PSC 100 - L13 - Concept Knowledge

rules, concepts

Intro

In a heterogenous and dynamic environment, acquisition of stable and persistent knowledge relies on our ability to learn what 2 things?

concepts

Intro

Sets of nodes connected w/ each other in semantic networks

every day knowledge

Understanding Concepts

What kind of knowledge is considered the building blocks from which all knowledge is created?

• Apply general knowledge to new cases

• Draw broad conclusions from exp

definitional approach

Understanding Concepts

This approach sets boundaries for what is considered “in” or “out.” The problem is that it is always possible to find exceptions

family resemblance

Understanding Concepts

According to Wittgenstein, members of categories share a blank, where there are no defining features. Instead, there are characteristic features across members

characteristic features

Understanding Concepts

Family Resemblance

The more blank of a category an object has, the more likely it is to belong to that category

prototype

An average of various category members that have been encountered; an example that possesses all the characteristic features

prototype theory, experience

Prototypes & Typicality Effects

This approach sets a central tendency. According to this theory, a prototype is specified and other objects are compared to that ideal to determine

• A category’s prototype will differ across individuals based on what?

typicality

Prototypes & Typicality Effects

Prototype theory suggests that category membership is judged based on what? This is how much something resembles the prototype

graded membership

Prototypes & Typicality Effects

Where objects close to a prototype are “better” members of the category than objects farther from the prototype

longer, high distortion, low distortion

Prototypes & Typicality Effects

Testing the Prototype Notion

In a sentence verification task, the more distant an item is from the prototype, the blank it takes to make a judgment

• Ppl make categorization errors for blank exemplars than for blank exemplars

the most typical

Prototypes & Typicality Effects

Testing the Prototype Notion

In a production task, PS generally name what category members first? (These also yield faster response times in sentence verification tasks)

closer

Prototypes & Typicality Effects

Testing the Prototype Notion

In rating tasks, items that are blank to the prototype are rated as more typical of the category. Also consistently seen in sentence & production tasks

basic level categories

Prototypes & Typicality Effects

Categories that seem more “natural to us (e.g. chair vs furniture or wooden desk chair)

• Represented by a single word

• Default for naming objects

• Easy to explain commonalities

• Learned first

exemplar based reasoning

Exemplars

Analogies from Remembered Exemplars

Where categorization relies on knowledge about specific category members (exemplars) rather than the prototype

• We may categorize an object based on our most frequent experience with similar objects

prototypes, exemplars

Exemplars

A Combination of Exemplars & Prototypes

Blank provides an economical summary of the category

Blank provides info about category variability; although it’s less economical, it’s easier to adjust categories based on this

conceptual knowledge, exemplars, prototypes

Exemplars

A Combination of Exemplars & Prototypes

What kind of knowledge is a mix of exemplar and prototype?

• Early learning often involves blank

• Experience often averages exemplars to get blank

• With more experience, we use both exemplars & prototypes to ascertain category membership & recognize objects

false

The Difficulties With Categorizing Via Resemblance

True or false? Category judgments and typicality are dependent on each other. Explain

typicality, prototypes and exemplars, resemblance, knowledge

The Difficulties With Categorizing Via Resemblance

The Broader Role of Conceptual Knowledge

• (1) Blank influences category judgments - how quickly and likely to categorize

• (2) #1 effects reveal the substantial role of blank and blank

• (3) When using #2 you reply on a judgment of blank

• (4) #3 depends on other blank - which attributes to pay attention to/ignore?

causes, possibilities, concepts

The Difficulties With Categorizing Via Resemblance

Category Knowledge Guides Your Thinking About New Cases

Theories we hold enable us to…

• Think about new blank

• Think about new blank for categories (e.g. can an airplane be made out of wood?)

• Learn new blank

general knowledge, broad conclusions

The Difficulties With Categorizing Via Resemblance

Category Knowledge Guides Your Inferences

Categorization enables us to…

• Apply blank to new cases

• Draw blank from prior experiences

typicality, theories/broader beliefs

The Difficulties With Categorizing Via Resemblance

Category Knowledge Guides Your Inferences

What two things are category based inferences guided by?

natural kinds

The Diversity of Concepts

Things that we believe to have relatively stable properties (e.g. skunks, raccoons)

artifacts

The Diversity of Concepts

Things that we believe to NOT have stable properties (e.g. toasters, coffeepots)

goal derived, relational, event

The Diversity of Concepts

Concepts can be characterized by features AND what 3 types of other categories?

embodied cognition

The Diversity of Concepts

A proposal that our concepts include representations of perceptual properties and motor sequences

• Sensory & motor areas active when thinking about certain concepts

hub and spoke knowledge, general, specific

The Diversity of Concepts

Embodied Concepts

Where a “hub” connects and integrates more specialized info (the “spokes”) from other brain areas

• Damage to anterior temporal lobes => loss of blank knowledge

• Damage to a “spoke” => loss of blank knowledge

propositions

The Knowledge Network

The smallest unit of knowledge that can be either true or false (e.g. “children” vs “children love candy”)

• Nodes can represent concepts

  • Links between nodes can form more complex concepts

  • Links between nodes can vary in strength

propositional networks

The Knowledge Network

Distributed Processing

Has local representations where each node represents on concept or idea

connectionist networks

The Knowledge Network

Distributed Processing

Has distributed representations where each idea is represented by a pattern of activation across the network

parallel distributed processing

The Knowledge Network

Distributed processing

Allows the identification of patterns despite the variability in pattern implementation; more powerful