2 CENTRAL METAPHORS OF THINKING & REASONING
PERCEPTION- automatic, follows predictable patterns
Where is thinking?
Where does your cognition end & a cultural tool begin? Ex: writing notes
We are learning A-RATIONAL stuff….????
GESTALT vs. SEARCH-INFERENCE FRAMEWORK
GESTALTISTS: PERCEPTION metaphor
SEARCH-INFERENCE FRAMEWORK: ACTION metaphor
(from computational tradition)
HOBBES: “by ratiocination I mean computation”…Turing
NEWELL & SIMON 1957-72
GPS = fundamental framework for AI & thinking.
Success: formalized a technical explanation for problem solving which led to further
P.S. FRAMEWORK - space, initial & goal states, operators, path constraints, solution
Solution: sequence of operations that can transform the initial state goal state while
obeying path constraints
Method: realized search space is combsplosionally vast. heuristic means-ends analysis?
featural, categorical view of probsolv
o transfer-specificity, domain-specificity
framing & formulation is a minor problem
o well vs. ill-defined problems
THORNDIKE cats in problem boxes. At no point do they really “get it”
Associationism: idea that simple learning mechanism (Law of Effect) splains intelligence. SO
greater intelligence = better functioning of law of effect?
KOHLER challenged this. Worked w/ Sultan. Said sultan & cat were doing very different things.
said Sultan was performing a massive restructuring of salient features – an ASPECT
SHIFT. This phenomenon was labeled, but not explained.
INSIGHT: THEORIES & DEBATES
Insight: a) a perception metaphor for problem formulation
b) ability to RESTRUCTURE ill-defined formulations
c) a qualitative change in learning
HEURISTICS: hill-climbing, trial & error, means-ends analysis
IS PROBFORM ANOTHER TYPE OF HEURISTIC?
Depends on which metaphor you use (WHY???)
IS PROBFORM PROCESS SAME AS PROBSOLV PROCESS? ****
Gestaltists – NO: formulation is like perception, solving is like action
POLYA: coined term “heuristics”
“provocative heuristics”: “tickle the memory”
Newell used these to MAKE the GPS, but did not make the GPS use these heuristics.
KAPLAN & SIMON?
WEISBERG & ALBA say insight is fake
gave 9-dot problem w/ VERBAL HINT – “go outside the box”
didn’t help to solve problems concluded insight was fake. Forgot to take into account
propositional/procedural distinction & transfer-appropriate processing
said fixation didn’t exist
compartmentalization of memory/search space into exhaustive domains …??? You
switch from one to another??? Memory Organization Thesis
o important idea, memory organization IS crucial, but mechanisms aren’t well-
Criticisms of compartmentalization:
o combinatorially explosive # of domains
o how to cross-reference between domains?
o How to label compartments
o Notion of what a domain/compartment is is vague
o how do you know when you’ve exhausted a domain
o How/when do you decide to switch domains BEFORE one’s exhausted?
Memory organization thesis: organized to help you solve problem
Magical memory thesis: ?
DOMINOWSKI – sufficiency vs necessity of breaking fixation
Breaking one’s fixation may be sufficient but unnecessary to “insight” probsolv.
This may be why Weisberg & Alba got these results.
Insight isn’t algorighmic. points to CASSELMAN & MAIER’S 9-dot-problem hints. our procedural knowledge
mechanisms are probs v. diff from propositional knowledge.
(dual processing model of cognition)
MAYBE WEISBERG & ALBA’S HINT HINDERED SOLUTION.
Assume a single solitary constraint to solving the problem.
In actuality, there are many mutually reinforcing ones … multiple simultaneous
Ex of perceived constraints: no crossing lines, staying within dots, making a non-dot
turn, having to make a closed figure…
Distinct from inference: a sequential process involving premises. This is generally more
propositional, & is applied to well-defined problems.
Maybe insight is so procedural cause it deals with MSCs, which require a different kinda
mechanism than the sequential inferential p