Problem solving - tutorial 2 Flashcards
Povinelli (2000)
see notes
The chimpanzees fail to understand which side to push from on the first trial.
Even when trained on one side then transferred to the other.
Chimps 50/50 on first trial
Not learning to avoid the trap – not understanding what they need to do – simply learning which end to push from
physics conclusions
McCloskey concluded that people have the wrong theory, a “naïve physics” such that when things curve they expect them to keep curving – at least for a while.
I don’t agree that they have a theory, but I do agree that they are influenced by their previous experience of the physical world.
- And, in that world, things that curve tend to keep curving…
The conclusion that I draw from these examples is that people often base their answers to these questions, in part, on their experience of the surface features of these problems.
- That is, how the problem looks rather than the underlying physics.
Thus people may also “rely strictly upon observable features” some of the time. But they do have other options…
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 other observers.
see notes
another Povinelli example
see slides
Whilst 5 year olds will pass the problem one slide back, chimpanzees do not perform satisfactorily on these tasks even after extensive training on the basics.
They cannot select the correct observer to beg from unless trained to do so over a number of trials, and then fail to transfer to a novel but conceptually related situation.
the false belief test (Wimmer and Perner, 1983)
The problem shown is a later development of the false belief task due to Baron-Cohen and associates.
The question posed to the child is “Where will Sally look for her marble?”.
Children of age 3 years old tend to answer (incorrectly) “Box”.
In fact Sally should look in the basket because as far as she knows it’s still there.
see notes
the false belief test (Wimmer and Perner, 1983)
The problem shown is a later development of the false belief task due to Baron-Cohen and associates.
The question posed to the child is “Where will Sally look for her marble?”.
Children of age 3 years old tend to answer (incorrectly) “Box”.
In fact Sally should look in the basket because as far as she knows it’s still there.
see notes
Children 5 and above tend to answer correctly, the question is - can we devise a similar test for chimpanzees?
No separation between what they know and what sally knows
false belief in chimps
Call and Tomasello (1999) devised this technique to test chimpanzees TOM.
The Communicator is allowed to see where the food is hidden, but then it is either moved without their knowledge or not.
The question is whether the chimpanzee can take this into account when deciding which box to pick to get the reward, on the basis of the communicator’s signals.
see notes
5 year old children solve this task, but chimpanzees fail to do so – eventually will learn
Communicator has a false belief
insight and problem representation (Knoblich et al., 1999)
“imagine these problems made of matchsticks - move just one matchstick, in order to make each statement true”
see notes
subjects much slower to solve type B problems than type A
harder to relax constraints of algebra
“insight occurs in context of impasse, unmerited in that thinker is competent to solve problem” (Ohlsson, 1992)
how?
- addition of new info (e.g. hint)
- “constraint relaxation” (e.g. I can change the =)
- re-encoding, where aspect of problem is re-interp (e.g. using pliers as weight in pendulum problem)
Duncker’s (1945) candle problem
subjects given candle, nails, matches
“please attach candle to wall so doesn’t drip on floor”
Re-evaluation of what the items are for to solve the problem
radiation problem (Gick and Holyoak, 1980)
imagine surgeon with patient who has stomach tumour
if tumour not destroyed, patient will die
condition such that operation impose: would also kill him
all have is machines producing rays that can destroy tumours; rays at sufficient intensity also destroy surrounding healthy tissue
fortress problem (Gick and Holyoak)
general trying to capture fortress
many roads led out from fortress
each road mined to only small groups of people could travel road safely
general needed forces to capture fortress
decided to split army and send small groups down each of roads so converged at fortress
radiation task perf
see notes
human poor at spontaneous use of analogy - many subjects also need relevance to be highlighted
use of analogy requires mapping of structure from source domain to target domain - infinite no. source domains
Underlying structure of problems is the same
analogical reasoning in chimps
Gillan, Premack and Woodruff (1981).
Here we have the conceptual analogical reasoning problem referred to in an earlier lecture.
Sarah can get these (mostly) right. By using two versions with the same alternatives, the experimenters rule out a simple associative explanation.
see slides
