Slide 37 updated Animals

1
Slide 37 updated!Animals InferenceDo
children have intuitive knowledge of biology?

• What does the child know about the world?
• Is the childs mind different from the adults
  mind, or does the child just know less?
• Does the child come factory equipped with any
  knowledge of the world?

2
What kinds of things are there?

• Natural kinds
• animate (all are living)
• people (definitely intentional agents)
• animals (which are intentional agents?)
• inanimate
• plants (inanimate but living)
• substances (water, quartz, pearls...)
• Artifacts (human-made machines, tools...)
• inanimate, not living
• Natural features? Mountains, rivers...

3
What kind of explanations are there?

• Intentional stance (for people, animals)
• intuitive psychology Theory of Mind Mech.
• Physical stance (for anything, but particularly
  inanimate)
• intuitive physics object mechanics
• Design stance (for artifacts but does it also
  apply to living things?
• Is there an intuitive biology?
• Essentialism (for substances, animals, people
  social groups? (ethnic, coalitions))

4
How do babies carve up the world into categories?
What do they know about those categories?

• Then and now...
• Then Perception drives everything (Piaget
  (Classical concepts), Rosch (Family resemblance)
• Perception precedes category formation, and
  drives it.
• Inferences will follow judgments of perceptual
  similarity
• Perceptually derived categories come in first
  superordinate ones (e.g., animal come last)
• Learning is bottom up categories built from
  inductive generalizations from perception
• No factory-equipped concepts (no innate ideas)

5
How do babies carve up the world into categories?
What do they know about those categories?

• Now Babies carve world using abstract categories
  
• Global domains (e.g, animal versus artifact)
  precede acquisition of basic level categories
  (tiger, car) in infants (Mandler)
• Categories not built by perception perceptual
  cues activate the appropriate evolved category
• Movement cues provide cut on world, skeletal
  knowledge with correlational learning filling in
  details (3-4 year olds, Rochel Gelman)
• Kind category can over-ride perceptual similarity
  in making inferences (3-4 year olds, Markman and
  Susan Gelman)

6
Habituated On
7
Infants. 3-4 month olds can form perceptual
categories

• E.g., Shown photos of cats, they distinguish
  tigers
• But does this mean they know what a cat is? Do
  they make different inferences about tigers and
  cats?
• Or do they merely pattern recognize?

Puzzle Given line drawings, dont distinguish
these...
8
Does it matter what you use in testing?

• Realistic photos?
• Line drawings?
• What about toy models?
• What about handling rather than looking time?

9
Jean Mandler Global domains first in infancy

• Habituation tasks
• Using looking time
• Using object manipulation
• Realistic toys
• Object manipulation tasks show different results
  from looking time tasks. E.g.
• Habituate on photos of cats, they dishabituate
  to photo of a tiger
• Habituate on handling a variety of cat toys, they
  do not dishabituate on tiger toy

10
Distinguishing global domains Animals versus
vehicles

• Habituation tasks, object manipulation
• E.g. Birds and planes perceptually very similar

• 7 months old babies distinguish birds from
  planes
• Do not distinguish among mammals
• do distinguish land animals from birds

11
Global domains distinguished (Some) Life forms
distinguished
12
Global domains before kind categories
13
Global domains before kind categoriesImitation
studies (9-14 months)

• Do global domains provide stopping rules for
  inductive inferences?
• Model gives dog a drink from a cup.
• baby has choice of rabbit or truck.
• Will model to rabbit, not truck
• if given choice of cat or dog, is just as likely
  to use cat as dog
• also aardvark, anteaters, birds, fish (never seen
  a fish drinking)
• will not give drink to a flying tiger airplane,
  even though tiger face on it. Global as stopping
  rule!

14
Global domains before kind categoriesImitation
studies (9-14 months)

• Is it the child just playing the experimenters
  game? (demand characteristics?)
• Will child model anything you doeven if it
  crosses a global domain boundary?
• If you model giving a CAR a drink
• infants reluctant to imitate this

15
Global domains before kind categoriesImitation
studies (9-14 months)

• What about vehicles?
• model puts key to a car
• baby has choice of rabbit or airplane
• Baby will model to airplane, not rabbit
• has never seen key put to an airplane,
  helicopter, etc.
• If you model keying an animal,
• infants reluctant to imitate this

16
Global domains before kind categoriesImitation
studies (9-14 months)

• Control condition
• What if you model something appropriate to both
  animals and vehicles, such as going into a
  building?
• no preference for animal or vehicle in imitation
• Shows that preferences for selective imitation
  with animals versus vehicles represent real
  distinctions and global-domain stopping rules.

17
Global domains before kind categoriesImitation
studies (9-14 months)

• Causal properties matter
• use mug to give a drink
• just as likely to use a pan, crib or bathtub
• use anything that can contain
• Selective imitation by 9 months, strong by 14
• By 20 months they start restricting to basic
  level categories.

18
Global domains before kind categories

• Relation to adult conceptual system?
• Semantic dementia reverse of developmental
  pattern
• First lose distinction between (e.g.) dog cat
  (land animals)
• Later lose mammal versus bird
• Last animal versus artifact
• Global domains primary. In development, finer
  distinctions made after global ones
• Global most buffered against assaults
• First in, last out

19
Global domains before kind categories

• Opposite of prediction for
• Classical view of concepts
• Family resemblance view
• On these earlier views, perception drives
  conceptual development
• First perception builds specific categories
• Generalizing to larger domain (animal or vehicle)
  comes later
• Mandler Global domains come first, and guide
  further inferences
• Kind domains (dog, cat) distinguished later
• Global domain guides inductive inference
  (imitation studies)

20
How do children generalize properties?

• by perceptual similarity?
• by basic object level?
• by kind? (e.g., bat vs. bird)
• by global kind only? (e.g., animal versus
  artifact)
• Within a natural kind category, which properties
  are generalized?
• all of them?
• only ones appropriate to a kind?
• What do we know from Mandler Markman?

21
How do children generalize properties?

• About animals, they generalized by kind
• internal organs, method of respiration, feeding
  habits, behavior (lt 14months global kind)
• But they did not use kind category to generalize
  weight, visibility at night, going into building
• Specificity of this pattern suggests they have
  intuitive theories about animals (and possibly
  minerals?)
• About natural kind substances
• internal structure, chemical properties, physical
  properties
• About vehicles, they generalize
• Keying, but not drinking... (lt14 months)

22
Can children use surface properties to predict
self-generated motion? Ages 3-4 (R. Gelman)

• Can it move itself up and down a hill?
• Color photos, unfamiliar things
• Animals Displaying lizard, spiny anteater
• Statues with familiar, animal-like forms parts
  (bird vessel, figurine w/ insect eyes
• Rigid complex objects (home gym)
• Wheeled objects (old bicycle)

23
(No Transcript)
24
(No Transcript)
25
Piaget thought No understanding of causal
mechanisms until 6 years

• But 3-4 year old children, 17/20 used these
  causal rules of motion
• all animals can move themselves and all inanimate
  objects require an external agent or source to
  move them
• choosy variant of above (youngest children)
  sometimes animal cant go up hill because there
  is a mismatch between it and task, e.g., the
  animal is too little, or hill too big. Details
  of context mattered but same basic rule (animal
  could do less strenuous things by itself) (5/17)

26
Piaget thought No understanding of causal
mechanism until 6 years

• But talk of 3-4 year old children
• Causal analysis of motion
• concern about causal conditions and whether
  animate.
• All inanimate talk about absence of agents or
  parts needed for self-generated motion.
• Animate some talk about movement-enabling parts.
  (Even claimed they could see limbs that were not
  visiblefeet on spiny anteater)

27
Children just focusing on what is salient?

• What do adults find salient?
• When adults asked to pair target items with ones
  most similar, paired statues of animals with
  animals
• No adult pairing based on causal basis of motion
• Children shift their answers depending on
  whether asked about motion or what looks most
  alike
• Would often claim they could see animates
  walking, jumping... Also took trouble to point
  out that inanimates were NOT walking moving (odd)

28
Learning predictive validity of cues in a domain?

• Motion-based skeletal theory allowing acquisition
  of knowledge of predictive surface features?
• Domain-specific principle (self-generated motion,
  contingent reactitivy) defines domain
• Then domain-general correlational mechanism
  computes features correlated with each domain???
• (Or some features (furry, scaly versus metalic)
  part of babys early domain-boundaries? No one
  knows...)

29
How many domain-specific inference systems do
young children have?

• Susan Carey Only 2!
• Theory of mind (intuitive psychology)
• Object mechanics (intuitive physics)
• Do children also have an intuitive biology?
• Domain-specific principles they apply to
  biological entitiesprinciples beyond what is
  found in intuitive psychology and intuitive
  physics?
• Springer, Keil, Barrett Yes

30
Stimuli with self-propulsion direction
Eye-like stimuli
ID
EDD
Dyadic representations (desire, goal)
Dyadic represent-ations (see)
Triadic representations
SAM
Full range of mental state concepts, expressed in
M-representations
Knowledge of the mental, stored used as a theory
ToMM
31
Intuitive biology Questions

• An understanding of kinship and inheritance?
• An understanding of physiological versus
  psychological causes? (e.g., disease versus
  upset?)
• Do they apply the design stance to living
  things?
• Living things have functional parts, parts that
  are good for them
• Do children have other biological concepts?
• Predator-prey interactions?
• Death?

32
Induction experiments

• Can tell us how children reason
• Do they have theories about how the world works?
• Animal 1 has trait X (Exemplar animal)
• 2 Target animals
• Does Animal 2 have trait X?
• Does Animal 3 have trait X
• Animal 2 may look like Animal 1 Animal 3 may
  look different but share some other relationship
• Test perceptual similarity versus other
  relationship

33
What do children understand about kinship and
biological inheritance? (4-5 years old)

• Induction experiments (Ken Springer)
• Do children induce properties using only
  perceptual similarity or also kinship?
• Does it matter which properties?
• Biological properties (stable, physical)
• Can see in the dark, has hairy ears, has tiny
  bones inside, eats small weeds
• Non-biological properties Scrapes on legs from
  running in bushes, likes to crawl in really tall
  grass, very dirty from playing in mud, knows a
  really good place to find food

34
Heres a duck. She can see in the dark. So when
it is nightime, she can see things outside. She
can see in the dark.
Heres another duck. (similar) Even though he
looks just like this duck, he comes from a
different family. Do you think he can see in the
dark, like she can?
Heres another duck. (dissimilar) Hes the baby
of this duck. So even though he doesnt look
like her, hes her baby. Do you think he can see
things in the dark, like she can?
35
Can kinship guide induction? For biological
properties...

• When no kinship specified, induction along
  similarity
• When kinship specified, induction along kinship
  lines, against similarity
• Is it just a social relationship specified, or
  specifically kinship?
• Heres another duck. He looks like this duck,
  and hes this ducks best friend. So they come
  from different families, but they are best
  friends. Do you think he can see in the dark,
  like she can?
• Still go with kinship

36
What about non-biological properties?

• Children do not consider kin any more or less
  likely to share non-biological properties
• So kinship implies shared biological properties,
  not shared properties in general
• Looking like some very specific knowledge of the
  biological world...

37
Carey Only psychological causality at 5
yearsKeil They can distinguish psychological
from physiological causes

• Disease What can you catch from someone after
  being in close contact for the week-end?
• Contrast psychological physiological (an
  unusual mental state vs. an unusual functional
  biological state)
• Little girl who suddenly developed the false
  belief that her hands were dirty, and kept
  washing them all the time. Hand became red and
  oozy
• Physiological outcome, but caused by a mental
  state (i.e. a psychological state)
• Even 4 year olds Not contagious!

38
Do children distinguish functional from
intentional explanations? (germs versus poison)
Frank Keil

• Functional description
• Theres this thing that needs to get inside you
  and use parts of your body to make you sick
• Intentional description
• Theres this thing that has goals and desires.
  It wants to get inside you and make you sick.
• Mechanical description
• Theres this thing that rubs around inside your
  body causing sickness through mechanical damage,
  such as abrasion

39
Can it have babies? Is it alive? Does it have
complex innards? Can it move on its own?
This looks like it favors Carey no biology. But
40
Does it know what it is doing?
Only intentional knows! (Favors Keil)
41
Applying intentional stance to everything?

• No. Thing with goals and desires knows what it
  is doing. Thing that needs to get inside and use
  body parts does not know what it is doing
• If only one domain (intuitive psych), both should
  know
• Not simply equating disease agents with
  prototypical animals or humans, which know what
  they are doing
• Even youngest know a thing can be biological
  (alive, move, reproduce, innards) without being
  sentient

42
Design stance and biology

• Animals and plants have adaptive properties that
  are functional
• The property is there because it solves a problem
  for the organism
• Teeth for eating, hands for grasping, stomach for
  digesting
• Tools have properties that are functional
• The property is there because it solves a problem
  for the person who designed the tool
• Do children apply the design stance to biological
  things?

43
Design stance Do living things have functional
parts? (5-7 years)

• Two people are talking about why plants/
  emeralds are green. This person says it is
  because it is better for the plants to be green
  and it helps there be more plants. This person
  says it is because there are tiny parts in plants
  that when mixed together give them a green
  color.
• Which reason is a better one for plants /
  emeralds?
• Plants better for it preferred 2.5 times as
  often
• Emeralds mechanical preferred 5 times as often

44
Design stance Do living things have functional
parts? (5-7 years)

• Show almost identical pictures
• Prickly plant and prickly mineral
• Almost only difference is the label
• Both are prickly, but only one is prickly
  because being prickly is good for it. Which one?
• Even 5 year olds prefer good for it for plants
  and animals
• Evidence of design stance (parts there because
  they serve a purpose)
• Good for the plant/animal, not good for a
  user / designer, as with tools

45
What counts as biology?

• Just physiology?
• Should our evolved inference mechanisms mirror
  divisions in a university?(!)
• What is important for a hunter-gatherer to know
  about living things?
• What about predator-prey interactions?
• What about adaptive properties of animals that
  are functional?

46
Is intuitive biology a unified domain? Or are
there islands of precocial inference?

• Predator-prey interactions 3-4 years old
• Same good understanding in urban Berlin and among
  hunter-horticulturalists in Amazon (Barrett)
• Claim Death is not really understood til age 10
• Understood as cessation of physiological
  properties
• Is this the concept of death an evolved organism
  should have?
• How about Death as the permanent cessation of
  the ability to act? (Death vs. sleep, etc)
• Test with animals rather than people

47
5 year olds Understand death as permanent
cessation of ability to act

• 3 4 year olds
• Understand run, eat alive
• Problems with time only

48
What kind of explanations are there?

• Intentional stance (for people, animals)
• intuitive psychology Theory of Mind Mech.
• Physical stance (for anything, but particularly
  inanimate)
• intuitive physics object mechanics
• Design stance (for artifacts seems to also apply
  to living things)
• Is there an intuitive biology? Or islands of
  precocial inference?
• Essentialism (for substances, animals, people
  social groups? (ethnic, coalitions))

49
Animals inference Is there an intuitive
biology?

• What does the child know about the world?
• Surprising amount! Kinship design stance some
  biological properties, not others
• How does the child come to know what she knows?
• Global domains first, stopping rules for
  inference
• Correlational after global domain is
  distinguished?
• Is the childs mind different from the adults
  mind, or does the child just know less?
• Does the child come factory equipped with any
  knowledge of the world?

50
Questions to think about throughout 142...

• What does the child know about the world?
• How does the child come to know what she knows?
• Is the childs mind different from the adults
  mind, or does the child just know less?
• Does the child come factory equipped with any
  knowledge of the world?

51
Questions to think about throughout 142...

• How does the environment affect development?
• How does maturation affect development?
• Why did scientists underestimate how much infants
  know?
• What is the competence/ performance distinction?
• Can one part of the brain know something that
  another part of the brain does not know?

52
Questions to think about throughout 142...

• What is the difference between studying natural
  competences and side-effects?
• What does learning mean?
• How many learning processes are there?
• Is instinct the opposite of learning?
• What is the design of the instinct that causes
  learning in a given domain?