Social intelligence, communication and meta-cognition Flashcards
mechanisms of social learning
Observational Conditioning
Stimulus or Local Enhancement
Imitation
Tool Use
Teaching
“social” learning in pigeons: it’s not what it seems to be
The pigeons are set up as shown, the bird on the left is the “Sender”, that on the right the “Receiver”.
The Sender can see two lights that cannot be directly viewed by the Receiver, the Receiver can, however, see the Sender.
The Receiver has two response keys, Left Hand (LH) and Right Hand (RH) to peck, and once again the Sender cannot directly view these keys.
The contingencies in play are such that the Sender and Receiver both get rewarded if the Receiver pecks RH when the green light is on and LH when the red light is on.
Can they communicate to solve this problem?
The answer is that they can solve the problem after a lot of training on it, but it’s not quite what most people would mean by “communication”.
observational/vicarious conditioning
e. g. fear of snakes in monkeys
- based on CC
- naïve monkey doesn’t fear snake
- sees fear reaction in other monkey when snake present
- fear of snakes conditioned
- may be particularly important in phobias
enhancement
e. g. blue tits and milk
- Behaviour spread across the country – are they learning from one another?
- Going over to stimulus as becomes interesting – get the cream which they like – reinforces behaviour
demonstrator’s behaviour draws observer’s attention to location or stimulus
imitation
when enhancement and observational conditioning can’t explain new behaviour
“Two-Action Test”
- Only reinforce one (either lever or pedal)
- Another observer animal watches
- Gets rewarded no matter which is presses
- Presses the one the demonstrator presses – imitation
with quail (+), rats (+), budgerigars (+), pigeons (+)
BUT: never a new behaviour?
Heyes et al. (1990, 1992, 1999)
The response used was pushing a hanging joystick either to the left or right (so not a typical rat behavior, and the bi-directional nature of the response controls for effects of enhancement, cf. Grindley’s demonstration of instrumental learning in the guinea pig).
Rats observed a demonstrator rat e.g. pushing the joystick to the demonstrator’s right, then were transferred to the demonstrator chamber and reinforced for pushing in either direction.
They preferentially pushed in the same direction as the demonstrator - imitation?
Scent mark when push with nose – only on side from where it pushed
More attractive to observer rat and so go to the same side
Need to replace joystick between demonstrator and observer
didn’t imitate when joystick changed
imitation in primates
potato washing in Japanese macaques:
- spreading in population too slow to be imitation (stimulus enhancement) – greater attention paid to certain stim – then giving it a go
Anecdotal evidence - problem of anthropomorphising - attribute human characteristics or behaviour to (a god, animal, or object)
Conclusion - evidence for true imitation at present inconclusive.
tool use and teaching
anvils to crack nuts, stick to fish for termites
teaching
An intentional facilitation of learning, with costs for the teacher.
Flavour preferences in rats or mobbing in birds are not examples of teaching
- May be that animals become familiar with certain stimuli – not necessarily imitation
Little evidence for intentional teaching in animals, only behaviour that helps the young to educate themselves and directs attention to an object, location or behaviour (e.g. Meerkats).
- Remove stinger from scorpion and given to youngster – teach animals how to cope with prey item
- Also a benefit – don’t have to deal with the youngsters anymore
comm/exploitation?
Advantage for the sender: manipulation of behaviour of others
Benefits for both sender and receiver ?
Altruistic behaviour?
- principle and moral practice of concern for happiness of other human beings and/or animals, resulting in a quality of life both material and spiritual
intention
“communication” = intention to send a message to the receiver
reliable effect on the receiver’s behaviour
can be between different species
referential comm
releaser/FAP communication: no decision, direct instruction to the receiver
referential communication: messages give varied information about matters external to sender and receiver
- vervet alarm calls
- other primate species
- The “dance language” of honey bees is referential
- Understand what the signal means and perform appropriate behaviour
Nim
His actual name was Neam Chimpsky
Terrace studied him intensively for years – he had tutors round the clock teaching him sign language.
Terrace was originally convinced that Nim could form sentences – i.e. “More banana” would be signed rather than “Banana more”.
But then he spotted something: Nim was using the signs provided by his tutors to produce his signs –he was effectively copying them or responding to them – but at such a short lag that they didn’t notice. You could only spot it on film.
His conclusion ultimately was that Nim did not acquire language. He was smart, he was fast, but he did not communicate in the way that humans can. He didn’t really generate his own sentences. He was simply good at producing responses that achieved the goal of getting food.
Terrace took a lot of flak for this conclusion and for this experimental work. There’s a recent documentary “Nim” which portrays him in an unflattering light. He stands by his conclusions, and feels that his portrayal in this documentary is unfair.
tactical deception
Dishonest signals might derive from operant conditioning
- E.g. monkey pretending to look at something so it can eat all the food – leads to good consequences
Anecdotes of primate dishonesty
Deception and intention
Understanding the intentions of the individual to be deceived
- some kind of Theory of Mind
theory of mind in primates
3 levels of evidence about animals’ apparently deceptive acts
- Level 0: Unintentional - the result is a “windfall gain”
- Level 1: Activities apparently directed towards a goal that can only be achieved if the receiver is deceived (but is learned, not understood) – did it because it worked not because the indv knew what it was doing
- Level 2: “Mindreading”, involving the attribution of intentions to another animal
‘mindreading’
Ability to infer what another animal can and cannot see
Attributes mental states to others
Acquired by age 5 in most children
ToM
A typical TOM problem is illustrated below.
In it two observers are depicted, one of whom knows what is in the box.
Which one should a child select to learn what is in the box?
The ability to answer this question correctly would suggest that the respondent had some idea that one observer possessed mental states that differed both from their own and from those of other observers.
see slides
self-concepts in primates
Theory of others’ minds theory of their own minds concept of the self
The mirror self-recognition experiment
- chimpanzees, orang-utangs, (some) gorillas = pass
- dolphins = pass
- parrots = not demonstrated so far
Gallup’s mirror test
Gallup (1970) studied the reactions of chimpanzees and macaques to their mirror reflections (8 hours per day for 10-14 days)
Over time, chimpanzees showed an increase in the number of self-directed behaviours that relied on the use of the mirror
- Grooming parts of the body that would otherwise be visually inaccessible
- Picking bits of food from the teeth whilst looking in the mirror
The monkeys, however, reacted to the mirror socially, as if treating it as a conspecific (as indeed had the chimpanzees during the first couple of days of exposure)
the mark test
After the exposure period, all subjects were anesthetised and bright red marks were placed on visually inaccessible locations on the body
- one eyebrow ridge
- the opposite ear
Upon re-exposure to the mirror, chimpanzees touched the mark more than other parts of their body
The monkeys did not touch the marks any more than any other part of the body
Chimpanzees understand reflection as being them
Macaques do eventually learn after many trials
MSR in other species
Epstein, Lanza and Skinner (1981) trained a pigeon to peck at a mark on its breast using a mirror
Gallup (1982) has criticised this experiment for the artificiality of the training and also the lack of any other evidence of mirror self-recognition
Thomson and Contie (1994) reported difficulty replicating this result
dolphin MSR
Reiss and Marino (2001) exposed two dolphins to reflective surfaces
Using a mark test procedure, they claim that the dolphins showed clear evidence of using the mirror to investigate marked parts of the body
self awareness and ToM
The mark test gives interesting results across different species - but is it truly an index of self-awareness?
- Able to show in a no. of species, just takes longer in some more than others
As I’ve remarked before, learning to control ones actions on the basis of feedback in a mirror is undoubtedly a skill (e.g. shaving) but does it imply self-awareness?
Each time you shave do you think about it? Do you have to go through some process of recognising “that’s me” before you can begin?
We might be better off trying to discover why some species pass the mark test and others don’t without getting bogged down in issues of awareness and self-concept.
As for ToM - there is little evidence of this in animals other than human - there is more on this in the Tutorial.
