Cognitive Psych Chapters 10-12

Tversky& Kahneman (1987)

-Asian Disease Problem: Outbreak of disease, expected to kill 600 people-Two groups -given different "Framing" of choices. •Positive Framing:-Program A:•200 people will be saved-Program B:•1/3 chance all 600 will be saved•2/3 chance 0 will be saved. Framed in terms of gain -- people are risk averse and take the safer choice (guaranteed saving 200 people). •Negative Framing:•Program A:•400 people will die•Program B:•1/3 chance all 0 will die•2/3 chance 600 will die. Framed in terms of loss, people tend to go with the riskier option. Willing to risk everyone dying for the chance of no one dying.

Causal reasoning

-Establishing cause and effect relations among events•I ate a warm egg salad sandwich from a gas station last night•Today I am sick•The sandwich caused me to be sick-We look for covariation (things that change together)•Change in life: ingested sandwich; change in life: sick-We also look for a mechanism that can link the two events•Bacteria grew in egg salad sandwich because not refrigerated -bacteria made me feel sick

Arkesand Blumer (1985)

-Procedure•Participants were to imagine that they had purchased tickets for two different ski trips: -Trip to Wisconsin for $50-Trip to Michigan for $100•Trip to Wisconsin was preferable because more enjoyable•Complication: two trips were on the same weekend and tickets are nonrefundable•Forced to choose one trip. So, you were out $150 regardless of what you choose. Do you go on the trip that would be more fun, or the one you paid more for? Participants opted for Michigan even though Wisconsin was touted as being more enjoyable. •Psychological accounting: not going to Michigan would "waste" more money-However, these costs were already "sunk" -The $150 had already been spent -should do what you would enjoy more

Problem solving process

1. Recognize and identify the problem 2. Define and mentally present the problem 3. Develop a solution strategy 4. Allocate mental resources for solving the problem 5. Monitor progress toward the goal and evaluate the solutions

Peter Wason

2,4,6 - cannot ask questions about the rule, but can come up with other sequences of 3 numbers and ask if they also follow the rule. Had parts talk aloud while reasoning. Identified 3 strategies: generate triplets to confirm their hypothesis (verification) - confirmation bias potential, generate triplets inconsistent with their hypothesis (falsification) - more likely to arrive at correct rule, and consider other variants of of their hypothesis.

Representativeness bias

A bias in reasoning where stereotypes are relied on to make judgments and solve problems (statistically more likely that Tom W is a social science major, but due to characteristics, you assume that he is a comp sci major)

Framing bias

A bias in reasoning where the context in which a problem is presented influences our judgment (contains 20% fat vs. 80% fat free). •Impact of framing (gain vs. loss)•People facing gain --risk averse-Avoid risk --protect what they have•People facing loss --risk seeking-Assume more risk to avoid loss. Imagine we flip a coin for $5.A) You win -- do you want to go double-or-nothing, or do you want to walk away with your $5 gain? B) You lose -- do you want to go double-or-nothing or do you want to walk away with your $5 loss?Most people will be more risk averse in A, and more risk seeking in B

Cognitive economy

A feature of some semantic network models in which properties of a category that are shared by many members of a category are stored at a higher level node in the network for storing efficiency. For example, the property "can fly" would be stored at the node for "bird" rather than at the node for "canary."

Conceptual hierarchies approach

A multilevel classification system based on common properties among items (game > card game > poker > texas hold 'em). Superordinate members and subordinate members. 'Is a' relationships. Subordinates inherent characteristics of superordinantes. We now know several things about a new game we've never played before because we understand games, board games, etc.) Also have transitive properties - if texas hold em is a kind of poker and poker is a card game, texas hold em is a card game.

Algorithm

A prescribed problem-solving strategy that always leads to the correct solution in problems with a single correct solution. You can consider the entire problem space, searching every possible solution

Tower of Hanoi problem

A problem involving moving discs from one set of pegs to another. It has been used to illustrate the process involved in means-end analysis.

Hill-climbing strategy

A problem solving strategy that involves continuous steps toward the goal state. Used when the problem solver poor understanding of the structure of a problem. Just keep moving toward goal. One of the problems with the hill climbing heuristic is that you can get stuck on a "local maximum", which isn't your ultimate goal state, but every possible move will be a away from your goal. Example: Rubik's Cube. There are often cases when you need to "backtrack" and move away from your goal briefly to actually reach the goal. If I blindfolded you in the middle of campus, and then asked you to find the highest point in Geneseo, you could use the hill climbing heuristic to do this. You could feel around you, and take a step in the direction that increases your elevation the most (i.e., the biggest step up). Then, you would do that again, and again, each time stepping in the direction that moves you up the most. Depending on where you start, you might actually find the highest point in Geneseo doing this, or you might get stuck on a local maximum.

Well-defined problem

A problem that has a clearly defined goal state and constraints. Know what your initial state is, know what your goal state is, lnow what operations/moves you have at your disposal. Most puzzles and games fall into this category. Often studied by psychologists because you can assess how optimal someone is in their problem solving (how close they are to most efficient solution)

Ill-defined problem

A problem that lacks a clearly defined goal state and constraints. May not know what the goal state is that you are trying to achieve, may not know all of the operations/moves at your disposal. Ex: Picking where to go to graduate school. More mentally difficult to represent and identify solution strategies for

Heuristic

A problem-solving strategy that does not always lead to the correct solution. Searches consider only part of the search space instead of considering all possible solutions, we mentally consider potential chains of subproblems, evaluating how each operator changes the current state. The approaches: means-end strategy, hill-climbing strategy, and working-backward strategy. Mental shortcuts we use to reduce the processing burden on our cognitive systems, but usually work by ignoring some information, which at times may result in making errors or biased conclusions: representativeness bias, availability bias, framing bias

Working-backward strategy

A problem-solving strategy that involves beginning with the goal state and working back to the initial state

Means-end strategy

A problem-solving strategy that involves repeated comparisons between the current state and the goal state. General Problem Solver program used this. With means-end analysis, we divide the problem up into subproblems that are easier to represent and handle. Each time we complete a subgoal, we are closer to achieving the final goal. Once we complete all of the subgoals, we have completed the final goal.

Smith and Osherson (1984)

A red apple was judged to be more typical of the combined concept 'red apple' than it was of either 'red' or 'apple', effect is even more so with atypical objects like a brown apple.

Typicality effect

A result where more common members of a category show a processing advantage. •Sentence verification --typical are faster -"A sparrow is a bird" vs. "A penguin is a bird" •Production --list types of birds -Robin listed before vulture •Picture identification (is this a bird?)-Wren faster than turkey •Explicit judgments (rate birdiness)-Wren more "birdy" than turkey •Categorization-Blue jay categorized as "bird" faster than flamingo •Family Resemblance (shared attributes)-Robin and wren more shared attributes (than ostrich)

Problem

A situation in which there is a difference between a current state and a desired goal state.

IDEAL framework

A step-by-step description of a problem-solving process. John Bransford and Barry Stein. Identify problems and opportunities, Define goals (think outside the box), Explore possible strategies, Anticipate outcomes and act, Look back and learn.

Insight

A sudden and often novel realization of the solution to a problem, the 'aha' moment. Gestalt psychologists suggested that we continue to unconsciously try to solve the problem once we've stepped away from it. Insight often occurs when particular barriers to problem-solving are overcome (you've suddenly restructured the definition of the problem and realized you can stack the pennies vertically).

Mental set

A tendency to approach a problem in one particular way, often a way that has been successful in the past - biases you not to see that Trial 3 of the water jug problem can be done by just A+C instead of the more difficult B-2C-A, but it's more work to come up with a new solution. A preconceived notion about how to approach a problem. Based on a person's past experiences with the problem (or similar problems). Water-jug problem: given mental set inhibited participants from using simpler solution (Luchins& Luchins, 1942). Sometimes we don't see a better solution to a problem because we are used to doing something a certain way. This "mental set" blocks us from finding more efficient solutions

Counterfactual thinking

Ability to reason about thing that could have happened but haven't. What if/if only.

General Problem Solver

Allen Newell and Herb Simon developed this computer program for problem solving. Proposed that problem solving typically proceeds by dividing the larger problem into smaller problems, searching for solutions to the smaller problems, then evaluating these solutions to see if they will bring you closer to solving the larger problem. Their approach radically changed the way the cognitive psychologists theorized about human problem solving. Searched through the problem space with algorithms

Feature comparison approaches

An alternative to the stored network approach, suggests that hierarchical relationships are computed using reasoning processes rather than being directly stored in a semantic network. Deciding how concepts are related involves comparing features of the two concepts.

Tversky& Kahneman (1974)

Anchoring bias was seen in the quick estimates of two number problems (8x7x6...x1 v. 1x2x3...x8). The initial number influenced the size of the estimated product, even though the actual products of the two series are exactly the same. A low initial number led to a low estimate, and a larger initial number led to a larger estimate.

Rosch et al. (1976, 1978)

Argued that basic-level objects are those at which the category members share the highest number of features, suggesting that these levels are more informative than others. Provide a lot of information about the category and are also distinct from other categories at the same level. The Goldilocks just right level

Berlin (1992)

Argued that conceptual hierarchies are a universal feature of all natural world categories.

Goel (2010)

Argued that performance patterns of brain-damaged patients (particularly with frontal lobe legions) suggest that there are neuropsychological differences between well defined and ill defined problems.

Cohen and Murphy (1984)

Argued that prototypes are better represented as schemata than as unstructured lists. Dimensions (often called slots) such as outer skin: feathers, number of legs: 2. Dimensions can be connected and limit on another - 0 legs means movement can't be walk.

Murphy and Medin (1985)

Argued that similarity-based theories of concepts fall short of adequately describing why concepts are coherent or meaningful because they don't take into account our theories of how the world works. Lightweight feathers and wings allow birds to fly - casual relationship is part of our general knowledge. Also take into account salient features (brightly colored comes to mind for birds, but less important for planes, although true).

Sloman (2005)

Argues that systematic manipulation of variables play a role in our everyday reasoning. We develop and update causal models through observing particular covariation and intervening in the causes and events.

Syllogistic reasoning

Aristotle developed. Process by which a conclusion follows necessarily from a series of premises (statements). If the premises are true, then the conclusion must be true. All A's are B's (first premise), all B's are C's (second premise), All A's are C's (conclusion). All is a quantifier (also no, some, some are not, and many). Syllogisms looking at relations among categories using quantifiers such as: All, Some, None

Category induction (inductive reasoning)

As we saw in chapter 10, our confidence in such inductions are influenced by the similarity between source and target and the typicality of the source. -Extending what you know about one member of a category to other members of a category-Discussed in the context of categorization in Chapter 10-If a certain pesticide harms squirrels, how likely is it that the pesticide will harm (chipmunks / coyote)?

Availability bias

Bias in reasoning where examples easily brought to mind are relied on to make judgments and solve problems. More words that start with the letter L or have L as the third letter - the latter, but harder to recall so we go with option 1 (Kahneman and Tversky)

Impasse

Blocked path; dilemma with no solution

Normative model for decision making

Break the decision down into independent criteria, weigh each criterion according to how important it is to the decision, list all the options and rate according to the list of criteria, option with the highest score is the decision we make. How we ought to think. Assumption: All other things being equal, people will try to maximize gain or minimize loss.

Similarity-Based Categorizations

Categorization involves judging the similarity between a target object and long-term memory standard. Prototype approach and exemplar approach. Rather than put things in or out of a category based on a list of features, proposes that we make category judgment by assessing how similar the to-be-categorized object is to some standard(s). Prototype and Exemplar approaches are going to differ in terms of what is used as the standard(s).

Chase & Simon (1973)

Chess experts have better memory for board positions than novices - had them remember the places of pieces and they could remember about 20 pieces. •Expert chess players have a better memory of chess piece positions than novices•Only for pieces "mid-game" -experts recognize strategies being used allowing them to chunk pieces•No advantage for experts with a random arrangement of pieces (no strategies to recognize -no chunks)

Concepts as definitions approach to categorization

Classical approach to concepts - Aristotle and Plato. Set of features in the definition are necessary (shape must have all these things to be a square) and sufficient (if something has all of these features, safe to assume it is a square). Difficult to categorize everything. you categorize as a 'game'. Failings include absence of clear necessary and sufficient features, no clear categorical boundaries, and typicality effects for category and non-category members ("apple" is a "better" example of fruit than "jackfruit"). Fuzzy boundaries - Is bowling a sport or a game?

Hierarchical Network Model

Collins and Quillian, 1969. One of the earliest models of how information might be organized in the mind. Each node is a concept, and the links are how concepts are associated (similar to the propositional representations we discussed in Chap 8). For example, the "tree" concept and the "bark" concepts are linked together with a "has" link. In other words, this represents the knowledge that you possess that trees have bark. Principle of Cognitive Economy. The idea that features are stored as high up in the hierarchy as possible leads to interesting predictions. If I ask you to verify whether something is true, the farther up the hierarchy you have to go, the longer it should take. Measure how long it takes people to say a sentence is true or false - salmon has gills (medium fast -- have to go one level up) vs. A canary is yellow (fast -- stores at same level). Findings were consistent with their model. This model can't explain typicality effects. A condor would have the same "IS A" connection to bird that a robin does. Can't explain why you would say a robin is a bird faster than you would say a condor is a bird.

Murphy and Wright (1984)

Compared feature lists generated for psychological disturbances by groups with levels of experience ranging from expert (PhD) to novice (undergrad). Results indicated that experts have richer conceptual representations and higher levels of agreement in feature lists.

Lesgold et al

Compared the ability of radiologists w/ experience to medical residents in finding tumors in xrays. Experts identified more subtle cues and relationships

Similarities between prototype and exemplar

Concept learning and categorization involve identifying features and marking comparisons, both place heavy emphasis on observable features and also largely ignore the role of prior knowledge in learning and using concepts

Deductive reasoning approaches

Conclusion interpretation approaches, representation-explanation approaches, surface (heuristic) approaches. Roberts (2005)

Kahnemanand Tversky (1983)

Conjunction fallacy. Presented participants with this description:-Linda is 31 years old; she's single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply consumed with issues of discrimination and social justice, and participated in anti-nuclear demonstrations.•Asked them to estimate the likelihood of various facts about Linda •Key comparison:•Likelihood of Linda being -(a) a bank teller-(b) a bank teller and active in the feminist movement. Many peopel will select (b), but that is impossible. The joint probability of two events (teller and activist) can't be greater than either individual probability.The Venn diagram below illustrates this. -People ignore basic probability principles-Two events in combination cannot be more likely than just one of the events

De Neys et al (2008)

Created scenarios designed to result in conflicts between our probabilistic and heuristic ways of processing (similar to Tom W comp sci major example), then asked to choose between 2 possibilities about that person based on the description, used fMRI scan to look at process in brain. Four conditions/2 cues: stereotype and base-rate cues. Incongruent condition pits base-rate cues (5 engineers, 955 lawyers) against stereotypical cues (likes math and puzzles), other 3 story types were controls. In congruent control condition one of the answers was consistent with both the base-rate and stereotype information, in the neutral control, there was no stereotype info in the story (base-rate would cue the answer), in the final control, base-rate information was the same for both groups (500 people in each group), so parts would base their answers on heuristic info. LOOK TO PPT FOR THIS

Galotti (2002)

Decision making is made up of 5 phases: setting goals (mental representations of a desired state of affairs), gathering information (options, likelihood of different outcomes, and criteria used to make your decision), structuring the decision (organize information in a way that will be useful for making the decision - pros and cons), making a final choice (choosing an option), evaluation (interpreting our choices and evaluating what went right or wrong)

McCloskey and Glucksberg (1978)

Demonstrated that category boundaries are not always so clear-cut. They presented participants with pairs of words with the second word being the category name. Participants had to quickly judge whether the first word was a member of that category. Some groupings were easy (chair, furniture) others were not (curtains, furniture) - these items had more disagreement across participants. There are levels to categorize with some examples being the best (sparrow = bird) and others being on the outskirts (penguin = bird)

Cummins (1995)

Demonstrated that familiarity of alternative causal mechanisms play a role causal reasoning. Colds can also be transmitted through touch, not just being sneezed on.

Garrod and Sanford

Demonstrated that reading time of an anaphor is faster if the antecedent it refers to is a typical category member.

Tversky (1972)

Described the elimination by aspect strategy - we dramatically limit the number of criteria we consider by first considering only the most important, if this criterion is sufficient to make our choice, then we do so, if not, we move to the next most important.

Elizabeth Warrington (1975)

Described three patients who had impairment of their conceptual knowledge reflected in deteriorated vocabulary (both production and comprehension) + their knowledge about properties of objects. Patients had difficulty naming pictures and describing characteristics of common objects.

Deductive reasoning

Determining whether a conclusion logically follows from premises. Making and evaluating arguments from general information to specific information. Vulcans are logical and Spock is a Vulcan, Spock is logical. Syllogistic (all, no, or some) and conditional. Deductive arguments can be assessed for their validity -- is the argument structure sound? Validity is different from "truth". If you have a valid argument structure, and the premises are true, then the conclusion is true. Validity is independent from truth. -Reasoning that is based on argument structure-Determine whether a conclusion logically follows from the premises-Using general principle to predict specific case•General Principle (Theory) Specific Case-Can assess validity of argument•Does the conclusion follow from the premises•Different from "truth" of the conclusion (depends on truth of premises) For both Syllogistic and Conditional reasoning:•Argument is validif conclusion follows logicallyfrom its two premises•If premises are true, the syllogism's conclusion mustbe true-Do not confuse "validity" with "truth"-Validity -argument structure-Truth -real world

Knowledge lean problems

Don't require extensive background knowledge. Again, often studied in Psychology labs because you can run these studies on college sophomores.

Pennies problem

Each penny can only touch 3 and only 3 other pennies. People see the problem in 2D, sliding horizontally is the only option, but in reality, you can stack vertically. People typically find the problem harder to solve when the initial state has some locations where the horizontally moved penny touches three other pennies, so people get hung up as they feel they're closer, where as with the other initial state, they're quicker to pivot to vertical stacks.

Prototype approach to categorization

Eleanor Rosch - sometimes referred to as the family resemblance approach or probabilistic approach. Prototype = "typical" or average member of a category. A prototype is a SINGLE representation of that category (may not be real). It is often thought to have the average or most typical features of all of the category members. Compare new object to prototypes of categories. Can account for typicality effects. In the prototype approach, typicality is basically the similarity to the prototype. Possesses characteristic features that describe what members of that concept are like (ex. prototypical bird). Views concepts as abstract representations (prototypes) that summarize the common and distinctive attributes of the members of the category that comprise the concept. The prototype of a category is essentially a weighted average of the important features of its members. Important features are those shared by most of the members (common) and not by members of other categories (distinctive). Problems: •Category representations include more specific information than prototypes would predict •We're sensitive to correlated information among category members •There are features that tend to co-occur together-Ex: big birds likely to squawk; small birds likely to sing-Wouldn't be captured by a single prototype. If your representation of "birds" was distilled down into a single representation, then features that are correlated across different types of birds would be lost when you calculated the "average" bird.

Conclusion interpretation approaches

Errors arise from general biases against making particular conclusions. People are reluctant to declare 'no valid conclusion' because they feel it is uninformative or because the number of terms can be reversed (called conversion). Some ants are insects and some insects are ants are logically equivalent, but ants are insects and all insects are ants are not.

Decision Making

Evaluate given information, arrive at a judgment, and, based on this, make a choice among alternatives. Research on reasoning and decision-making focuses on the mistakes people make

Medin, Lynch, Coley, and Atran (1997)

Examined categories and inductive reasoning in 3 types of tree experts and found that different group experts structured their conceptual systems differently (landscapers were goal derived, taxonomists and maintenance workers were more scientific and taxonomic). Across types of experts, inductive reasoning suggested that the genus-level categories were treated as the basic level of their hierarchy.

Knoblich et al (1999)

Examined how people solve the matchstick problem. We construct our initial problem space from our knowledge of the rules of math (correct the formula by using only one matchstick). Argued that the inclusion of math rules and chunks is what makes it difficult to solve the problem. Decomposing the equals sign is very tightly chunked, roman numeral III is less tightly chunked (you see the 3 Is). Level of rule and degree of chunking determined how quickly the participant solved the problem.

Lin and Murphy (1997)

Examined the influence of knowledge within a categorization task. Had 2 groups learn about an artificial tool and hear a story about the tool being used. Manipulated the functional importance of some of the tools features between groups. Participants were more likely to categorize exemplars that aligned with the important functions from the story. Whether they categorized a new item as a TUK depended on their "model or theory" of a TUK . Subjects were shown the same features at learning, but which ones were deemed critical or essential depended on their model of why a TUK existed. Results suggest the importance of general background causal knowledge for our conceptual system.

Chi and Snyder

Examined the role of past experience for solving the nine-dot problem. The used tDCS to temporarily inhibit the right anterior temporal lobe three minutes before stimulation, three minutes during stimulation, and three minutes immediately after stimulation. Those stimulated were much better at solving the problem as they weren't inhibited by past experiences that made them see the dots as a square with boundaries

Barsalou (1985)

Examined the typicality of taxonomic concepts like those used in Rosch and Mervis's (1975) study with a set of goal-derived concepts (how well items satisfy a particular purpose - birthday presents, food not to eat on a diet). Measure three variables: central tendency (family resemblance), frequency of instantiation (how often an item was considered a member of that category), and how well an item satisfied that goal (the ideal). All 3 variables were important, poking holes in the exemplar and prototype approaches that relied heavily on observable features.

Representation-explanation approaches

Focus on how we represent that arguments. The difficulty of an argument and the likelihood of making an error are the result of either incomplete information or incorrect representation of the argument. You maybe did not correctly represent your roommate's priority of getting food > leaving without you. Difficulties with working memory. Rule-based approaches (abstract rules, learned schemas, evolutionary), and model based approaches (Mental Models: e.g., Johnson-Laird (2001)).

Hsee (2002)

Focus on the criteria that are easy to evaluate - fairly easy to imagine laptop screen size, computational power less so

Associationist approach to problem solving

Focused primarily on trial and error. Over time, as we see patterns in what approaches work for certain problems, we use these associations when encountering new problems. This approach works well when there are relatively few possible solutions.

Functional fixedness

Focusing on how things are typically used and ignoring other potential uses in solving the problem (the screwdriver as a weight on a pendulum for the string of lights problem. Dunker's Candle problem: Attach candle above a table so that wax doesn't drip on table. Seeing boxes as containers inhibited using them as supports. Gestalt psychologists first identified this problem

Chi, Feltovich, and Glaser (1981)

Found similar results for physics problems comparing advanced doctoral students to undergrads. Found that experts were able to see past surface problems and perceive the underlying problem. The experts in this case knew to attend to the underlying principles (the structure of the problem), and were able to ignore the physical stimuli being used to illustrate them.

Tanaka and Taylor (1991)

Found that within areas of expertise, experts' basic level shifted to a lower level of the hierarchy, but when tested outside of their area of expertise, they went back to the basic level.

Barsalou

Found typicality effects for exemplars outside of categories - a chair is less of a bird than a butterfly is

Wason (1968)

Four-card task used to study inductive/conditional reasoning. Presented with four double-sided cards, letter on one side, number on another, if a card has a vowel on one side, then it has an even number on the other side. You can test this by turning over 2 cards to check - you are given A, D, 4, and 7. You should check A (p=q for p) and 7 because if 7 has a vowel on the other side then we can say the claim that a vowel resulted in a non even number (p != q for not q). People have trouble with modus tollens. People can do better with less abstract versions. -E and 7: Application of modus ponens (check E) and modus tollens(check 7)•Affirm antecedent and deny consequent-People make errors due to biconditional thinking and confirmation bias. Confirmation Bias: Tendency to seek out evidence consistent with a particular hypothesis Misinterpret as bidirectional (especially with abstract problem)-"If Vowel then Even" interpreted as "If Even then Vowel"•If bidirectional, it would be valid to check the 6•But it is not bidirectional, so it is invalid to check the 6. Why do people struggle on this task? If the if-statement were bidirectional, then you should check the 6 (but it is not so you shouldn't) People have a bias to look for information consistent with a hypothesis. When people pick the 6, they are expecting to find a vowel, which would confirm the hypothesis. Most people don't pick the 7 -- doing this is an attempt to disprove the hypothesis by finding a vowel. Again, people don't naturally do this.

The mutilated checkerboard problem

Full checkerboard can be covered with 32 dominoes occupying two squares each. Two squares from opposite corners are removed. Can the remaining 62 squares be filled with 31 dominoes? Typically requires a change in problem representation. This problem is hard if you try to solve it spatially using imagery. We don't have the working memory capacity to represent the location and orientation of 31 different dominoes.The problem becomes easier if you represent it differently. Rather than representing it spatially, represent it numerically. Regardless of the orientation or location, each domino will cover one black and one white square. . . . . . . so there needs to be an equal number of black and white squares.What color were the squares that were removed? Is it possible to tile the 62 squares?

Chi & Snyder (2011)

Further investigated the role of the right ATL by having participants solve the matchstick problem. Used tDCS to selectively stimulate different areas of the brain. One had R+, L- another had R-, L+, the third was control. R+, L- patients solved more insight problems than the other two groups, probably due to diminished top-down information.

Wason & Shapiro

Gave a version of the four card problem with a traveling task - location and one side and method of travel on the other. Condition to test is 'every time I go to Chicago, I take the train.' You turn over the Chicago card to check (not the NYC card bc it doesn't matter how you get to NYC) and the plane card (if the other side is Chicago, you know the claim is false) not the train card (because you can go other places on the train too).

Positive transfer

Gick and Holyoak (1980) - past experience with analogous problems can be a powerful strategy to solve new problems.

Rosch and Mervis (1975)

Had one group of subjects rate the typicality of different members of categories. Ordering shown for 3 of the categories (pants was the most typical clothing, hat was #16 in clothing). They then had a different group of subjects list the features shared by the top five members of category, and the bottom five. Common features of: pants, shirt, dress, skirt jacket? Common features of: hat, apron, purse, wristwatch, necklace More typical members had more things in common

Griggs & Cox (1982)

Had parts imagine that they were a police officer assigned to enforce a law requiring those who drink alcohol to be at least 21 years old. Cards representing different patrons of the bar. No point in checking the soda drinker (doesn't matter how older they are) or the person who was over 21 (doesn't matter what they drink). 75% made the correct choice. Demonstrates that we aren't good at logical reasoning, but do use it in some contexts.

Hobbits and Orcs Problem

Hobbits and OrcsProblem -Three Hobbits and three Orcs on one side of river -All want to cross, but boat will only hold one or two creatures -The boat cannot travel by itself -Orcs cannot outnumber Hobbits on either side of river •(Hobbit stew) -How can everyone get across safely? Solution: Fairly linear•Initially, each step moves creatures closer to their goal states (other side of river)•People slow down at step with circled Orc-Orc who has been in the goal state must move away from goal state-People hesitant to do this•They take longer to do this•Use other (incorrect) moves The Hobbits and Orcs problem demonstrates that people have a bias to use hill climbing -- choosing moves that move you closer to the goal and not wanting to move away from the goal.

Murphy and Ross (2005)

How certain we are that something is a member of a category impacts the likelihood of making inductions (not sure whether or not a bat is a bird but certain that a robin is a bird).

Descriptive Approach

How we actually think

Galotti (1989, 2002)

Identified many potential differences between reasoning in the lab and reasoning in everyday life. Lab problems are well defined, with clear premises supplied, a single correct answer, and arguments that are evaluated because the researchers ask you to evaluate them. In everyday life, much less defined, not always easy to know what the premises are, arguments are probably personally relevant (aimed at a goal), and there may be several answers that vary in quality.

Hypothesis Testing

If we think there is a causal relationship between two things, we might try to test that hypothesis by running an "experiment". -One way to assess whether there is a relationship is to manipulate situation and observe outcome-People tend to focus on confirming hypotheses-People have a harder time devising ways to falsify their hypotheses-For example, in the Wason 4 card task:•Selecting the E is trying to confirm the hypothesis that there is an even number (people readily do this)•Selecting the 7 is trying to disconfirm the hypothesis by seeing if there is a vowel (people tend to not do this)

Knoblich et al (2001)

In a follow up study, examined eye movements of participants trying to solve the matchstick problem, predicted that their eye movements would indicate how they were trying to solve the problem, as people tend to stare at what they are thinking about. Parts focused on the roman numerals, not the operators, and fixations got longer as time went on (they had eliminated other options and only stared at the options they still deemed viable). Initial representation of the problem led to an impasse, participants needed to re-represent the problem.

Conjunction Fallacy

In addition to ignoring base rates, people also ignore other statistical principle when they can use the representativeness heuristic.

Ohlsson (1992)

In order for insight to occur, we need to change the problem representation in one of three ways: -Constraint relaxation (Davidson's selective encoding): Inhibitions on what is regarded as permissible are removed, nine-dot problem example -Re-encoding (Davidson's selective combination): Some aspect of the problem representation is reinterpreted, mutilated checkboard example -Elaboration (Davidson's selective comparison): New problem information is added to the representation, what if Cleopatra is a goldfish example

Analogical reasoning

Inductive reasoning. Process of using the structure of one conceptual domain to interpret another domain. A tree is to a forest as a solider is to ___. The part-whole structure between tree and forest and that then gets mapped onto the second part so the soldier fits the part and you identify the whole. Using what we know about one domain to predict another domain. How successful this is depends on how well the underlying structures match .For example, trying to apply how we disinfect surfaces to how we might disinfect people is a bad analogy (and dangerous). People can solve analogical reasoning problems like this if they recognize the underlying structure.

Problem solving as problem space searches

Inspired by the General Problem Solver computer program. The problem space consists of the mental representation of the set of intermediate states, subgoals, and operators (the actions that can be performed to change states). The General Problem Solver searched through the problem space with algorithms. Problem space-Initial state-Intermediate state(s)-Goal state-Ex: Tower of Hanoi

Distributed Representations

Instead of a concept being stored in a single location, the different properties of the object (such as the color, shape, motion, action/use, name, sound of an object) would be represented in specialized areas. Not all areas are used for all types of categories. Representation of a hammer would contain links between areas representing relevant features of a hammer: shape, action, name, sound (color wouldn't matter). Representation of a sheep would contain links between areas representing relevant features of a sheep: shape, motion, name, sound, color (action isn't relevant)

Tweney et al

Instead of classifying triplets participants generated as yes or no in Peter Wason task, they classified rules as either Med or Dax. 2,4,6 = Dax. Performance increased dramatically, as having 2 rules to consider may have reduced confirmation bias.

Mental models

Johnson-Laird (2001). Mental Models: e.g., Johnson-Laird (2001)•We create models of the world in order to evaluate arguments•Rather than use "rules" we use "simulations". -Johnson-Laird proposed a three step process:•Model Construction: -Build mental models of the world described by the premises•Conclusion Formulation: -Combine the models of the separate worlds into unified representations -Keep the models that are consistent with both premises•Conclusion Validation: -Look for models inconsistent with our conclusions. •Example: -All of the Sithwear black-None of the Jedi are Sith-Therefore, none of the Jedi wear black All of the Sith wear Black (2 possible models). There are 2 possible "worlds" that could correspond to the first premise. The Sith are a subset of the black-wearers OR The set of entities that are Sith and wear black perfectly overlap. None of the Jedi are Sith (only 1 model). There is only 1 possible world that corresponds to the second premise. When you COMBINE the worlds created for each premise, there are different possible worlds that satisfy both premises (I left out one for simplicity). In this model world, all of the Sith are wearing black, none of the Jedi are Sith, but yet, there are Jedi wearing black. Therefore the argument must be invalid.

Prospect theory

Kahneman & Tversky (1979) - explained many of the heuristics and biases within this framework. Descriptive Theory (what people actually do). Biases in decision making often resulted from the fact that we do not treat gains and losses equally (loss aversion takes favor). Also assumes diminishing returns, a gain or loss of $100 matters a lot if we have a balance of $1,000 in our bank account, but matters very little if we have 100X that. People tend to also overweight low-probability outcomes and underweight high-probability outcomes - people fear flying more than driving. •Key Points-People identify a reference point generally representing their current state -think about new state as gain or loss-People are more sensitive to potential losses than potential gains (loss aversion) •"Value" for losses grows more quickly-$5 bill in storm sewer. If you declined the bet I offered before, it is because the loss of $10 is represented as a greater change in value than a gain of $20.

Kahnemanand Tversky (1984)

Kahnemanand Tversky(1984) •Observed that people often violated expected utility theory•I offer you the following bet: I flip a (fair) coin•Heads: I give you $20•Tails: You give me $10•Would you take the bet?•Should you take the bet?•Take Bet: (.5 x $20) + (.5 x -$10) = $5 gain•Decline Bet: 1.00 x $0 = $0 gain

Conceptual knowledge

Knowledge that enables us to recognize objects and events and to make inferences about their properties

Basic-level concepts

Level of concept hierarchy where common objects reside - these levels are privileged over other levels (dog more likely than border collie and also more likely than mammal - middle level). Children learn these basic categories and their names before other levels. Basic categories typically share common shapes and movements, allow for faster categorization of pictures and are used more frequently in naming.

Heit and Rubenstein (1994)

Likelihood on induction depends on how relevant it is to the kind of categories being compared (more likely to infer something about heart rate (anatomical property) than a migration pattern (behavioral property) between two species that are both mammals). However, if the categories are related by virtue of their environment (a tuna and a whale both live in the sea), then you are more likely to make an induction about a behavioral property.

Hypothesis testing

Make and test an educated guess about a problem/solution.

Inductive reasoning

Making and evaluating arguments from specific information to general information. Going from specific cases to general principles. Cannot assess inductive arguments as valid or invalid. Strength of argument -assess as more or less likely to be true. Dr. Watson had recently been caught in a rainstorm based on his observation of his shoes, several parallel cuts on the leather must have resulted from careless scraping of mud from the sole and the mud resulted from a torrential rainstorm. Examine the likelihood of a conclusion being true, rather than its absolute truth Types: analogical reasoning, category induction, causal reasoning, hypothesis testing, counterfactual thinking, everyday reasoning. We judge inductive arguments as being more or less likely to be true. •Dr. Kirshis funny•Dr. Mounts is funny•Therefore all GeneseoPsychology professors are funny -Reasoning that is based on observation-Drawing conclusions from evidence-Use specific case to postulate general principle•Specific Case General Principle (Theory)-Product of inductive reasoning is not necessarily "correct"-Inductive arguments are assessed in terms of their strength, rather than validity-Ex:•Dr. Kirshis funny•Dr. Mounts is funny•Therefore all Geneseo Psychology professors are funny

Category induction

Making inferences -use knowledge of category to make predictions about category members you haven't encountered before. We use concepts to make predictions about new objects and categories. You know what taking car of a cat will entail even if you've never seen the cat or taken car of one before. One of the most important functions of our conceptual system.

Reingold, Charness, Pomplun, and Stampe (2001)

Measured the eye movements of chess experts and novices. Experts spent more time looking between pieces as they were focused on the overall structure of the board vs. novices with individual pieces. Can focus on higher order problems and goals.

Concept

Mental representation used for a variety of cognitive functions

Representing Problems

Mentally representing the current and goal states, the roles or constraints, and the allowable operations available to solve the problem. Incorrect representations of constraints and/or the allowable operations hinder problem solving. You need to assess where you are now, what end state you want to be in, and what types of changes you can make to move from one state to the other.

Moreno et al (2002)

Monitored the eye movements of experts and novice gymnastics coaches. while they viewed routines - experts had more and longer fixations on regions that were critical to performance and shorter fixations on non-relevant areas.

Neuroscience inspired approaches

Non of these models include explicit representation of hierarchical conceptual structure (and in some cases, no direct representation of concepts). It is widely believed that our conceptual knowledge is distributed across multiple areas of the brain involving both perception and action (how an orange tastes, smells, how it can be dissected, likely all live in different areas of the brain).

Roger Brown (1958)

Observed that parents typically refer to middle level/basic-level concepts more often when speaking to their children.

Expert vs. novices:

Perception and attention (know what to focus on and what to ignore), memory (mentally group aspects of problems differently, more likely to focus on the underlying structure due to analogical transfer), better strategies (more time spent analyzing = more relevant knowledge to their strategies - better defensive structure in chess vs. just accounting for each move)

Kounios & Beeman (2009)

Performed a series of experiments using both EEG and fMRI to examine insight problem solving. Their studies indicate that insight is the result of a series of brain states that operate at different time scales. They implicate an important role of the anterior temporal lobe for solving insight problems.

Decoy effect

Preferences for either option A or B changes in favor of option B when option C is presented, which is similar to option B but in no way better. Small vs. large you would not be drawn to large option, but with medium added in priced close to the large size, medium is the decoy as it pails in comparison to the large for the price.

Gick and Holyoak (1980)

Presented participants with a similar problem before presenting them with the Duncker radiation beam problem. Attack the fortress by sending small armies over each road vs. entire army through one road. Surface features of this problem are different from the beam problem, but underlying problem is the same. Found that 70% of people solved the second problem after seeing the first vs. 10% that didn't - but only when given a hint that the two problems might be related. Demonstrates positive transfer.

fugelsang, thompson, and dunbar (2006)

Presented participants with brief stories containing an event and possible cause of the event, they were then asked to rate how likely it was that one caused the other. Results showed that strength of the inferred causal relationship were dependent on belief and covariation (low and high). High B/High C: Slippery roads may be due to ice storms Low B/High C: Slippery roads may be due to slippery sidewalks High B/Low C: Slippery roads may be due to rainfall Low B/Low C: Slippery roads may be due to excessive traffic

Allen and Brooks (1991)

Presented participants with digger vs. builder animals that were distinguished by 5 features, 3 that were relevant (leg length, angularity of body, spotted or not), and 2 that were irrelevant (number of feet and length of neck). Researched created cartoon characters that were 'good' matches (changing a feature that was irrelevant) or 'bad matches' (one that was relevant). Subjects learned the 2 categories in the learning phase where they were presented with 8 characters and were then tested with new examples. Parts were slower and made more errors when categorizing 'bad' matches. Suggested that participants were relying on the similarities to the specific learned exemplars rather than relying on an abstraction like a prototype.

Novick (1988)

Presented students with a series of math word problems - some structurally similar, others surface. Math experts showed greater positive transfer between analogous problems relative to novices. Also show less negative transfer between nonanalogous problems that share surface features.

Mental logic theories

Propose that our deductive reasoning proceeds by applying a set of rules. Some of these theories propose context-free (what the argument is about doesn't matter) rules that operate on on propositional representations of the of the premises. Propositional statements are either true or false. Other theories propose that the context of the rules do matter (Cheng and Holyoak proposed that we reason using sets of rules defined with respect to particular goals) learned through ordinary day-to-day experiences. Cosmides by contrast, argued that we have evolved to reason using rules related to social exchanges (we are born knowing if I do something for you, you will do something for me).

Cheng and Novick (1992)

Proposed a model in which our causal reasoning is based on probabilities of an event happening (getting a cold) with and without the causal event (being sneezed on or not).

Smith, Osherson, Rips, and Keane

Proposed a model wherein concepts are represented as prototype schema with dimensions and values. Further research has shown limitations of the model (sometimes it is difficult to predict which dimensions of combined categories modify each other - apple and farmer with organic apple and organic farmer).

Philip Johnson Laird

Proposed one of the most influential theories of reasoning - reasoning proceeds through three stages: model construction (building a mental model of the world described by the premises), conclusion formulation (mental model of the premises are integrated such that consistent models are conjoined and inconsistent ones are discarded), and conclusion validation (we look for models that falsify the conclusion and reason a conclusion is valid if it doesn't). Working memory limitations interact with the reasoning process - the more mental models required, the more we are prone to error.

Barsalou (1999)

Proposed perceptual symbols theory of conceptual representation that has its roots grounded in an embodied theoretical framework. Proposes that our conceptual system is largely perceptually based rather than based in amodal symbolic representations. Apple = what it looks like, smells like, tastes and sounds like when we bite it.

Social stereotypes

Proposed that they are part of a 2 phase process - automatic activation of stereotypical knowledge (when we encounter people, we quickly sort them into a category based on readily available features), followed by more deliberate and controlled processing - we can overcome our initial categorization.

Chater and Oaksford

Proposed the probability heuristics model. Everyday reasoning is not based on logic but rather on probability. Rather than treating the premises as statements of truth, we analyze the probability of the premise and the strength of an argument (what is the probability of something being an any knowing that all ants are insect - 100% chance something is an insect if it's an ant, if no ants are insects, 0% chance. We rarely encounter 100% or 0% in life.

Janet Davidson et al

Proposed three mental processes involved in insight: selective encoding (restructuring so that information originally viewed as irrelevant becomes viewed as relevant), selective combination (a previously non-obvious framework for relevant features becomes identified - pennies can be moved in 3D), and selective comparison (you discover a non-obvious connection between new information and prior knowledge).

Conditional reasoning

Propositional reasoning. Inclusion of connective words like if and then (also and or and not). Statements are either true or false. Antecedent (premise) and consequent (valid conclusion). If it's sunny, I will walk to class = if it's not sunny, I might still walk to class. Knowing that you walked to class won't tell you anything about what the weather is like. With conditional reasoning, we are going to see that there are only 4 forms that arguments can take. The antecedent is the "If" part of the premise. The consequent is the "then" part of the premise. Goal is to determine the validityof the conclusion, based on the premises. People tend to interpret conditional statements as bidirectional but they are not. If P then Q ≠ If Q then P. If it is raining, then I will wear boots ≠ If I am wearing boots, then it is raining Four Argument Forms: -Affirming the antecedent (modus ponens): Valid-Affirming the consequent: Invalid-Denying the antecedent: Invalid-Denying the consequent (modus tollens): Valid "If-then" reasoning-Evaluating whether a conclusion is valid, given that certain premises hold

Surface approaches

Reasoning relies primarily on general heuristics focused on the surface properties of the quantifiers in the argument rather than on reasoning analytically. IF the premise contains all or no, then the conclusion probably will as well, same with (no or some... not) then the conclusion will be negative.

Metcalfe and Weibe (1987)

Recorded participants' feelings of warmth while engaged in insight and non-insight problems. Recorded "warmth" every 15 seconds while solving problem. While solving the problems, subjects had to give a 1-7 rating every 15 seconds on how close they thought they were to the solution.When they start, they should be at 1. When they are about to solve it, they should be at 7. After Solution, looked at warmth: 60, 45, 30, 15, 0 seconds before solution. There was a progressive increase in warmth during non-insight problems. With insight problems, warmth ratings remained at the same low level until suddenly increasing right before the solution was reached Non-insight problems (you get closer and closer to the solution): •Given containers of 163, 14, 25, and 11 ounces, and a source of unlimited water, obtain exactly 83 ounces of water •If the puzzle you solved before you solved this one was harder than the puzzle you solved after you solved the puzzle you solved before you solved this one, was the puzzle you solved before you solved this one harder than this one? •Tower of Hanoi -change from start (left) to goal state (right) Insight problems (path to solution would occur all at once): •A prisoner was attempting to escape from a tower. He found in his cell a rope that was half long enough for him to reach the ground safely. He divided the rope in half, tied the two parts together and escaped. How could he have done this? •The triangle points to the top of the page -move three circles so that it points to the bottom of the page •A landscape gardener is given instructions to plant four special trees so that each one is exactly the same distance from each of the others. How is he able to do it?

Knowledge rich problems

Require extensive expertise in an area.

Conceptual combination

Research suggests not simply the intersection of two categories (game night is not simply game and night).

Thinking aloud approach to research

Researchers may gain insight into how people represent a problem, but not all problem solving abilities are consciously accessible + people may not report all of what is consciously accessible (the strategies that came time mind but were quickly rejected)

Neuroscience approach to decision making

Researchers suggest that there is likely no single, unitary reasoning or decision-making system in the brain, but instead distributed systems that dynamically respond to particular task demands and environmental cues (Goel, 2007)

Patterson et al (2007)

Revied two concepts related to conceptual knowledge storage in the brain: one that suggested that our concepts are directly represented within the connections between sensorimotor areas (distributed-only view). The other suggests that distinct areas of the brain (convergence zones or hubs) bind features together such that there are conceptual representations distinct from sensory motor areas (distributed-plus-hub view).

Collins & Loftus (1975)

Revised the semantic network model's proposed changes to how activation spreading occurs and the addition of variables weighting on the connections between concepts (link between bird and robin is stronger than penguin and bird). Activates a node-activation then spreads to related concepts. Nodes represent specific concepts. These concepts are associated with one another through links (with different relationships). What you know about a concept is represented by what that concept is linked to. For example, find Shark in the network -- what you know about sharks would be linked to sharks or other associated concepts (like fish). If I ask you to list birds, activation would spread out from the bird node. Robin would become activated by Bird before Condor would. The strength of the connections between concepts is shaped by your environment

Using Analogies to Solve a Problem

Russian marriage problem (Hayes, 1978) In a small Russian village, there were 32 bachelors and 32 unmarried women. Through tireless efforts, the village matchmaker succeeded in arranging 32 heterosexual marriages. The village was proud and happy. However, two of the bachelors were killed in a hunting accident. Can the matchmaker redo the matches to come up with 31 heterosexual marriages? Same underlying structure as the mutilated checkerboard. Analogical transfer: The transfer from one problem to another Gickand Holyoak in order to use analogy to solve a problem: -Notice relationship -Map correspondence between source and target -Apply mapping

Wilson & Schooler (1991)

Selected five varieties of jams that had been independently rated by experts for quality ranging from top-ranked to one of the worst and then asked college students to say what they did and didn't like then rank (rankings were very different from experts - used system 2 analysis). In another condition, asked questions about why they selected their college majors then rate the jams (their ratings more closely matched the experts - only used system 1 automatic).

Brewer Dull and Lui (1981)

Stereotypes of the elderly may be represented hierarchically and that within it most stereotypical behaviors tend to operate at a basic level. Findings support the idea that stereotypes come from our general conceptual system.

Wilkins (1928)

Suggested several factors impact how likely we are to follow the rules of logic: phrasing of premises (arguments with some are harder than those with all), how people understand language (some does not means some mammals are docs and some are not dogs, logically, it means at least one mammal is a dog, but there may or may not be other mammals that are not dogs), and we are influenced by the content of what we are reasoning about (mammals being dogs vs. A, B, C)

Dijksterhuis (2004)

Suggested that system 1 thinking that is more automatic and unconscious can result in better reasoning. Asked to select apartments and roommates from a list, then were presented with alternatives with pros and cons for each one (designed to have one best option and one worst option based on number of relative pros and cons). Subjects who were in the unconscious consideration condition made the best decisions. These results suggest that everyday reasoning may be better when it involves more unconscious processing > conscious, deliberate thought.

Gestalt approach to problem solving

Suggests that people go beyond past associations, and that solutions arise out of new productive processes (creating new mental representations of information structured to achieve particular goals). Focused on the structure of the representation of the problem. Thought that associationist approach suggested that problem solving was a gradual process and that many nonsensical errors would occur every time people try to solve problems. Thinking aloud research showed that people appeared to use more systematic strategies, rather than trial and error. Solutions arise out of new productive processes, creating mental representations of information structured to achieve particular goals, insight processes, analogical transfer

Mayberry, Sage, and Lambon Ralph

Support for typicality effect. Demonstrated that impairments of conceptual knowledge (semantic dementia) are constrained by concept typicality. When asked to match members to categories, subjects made more mistakes with atypical members.

Duncker (1945)

Surgeon with patient with inoperable stomach tumor. Beam of radiation might work, but high intensity beam with destroy the healthy tissue - you can adjust the intensity of the beams, you can have more than one beam. This is the 'dispersion' solution. Classic example on an insight problem (aha).

Dual process framework

System 1 processes are assumed to be automatic, system 2 processes are slowed, controlled, and often conscious. Has been proposed to explain why we may reason differently at different times. Wilson and Schooler and 5 varieties of jams problem.

Rips 1975

Systematically examined how we make inferences by asking participants the likelihood that a disease would spread from one species to another on an island. Found that ratings depended on 2 factors: typicality of the diseased species (typicality of the target species had no effect) and similarity between diseased and target species. If something is true of a typical member of a category, we are more likely to generalize to other members of the category.If a Toyota Camry has a certain feature, we would expect all cars to have this feature.If a Tesla Model 3 has a certain feature, we would be less likely to expect all cars to have this feature. More likely to generalize between exemplars that are similar.If a Tesla Model 3 has a feature, we would be inclined to think that a Nissan Leaf would have this feature(note: both are all-electric cars)

Stored-network approach/Hierarchical Network Model

Taxonomic hierarchy approach. Hierarchies are stored in memory as networks of relationships. Collins and Quillian proposed that when an object is categorized, 'activation' spreads from that object's corresponding concept node to other associated nodes. A canary is a bird would spread activation from canary to bird, if spreading activation intersect, the answer is yes. Major advantage is cognitive economy. Drawbacks: it is not able to account for typicality effects and has no mechanism for explaining why some subordinates were considered better than others. Also transitive property sometimes fails (lamp = furniture, car headlight = lamp, car headlight != furniture). Among the most widely adopted and persistent theoretical approaches within cognitive psychology.

Negative transfer

Using problems that have strong surface structure similarities but different underlying structure leads individual to attempt the wrong solution. Glick and Holyoak participants needed the hint of using the army problem to solve the surgeon problem because they were too focused on the differences in surface structure to see the similarities in underlying structure.

Superordinate concepts

The broader concepts in a three-level hierarchy of concepts. Tend to be distinctive but not as informative.

Distributed + Hub View

The distributed plus HUB view says that a central area is responsible for coordinating activation of the features in the different areas. Different areas of the brain are specialized for representing different features, such as the color, shape, motion, action/use, name, sound of an object. In addition, there is a "hub" area that ties these different features together. The anterior temporal lobe (shown in red above) has been suggested as the potential hub

Reasoning

The evaluation of a conclusion based solely on given information. Research on reasoning and decision-making focuses on the mistakes people make

Nine dot problem

The goal of the puzzle is to link all 9 dots using four straight lines or less, without lifting the pen and without tracing the same line more than once. As you are trying to solve the problem, are you keeping you lines "inside the bounds of the square"? People are restricted by the perceived 'borders' given the arrangement of the nine shapes. Problem solving times are often sped up.

Exemplar approach to categorization

The idea that concepts are represented based on exemplars of the category that one has experienced previously, rather than to abstract "average" - categorization of an object is accomplished by comparing it to all of your memories of similar things. Items are more typical if you encounter them more often. Takes into account atypical cases-Ex: Emu similar to an ostrich and deals with variable categories-Ex: chess not similar to dodgeball, but is similar to checkers. Routed in comparison to memories of actual experiences as opposed to comparison to abstractions of those experiences. Most results from experiments using artificial concepts favor this over the protype model, but we could be using both. There is nothing saying we couldn't use both approaches, and evidence suggests that we do. In some situations, exemplar theories provide better "fits" to behavioral data. In other situations, prototype theories provide better fits.

Concepts based on world knowledge approach

The idea that the representation of concepts includes how well their members satisfy a particular purpose based on world knowledge (casual relationships, salient features, etc.)

Subordinate concepts

The narrower level of concepts in a three-level hierarchy of concepts. Tend to be informative but not as distinctive

Categorization

The process by which things are placed into groups. Categories are all possible examples of a particular concept. Useful for helping us understand individual cases not previously encountered, act as pointers to knowledge (categories provide a wealth of general information about an item), and access your semantic knowledge

Problem solving

The process of developing a solution or set of solutions designed to change the state of affairs from the current state to the goal state

Defining and representing problems

The process of stating the scope and goal of the problem and organizing the knowledge needed for addressing the problem, includes mentally representing the current and goal states, rules or constraints, and allowable operations available to solve the problem.

Defining Problems

The process of stating the scope and goal of the problem and organizing the knowledge needed for addressing the problem.

Belief bias

The tendency to think that a syllogism is valid if its conclusions are believable or thought to be true. •Example: -Some A are B-C is an A-Therefore, C is a B•Invalid Conclusion •Example: -Some professors are funny-Dr. Mounts is a professor-Therefore, Dr. Mounts is funny•Invalid Conclusion - Here, the argument structure is invalid, but you might be tempted to think it is valid because you agree with the conclusion.

Evans (1984, 2006)

Theory within the Dual process framework that we use one system based on heuristic processes (type 1) and another based on analytical processes (type 2). Type 2 operate logically based analyses with the representations activated from type 1.

Family resemblance

Things belonging to a category are related by virtue of having a set of overlapping similar features. Children look like their parents, but difficult to identify the exact features that make this the case.

Knowledge based categorization

This approach suggests that the different features of a concept are related to one another. Say that our concepts of a category include not just what features category members tend to have, but why they have them and how they are related. So, our concept of a pickup truck isn't just the features a truck has, but why it has those features -- how they help it satisfy its "reason for existence". Not an arbitrary list of features, but related to the purpose/behavior of the category

Casual reasoning

Two important factors when we draw conclusions: identifying covariation (how often 2 events co-occur) between the two events and believing that there is a mechanism (germs - this is what distinguishes cause from correlation) for the casual relationship. Your friend sneezing on your results in you feeling like you're getting sick the next night.

Mervis, Catlin, and Rosch

Typical items are produced first when prompted to produce category members

Meints, Plunkett, and Harris

Typical items are usually learned first

Pobric, Jefferies, and Lambon Ralph (2010)

Used TMS to test the distributed-only view and the distributed-plus-hub view. They were able to temporarily induce category-specific picture-naming deficits in normally unimpaired individuals. Applied the TMS to the anterior lobe (thought to be an amodal conceptual hub), the inferior parietal lobe (thought to be involved with processing concepts involving manipulable man-made objects), and occipital pole (served as the control condition). Findings saw that interference with IPL slowed naming across category-specific objects and ATL slowed naming across all types of concepts. Results consistent with the distributed plus hub approach. Results consistent with ATL serving as hub. TMS to the ATL disrupted naming of all types of objects. This is what would be predicted if the ATL is serving as a hub for integrating the features of objects.

Rips, Shoben & Smith (1973)

Used a speeded category verification task in which they presented subjects with sentences like 'a robin is a bird' or 'an elephant is a bird' and participants had to respond T/F as quickly as possible. Suggested the typicality effect.

Artificial concepts

Used in studying the difference between the exemplar approach and the prototype approach. The advantage is that researchers can tightly control the features involved in this category and can examine how the concepts are initially acquired. Limited in that they can be very simple compared to real-world concepts

Malt (1989)

Used pictures of real animals in a priming task that allowed her to investigate whether exemplars or prototypes were activated during a categorization task. Her results suggested that we may use both

Analogical transfer

Using the same solution for two problems with the same underlying structure. Gick and Holyoak (1980). Requires recognition of the similarities between the two problems. Participants must first compare the surface structure (medical vs. military/tumor, doctor, radiation beam, healthy tissue vs. dictator, army, general, village), then compare the underlying structure of the problem (dictator = tumor, army = radiation beam). Applying the solution from one problem to another that has a similar underlying structure. General story group: 40% didn't need a hint-35% solved with a hint, able to use analogical transfer vs. 10% without hint in control group.

Judgment

Very often, in making decisions we have to make judgments about the likelihood or frequency of an event. For example: do you need to do your readings for a class tomorrow? Your decision might rest on how likely it is that you will have a pop-quiz in that class.How do you make such judgments? •An extension of inductive reasoning-The human ability to infer, estimate, and predict the character of unknown events•In almost all judgment situations:-We do not have all the information to reach an accurate conclusion-Even if we did, we don't have the computational power required to combine the information-So often we make educated guesses using heuristics•Heuristics: "Rules of thumb" that are likely to provide the correct answer to a problem, but are not foolproof•Why do we use heuristics-Fast, cognitively easy, typically right

Ludwig Wittgenstein

Was an Austrian philosopher and a logical empiricist who argued against that concept as definitions approach using 'games' as the example. Introduced the concept of family resemblance

Luchins (1942)

Water jug problem - 3 containers with 100 cups. Solution is B-2C-A. When given simpler problem (7 & 8), kept complicated subtraction approach

Kelly, Bock, and Keil

When mentioning two members together, the more typical member is usually mentioned first

Mack, Preston , and Lowe

compared computational models of the exemplar and prototype approaches with fMRI scans of people's brains as they performed a categorization task. Prior to scan, they were told to categorize news objects into two categories. Following this, they categorized old and new objects while in the scanner. Both exemplar and prototype approaches accounted well for the behavioral data. Researchers then used the two models to compute the representational match between the test objects and the different representations (exemplar versus prototypes). They then compared these representational matches with the brain response data. Their results indicated that the exemplar model provided a better prediction than the prototype model for most participants.

Anchoring

•Anchoring & Adjustment heuristic•We quickly latch on to an estimate•Then we make adjustments up or down from this estimate•But --we won't adjust too far away from that estimate

Evolutionary rules: Cosmides(1989)

•Built-in cognitive program for detecting cheating(someone under 21 drinking beer is cheating)•Evolved to evaluate social contracts -is someone entitled to an outcome?•Evolution: Someone taking something they are not entitled to hurts your survival chances This is similar to the schema idea, but Cosmides claims that this is hard-wired through evolution. We are set up to detect cheaters because cheaters hurt your chances for survival. If a group member is taking more resources than they are entitled to, this negatively impacts the other members of the group. People who couldn't detect such cheating were less likely to survive and pass on their genes.

Learned Schemas: e.g., Cheng & Holyoak(1985)

•Humans learn schemas related to situations such as permission, obligation, and causation•Ex: In order to do something, you must first receive permission-Being 21 grants you permission to drink beer•Ex: If you do something, the other party is obligated to do something-If you go to work, your employer must pay you Rather than representing rules abstractly, we represent rules related to social transactions with other people. These are schemas that we can apply to different situations. So, the beer drinking version of the Wason 4 card task is easier because you can represent it as a permission schema. In order to do something, you need permission to do so.

Abstract Rules: e.g., Rips (1994)

•Humans possess knowledge of the rules of logic and apply them when reasoning•We create an abstract representation of the problem and evaluate using these rules Basically, we understand the rules of logic and can apply these rules using abstract representations.

Bernoulli: St. Petersburg Paradox

•People don't always do what expected value theory proposes they should•Expected value theory --Should make choice to maximize expected value. •St. Petersburgh Paradox-Choice A: $10,000 with probability = 1.0-Choice B: Flip a coin•Heads: win $2•Flip again --for every successive head --double amount-2 heads: $4-3 heads: $8-4 heads: $16-5 heads: $32-Etc. •Should choose B: $ ∞ > $10,000. Nobody does this, suggests that people don't use expected value theory when making choices.

Human Limitations

•Usually cannot have full representation of problem space•Aware of what some possible moves are and the resulting states. In most cases, humans don't have the processing capacity for many algorithmic approaches. For instance, you can't imagine the entire problem space for the Tower of Hanoi problem.