Chimps vs Dogs

1
Chimps vs Dogs

• Greg Cogan, Cyprian Laskowski Justin Quillinan

2
Bonobos Phylogeny

• Share 98 of our genes
• Split between chimps and bonobos occurred later
  than the split between humans and chimps (8
  million years ago)
• Believed to have never left the trees less
  evolutionary change
• Body proportions have been compared to
  australopithecines

3
Role of Sex for Bonobos

• Sex plays a prominent role in social relations
• Separate sex from reproduction
• At the San Diego Zoo Bonobos when given food,
  the males would immediately get erections and
  invite females for sex (females would also invite
  males)
• Not only food that excites bonobos if 2 bonobos
  approach a cardboard box thrown into the cage,
  they will briefly mount before playing with it
  (de Waal 2005)
• A jealous male will chase another male away from
  a female, after which the 2 males will engage in
  scrotal rubbing
• Sex appears to be a casual intrinsic part of
  social relations as a way to avoid conflict

4
Matriarchal Organization

• Live in fission-fusion societies move alone in
  or in small groups whose makeup varies all
  relations are fleeting except mother and
  dependant child
• Males tend to stay in their natal groups, whereas
  females tend to leave around adolescence
• Females then seek out attention from senior
  female members and engage in GG rubbing and
  grooming
• Males remain close to their mothers throughout
  their lives, so the important males tend to be
  son of important females

5
Chimpanzees Culture

• Pooling of chimp research has revealed 39
  different traditions, 19 in orangutans
• Grooming hand-clasp chimps at Ngogo in Uganda
  will scratch each others backs using short
  jabbing technique, while at Mahale, they use a
  long raking style
• Goualougo use two-part tool set for termite
  fishing
• Key difference in humans ability for cumulative
  culture
• Could be due to social learning mechanisms
  available to each species

6
Angry Chimps

• Interactions with neighbouring groups tend to be
  negative
• Chimps sneak into neighbouring groups and kill
  one member
• Intragroup aggression not uncommon for
  combatants to reunite with a kiss and an embrace
  thought to strengthen relationships
• Males more aggressive, but also more
  reconciliatory
• Chimps observed catching colobus monkeys and
  tearing them apart and eating them alive

7

8
Chimp Psychology

• Folk Psychology A system of knowing that enables
  individuals to infer what others believe, desire
  and want
• In humans by 4-5 years can use eye gaze as a
  clue
• Begging experiment with chimps one experimenter
  could see the chimp, the other couldnt
• Chimp begged equally from both (Hauser 2005)
• Both chimps and bonobos use facial expressions to
  indicate emotion

9
Chimp Psychology Competition Study

• Condition 1 One banana visible to both
  competitors, and one banana hidden behind an
  opaque barrier and visible only to the
  subordinate.
• Result Subordinate retrieved about half of the
  food, typically moving to the barrier before the
  dominant reacted

10
Chimp Psychology Competition Study

• Condition 2 Involved placing one banana visible
  to both and one behind a transparent barrier, so
  that both bananas were in view.
• Result Subordinate allowed dominant to get both.

11
Chimp Psychology Competition Study

• Condition 3 2 opaque barriers. While
  subordinate watched and dominant looked away,
  experimenter hid one banana on the subordinates
  side of the barrier.
• Result Subordinates obtained more food than
  dominants
• Hare et al. (2001)

12
Chimp Psychology Imitation versus Emulation

• Experiment 1 Young children and chimpanzees
  watched a human first stab a tool into a small
  hole in the top of an opaque box and then remov
  it and applied to it a second lower hole to
  recover food
• Human children typically performed both pokes in
  order.
• Experiment 2 Used a transparent box and it was
  shown that the first poke was ineffectual
• The children continued to poke two holes while
  the chimps only focused on the bottom
• (Horner Whiten 2004)

13
Human Social Intelligence

• Is either a Bonobo or Chimp a good model for
  human social intelligence?
• Culture Both Bonobos and chimpanzees are
  extremely closely related to humans, yet are
  quite different in terms of behavior
• Humans seem to share elements of both cultures

14
Human Social Intelligence

• Social Intelligence
• Chimps pay attention to role in society
  (subordinate versus dominant)
• Do have some knowledge about others states of
  mind
• Use facial expressions
• Have the ability to switch between emulation and
  imitation (young children do not)

15
If not apes, then who?

• What other species can we look at for signs of
  human social intelligence?
• How about dogs?
• In general, chimpanzees perform better than dogs
  on most intelligence tasks, BUT there are
  significant exceptions (Hare et al. 2002).

16
Object Choice Tasks

• Food is hidden under 1 of 2 opaque containers
• Human gives communicative cue (e.g., looking,
  pointing, tapping, marking)
• Animal chooses 1 container

17
Chimpanzees vs Dogs

• Chimpanzees perform at chance levels, though
  sometimes learn after dozens of trials
• Dogs generally very adept, even with more
  complex cues (e.g., human moves to wrong choice
  while pointing at right choice)
• Why? Hare and Tomasello (2005) offer 3
  hypotheses.

18
Hypothesis 1 Ontogenetic Experience

• Claim Dogs do well because they learn these
  skills through their lifetime experience
• Prediction Performance should correlate with a
  dogs previous human exposure
• Test Human-adopted vs kennel-reared dogs, and
  younger vs older puppies
• Results Predicted correlation was not found in
  either test
• Conclusion Forget hypothesis 1

19
Hypothesis 2 Wolf Ancestry

• Claim Dogs have inherited these skills from
  their Old World wolf ancestors (who were social
  pack hunters)
• Prediction Wolves should perform as well as dogs
• Test Wolves vs dogs
• Results Wolves generally performed only at
  chance levels, except on gaze-and-point cue
• Conclusion Forget hypothesis 2

20
Hypothesis 3 Domestication

• Claim Dogs have evolved these skills during
  process of domestication
• Prediction Dogs should perform better than
  wolves
• Test Wolves vs dogs
• Results As we have seen, dogs perform much
  better than wolves
• Conclusion Hypothesis 3 seems correct

21
Domestication and natural selection

• But why should domestication lead to changes in
  social intelligence? What were the selective
  pressures?
• Did selection favour dogs (or, rather, wolves)
  who could follow human communication? Or was it
  a genetic side-effect?
• There is some evidence that it was a side-effect
  (among many others) of selection for tameness.

22
Belyaevs fox experiment

• Belyaevs (1978) conception of domesticated
  behavior
• the ability to have direct contact with man,
  not to be afraid of man, to obey him, and to
  reproduce under the conditions created by him,
  which constitutes the necessary condition for the
  economical use of animals. It is obvious that
  selection for behavior has been unconsciously
  carried out by man since the earliest days of
  animal domestication. (p. 301)

23
Belyaevs fox experiment

• In 1959, Belyaev selectively bred an experimental
  population of foxes based solely on the criterion
  of tameness (as well as a control population)
• Very rapidly, the selected population became
  friendlier and less afraid of humans, and even
  acquired morphological, physiological and
  reproductive changes
• Hare et al. (2005) found that in object choice
  tests, experimental foxes performed like dogs,
  and control foxes like wolves

24
Dogs, chimps and humans

• We must be careful about making a priori
  conclusions based on phylogenetic relationships
  alone.
• Of course, chimpanzees and humans can use tools
  (unlike dogs)
• But dogs and humans can follow pointing (unlike
  chimpanzees)
• And dogs and chimpanzees go around naked (unlike
  humans)

25
So what about human social evolution?

• Hare and Tomasello (2005) suggest one might
  seriously entertain the hypothesis that an
  important first step in the evolution of modern
  human societies was a kind of self-domestication
  (selection on systems controlling emotional
  reactivity) in which a human-like temperament was
  selected (e.g. individuals within a social group
  either killed or ostracized those who were
  over-aggressive or despotic).

26
Should linguists now start working on dogs
understanding of human language?
27
Why dogs?

• Human and canine co-evolution
• Possibly also selected for attending to the
  communicative intentions of humans
• In addition (or after) selection for tameness
• Similarities in social cognition

28
What chimps cant do

• Fast mapping
• Avoidance of lexical overlap
• Inference by exclusion learning

29
What chimps cant do

• Possibly gloss
• I already know what these are, so the new word
  must refer to the new item
• OR
• If they had wanted one of those familiar items,
  they would have asked for one, so they must mean
  the new one
• Important difference - Theory of mind

30
Rico

• Vocabulary of over 200 words
• Retrieve familiar items from adjacent room
• 37 out 40 items retrieved correctly
• Fast mapping retrieve novel object from
  adjacent room
• In 7 of 10 sessions novel items retrieved
  correctly
• One month later
• 3 out of 6 sessions, novel items retrieved
  correctly
• In other cases, always retrieved an unfamiliar
  item

31
Conclusions?

• Fast mapping is mediated by general learning
  mechanisms also found in other animals and not by
  a LAD that is special to humans (Kaminski et al.
  2004)
• Question of degree?
• Rico 200 items humans 60,000 words
• Or are there more mechanisms involved?

32
Can dogs really learn words?

• Words vs. holophrases

33
Can dogs really learn words?

• Reference vs. associative learning
• Different contexts
• Shared attention
• Grammatical status
• Referential (Noun, Verb, Adjective etc)
• Non-referential categories (articles,
  conjunctions etc)

34
What could linguists study in dog perception?

• Knowledge of words for non-fetchable objects
• Knowledge of words in different contexts
• Understanding of different instructions
• Dont fetch an item
• Using attentional cues to disambiguate words
  monitor a speakers referential intent

35
Implications and (wild) speculation

• Perception divorced from production
• Saussurian signs
• Theory of mind
• Does word learning require a theory of mind?
• In what way might dogs be restricted if they
  dont have a theory of mind?
• Perception before (and independently of)
  production
• Reference in hearer not speaker
• What is left thats uniquely human?

36
(No Transcript)
37
References

• Belyaev (1979) Destabilizing selection as a
  factor in domestication, The Journal of Heredity,
  70, 301-308.
• Bloom, P, (2004) Can a Dog Learn a Word? Science,
  Vol. 384, pp. 1605-1606.
• Hare et al. (2002) The domestication of social
  cognition in dogs. Science, 298, 1634-1636.
• Hare et al. (2005) Social cognition evolution in
  captive foxes is a correlated by-product of
  experimental domestication. Current Biology, 15,
  226-230.
• Hare Tomasello (2005) Human-like social skills
  in dogs? Trends in Cognitive Science, 9(9),
  439-444.
• Hare, B., Call, J., Tomasello, M. (2001). Do
  chimpanzees know what conspecifiers know? Animal
  Behavior 61, 139-151.
• Hauser (2005) Our chimpanzee mind. Nature, 437,
  60-63.
• Horner, V. Whiten, A. (2005). Causal knowledge
  and imitation/ emulation switching in chimpanzees
  (Pan troglodytes) and children. Animal
  Cognition, 8, 164-181.
• Kaminski et al. (2004) Word learning in a
  domestic dog evidence for fast mapping.
  Science, 304, 1682-1683.
• Markman, E.M. and Abelev, M. (2004) Word learning
  in dogs? Trends Cogn. Sci. 8, 479-480.
• de Waal (2005) A century of getting to know the
  chimpanzee. Nature, 437, 54-59.
• Whiten (2005) The second inheritance system of
  chimpanzees and humans. Nature, 437, 52-55.