II: Part 4

What is a category? Give an example.

A set of all possible examples of a particular concept > The category chair has a variety of chairs

In relation to the classical view of theory of categorization, what should all object meet in order to be categorized in a specific category? What two things should the object meet?

All object should meet all the definitions of the category they belong in > Should have all the necessary and sufficient condition

In relation to one of the principle of Semantic Network, what does cognitive economy mean?

Avoid redundancy by storing the shared attributes at the highest level possible > Only the executions or unique attributes are stored in the lower levels (subordinate level)

In relation to the probabilistic view and the prototype theory, when writing down as many characteristics as we can we assigned many of the same attributes to chair and sofa than to mirror or rug. What does this mean? What does this show? What relationship exists between pro typicality and family resemblance?

"Good" prototypical examples of a category (chair, sofa) share many attributes with other members of the category > SHOWING prototypical "good" examples are high family resemblance (the higher the prototypically, the higher the family resemblance)

In relation to theory-based categorization, when given a category of bird, the feature "has wings" causes the feature "can fly" which causes the feature of "nests in trees." These links can help people develop a theory. Which feature is more important than nests in "nests in trees"? Why? What does this show?

"Has wings" because it is the cause while "nests in trees" is the effect > Features that cause other features are more important Easier to construct concept based on casual links between different features

In relation to the probabilistic view and the prototype theory, we organize an object on the basis of a prototype. What is the prototype? How are the categories defined?

-A "typical" or more reperesentative category member -"Average" of all the category members -Categories are defined by the "middle" instead of a definition that defines that boundaries of the category

What are the 3 theories of categorization? What do they categorize?

-Classical view (definitional approach) -Probabilistic view (which includes two different theories) > Prototype > Exemplar -Theory-based (explanatory approach) These theories explain the categorization of objects

In relation to the probabilistic view of theory of categorization, there are 3 solutions to getting around the issues of classical view. What are they? (these are features of the probabilistic view)

-Features of a concept are "characteristics" but not necessarily "defining" -An item belongs to a category if it is "similar" to members of the category -Some member have more characteristic properties than other > Unequal membership among member in the category

In relation to the different levels of categories, why is the basic level special? 4 reasons

-They are the first words that come into mind -They are the fist words that kids learn -Highest level of category with a single visual image > EX: saying dog instead of saying animal -Shortest word (usually a single word)

In relation to Connectionist Network model and learning, how can connection rates be changed?

-Two nodes are activated/firing together strengthen their connection (synchrony)

In relation to the Semantic Network, what are 3 of its weaknesses?

-Typicality effect: although some items have the same number of links, some are verified faster than others -Sometimes a link from the subordinate to superordinate link is made faster than the subordinate to basic link > "A pig is an animal" (superordinate) reaction time is faster than "a pig is a mammal" (basic)

In relation to the probabilistic view and the prototype theory, during a semantic priming test there was a good example of the color red (bright red) and a poor example of the color red (light pink). Before being tested, individuals were primed with the word "red". Results showed that the good examples were judged much faster than the poor examples. What do these results suggest? Why does this occur?

After semantic priming, prototypes are judges more quickly than nonprototypes > Occurs because when we see or hear the word red, it bring up the prototypical red in our minds. Then when we see the color red it matches with the red prototype that we have in our mind

In relation to theory-based categorization, what does essence have on categorization? (3)

All-or-none membership > If a thing possesses an essence of being something, it will definitely be categorized as that thing. If a thing does not have the essence, it will be excluded from the category In order for an object to change membership, its essence for being in its original category must be removed > As long as the item has the essence, it will be categorized in the same category People believe everyday categories also have essences > Social groups (gender, racial, ethnic), health categories

In relation to the prototype theory, how would you explain typicality and graded membership of an item?

Average of a category and Based on an item's similarity to the "average"

In relation to the different levels of categories, which level do most people use?

Basic level (bird or dog)

In relation to theory-based categorization, what is the psychological definition of essentialism?

Belief or view that an essence makes an object what it is > Members of a category share a deep underlying unobservable features (essence) of a category that constrains observable features > Essence of a category is immutable and unchangeable

Are everyday concepts a mix of exemplar (exemplar) or prototypes (prototype) or both? Explain. Also, which is suitable for small categories? And which is suitable fro large categories?

Both > Early learning of a category involves exemplar then experiences with examples will allow us to average them creating prototypes -Exemplars --> more suitable for small categories with relatively few members -Prototype --> more suitable for large categories with more members

In relation to the exemplar theory, how is a category represented? What does this mean?

By a specific set of examples > Not by an average representation of the category Means that each member of a category is an exemplar > There is no single best representation

In relation to Connectionist Network model, how does learning take place in connectionist/PDP networks?

By changing the connection weights (strength of connections) > Learning involves changes in the strength of connections between two nodes

In relation to theory-based categorization, what are ad hoc categories? What is an example? What is it dependent on?

Categories that are created at the time of categorization based on one's needs at the moment > EX: things to take on a camping trip (greatly dependent on our knowledge)

In relation to one of the theories of categorization, what does the classical view (definitional approach) state? What are they determined by?

Categorize have defining properties > Determined by necessary and sufficient condition

What two principle are assured in the Semantic Network?

Cognitive economy and inheritance

In relation to one of the theories of categorization --> Theory-Based Categorization, how are concepts defined? How is a category represented? What is the key point?

Defined by constructing theories about people, things, and events We represent categories based on explanatory theories that are driven by a set of briefs or knowledge about the world Key point: We use background knowledge, beliefs, and casual relationships to determine category membership

In relation to the classical view of theory of categorization, it was proposed that members of a given category have "family resemblance". What does this mean? What does this allow for? Do features define members? If not, what does?

Definitions do not apply to all members of the category > The idea that things in a particular category resemble one another in a number of ways > Allows for more variation Feature do not define members, but members resemble one another in one way or another

In relation to one of the theories of categorization, what does the exemplar theory state? What is an exemplar?

Draws a knowledge of specific category memory or exemplars from a category instead of specific information about the category Exemplar: an exemplar/instance of a category previously seen

In relation to the exemplar theory, how would you explain typicality and graded membership of an object?

Encountered more often and Based on how often an item is encountered

In relation to Connectionist Network model and learning, connection weights change when there is an error and back propagation occurs. What does this mean?

Error signals are transmitted back through the circuit to decrease the connection weights in the nodes that led to the error

In relation to theory-based categorization, we see categories as having deep, non obvious important properties. In other words.....

Essences > What makes category members that kind of things that they are

In relation to organization of concepts in the mind and the semantic network, what is the structure of the network?

From specific to more general level -Subordinate level (specific) --> basic level --> superordinate level (general)

In relation to the probabilistic view and the prototype theory, what two characterizations do they have? Define them. (related to their membership)

Graded membership: some members are more closer to the prototype of a category than others Fuzzy boundaries: no clear dividing line for membership

In relation to organization of concepts in the mind, what does the semantic network explain?

Hierarchical model for knowledge representation

In relation to Connectionist Network model, what are two ideas support this model?

High resembles of how the brain actually work > Damage to one part of the brain doesn't cause the whole brain to suffer similar to when one node is damaged, it does not completely destroy the other nodes Can explain generalization of learning > Similar concepts have similar patterns > We can make prediction about new category examples we've never seen

In relation to Connectionist Network, there are input units, output units, and hidden units. Incoming stimulus activates some of the input units which travels through the whole network activating the hidden and output units. What determines the activity patten of the hidden and output units?

Initial strength/activity of the input units AND Connection weights: the strength of relationship of different nodes --> determines how much activation can one node pass to one node

In relation to the probabilistic view and the prototype theory, how are items judged to belong to a specific category?

Judged on their similarity to the prototype > The more similar to the prototype, the more likely they will be judged to belong to that category

In relation to the different levels of categories, what is the exception to using the basic level? What affects categorization? Explain

Knowledge can affect categorization > Although people use the basic level, with more knowledge people tend to use more subordinate category words

In relation to organization of concepts in the mind, what is the Connectionist Network how is knowledge represented? Are there still nodes? What is the pattern of activation?

Knowledge is represented by distributed activity of many different nodes in the network (distributed representation) > Not locally represented by nodes > There are still nodes, but an individual node doesn't carry meaning by itself Involve parallel distributed processing (PDP): each node processes information at the same time

In relation to one of the principle of Semantic Network, what does inheritance mean?

Lower-level items share attributes to higher-level nodes that they are connected to

What is a concept? Given an example.

Mental representation of a class or an individual > Each of us have an idea of what a chair is. That is, we have a concept of a chair

What time to retrieve information based on? How does this support the semantic network model?

The number of links to travel -The more links you have to travel, the longer it take to retrieve the information Support semantic network: the farther they are apart, the longer it takes us to discover that they are related

What are the different levels of categories?

Top to bottom: > Superordinate level (animal) --> basic level (bird or dog) --> subordinate level (robin and sparrow) --> sub-subordinate level (American robin) Top to bottom --> general to specific

In relation to the classical view of theory of categorization, one of the problems is that there are exceptions to the rule. Is this true or false?

True

In relation to difficulties with categorizing via resemblance, what two things can sometimes be independent of each other?

Typicality and category membership

In relation to independence of typicality and category membership, participants were given a list of even numbers and asked to rate each number by how typically even they are (how even each number is). They saw that typicality did differ, but they had no problem categorizing them into even vs odd numbers. What does this mean?

Typically and category membership can sometimes be unrelated

In relation to the exemplar theory, how do we classify new items to our exemplars? What is it based on?

We first learn specific examples of a concepts, and then classify new items by comparing the resemblance of the new items to our exemplars > Based on how much it resembles the other exemplars in the category

In relation to the classical view of theory of categorization, one of the problems is that the classical view definition is inefficient. Why?

We would need lots of features to define a category so that it includes all embers we assign to that category > This is more information that we can process

In relation to the probabilistic view and the prototype theory, do memberships have fuzzy boundaries? Give an example.

Yes, there is no clear diving line for membership EX: the boundaries for all the shades of green are fuzzy

In relation to the classical view of theory of categorization, one of the problems is "defining properties" does not hold for all members. What does it mean? What does the classical view not allow?

You can often find exceptions in members of the same category > Does not allow for exceptions, does not allow for variability, does not allow for unequal membership in a category

In relation to theory-based categorization and the "Roobans" study, there were 2 conditions: without casual background --> shown 3 features without casual background (eats fruits, has sticky feet, and nests in trees) about roobans without any explication of the causal links among the features. In the other conditions, the researcher provided participants with casual background within the given features. The results showed that when no background information was given, whether the missing feature was x, y, or z it did not matter. Why did this occur? However, when provided with background knowledge and when the first causal feature was missing, category membership rating was the lowest. What does this experiment show?

No background: People did not put differential imporance for z, y, z features Background: Providing people with causal features can change their categorization patterns

In relation to difficulties with categorizing via resemblance, does a atypical feature (low resemblance) get excluded from a category membership?

No, we need the more mere typicality to be able to account for how our minds places different items in different categories

In relation to organization of concepts in the mind and the semantic network, what does it consist of? What can it be conceptualized as? What happens when one node is activated? What happens as a result?

Nodes and connection between those are links Can be conceptualized as a fishnet where everything is connected to everything else When one node is activated it spreads to all the links. As a result, the nodes are primed -- making them easily accessible from memory

In relation to the probabilistic view, we place items in categories based on their resemblance to what two theories? What are they based on?

Prototype theory or exemplar theory Based on similarity or resemblance

Comparing the prototype theory and exemplar theory: four circle items are shown --> 2 red and 2 blue. What would the prototype theory force you to do? What would the exemplar theory force you to do? Which is more economical and flexible? Which is less?

Prototype: -Take the average (purple) -Economical -Less flexible Exemplar: -Actual examples (take one of each) -Flexible -Less economical

In relation to difficulties with categorizing via resemblance, what can we conclude?

Resemblance cannot fully account for how we categorize things

In relation to the probabilistic view and the prototype theory, during a production task when asked to name as many birds as possible, category members closest to the prototype were first ranked. What does this show?

SHOWS that the prototypical items are readily available in our mind

What are the two organization of concepts of mind?

Semantic Network and Connectionist Network

In relation to the probabilistic view and the prototype theory, a category will have a graded membership. What does this mean? What word as associated with this idea? Are categories created equal?

Some members of a category are closer to the prototype > Some members are more "prototypical" (good example) to a category while some are non prototypical > Member of categories are not created equally , but differ in their prototypically

In relation to the probabilistic view and the prototype theory, during a sentence verification task, atypical and typical statements about categories were asked. When were the responses slower? When were there response the fastest?

Statements about typical category members such as "is apple a fruit" took less time Slower response when sentences included atypical member of the category like "is avocado a fruit"