Conceptual Knowledge II: Concepts Organized in the Mind

Question 1

Concepts Definition?

Answer 1

“categories of objects, events, and abstract ideas”

Broader than a category

Question 2

Are concepts really organized in hierarchies? The semantic network approach says…

Answer 2

The semantic network approach says yes!
- Concepts are arranged in networks
- Networks are hierarchical

Question 3

Semantic Network Approach - Collins & Quillian(1969) hierarchical model? *Cognitive economy

Answer 3

• Categories/concepts are called nodes
• Properties of each concept are indicated at nodes
• Move up in the network to find more properties
  Cognitive economy:
  Shared properties are only stored at higher levels to avoid redundancy
• Exceptions stored at lower nodes

Question 4

Semantic Network Approach - Collins & Quillian(1969) experiment?

• Participants respond to true or false statements
• “A canary is a bird”
• Each statement contains two nodes
• IV: distance between nodes

Answer 4

Results suggest people respond faster when nodes are closer together

Supports Collins & Quillian’s hierarchical model

Question 5

Collins & Quillian’smodel intended to explain how concepts are functionally linked, not physiologically linked… Spreading activation?

Answer 5

Spreading Activation:
Activation of one node spreads to other connected nodes

Question 6

(Priming by Spreading Activation Example) Seeing a list of multiple words relating to one topic but….

Answer 6

Not the word itself, however one think they do

Snooze
Pillow
Moon
-> Think we saw sleep

Question 7

Limitations of Collins & Quillian’s model of semantic network approach?

Answer 7

• Doesn’t explain typicality effect
• Uncommon nodes are “skipped”, do not benefit from spreading activation

Question 8

Connectionist Approach?

Answer 8

• Connectionist models are implemented on computers, intended to mimic the way brains work and learn
• Modern machine-learning algorithms use this (esp. “neural network” algorithms)
• Concepts are represented by a distributed network of activation

Question 9

Connectionist Approach… Instead of “nodes,” there are “units” (like neurons)?

Answer 9

• Units are connected (like axons)
• Connection strength is assigned a weight
  Weight: degree to which next unit is activated or inhibited

Concepts are learned by experience in identifying them
- Errors cause readjustment of connection weights

Question 10

Support for the Connectionist Approach?

Answer 10

• Connectionist models really work!
• Computers can learn to identify things (computer vision loves this)
• Concepts are represented in a functional way

Question 11

Support for the Connectionist Approach - Evidence from brain damage?

Answer 11

• Damage usually only causes partial loss of concepts
• Supports idea that concepts are distributed widely

Question 12

Support for the Connectionist Approach - Evidence from learning?

Answer 12

• Generalisation occurs when e learn concepts that are similar to previously known ones
• Connectionism has generalisation built in

Question 13

How Concepts are Represented in the Brain - Sensory-functional hypothesis?

Answer 13

Sensory-functional hypothesis:
Living and nonliving things are represented by different systems

Evidence by some patients having category-specific memory impairment
- People could identify tolls, but not animals

But there are many exceptions; distinction between sensory and functional representation is too simple

Question 14

How Concepts are Represented in the Brain - Multiple-Factor approach?

Answer 14

Multiple-Factor approach:
Like sensory-functional hypothesis, but more factors use to distinguish concepts
E.g.: colour, smell, type of motion, how to use

Question 15

How Concepts are Represented in the Brain - Semantic-Category approach?

Answer 15

Semantic-Category Approach:
Semantic categories are mapped to distinct brain networks

Distributed, but some are visible with FMRI scans

Question 16

How Concepts are Represented in the Brain - Embodied approach?

Answer 16

Embodied Approach:
Concepts (esp. sensory and action concepts) are represented by the same networks responsible for sensing them and motor interaction with them

FMRI studies validates this

Question 17

How Concepts are Represented in the Brain - Hub-and-Spoke Model?

Answer 17

• Anterior temporal lobe (ATL) is the “hub” integrating information
• ATL is the nexus of other types of representations distributed throughout the brain
• Concepts are represented by linking the specific networks (e.g., visual, motor, auditory) together by a central point in the ATL

Question 18

Support for the Hub-and-Spoke Model?

Answer 18

• Damage to specialized brain areas can cause domain-specific conceptual deficits
• Damage to the ATL can cause general conceptual deficits, called semantic dementia
• Evidence from Transcranial Magnetic Studies