputs vary, not the computation (swapping regions in ferrets; repurposing of cortez in congentially blind or stroke patients

Answer 192

- that perception is inference - we perceive a simulation of the world, not the world itself - properties of perception that suggest this is true: filling-in, one-at-a-time, can't-unsee

Answer 193

- the neocortex may be implementing a generative model in the same spirit as generative AI systems - the “Helmholtz machine” model for recognizing handwritten digits

Answer 194

- we hallucinate a possible model of the world, which is compared to (contrained by) sensory input - neocortex appears to be continuously predicting its sensory data – i.e., simulating the entire world around the animal

Answer 195

- charles Bonnet syndrome – visual hallucinations in people whose eyes stop sending signals to neocortex - perceptual hallucinations during sleep (“dreaming”) - imagination and perception compete for cognitive resources - the same neurons are active during imagination and perception

Answer 196

1. vicarious trial and error 2. counterfactual thinking 3. episodic memory 4. fine-grained motor control 5. models of one's own behaviour 6. theory of mind (at least in primates)

Answer 197

imagining the reward from different possible actions

Answer 198

- when rats reach forks in a maze where the choice isn’t obvious, they toggle their heads back and forth – like they are imagining each option - when they do this, the “place” cells in the hippocampus corresponding to the imagined locations are activated

Answer 199

- works even when past actions were not rewarded (e.g., learning the path around a barrier vs using that knowledge to seek reward) - works even when past rewards are not predictive of current rewards (e.g., rats ordinarily avoid salt, but a salt-deprived rat will approach an area of a maze with salt)

Answer 200

imagining what would have happened had a different action been taken

Answer 201

- “restaurant row” experiment: a tone at each “restaurant” indicates the length of delay; the food options vary in their desirability. rats look back (suggesting “regret”) if they make a choice that does not turn out well - rock-paper-scissors in monkeys: if they lose, they choose the move that would have won in the previous game

Answer 202

can assign credit to the actions that, had they not been taken, would have prevented the outcome

Answer 203

- memory for specific past episodes in life - remembering is simulating - the same brain areas for imagining the future and recalling the past - suggestibility of memory (inferential, like perception)

Answer 204

rats can learn to run down one path if they had encountered food a few minutes previously and another path if not

Answer 205

a partnership between the hippocampus (rapid pattern recognition for storing new memories) and neocortex (retrieves memories via simulation, reactivating the hippocampus)

Answer 206

motor cortex

Answer 207

all animals can control motor movements, but only (some) mammals have a motor cortex

Answer 208

motor cortex is only needed for learning new movements

Answer 209

the “active inference” theory

Answer 210

motor cortex is learning to predict the body movements from the adjacent somatosensory cortex

Answer 211

- the motor cortex predictions flow to the spinal cord to control movement – to make the predictions true - these simulations train the basal ganglia, which can eventually learn the movements as an automated (rather than planned) behaviour

Answer 212

because it is not always clear when or what to simulate

Answer 213

chimps can hide food from one another and mislead about its location

Answer 214

- in sensory neocortex, inputs from external sensors - in apfc, inputs from hippocampus (places, memory), hypothalamus (valence), and amygdala (internal affective states) - the apfc seems to be inferring intentions - a model of one's behaviour - possibly different columns each make a prediction, and simulation is triggered when predictions diverge which simulations are triggered? the ones that were predicted by different columns

Answer 215

intentional and accidental actions (e.g., a box marked accidentally with a line that ordinarily signals food)

Answer 216

prefer an experimenter who is unable to give food (can’t see it) versus unwilling to give it (has it but won’t give it)

Answer 217

strategic alliances and deception help to navigate and climb social hierarchies

Answer 218

- in primates, new cortical regions (e.g., gpfc) take input from apfc (the model of behaviour) - inferring mental states – a model of one’s intentions (themselves a model of behaviour) - evidence: the same regions are also active when completing theory of mind tasks – we use the same computations to model our own minds and others’ minds - these simulations of minds also appear to have given rise to imitation learning (e.g., to use tools) and anticipating future needs

Answer 219

- steering - associative learning - reinforcement learning - pattern recognition - simulation

Answer 220

we sometimes can’t get to one step without another - affect isn't useful without steering around an environment - simulation-based trial and error won't work without reinforcement learning - motor prediction, self-models, theory of mind all depend on similar neural architecture

Answer 221

neurological constraints interact with selection pressures to determine what brain structures and resulting forms of intelligence evolve

Answer 222

- neurons: with the opportunity to adopt “kill” vs “wait” strategies when respiration became viable - simulation: with warm-bloodedness and the first-mover advantage - theory of mind: with group living and social hierarchies

Answer 223

- william james: the baby, assailed by eyes, ears, nose, skin, andentrails at once, feels it all as one great blooming, buzzing confusion - john locke: let us then suppose the mind to be, as we say, white paper, void of all characters, without any ideas:—How comes it to be furnished?... To this I answer, in one word, from EXPERIENCE. - plato: we do not learn, and that what we call learning is only a process of recollection

Answer 224

when a silhouette is towed over a yard of chicks, they scatter if its movement implies it is a hawk (but not a goose)

Answer 225

chicks prefer to associate with objects that appear to be 'self-propelled'

Answer 226

infants will (sometimes) imitate facial expressions just 1 hour after birth

Answer 227

6-month-old infants preferentially look to a display with a number of objects that matches the number of sounds they heard

Answer 228

4-month-old infants understand that things that share a “common fate” tend to be part of the same object

Answer 229

Tests whether infants look longer at an event that violates expectations

Answer 230

4-month-old infants understand - object permanence – objects continue to exist over time (even when occluded) - object solidity – rigid objects can’t pass through other rigid objects

Answer 231

15-month-olds understand that a person will act on a false belief – they distinguish between others’ mental states and what they know to be the true state of the world

Answer 232

examines variation in traits. it quantifies the role of: - heredity (inherited genes from parents), - shared environment (e.g., parenting style; neighborhood) - nonshared environment (e.g., idiosyncratic life experiences) - this doesn’t work for traits with very little variance – the ability to understand simple spoken sentences, fall prey to visual illusions, or carve the world into categories

Answer 233

- if genes explain all variation in a trait: MZ twins should be very similar to each other and DZ twins should be less similar to each other - if shared environment explains all variation in a trait: MZ twins and DZ twins should both be very similar (and equally so) - if unshared environment explains all variation in a trait: MZ twins and DZ twins should both be very dissimilar (and equally so)

Answer 234

- if genes explain all variation in a trait: RT twins and RA twins should both be very similar (and equally so) - if shared environment explains all variation in a trait: RT twins should be very similar to each other and RA twins should be less similar to each other - if unshared environment explains all variation in a trait: RT twins and RA twins should both be very dissimilar (and equally so)

Answer 235

large impact of genetics and nonshared environment

Answer 236

virtually every individual difference trait that has been studied is genetically heritable (i.e., mz > dz similarity)

Answer 237

- cognitive abilities (e.g., verbal intelligence) - personality traits (e.g., extraversion) - beliefs and attitudes (e.g., conservatism, religiosity) - occupational interests (e.g., interest in art vs science vs business) - behavioral disorders (e.g., depression, alcoholism)

Answer 238

- role of shared environment varies more across traits (e.g., low for cognitive ability; higher for educational attainment) - role of nonshared environment high for most traits - genetics ≠ destiny

Answer 239

is the environment of mz twins more similar than dz? probably not – when parents think their mz twins are dz, heritability estimates similar

Answer 240

look at which genes correlate with which behaviours

Answer 241

- each gene typically explains a tiny proportion of the variance (e.g., 0.1%) - can look across many genes to create a polygenic score that quantifies an individual’s “genetic intelligence” or “genetic extraversion”

Answer 242

- confounding – some of these genes are associated, not causal - genetic nurture – a heritable trait (e.g., educational attainment) affects the environment - population stratification – people don’t mate at random, so traits can become correlated due to shared ancestry and then appear to be linked to a trait common in that group - assortative mating – people tend to mate with similar people (e.g., of similar attractiveness), so desirable traits can become correlated with each other; missing genes - huge gap between heritability estimate from twin studies and contribution from polygenic score (e.g., 40% vs 5% for educational attainment) - can’t detect mutations or rare variants, or interactions among genes

Answer 243

tests whether genes affecting x are linked to y

Answer 244

are intelligent people happier? lots of studies show a moderate positive link

Answer 245

intelligence affects many life outcomes, including health and income, which themselves might affect happiness

Answer 246

more mixed results. these studies also have problems - confounding (as for GWAS) - pleiotropy – some genes do lots of things

Answer 247

- adaptive trade-offs – some traits have both benefits and costs (examples: extraversion – more sexual partners; more accidents. agreeableness – valued as coalition partners; can be taken advantage of) - mutational load – very broad traits are affected by many genes (i.e., highly polygenic), each of which can be randomly mutated; everyone will have some of these mutations, so these phenotypes can become fitness indicators that signify genetic quality

Answer 248

symmetry – almost any deviation from symmetry is deleterious, so people tend to find symmetrical people more attractive; intelligence – actually turns out to be correlated with symmetry!

Answer 249

brain is adapted for information processing - behaviour is computed by mental programs in response to information from the environment

Answer 250

over generations, these programs were sculpted by natural selection to be fitness-enhancing for our ancestors - this is called the environment of evolutionary adaptiveness (EEA) - the modern environment differs in many ways from the EEA

Answer 251

our ancestors faced a range of adaptive problems, each providing a distinct selection pressure - finding a mate, cooperating with others, hunting, avoiding predators, etc.

Answer 252

our minds comprise a suite of specialized mental programs that are adapted for information-processing problems that were relevant to the survival and reproduction of our ancestors

Answer 253

- domain-specificity: responds only to a specific kind of input - limited accessibility: output is accessible to other parts of the mind, but operational details are not - informational encapsulation: does not take account of other knowledge - obligatory processing: cannot be deliberately ‘turned off’

Answer 254

a neuropsychological condition resulting in 'face blindess'

Answer 255

- object individuation: distinguishing among different items within a category - holistic processing: recognizing objects that allow little shape decomposition - configural processing: representing the spacing between features - curvature processing: representing objects with curved lines or surfaces - perceptual expertise: learning how to recognize individual items within a visually homogeneous class

Answer 256

- 53 year old physics PhD had a lifelong trouble recognizing faces - yet his other perceptual abilities are perfectly normal

Answer 257

- responds more to faces than to objects or even to inverted faces - activated by face stimuliwith very different low-level features (drawings, cats) - training on novel stimuli does not activate it

Answer 258

- wason selection task: a classic logic puzzle that reveals how people apply rules, often failing at abstract versions due to confirmation bias but succeeding more often with concrete, real-world scenarios - cheater detection module

Answer 259

- input/output systems: info-processing systems that interface with the external world – e.g., vision, language, motor control - central cognition: info-processing that mediates between input/output systems

Answer 260

- generally thought to be modular (e.g., opaque to consciousness, informationally encapsulated) - can comprise chain of processes, each feeding into the next (e.g., feature detectors, edge extraction, object parsing)

Answer 261

- combining information from different input systems - updating beliefs in light of new information from input systems - drawing inferences from one belief to another within central cognition - choosing among several courses of action for output systems to carry out - controversial how much is modular: aspects of ‘mind-reading’ (inferring others’ beliefs and goals) may be automatic; reasoning seems to be better in evolutionarily relevant tasks (cheater-detection); but pretty clear that much of everyday ‘thinking’ is effortful and introspectively accessible

Answer 262

the idea that higher-level cognitive functions (e.g., beliefs and desires) impact what we see - most, if not all, of these studies have methodological problems

Answer 263

- desired object seen as closer - grayscale objects appear tinged with the object’s actual colour - accurately throwing darts onto a target makes the target look bigger - positive words seen as lighter than negative words - ...and hundreds more

Answer 264

- when asked to wear a backpack, people judge a hill as steeper than when not wearing one - claim: this is because people estimate the effort of climbing the hill as part of their perception of steepness - but this turns out to be due to demand characteristics: when given a plausible explanation for why they are wearing a backpack (e.g., to carry heavy monitoring equipment while climbing), this effect disappears. participants in the original study were adjusting their responses based on assumptions about the experiment’s purpose

Answer 265

- when determining whether a sequence of letters is a word vs non-word, people can do this faster if the word is related to morality (e.g., “blame”, “justice”) than if it is unrelated (“limited,” “saucer”) - claim: moral stimuli are privileged in the mind - but this turns out to be due to semantic priming: words “prime” memory for related words - DOG -> “C_T” KNIFE -> “C_T” - researchers demonstrated similar “pop out” effects for fashion and transportation words

Answer 266

- even when matched on luminance (i.e., actual darkness), black faces look darker than white faces - but this turns out to be due to other low-level differences - even when blurred so that one can’t tell which face is white or black, black faces still look darker

Answer 267

- larger animals have larger brains - humans are outliers - we have a high brain-to-body size ratio

Answer 268

- obstetric geomtery: big heads may not fit through the pelvic canal (especially for a bipedal animal) - metabolic cost: big brains pose a high energy cost to the mother - are born with brains that are smaller relative to their adult size --> leads to longer juvenile periods

Answer 269

- infant searches for the hidden object under cover A, where it was initially hidden and recovered several times, even though it was last hidden in location B - traditional interpretation: failure to understand object permanence

Answer 270

- infants are suprirsed if an object shows up in the wrong location, including infants who fail the a-not-b task - this suggests that it is a motor control deficit, not an object permanence deficit

Answer 271

- the verbal task requires greater (slow-developing executive control: inhibiting the child's own knowledge, and storing and retrieving both locations in working memory - when the task is simplified to reduce these demands, younger children (2.5 years old) can succeed on the verbal task

Answer 272

- reproduction now versus reproduction later (and maintenance of tissue) - quality vs quantity of offspring (e.g., effort to mating vs parenting

Answer 273

- invest in growth during maturation period, in offspring during reproductive period - invest in offspring (versus finding more partners) after parenthood

Answer 274

sequence of developmental changes are evolutionary programmed to benefit the organism to each life stage - can be either pre-programmed (canalized) or responsive to input (plastic) - sometimes implemented through neurochemical (e.g., hormonal) trajectories: testosterone causes risk-taking + mate-seeking and low in childhood, rises during adolescence (esp. in males), and falls after parenthood

Answer 275

young children know that they should share, but actual sharing behaviour develops much later - LHT explanation: this optimally navigates a tade-off: receiving resources from caregivers when young and learning to cooperate with non-kin when older

Answer 276

children who grow up in resource-abundant households grow up to be more tolerant of risks and long term oriented - LHT explanation: childhood resources are signal of future resource-abundance/scarcity - calculated risks and future rewards are only worthwhile when current needs are met

Answer 277

children actively construct their knowledge by integrating new experience with pre-existing mental structures - plainly something is right about this: we're not brorn with concepts like 'telephone', 'natural selection', or '717'

Answer 278

jean piaget

Answer 279

- core knowledge: a set of innate knowledge systems - numbers, objects, agents, space, causality - bootstrapping: using more basic concepts to learn more sophisticated concepts, using core knowledge as a starting point

Answer 280

1. the approximate number system 2. object tracking system

Answer 281

- sensitive to the ratio of number, ot the absolute number - 100 vs 50 or 50 vs 25 is easy (same ratio, smaller absolute difference) - 1000 vs 950 is hard (same absolute difference, smaller ratio)

Answer 282

after repeating an old stimulus, if a new stimulus causes infants to regain interest, they must distinguish between the old and new stimuli

Answer 283

rats - trained to press a lever 'N' times to receive a food pellet - measurement is the number of presses before the rat checks the food alcove

Answer 284

- ots is sensitive to the absolute number of items - can keep track of up to ~4 objects

Answer 285

- groups of lions were played recordings of either one lion roaring or three different lions roaring. researchers measured whether the lions approached the loud speaker - ponies were trained to mathc the same option as a sample. this was used to study ability to match numerosity - rhesus macaque monkeys were shown apple slices being put into 2 opaque boxes. researchers measured which box the monkey approached.

Answer 286

neither system is sufficient to represent the positive integers: ANS cannot represent exact numbers and OTS system cannot represent large numbers then how do we learn to represent numbers like 6, 77, 10,586?

Answer 287

different representational systems are combined to learn a counting system which includes: the cardinality principle and successor function

Answer 288

when counting a set of items, the numerosity of a set is given by the last number counted (e.g., one, two, three)

Answer 289

from every number N, another number N+1 can be derived (e.g., from 20, we can derive 21)

Answer 290

- the first number words are learned from natural language quantifiers like “a” - “one-knowers” can retrieve “one” object, but respond randomly with larger numbers

Answer 291

- most children go through stages of being "one-knowers," then “two-knowers,” then “three-knowers” and occasionally “four-knowers” - after knowing enough numbers, the child notices the mapping between the OTS and the known numbers – they discovered the cardinality principle

Answer 292

- meanwhile, children have learned the sequence of number words (“one”, “two”, “three”, “four”, “five”), but it’s like any ordered, ritualized list (“eenie”, “meenie”, “minie”, “moe”) – a placeholder - once they combine the cardinality principle with the placeholder list of number words, they have constructed the positive integers

Answer 293

fabricating behaviours (canalization) versus letting them vary through development (plasticity) - canalization --> more energy-efficient, less prone to random error - plasticity --> more flexible - canalization usually wins when environment changes slowly; plasticity when it changes rapidly

Answer 294

innate learning algorithms: associative learning, reinforcement learning

Answer 295

a window of developmental time during which it is easier to learn something

Answer 296

- nightingales can reproduce a set of 60 songs after being exposed to each song once per day for 20 days - sparrows won’t incorporate songs heard after 100 days into their adult song – and won’t ever produce normal songs if they don’t hear song before 100 days

Answer 297

- children with untreated congenital cataracts usually do not develop binocular depth perception as adults even if the cataract is later removed - this has been shown experimentally in cats by covering one eye early in life

Answer 298

a predisposition to learn certain associations. children aren't born afraid of snakes but naturally pay more attention to them - they could more easily spot a pic of snake amongst flowers than a pic of flowers amongst snakes

Answer 299

8-month-olds were exposed to 2 minute strings of nonsense syllables: e.g.,“bidakupadotigolabubidakudo...” - some syllables (”bi”) are always followed by the same syllable (“da”), whereas others (“ku”) were followed by different syllables - test trials: infants were given repetitions of familiar ‘words’ (e.g., "bida”) or novel non-words (e.g., “biku”) - infants listened longer to the novel non-words than to the familiar words

Answer 300

- learning causual relationships among 30-month-old infants - one-cause condition: 78% identified A (vs B) as the blicket - two-cause condition: 47% identified A (vs B) as the blicket

Answer 301

4- and 5-year old children distinguish between confounded and unconfounded evidence

Answer 302

they are more likely to gather more evidence (i.e., playing) with the box when evidence is confounded and therefore more informative

Answer 303

our social learning abilities

Answer 304

1. our brains encode grammatical rules 2. these grammatical rules are unconscious 3. learning these rules relies on innate knowledge

Answer 305

- our brains encode grammatical rules - there is an effectively infinite number of understandable sentences in any natural language

Answer 306

- a rule can feed into itself — e.g., if s is a sentence, then “x verbs that s” is a sentence - e.g., “bill thinks that beth is a genius,” “sue suspects that bill thinks that beth is a genius,” “charlie said that sue suspects that bill thinks that beth is a genius.”

Answer 307

- these grammatical rules are unconscious - expletive infixation: native speakers know where to insert “bloody” in “elephant,” but it was never taught

Answer 308

relies on innate knowledge

Answer 309

- linguists take decades to describe a single grammar, yet children master theirs effortlessly - language acquisition is domain-specific – a mental faculty specific to language, not just general learning or pattern recognition

Answer 310

a hypothesized set of commonalities across all human languages that is part of our genetic endowment

Answer 311

- principles are the commonalities or constraints - parameters are the settings of a given language within those constraints

Answer 312

- absolute: no language forms questions by reversing the order of the sentence - statistical: 90% of languages follow subject–object (so) order – most are svo, sov, or vso - conditional: vso --> question words at beginning; if a language has derivational and inflectional suffixes, derivational come first (“darwin + s + ism”)

Answer 313

up to age ~1 for distinguishing phonemes; ages 3–7 for producing grammatical sentences

Answer 314

- babbling stage (1st year): playing with sounds - one-word stage (beginning of 2nd year): holophrases like “doggie” or “mommy.” - two-word stage (end of 2nd year): rudimentary syntax, e.g., “man pick brush.” - mature syntax (by 3rd birthday): most grammatical rules followed most of the time.

Answer 315

syntax seems unlearnable based on the input children receive - ‘motherese’ isn’t much simpler syntactically - some cultures lack child-directed speech - input is mostly positive examples, not negative ones

Answer 316

- children learn language through parental correction - many cultures lack child-directed speech - parents correct content, not grammar - correction attempts fail

Answer 317

a simplified linguistic system used to communicate between groups without a common language

Answer 318

- a fully language created by a new generation of children exposed to a pidgin - have standardized word order and grammatical markers, often similar across different creoles

Answer 319

deaf children created a sign-language pidgin (lsn), then younger children transformed it into a creole (isn) with systematic grammar

Answer 320

they are discrete and combinatorial, with finite recombinable parts

Answer 321

not pantomimes or one-to-one cyphers of the surrounding spoken language, but rather autonomous languages with their own grammatical systems

Answer 322

a genetically inherited language impairment with intact cognition and normal iq, causing slow, effortful, ungrammatical speech - could be consistent with a domain-general story about language acquisition (maybe language learning is just the hardest task faced by general intelligence)

Answer 323

a genetically inherited cognitive impairment (average iq for ~50) with intact, fluent language and rich conversational ability - struggles with basic tasks (numeracy, motor skills, navigation)

Answer 324

two cognitive systems are shown to be independent if one impairment affects X but not Y, while another impairment affects Y not X - e.g., prosopagnosia (intact object recognition but impaired face recognition) and visual object agnosia (intact face recognition, impaired object recognition) | evidence of domain-specific mechanisms for face and object recognition

Answer 325

impairment in language production, both written and spoken with comprehnsion preserved (except some syntactic subtlety)

Answer 326

impairment in language comprehension - production can be fluent, but will not make sense - frequently confuse words with one another

Answer 327

1. categorical perception 2. constraints on meaning 3. generalization of rules

Answer 328

phonemes, then morphemes, then syntax

Answer 329

our perceptual system draws category boundaries over a continous quantity - even newborns' perceptual systems are attuned to the same discrete categories

Answer 330

we lose the ability to distinguish phonemes that are not meaningful in our native language - in the first year of life, English-speaking infants lose the distinction between two different d sounds that exist in Hindi

Answer 331

simplifies learning by ignoring irrelevant distinctions, but makes foreign languages harder to learn

Answer 332

determining what part of a situation a word refers to (e.g., “rabbit,” “fur,” or “hopping”)

Answer 333

- whole-object bias: words refer to whole objects - taxonomic bias: words refer to similar kinds, not themes (e.g., policecar extended to car not policeman) - mutual exclusivity: two words don’t refer to the same thing - syntactic bootstrapping: using sentence syntax to infer meaning (e.g., a transitive verb (she is gorping her) describes an action that involves two individuals | assumptions we use to rule out most hypotheses about word meaning

Answer 334

most English nouns are pluralized by adding: - -’s’ (pronounced ‘s’) as in ”boots” - -’s’ (pronounced ‘z’) as in ”mugs” - -’es’ (pronounced ‘ez’) as in “glasses” - preschool children can pronounce the correct form for novel words (e.g., “wugs”)

Answer 335

most English verbs add ‘-ed’ for the past tense but there are many irregular verbs - “saw,” not “seed” - “went,” not “goed

Answer 336

we get worse at this before we get better! why? kids start by memorizing each verb separately, but then begin to overregularize verbs, before learning both the rule and the exceptions

Answer 337

compositional syntax - fixed set of calls (e.g., monkey alert calls) - represent a continuous magnitude (e.g., bee dances) - theme with variation (e.g, birdsong)

Answer 338

alarm call of Campbell’s monkey - discrete parts of the call (including an ‘-oo’ suffix) produce different meanings - but the repertoire is very small and is are not generative (much less recursive)

Answer 339

- raising a chimp (Gua) and child (Donald) together in the hopes the chimp would learn spoken English - training chimps and gorillas to use sign language - pairing a bonobo (Kanzi) with a 2-year-old to press series of symbols | would be evidence against domain-specific language modules

Answer 340

no --> - repeating old phrases, rarely creating new ones - language acquisition lags dramatically behind human children - little evidence of sensitivity to word order - no complex syntactic structures

Answer 341

false - it’s hard to pinpoint exactly what they don’t have

Answer 342

it's all mechanisms required for language

Answer 343

mechanisms that are specific to language and uniquely human - possible that FLN is empty – humans only be unique in having all components of FLB in the same species

Answer 344

vocal imitation (“hey!, get ova’ here”) in parrots, seals - rule-based ordering (“chickadee” vs “chickadeedee”) – Songbirds, capuchins - distinguishing speech sounds (e.g., “ba” vs “pa”) – Chinchillas - learning constraints (e.g., mutual exclusivity) – Dogs, sea lions - planning (e.g., tool use) – Crows, chimpanzees - feference – Monkeys, bees, dolphins ## Footnote some of these species are more distant than primates – convergent evolution

Answer 345

no animal woke up one day and started speaking it’s unlikely that a single mutation would produce a qualitative shift from non- linguistic to linguistic primates - even if that animal existed, it would have no one to speak to! so what adaptive value would this change have?

Answer 346

a transitional form between the communication systems of non-human primates and modern humans | most cognitive ethologists believe that early hominids had this

Answer 347

1. holophrastic: single symbolic word like 'food'. lacks combinatorial structure, compositionality, recursion. (e.g., pidgins, one-word stage, broca's aphasia, apes' single-word use) 2. gestural: communication through pantomime and pointing. verbal channel evolved later to free hands + allow comunication in darkness/at a distince. (e.g., apes' flexible gestures, mirror neuron activation, full syntax in sign languages) 3. musical: emotional vocalizational for social bonding that evolved into syntax. repetition, rhythm (e.g., parallels between music and speech, overlap in broca's area, neanderthal instruments, whales, birds)

Answer 348

hierarchical syntax enabling recursion and productivity

Answer 349

martin track the great eskimo vocabulary hoax through successively more careless repetitions and embroiderings in a number of popular books on language

Answer 350

once more popular sources start to get hold of the example, all constraints are removed: arbitrary numbers are just made u as the writer thinks appropriate for the readership

Answer 351

horsebreeders have various names for breeds, sizes, and ages of horses; botanists have names for leaf shapes; interior decorators have names for shades of mauve; printers have many different names for different fonts

Answer 352

the prevalence of the great Eskimo snow hoax is testimony to falling standards in academia, but also to a wider tendency … toward fundamentally anti-intellectual ‘gee-whiz’ modes of discourse and increasing ignorance of scientific thought

Answer 353

- pattern recognition is hard - many animals alive today, even after half billion yeards of evolution, never acquired this ability

Answer 354

- discrimination: how to recognize overlapping patterns as distinct - how to generalize a previous pattern to recognize novel patterns that are similar but not the same

Answer 355

create a network of neurons by adjusting the weights of the connections between neurons, you can make the network perform a variety of operations on its input

Answer 356

they propogate the error at the end back throughout the entire network, calculate the exact error contribution of each synapse and nudge that synpase accordingly (aka supervised learning)

Answer 357

the brain does other not do supervised learning - you are not given labeled data when you learn that one smell is an egg and another is a strawberry

Answer 358

pyramidal neurons

Answer 359

some theorize that the vertebrate brain's ability to solve the invariance problem derives not form the unique cortical structures in mammals, but from the complex interactions between the cortex and the thalamus

Answer 360

tendency for artificial neural networks to abruptly and drastically forget previously learned information when trained on new data

Answer 361

learners should use three cultural learning cues to target their learning: age, success, and prestige

Answer 362

people are more inclined to copy more successful others - we are prestige biased

Answer 363

automatically and unconsciously, people also use cues of self-silimarity, like sex and ethnicity, to further hone and personalize their cultural learning - children learn their sex roles because they copy same-sex models not vice versa

Answer 364

by exploiting large data set ... being taught by instructors whom you match on ethnicity/race reduces your dropout rate and raises your grades

Answer 365

in addition to using model-based mechanisms for cultural learning, we should also expect natural selection to have equipped us with psychological abilities and biases for learning about certain predictable content domains

Answer 366

any learners who miss the boat on mentalizing and cultural learning … will be at a serious disadvantage

Answer 367

infants have language abilities, perhaps even before birth, suggesting some aspects of language are innate rather than learned

Answer 368

by 10 months, babies lose the ability to distinguish all phonemes and respond only to those used by their caregivers

Answer 369

pinker proposes a speech analysis module that sorts sounds before meaning is learned and plays a key role in acquiring words and grammar

Answer 370

all babies babble the same sounds, and their first words tend to be objects, actions, and social routines

Answer 371

it allows children to rapidly expand sentence length and complexity as they develop

Answer 372

by age three, children follow most grammatical rules and make errors consistent with those rules, showing internalized grammar

Answer 373

because “languages are infinite [and] childhoods finite,” children must generalize from limited input to create novel sentences

Answer 374

the brain provides a blueprint for the UG that limits possible generalizations, guiding children toward valid sentence structures

Answer 375

1. parents’ speech modeling the basic design of phrase structure, and 2. the context of parental speech helping children infer structure

Answer 376

- described children as 'little scientists' - they form hypotheses about how the world works, test them thru exploration and play, and revise their understanding when faced with new evidence - mirrors scientific reasoning - children design and run experiments every day as part of natural learning

Answer 377

- parents and caregivers create the “laboratory environment” in which children’s experiments take place - through interaction, they provide both “data” and feedback that help children refine their mental models of the world - social engagement isn’t just background—it’s a crucial part of how children generate and test ideas

Answer 378

- play is the child’s version of scientific experimentation - it allows children to test cause-and-effect relationships in a flexible, low-risk way, generating observations about how objects, people, and actions behave - even apparent “mistakes” in play are valuable because they guide the child toward more accurate theories

Answer 379

- Dr. P. was a distinguished musician and teacher who developed visual agnosia—a neurological disorder that left his vision intact but destroyed his ability to recognize faces and objects - he could perceive details, colors, and shapes but could not interpret or synthesize them into meaningful wholes

Answer 380

- he could describe isolated features—such as a nose, color, or contour—but failed to perceive integrated forms - he saw the world as disconnected fragments - his inability to recognize faces led him to mistake his wife’s head for a hat, a symbol of how his visual world had become abstract and depersonalized

Answer 381

- music remained the one domain where Dr. P.’s faculties were fully preserved - he could perform, teach, and perceive musical structures flawlessly - rhythm and sound provided an organizing framework for his life, allowing him to navigate and act coherently in situations that would otherwise overwhelm him visually

Answer 382

- sacks uses the case to illustrate how specific neural systems create meaning from perception - when these systems fail, a person can lose access to the world of faces, expressions, and personal connection—yet still retain intellect, creativity, and emotional life - the story highlights both the fragility and adaptability of cognition

Answer 383

- pinker argues that evolution is not a ladder leading inevitably to intelligence or progress - it is a branching process shaped by adaptation to environments. Intelligence, like the elephant’s trunk or a spider’s web, is a specialized evolutionary solution—not a sign of cosmic advancement or purpose

Answer 384

- natural selection operates through replication, variation, and selection - random genetic changes create variation; those that improve survival and reproduction persist - over vast timescales, this process yields complex organs and minds that appear designed for purpose—without any foresight, intention, or guiding force

Answer 385

- pinker dismisses these theories because they fail to explain how adaptive complexity arises - lamarckism assumes acquired traits can be inherited, which evidence disproves - complexity theory describes patterns but not function - only natural selection accounts for how organisms—and minds—develop purposeful structure from blind, mechanical processes

Answer 386

- the brain is designed to be plastic — it can adapt its structure and function in response to experience - in Pinker’s terms, evolution didn’t give us fixed behaviors but flexible learning mechanisms that allow the mind to adjust to the environment - this plasticity ensures that universal mental structures can still produce different outcomes depending on the individual’s upbringing, culture, and experiences

Answer 387

1. cultural information: beliefs or behaviours that one individual learns from another (e.g., social norms, religious beliefs and practices, etc.) 2. cultural groups: social groups that share a set of common beliefs and behaviours (e.g., countries, regions, small-scale societies)

Answer 388

- aspect of cognition or behaviour that are observed in all (or the vast majority) of cultural groups) - likely to be innate and biologically evolved - could have been discovered once and dispersed widely (e.g., the wheel) - or frequently reinvented solution to a common problem (e.g., fishing nets)

Answer 389

- aspects of cognition or behvaiour that vary across cultural groups - must be culturally transmitted - differs due to cultural (not biological) evolution

Answer 390

- every language has nouns and verbs -> we don't need to learn this (universal, biological evolution) - the specific words used to refer to things vary across languages (variable, cultural evolution)

Answer 391

- cultural - social - behavioural - mental

Answer 392

a social suite of adaptions common to all cultures: individual identity, love for partners partners/offspring, friendship, social networks, cooperation, in-group bias, mild hierarchy, social learning

Answer 393

- cooperation - individualism/collectivism

Answer 394

- some variance seems idiosyncratic: some societies have a norm wherein sharing is seen as creating obligations, so high offers are rejected - but other variance is systematic: societies with more experience with markets and greater payoffs to cooperation are more cooperative

Answer 395

- collectivist cultural norms: occur when a group maintains extensive networks of distant relatives. behaviour highly constrained by context and relationships. associated with in-group conformity, deference to authorities, policing the behaviours of close associates. success tied to how well one promotes the group’s collective success - individualist cultural norms: occur when there are relatively few inherited ties (outside the nuclear family). success tied to how well one develops individual attributes and attracts friends and partners with those attributes

Answer 396

- collectivism: roles and relationships (josh's dad) - individualism: personal attributes (a scientist)

Answer 397

- modernization: wealth, education, and market institution promote individualism - pathogen prevalence: high prevalence of infectious disease makes it dangerous to deal with strangers, promoting collectivism - family structure: greater genetic relatedness within marriage (e.g., marrying less distant cousins) increases kinshipp ties, promoting collectivism - subsistence style: some forms of subsistence require more interdependence than others, promoting collectivism

Answer 398

- rice farming requires elaborate irrigation systems and many more hours of work – so rice farming increases the value of cooperation - regions in China that cultivate rice tend to be more collectivistic - this was even true in a “natural experiment” in which the Chinese government randomly assigned people to wheat or rice farms after WWII

Answer 399

- some cultures, particularly Western ones, tend to have an analytic cognitive style - this involves focusing on objects as separate from their context and using formal logic - in contrast, cultures in East Asia often use a more holistic cognitive style, which emphasizes the relationships between elements and the overall context. | rabbit = cat or carrot?

Answer 400

- joint attention: facilitates other learning mechanisms via shared knowledge - imitation: learning by copying. children will sometimes overimitate. - pedagogy: learning by teaching - testimony: learning by explicit linguistic communication

Answer 401

knowing whose testimony can be trusted

Answer 402

copying the outcome of an action, but not the action itself

Answer 403

a cognitive adaptation for learning opaque information from intentional instruction - requires sensitivity to obstensive cues (e.g., eye contact, direct speech) - expectation that ostensive context involve reference to objects (e.g., pointing, gaze direction) - assumption that the communication conveys generalizable information

Answer 404

- pedagogical condition involves a social signal from a teacher to a learner that guides the learner's expectations and behavior, making learning efficient for the demonstrated task but potentially limiting broader discovery - the naive condition lacks this signal, prompting more extensive, open-ended exploration - 4-6 year olds played with the toy longer in the naive conditions and discovered more of its function

Answer 405

humans play host to two sets of replicators: gene and idea 1. the capacity for culture is itself a biological adaptation: social learning mechanisms 2. culture evolves along similar principles of selection 3. ones culture becomes part of the environment in which biological evolution occurs

Exam 2 Flashcards

(429 cards)