PSY370H1 Lecture 2: Study Guide Week 2-5
Week 2
Review three kinds of theories which are all inter-connected and inter-meshed
Descriptive
Does it exist? Can it be studied scientifically? What are the appropriate conceptual fameowrks
Normative
Generated by rational reflection, layout certain norms that can be used to obtaining truth
Explicit or implicit theory? Is it ecologically or externally valid?
Prescriptive
How can one design and teach?
Often a driving motivation for many psychological researches
Best methods for enhancing learning and transfer
Today
The search inference framework
Computational tradition: thinking=info processing, which is best understood as computation
Newell & Simon (1957)
Gestalt tradition: long debate w/ the search inference
Problem solving = a new search through problem space
Baron (1994) search for 3 kinds of things
1. Possibilities
2. Goals: the criteria by which you can evaluate possibilities
3. Evidence: belief or potential belief that helps you determine how well a possible X can
actually achieve a goal
Inference is the used of evidence to alter beliefs about possibility, goal, and evidence
Judgement is the evaluation of the interconnections between goals and possibility
The nature of search: they argue that problem’s formally represented by 4 kinds of elements
1. Description of initial state
2. Goal state
3. Set of operators, which are actions that can be altered
4. Path constraints, additional constraints on the solution beyond the goal (eg. Finding the
solution using the fewest steps)
A solution = a sequence of operators that can transform the initial state into the goal state while obeying
path constraints
Problem solving method = a procedure for finding a solution
FD =the search space.
F = the number of operators as any step and D = the number of steps
Combinatorial explosion
In order to solve the problem, one has to formulate the problem as such that most of the search space is
excluded
General Problem Solver (GPS)
www.notesolution.com
Try to develop a GPS that can shed light on how ppl actually solve problems
Research failed because of 2 fundamentally flawed assumptions
1. Problems are in some sense the same and form a unified category, which means that they
share a certain essential features
2. Formulating the problem is an easier job than solving the problem
Why is 1 wrong?
Often, how one solves a set of problem doesn’t transfer to how they solve other problems
Why is 2 wrong?
There are 2 kinds of problems: well-defined and ill-defined. Ill-defined problem means that the
representation of all elements (eg. Path constraints and operators) are incomplete, thus require
reformulation first
Mutilated chess board: almost everyone solves it topographically in the beginning, but once reformulated,
then the problem was solved in 3 steps
Insight: Is it a real psychological construct? Is there even a thing called insight?
The distinction between algorithm and heuristic
Algorithm:
A completely reliable routine or procedure that can be carried in a finite number of steps to solve a
problem
Heuristic:
A rule of thumb that do not guarantee a solution but often help in dealing w/ a problem
Most of our problem solving skills cannot be algorithmic (logical) in nature
Heuristic search: since heuristic doesn’t guarantee a solution, then mistakes are unavoidable
Bias is just a heuristic that showed up in the wrong context
Trial & Error: a myth, human beings don’t actually engage in trial and error
“Hill climbing” heuristic:
You have some way of knowing whether you are getting closer to the goal
(eg. Hill climbing method in dosage prescription by doctor)
Metaphor: the strategy isn’t really realistic, nor is it that powerful and useful
Means-ends analysis:
1. Comparison between current and goal state: identify salient differences between them (how to
identify salience?)
2. Select operator to reduce difference (which requires well-defined operators)
3. Apply operator if possible. If not, select a new subgoal under which the operators can be applied
However, this is recursive because of the problem of infinite regression. The number of attempts at sub-
goal is an ill-defined problem
4. Return to step 1
Productive Heuristics = heuristic procedures that can be used to resolve a problem
Provocative Heuristics = “tickling memory” or insight
www.notesolution.com
Are productive and provocative heuristics the same?
Kaplan & Simon (1990) argued that the two are the same, though slightly different processes, but “not too
different”
Week 3
Gestalt tradition
• 2 central ideas are that thinking is like perception and that the organizational structure is greater
than the sum of its parts
• the concept of restructuring is central to gestalt’s understanding of problem solving
Thorndike, Edward
• he framed and conceptualized intelligence as the ability to solve problem (which is reasonably
plausible conceptualization according to the prof)
• he tested this idea on cats initially and timed their escape time and the number of trials
• he found that the escaped time decreased with increasing number of trials, however, the graph
was jagged and gradual
• he did not find an insight or revelation where the escape time dramatically decreases in one trial
and form an S-curve
• he thus concluded that everything’s just gradual learning
However, the gestalt tradition challenged 2 things in Thorndike’s conclusion:
1. that problem solving is just bringing past experience into the present
2. that all problem solving all look like the graph Thorndike found for cats
Kohler, Wolfgang
• worked with Chimpanzee “Sultan” in an experiment structurally similar to Thorndike’s and found
that “Sultan” experiences an insight where he just gets it, much like the S-curve of learning
• however, his findings received methodological and cultural criticisms
• he introduced the idea that insight equals to a restructuring of the given material
Duncker, Karl
• coined the term “functional fixedness” in his book “On problem solving”
• in his candle problem, he showed that past experience can hinder problem solving just as easily
as it can facilitate, mainly through the mechanism of “functional fixedness”
• “functional fixedness” describes the difficulties in visual perception and in problem solving that
arise from the fact that one element of a whole situation already has a (fixed)function which has
to be changed for making the correct perception or for finding the solution to the problem
• functional fixedness is also demonstrated in Maier’s 2 string problem (1930) and Duncker’s
candle problem
• Duncker states counters Thorndike and states that functional fixedness is the blinding effect of
prior experience
• thus, problem solving requires a restructuring of problem formulation
• he suggested 2 methods of restructuring:
1. “suggestions from above”
2. “suggestions from below”
www.notesolution.com
Document Summary
Review three kinds of theories which are all inter-connected and inter-meshed. Generated by rational reflection, layout certain norms that can be used to obtaining truth. Often a driving motivation for many psychological researches. Gestalt tradition: long debate w/ the search inference. Problem solving = a new search through problem space. Baron (1994) search for 3 kinds of things. Inference is the used of evidence to alter beliefs about possibility, goal, and evidence. Judgement is the evaluation of the interconnections between goals and possibility. A solution = a sequence of operators that can transform the initial state into the goal state while obeying path constraints. Problem solving method = a procedure for finding a solution. F = the number of operators as any step and d = the number of steps. In order to solve the problem, one has to formulate the problem as such that most of the search space is excluded.
